Split (1)

train · ~9.11M rows (showing the first 93.5k)

identifier stringlengths 11 32	pdf_url stringlengths 17 4.62k ⌀	lang stringclasses 120 values	error stringclasses 1 value	title stringlengths 2 500 ⌀	source_name stringlengths 1 435 ⌀	publication_year float64 1.9k 2.02k	license stringclasses 3 values	word_count int64 0 1.64M	text stringlengths 1 9.75M
https://openalex.org/W4287509495	https://publikationsserver.tu-braunschweig.de/servlets/MCRFileNodeServlet/dbbs_derivate_00048592/Vogt_energies-14-07271.pdf	English	null	Sensor-Based Machine Learning Approach for Indoor Air Quality Monitoring in an Automobile Manufacturing	RePEc: Research Papers in Economics	2,021	cc-by	9,321	Keywords: machine learning; HVAC; control system; body shop; automotive industry Keywords: machine learning; HVAC; control system; body shop; automotive industry Sensor-Based Machine Learning Approach for Indoor Air Quality Monitoring in an Automobile Manufacturing dy 1,, Marcus Vogt 2 , Rushit Kansara 2 , Clemens Felsmann 1 and Christoph Herrmann 2 Yose Wandy 1,, Marcus Vogt 2 , Rushit Kansara 2 , Clemens Felsmann 1 and Christop 1 Chair of Building Energy Systems and Heat Supply, Institute of Energy Technology, Technische Universität Dresden Helmholtzstr 14 01069 Dresden Germany; clemens felsmann@tu dresden de 1 Chair of Building Energy Systems and Heat Supply, Institute of Energy Technology, Technische Univers Dresden, Helmholtzstr. 14, 01069 Dresden, Germany; clemens.felsmann@tu-dresden.de 1 Chair of Building Energy Systems and Heat Supply, Institute of Energy Technology, Technische Universität Dresden, Helmholtzstr. 14, 01069 Dresden, Germany; clemens.felsmann@tu-dresden.de 2 Chair of Sustainable Manufacturing and Life Cycle Engineering, Institute of Machine Tools and Production Technology, Technische Universität Braunschweig, Langer Kamp 19b, 38106 Braunschweig, Germany; marcus.vogt@tu-braunschweig.de (M.V.); rushit2011@gmail.com (R.K.); c.herrmann@tu-braunschweig.de (C.H.) * Correspondence: ywandy@ymail.com Dresden, Helmholtzstr. 14, 01069 Dresden, Germany; clemens.felsmann@tu-dresden.de 2 Chair of Sustainable Manufacturing and Life Cycle Engineering, Institute of Machine Tools and Production Technology, Technische Universität Braunschweig, Langer Kamp 19b, 38106 Braunschweig, Germany; marcus.vogt@tu-braunschweig.de (M.V.); rushit2011@gmail.com (R.K.); c.herrmann@tu-braunschweig.de (C.H.) * Correspondence: ywandy@ymail.com * Correspondence: ywandy@ymail.com Abstract: The alternative control concept using emission from the machine has the potential to reduce energy consumption in HVAC systems. This paper reports on a study of alternative inputs for a control system of HVAC using machine learning algorithms, based on data that are gathered in a welding area of an automotive factory. A data set of CO2, ﬁne dust, temperatures and air velocity was logged using continuous and gravimetric measurements during two typical production weeks. The HVAC system was reduced gradually each day to trigger ﬂuctuations of emission. The data were used to train and test various machine learning models using different statistical indices, consequently to choose a best ﬁt model. Different models were tested and the Long Short-Term Memory model showed the best result, with 0.821 discrepancy on R2. The gravimetric samples proved that the reduction of air exchange rate does not correlate to escalation of ﬁne dust linearly, which means one cannot rely on just gravimetric samples for HVAC system optimization. Furthermore, by using machine learning algorithms, this study shows that by using commonly available low cost sensors in a production hall, it is possible to correlate ﬁne dust data cost effectively and reduce electricity consumption of the HVAC. Citation: Wandy, Y.; Vogt, M.; Kansara, R.; Felsmann, C.; Herrmann, C. Sensor-Based Machine Learning Approach for Indoor Air Quality Monitoring in an Automobile Manufacturing. Energies 2021, 14, 7271. https://doi.org/ 10.3390/en14217271 energies energies Keywords: machine learning; HVAC; control system; body shop; automotive industry 1. Introduction Received: 29 July 2021 Accepted: 20 October 2021 Published: 3 November 2021 The digitalization of manufacturing is being driven forward by the decreasing costs of information and communication technology. Cyber-physical systems (CPS), or in the manufacturing context Cyber-Physical Production Systems (CPPS), are an important tech- nological element in the realization of the 4th industrial revolution [1]. CPPS and their applicability in industrial environments are of increasing interest in current research and industry. A CPPS is clustered into subsystems: the physical world and the cyber world. These subsystems interact with each other through data acquisition and decision-making support and control functionalities, respectively [2]. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. In 2018, the World’s energy consumption has increased by almost 40% compared to the consumption in 1990 [3]. The industrial share of total ﬁnal consumption (TFC) stayed at the ~30% [4]. Although the absolute consumption increased to ~57% compared to the consumption of industrial sector in 1990 [5]. In times of increasing energy demand and decreasing energy resources, a further optimization and revaluation in industrial sector is imperative. Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). p The Heating Ventilating and Air-Conditioning (HVAC) system is responsible for a large part of the energy consumption in industry. According to internal Volkswagen data [6], it accounts on average for about 40% of the total energy consumption in industrial buildings. Automobile production can be classiﬁed into press shop, assembly, paint shop https://www.mdpi.com/journal/energies Energies 2021, 14, 7271. https://doi.org/10.3390/en14217271 Energies 2021, 14, 7271 2 of 16 and body shop. Body shop is responsible for almost one-quarter of the total energy consumption, preceded only by the paint shop [6]. Within the case study of a body shop, the HVAC system has the highest energy demand, succeeding the other infrastructure consumption: the hall lighting and workplace lighting. In the body shop, the HVAC consumption exceeds 60% of the total energy demand, succeeding the consumption from the production itself [6]. 1. Introduction p The main function of an HVAC system is to satisfy and maintain the requirements with respect to air and building environment in general, for example, thermal comfort, air pollution control and the hygienic aspect, which is achieved by conditioning outdoor air to the desired levels in occupied buildings or for product processing and transporting the air in the room through an exhaust duct. It controls and maintains temperature, humidity, air movement, air cleanliness and pressure differential within a deﬁned space. Wang [7] classiﬁed HVAC systems into comfort HVAC systems and process HVAC systems depend- ing on the application. Comfort HVAC systems provide occupants with a comfortable and healthy indoor climate and process HVAC systems provide and maintain the necessary air conditioning for production, storage or other deﬁned processes [7,8]. In order to guarantee that the indoor climate in the body shop is up to the user requirements, the production’s emissions need to be transported out of the conditioned space using the HVAC system, this is done using the technical exhaust air vent. One of the main emissions in a body shop is ﬁne dust, which is a byproduct of the welding process. In the geographical location of this study, Wrzesnia, Poland, to be precise, an occupational exposure limit is regulated under TRGS 900 [9]. The limit values are 1.25 mg/m3 for PM2.5 and 10 mg/m3 for PM10; this is then taken up again in Chapter 2 and elaborated on in more detail. Fine dust can be measured using speciﬁc measuring technology, speciﬁcally using an optical sensor that measures and monitors ﬁne dust concentrations with preconditioned air. Within the study, the ﬁne dust is measured by this optical scattered light measurement [10]. However, this method is highly cost-intensive [11]. Digital technologies, such as the Internet of Things (IoT), offer a wide number of ap- plications in various sectors [12]. One of the examples is to use cost-effective smart sensors, which enable users to measure ambient conditions and upload them to the cloud [13]. This enables users to experiment or use a more modern control concept, for example, machine learning algorithms. Li [14] presented a hybrid model based on long short-term memory (LSTM) network and attention mechanism (LSTM-attention) applied to the prediction of Total Suspended Particulate Matter (TSP) concentration. 2.1. Particulate Matter (PM) Fine dust or Particulate matter (PM) is the term used to describe particles in the air that do not sink to the ground immediately, but remain in the atmosphere for a certain period of time. Particulate matter can come either from natural sources or from human activities. Natural contributions include emissions from volcanoes, oceans, soil erosion, forest and bush ﬁres, as well as dead and eroded (abraded) organic remains, pollen, spores, bacteria and viruses. Sources from human activities include transport, power plants, district heating plants, waste incineration plants, private and commercial heating systems and certain industrial processes such as metal production, especially steel production [15,16]. Airborne particles range in size from a few nanometers (nm) to about 100 microme- ters (µm). This depends on the process and material used in the working environment. Madanchi [17] presented an experimental approach to evaluate the inﬂuence of process parameters on the formation of emissions and particle size. The work showed that the process parameter (cutting speed and cooling lubricant volume ﬂow) has linear impact not only towards the material removal process, but also the emissions and particle size. y p p The size of the dust particles, characterized by “aerodynamic diameter” as a model, is the decisive property for their classiﬁcation [15,18]: • The term ’Total Suspended Particulate Matter’ (TSP) covers all suspended particles with an aerodynamic diameter smaller than ~60 micrometers (µm), y • PM10 refers to all particles with aerodynamic diameter less than 10 micrometers. • PM2.5 refers to particles with an aerodynamic diameter of less than 2.5 micrometers • PM2.5 refers to particles with an aerodynamic diameter of less than 2.5 micrometers. • The smallest categorized particles are called ultra ﬁne particles. Their aerodynamic diameters are less than 0.1 micrometers. p y • The smallest categorized particles are called ultra ﬁne particles. Their aerodynamic diameters are less than 0.1 micrometers. The nose, mouth and throat can ﬁlter back particles larger than 10 micrometers. A small percentage of particles between 10 and 2.5 µm (PM10-2.5) are able to reach the bronchiole and alveoli. Particles that have a diameter smaller than 2.5 µm (PM2.5) have a higher probability to reach and stay in the alveoli. These particles can trigger inﬂammatory processes in the lung tissue [19]. In the alveoli, respiration and blood circulation are functionally and anatomically very closely connected [15]. 1. Introduction Li [14] showed that the model approach can be applied to the prediction of dust concentrations in open-pit mines, thus helping in decision support on when to carry out dust suppression work. The main scope of this paper is to study different model approaches using data measured in an industrial welding area of an automobile body shop. The paper starts with a literature review concerning ﬁne dust or particulate matter and the national regulations and technical rules concerning Particulate Matter (PM) and industrial welding. It then explains the methodology and approach used in this paper. This is followed by the factory data collection setup, data analysis and data correlations. Subsequently, the created data models or algorithms are presented with the results and the statistical comparisons. Finally, the conclusion and future scope or applicability of the paper are outlined. The project shows the possibilities of using an alternative data model correlation from more common and less expensive parameters, such as air velocity, air temperature, relative humidity, the CO2 content around the perimeter and the electric current of the welder. By using machine learning algorithms, we show that by using commonly available sensors in a production hall, it is possible to correlate ﬁne dust data cost-effectively. Energies 2021, 14, 7271 3 of 16 2.1. Particulate Matter (PM) This is why abnormalities in one system, such as inﬂammations in the respiratory tract, can also additionally affect other systems, for example, the heart or circulatory system [15]. 2.2. National Regulations and Technical Rules concerning Particulate Matter (PM) and Industrial Welding The main focus of welding processes lies on materials (ﬁller and base material). How- ever, note that neither base materials nor ﬁller materials (for example, welding materials) are considered “hazardous substances”. “Hazardous substances” are only released as a re- sult of a welding process, which consist primarily of particulate and/or gaseous hazardous substances. These hazardous substances are classiﬁed within the meaning of the Regulation on Hazardous Substances (GefStoffV). An occupational exposure limit was included in TRGS 900 [9]. TRGS 900 shows a list of the quantities’ threshold for substances with speciﬁc toxic properties, for example, for dusts containing heavy metals or for wood dusts and ﬂour dusts. If the substance in use is not listed, then a general threshold of ﬁne dust is applicable. These thresholds are compulsory in Germany from beginning of 2019, after a transitional period of ﬁve years (until 2018) was established. However, welding fumes are not dusts in the sense of the deﬁnition, as dusts are produced during mechanical processes, for example grinding. Respectively, welding fumes are produced by thermal processes at very high temperatures, for example, arc welding. Welding fumes can therefore not be directly equated with dust. It consists of mainly from particles that are alveolar. In comparison to ﬁne dust, fumes are smaller in size, approximately between 0.2 to 1 µm [20]. This means that the welding fume particles and in unfavorable workplace situations to impair the ability of the person exposed to the respiratory tract or damage it. Furthermore, other negative health effects on the human organism are also to be considered [21]. A direct Energies 2021, 14, 7271 4 of 16 speciﬁc threshold for welding fumes has not yet been deﬁned. The metal compound that can be found in welding fumes has a dust density higher than 0.0025 mg/m3, approxi- mately between 0.005 mg/m3 and 0.007 mg/m3 [22]. In reality, this is the reason for the risk assessment: the dust limit values are measured and converted mathematically to get the density of the metals from the fumes [22]. y Basically, the dust limit value for alveolar dust is (<2.5 µm) of 1.25 mg/m3 [9]. DIN EN 689 [23] deﬁned the limit further, including the area analysis, control measurements, protocol and processing of measurement data. 2.2. National Regulations and Technical Rules concerning Particulate Matter (PM) and Industrial Welding In the automotive industry, companies such as Volkswagen AG have to establish their safety measures and deﬁne their time interval between control measures, based on their measurements data and national regulations. Ac- cording to DIN EN 689 [23], the time interval between measurements should be determined taking into account the following factors: • process cycles taking into account normal working conditions; • the effectiveness of process controls; • time span until the condition is under control again; • the variation over time of the measurement results. Based on these factors and the upper limit deﬁned by the national regulations, Volk- swagen AG detailed their guidelines concerning the threshold and control measurements in welding area for alveolar dust. Using gravimetric analysis, the expositions in welding areas are measured and controlled. 3. Methodology and Approach As methodological framework a Cyber-Physical Production Systems (CPPS) is used; Figure 1 shows an adapted CPPS framework for this application. Data Acquisition Data Gathering Decision Support Control Visualisation Human in Focus Data Treatment & Storage LSTM PREDICTED ACTUAL Normalized PM10 Body Shop ሶ𝑉: 50 ሶ % 𝑉𝑁 Worker Tm: 22 °C CO2m: 600 ppm … Welding Area ሶ𝑉𝐻𝑉𝐴𝐶: 100 % ሶ𝑉𝑁 ሶ𝑉: 50 % ሶ𝑉𝑁 Physical World Design & Control Parameters State & Disturbance Variables Cyber World Modeling & Simulation Supervised Learning Regression Classification Figure 1. Schematic of CPPS framework for the use case based on the work in [2]. Physical World Design & Control Parameters Figure 1. Schematic of CPPS framework for the use case based on the work in [2]. The approach to integrate a CPPS in the body shop is driven by the goals of improving energy efﬁciency while maintaining air quality in the production environment and being cost efﬁcient. The CPPS system provides a decision support by calculating the ﬁne dust concentration in the production environment. This enables the HVAC operator in the physical world to use conservative sensors, such as temperature, relative humidity and Energies 2021, 14, 7271 5 of 16 CO2 to operate the HVAC system. Further, the CPPS enables a model-based control in the future. CO2 to operate the HVAC system. Further, the CPPS enables a model-based control in the future. With this cyber-physical HVAC system, energy efﬁciency, air quality and thermal conditions can be improved in the production environment by providing decision support, greater transparency and enabling model-predictive control in real-time. With the proposed approach, one can measure ﬁne dust in the air without using expensive ﬁne dust sensor in each welding area in a body shop. This can lead to a high cost savings when monitoring a large body shop hall in a factory. The low-cost IoT sensor technology combined with machine learning enables large-scale deployment with moderate costs compared to the high-precision ﬁne dust sensors on factory scale. 3.1. Setup for Collecting Data in the Factory Hall In order to ﬁnd the correlation between the HVAC-System, air quality and production activity, it is necessary to collect and analyze the variables on site. An experiment within the scope of a collaboration project between TU Braunschweig and Volkswagen AG is thus commenced. To investigate seasonal inﬂuences, experiments were conducted in winter and summer at ﬁve consecutively working days, to be exact from 6th to 8th of August 2018 and from 26th to 30th of November 2018. This allowed us to measure the experiment under 5 different air exchange rates, which were set 8 h before the start of each measurement. For the ﬁrst three days, the reductions were concentrated on the main exhaust air vents, the openings of which can be found directly under the ceiling of the production hall, this is detailed Table 1. Moreover, for the last two days, we set to measure the effect on the reductions on the technical exhaust vents, which is directly responsible for exhausting the welding emission. Figure 2 shows the difference on the positions of the two exhaust vents and the HVAC system used in the factory hall. Silencer Ventilator Filter Shutter Flap Weather Prot. Grid Deflector Hood Heater (Gas) Heat Recovery Outdoor Air Exhaust Air Supply Air Exhaust Air Technical Exhaust Exhaust (Hall) Figure 2. Schematic representation of the HVAC system and the difference between the two exhausts. Figure 2. Schematic representation of the HVAC system and the difference between the two exhausts. The reduction limit was set to 50% because the technical exhaust vents are the main way to exhaust the byproducts of the welding. Table 1 lists the experimental parameters for the work days with the expected energy saving potential in compare to the reference day (monday), which is based on the assumption of constant fan efﬁciency. A change in volume ﬂow rate affects the fan power requirement as follows [8]: P2 = P1 V2 V1 3 Energies 2021, 14, 7271 6 of 16 for example, 10% less volume rate will lower power demand by 33%. The supply air fan has a nominal power of 40 kW, while the exhaust air fan and the technical exhaust fan operate at nominal power of 30 kW and 60 kW. Table 1. Parameters for the experiments in winter and summer with the theoretical electricity savings. for the experiments in winter and summer with the theoretical electricity savings. Table 1. 3.1. Setup for Collecting Data in the Factory Hall Parameters for the experiments in winter and summer with the theoretical electricity savings. Day Supply Air Energy Saving Exhaust Air Energy Saving Techn. Exhaust Energy Saving Monday 75.000 m3/h - 22.137 m3/h - 24.000 m3/h - Tuesday −13% 34% −46% 84% −0% 0% Wednesday −30% 66% −100% 100% −0% 0% Thursday −16% 41% −0% 0% −30% 66% Friday −17% 43% −0% 0% −50% 88% The samples were taken during welding corrections; welding of ﬂoor elements (gal- vanized steel) with the welding torch SKS Power Feeder PF5 (MAG method); coated with steel wire with copper with a diameter of 1.0 mm; Argon; and CO2 inert gas welding, sanding and cleaning of welding seams. An oscillating time of at least 6 h was taken into account, so that the air conditions in the welding area settle before the measurements were commenced. 3.2. Data Acquisition and Analysis In order to ensure that the regulatory thresholds are well maintained, certiﬁed gravi- metric measurements from the work safety department accompanied the experiment. This gravimetric measurements were executed according to PN EN 689:2018, a Polish version of EN 689 [24]. Tables 2 and 3 show the result of the gravimetric measurements respectively in summer time between 6 and 8 August 2018 and winter time from 26 to 30 November 2018, measured over the time period of 8 h. The gravimetric measurement devices were hung directly on the workers shown in Figure 3. The results of the gravimetric measurements are shown in Tables 2 and 3. Table 2. Results from the gravimetric measurements in summer. Table 2. Results from the gravimetric measurements in summer. 06.08.2018 07.08.2018 08.08.2018 09.08.2018 10.08.2018 PM2.5 (mg/m3) 0.29 3.24 0.89 0.36 0.31 PM10 (mg/m3) <0.11 1.42 1.52 0.42 0.2 The exposition limits (PM2.5 = 1.25 mg/m3 and PM10 = 10 mg/m3) were held under the regulatory threshold during the experiments, except for the PM2.5 on Tuesday 7 August 2018. This outlier, however, was not backed by the PM10 result on the same day, which stayed far below the limit of 10 mg/m3. The measurements with the same air exchange rate settings were also repeated in winter and shown in the Table 3. The measurement results are more dependent on workers behavior, movements and routines, rather than the reduction in air exchange rate shown in Table 1, which might explain the spike and the nonlinear results. Table 3. Results from the gravimetric measurements in winter. 26.11.2018 27.11.2018 28.11.2018 29.11.2018 30.11.2018 PM2.5 (mg/m3) 0.43 0.26 <0.12 0.39 <0.12 PM10 (mg/m3) 0.58 0.38 0.38 0.58 0.22 Table 3. Results from the gravimetric measurements in winter. 3.3. Correlation Study between Fine Dust and Other Parameters 3.3. Correlation Study between Fine Dust and Other Parameters In order to be able to design a system which can predict the amount of values of PM10 and PM2.5 with the combination of other sensors, reference data sets are measured and collected. 7 of 16 Energies 2021, 14, 7271 Worker Worker V CT CO2 T RH PM2.5 PM10 V CT CO2 Figure 3. Schematic representation of the experimental measurements (V = Air Velocity, CT = Electric Current used for welding, T = Air Temperature, RH = Relative Humidity, CO2 = Carbon Dioxide). Figure 3. Schematic representation of the experimental measurements (V = Air Velocity, CT = Electric Current used for welding, T = Air Temperature, RH = Relative Humidity, CO2 = Carbon Dioxide). Figure 3 shows a representative schematic of the measurements setup in the hand welding area on-site. The places chosen for the sensor installations represent the spots where workers operate manually using a hand welder during the productions, which means the resulting air velocity, Temperature, Relative Humidity and CO2 depend greatly on workers actions and activities. The ﬁne dust PM2.5 and PM 10 are measured using optical sensors as well providing a dataset for the purpose of validating the designed algorithm. The sensor used was ﬁne dust FDS 15 from Dr. Födisch Umweltmesstechnik AG which has accuracy of ±5 µg/m2 and a measuring range from 2 µg/m2 to 3.000 µg/m2 [10]. y g g g g g Table 4 shows the results of the measurements data in winter time on 8-h average, the data of PM2.5 and PM10 were measured using optical sensors FDS 15. With the proper use of the data from these sensors combined with an algorithm, the amount of PM10 and PM2.5 in the air can be predicted to a certain extent. Which variables are to be selected for input can be decided based on the correlation analysis. The best combination of the inputs, that gives the lowest error in the prediction, can be chosen as input variables for the model. This selection is known as feature selection [25]. From the correlation analysis, the dependency of PM10 and PM2.5 on other available variables is noticeable. Figures 4 and 5 show the calculated linear coefﬁcients for each variable using Pearson and Spearman’s Rho Correlation. The correlation between and input variables and target variable provides the basis for the feature extraction. 3.3. Correlation Study between Fine Dust and Other Parameters Current CO2 RH T PM2.5 PM10 [m/s] [A] [ppm] [%] [◦C] [mg/m3] [mg/m3] 26 Nov 2018 08:00 0.15 2.07 855.2 34.9 21.8 0.063 0.084 26 Nov 2018 16:00 0.14 1.27 877.7 31.5 21.3 0.062 0.088 27 Nov 2018 00:00 0.13 1.30 966.5 29.8 21.3 0.061 0.091 27 Nov 2018 08:00 0.13 1.15 876.4 30.7 21.5 0.061 0.081 27 Nov 2018 16:00 0.13 1.24 847.1 30.3 21.6 0.061 0.083 28 Nov 2018 00:00 0.13 1.09 869.3 31.5 21.6 0.061 0.086 28 Nov 2018 08:00 0.13 1.25 907.4 31.1 21.6 0.064 0.079 28 Nov 2018 16:00 0.15 1.72 911.9 29.5 21.6 0.080 0.086 29 Nov 2018 00:00 0.18 1.31 999.2 26.9 21.7 0.077 0.083 29 Nov 2018 08:00 0.18 1.17 943.3 25.0 21.8 0.076 0.082 29 Nov 2018 16:00 0.21 1.57 944.2 24.6 21.8 0.076 0.082 30 Nov 2018 00:00 0.20 1.68 957.9 25.6 21.8 0.075 0.080 30 Nov 2018 08:00 0.24 3.31 890.7 26.2 21.9 0.077 0.081 V CT CO2 T RH PM 2.5 PM 10 V CT CO2 T RH PM 2.5 PM 10 1 0.083 0.044 0.073 -0.025 0.067 0.059 0.083 1 0.039 0.032 -0.13 0.14 0.15 0.044 0.039 1 0.024 -0.55 0.38 0.34 0.017 0.032 0.024 1 -0.48 0.17 0.075 -0.025 -0.13 -0.55 -0.48 1 -0.6 -0.36 0.067 0.14 0.38 0.17 -0.6 1 0.68 0.059 0.15 0.34 0.075 -0.36 0.68 1 1.00 0.75 0.50 0.25 0.00 0.25 0.50 0.75 1.00 Figure 4. Correlation analysis using Pearson Correlation. Table 4. 8-h average values of the ﬁeld data measurements. V CT CO2 T RH PM 2.5 PM 10 V CT CO2 T RH PM 2.5 PM 10 1 0.083 0.044 0.073 -0.025 0.067 0.059 0.083 1 0.039 0.032 -0.13 0.14 0.15 0.044 0.039 1 0.024 -0.55 0.38 0.34 0.017 0.032 0.024 1 -0.48 0.17 0.075 -0.025 -0.13 -0.55 -0.48 1 -0.6 -0.36 0.067 0.14 0.38 0.17 -0.6 1 0.68 0.059 0.15 0.34 0.075 -0.36 0.68 1 1.00 0.75 0.50 0.25 0.00 0.25 0.50 0.75 1.00 Figure 4. Correlation analysis using Pearson Correlation. Figure 4. Correlation analysis using Pearson Correlation. 3.3. Correlation Study between Fine Dust and Other Parameters Current CO2 RH T PM2.5 PM10 [m/s] [A] [ppm] [%] [◦C] [mg/m3] [mg/m3] 26 Nov 2018 08:00 0.15 2.07 855.2 34.9 21.8 0.063 0.084 26 Nov 2018 16:00 0.14 1.27 877.7 31.5 21.3 0.062 0.088 27 Nov 2018 00:00 0.13 1.30 966.5 29.8 21.3 0.061 0.091 27 Nov 2018 08:00 0.13 1.15 876.4 30.7 21.5 0.061 0.081 27 Nov 2018 16:00 0.13 1.24 847.1 30.3 21.6 0.061 0.083 28 Nov 2018 00:00 0.13 1.09 869.3 31.5 21.6 0.061 0.086 28 Nov 2018 08:00 0.13 1.25 907.4 31.1 21.6 0.064 0.079 28 Nov 2018 16:00 0.15 1.72 911.9 29.5 21.6 0.080 0.086 29 Nov 2018 00:00 0.18 1.31 999.2 26.9 21.7 0.077 0.083 29 Nov 2018 08:00 0.18 1.17 943.3 25.0 21.8 0.076 0.082 29 Nov 2018 16:00 0.21 1.57 944.2 24.6 21.8 0.076 0.082 30 Nov 2018 00:00 0.20 1.68 957.9 25.6 21.8 0.075 0.080 30 Nov 2018 08:00 0.24 3.31 890.7 26.2 21.9 0.077 0.081 V CT CO2 T RH PM 2.5 PM 10 V CT CO2 T RH PM 2.5 PM 10 1 0.083 0.044 0.073 -0.025 0.067 0.059 0.083 1 0.039 0.032 -0.13 0.14 0.15 0.044 0.039 1 0.024 -0.55 0.38 0.34 0.017 0.032 0.024 1 -0.48 0.17 0.075 -0.025 -0.13 -0.55 -0.48 1 -0.6 -0.36 0.067 0.14 0.38 0.17 -0.6 1 0.68 0.059 0.15 0.34 0.075 -0.36 0.68 1 1.00 0.75 0.50 0.25 0.00 0.25 0.50 0.75 1.00 Figure 4. Correlation analysis using Pearson Correlation. Table 4. 8-h average values of the ﬁeld data measurements. Time Air Vel. 3.3. Correlation Study between Fine Dust and Other Parameters Strong correlation means that the changes in chosen features (variables) causes higher changes in the target variable. Pearson coefﬁcient suggests how two variables are correlated linearly. On the other hand, Spearman’s Rho coefﬁcient shows the nonlinear monotonic relation between two variables. The formula for the Spearman rank correlation coefﬁcient is [26] rs = 1 − 6 ∑d2 i n(n2 −1 ! , where di = R(Xi) −R(Yi) is the difference between two ranks of each observation and n is the number of observations. Energies 2021, 14, 7271 8 of 16 8 of 16 Table 4. 8-h average values of the ﬁeld data measurements. Air Vel. Current CO2 RH T PM2.5 PM10 [m/s] [A] [ppm] [%] [◦C] [mg/m3] [mg/m3] 0.15 2.07 855.2 34.9 21.8 0.063 0.084 0.14 1.27 877.7 31.5 21.3 0.062 0.088 0.13 1.30 966.5 29.8 21.3 0.061 0.091 0.13 1.15 876.4 30.7 21.5 0.061 0.081 0.13 1.24 847.1 30.3 21.6 0.061 0.083 0.13 1.09 869.3 31.5 21.6 0.061 0.086 0.13 1.25 907.4 31.1 21.6 0.064 0.079 0.15 1.72 911.9 29.5 21.6 0.080 0.086 0.18 1.31 999.2 26.9 21.7 0.077 0.083 0.18 1.17 943.3 25.0 21.8 0.076 0.082 0.21 1.57 944.2 24.6 21.8 0.076 0.082 0.20 1.68 957.9 25.6 21.8 0.075 0.080 0.24 3.31 890.7 26.2 21.9 0.077 0.081 V CT CO2 T RH PM 2.5 PM 10 V CT CO2 T RH PM 2.5 PM 10 1 0.083 0.044 0.073 -0.025 0.067 0.059 0.083 1 0.039 0.032 -0.13 0.14 0.15 0.044 0.039 1 0.024 -0.55 0.38 0.34 0.017 0.032 0.024 1 -0.48 0.17 0.075 -0.025 -0.13 -0.55 -0.48 1 -0.6 -0.36 0.067 0.14 0.38 0.17 -0.6 1 0.68 0.059 0.15 0.34 0.075 -0.36 0.68 1 1.00 0.75 0.50 0.25 0.00 0.25 0.50 0.75 1.00 Figure 4. Correlation analysis using Pearson Correlation. V CT CO2 T RH PM 2.5 PM 10 V CT CO2 T RH PM 2.5 PM 10 1 0.38 0.35 0.37 -0.32 -0.057 -0.079 0.38 1 -0.003 0.32 -0.46 0.14 0.13 0.35 -0.003 1 0.02 -0.45 0.18 0.14 0.37 0.32 0.02 1 -0.68 0.13 -0.073 -0.32 -0.46 -0.45 -0.68 1 -0.008 0.17 -0.057 0.14 0.18 0.13 -0.008 1 0.83 -0.079 0.13 0.14 -0.073 0.17 0.83 1 1.00 0.75 0.50 0.25 0.00 0.25 0.50 0.75 1.00 Figure 5. Correlation analysis using Spearman’s Rho Correlation. Table 4. 8-h average values of the ﬁeld data measurements. Time Air Vel. 3.3. Correlation Study between Fine Dust and Other Parameters V CT CO2 T RH PM 2.5 PM 10 V CT CO2 T RH PM 2.5 PM 10 1 0.38 0.35 0.37 -0.32 -0.057 -0.079 0.38 1 -0.003 0.32 -0.46 0.14 0.13 0.35 -0.003 1 0.02 -0.45 0.18 0.14 0.37 0.32 0.02 1 -0.68 0.13 -0.073 -0.32 -0.46 -0.45 -0.68 1 -0.008 0.17 -0.057 0.14 0.18 0.13 -0.008 1 0.83 -0.079 0.13 0.14 -0.073 0.17 0.83 1 1.00 0.75 0.50 0.25 0.00 0.25 0.50 0.75 1.00 Figure 5. Correlation analysis using Spearman’s Rho Correlation. V CT CO2 T RH PM 2.5 PM 10 V CT CO2 T RH PM 2.5 PM 10 1 0.38 0.35 0.37 -0.32 -0.057 -0.079 0.38 1 -0.003 0.32 -0.46 0.14 0.13 0.35 -0.003 1 0.02 -0.45 0.18 0.14 0.37 0.32 0.02 1 -0.68 0.13 -0.073 -0.32 -0.46 -0.45 -0.68 1 -0.008 0.17 -0.057 0.14 0.18 0.13 -0.008 1 0.83 -0.079 0.13 0.14 -0.073 0.17 0.83 1 1.00 0.75 0.50 0.25 0.00 0.25 0.50 0.75 1.00 Figure 5. Correlation analysis using Spearman’s Rho Correlation. Figure 5. Correlation analysis using Spearman’s Rho Correlation. Energies 2021, 14, 7271 9 of 16 9 of 16 Table 5 shows exemplary di as the difference between ranks of each observation of PM2.5 and relative humidity (RH), where n deﬁned the total number of observations, which means total number of samples, in the case of this study approximately 350,000. Table 5. Spearman’s Rho Correlation Formula on PM2.5 and Relative Humidity (RH). Observations PM2.5 Rank RH Rank d 1 10 7 3 2 124 80 44 ... ... ... ... Table 5. Spearman’s Rho Correlation Formula on PM2.5 and Relative Humidity (RH). Each variable has correlation a coefﬁcient of 1 to itself. Because both substances are categorized as ﬁne dust, PM10 and PM2.5 have a strong positive linear correlation to each other. Both substances have negative intermediate correlations with relative humidity (RH) and positive intermediate correlations with CO2. The electricity current (CT) shows a weak positive correlation to both PM10 and PM2.5. These variables were chosen, as those are the potential quantities which might affect the amount of PM2.5 and PM10 in the welding area. 3.4. Data Model Spearman’s rho coefﬁcient test is used to show the monotonic behavior of the variables to each other. PM10 and PM2.5 show a strong increasing monotonic relation with each other and also have sufﬁcient monotonic relation with the current (CT), CO2 and relative humidity (RH). This result also shows that PM10 and PM2.5 have decreasing monotonic behavior with Air Velocity (V). The combination of Pearson and Spearman’s coefﬁcients show that the Current (CT), CO2, Relative Humidity (RH) and Air Velocity (V) are more suitable to serve as input variables for the model. The aim is to exclude PM2.5 as an input and make a model which would give better prediction for PM10. In order to select the best ﬁt model, different machine learning models were designed, trained and tested using the available data. These includes linear regression and feedfor- ward neural networks for regression and classiﬁcation (Probabilistic Neural Network), which contain the whole interval of possible values separated in high number of intervals. This is a particular case of application of classiﬁcation. Aside from that, the design of the models use supervised learning, which means the output target variable is already available for the training. Different features and respectively variables were considered for the models and the performances were compared. Current (CT), CO2, Relative Humidity (RH) and Air Velocity (V) Features are the variables chosen as inputs for the model. The splitting of the measured data is set at random, accordingly, the ratio of the training and test data sets was set manually. Figure 6 shows the schematic of the Feedforward Neural Network (FFNN) Model for regression using Current (CT), CO2, Relative Humidity (RH) and Air Velocity (V) as Inputs, feeding 2 Hidden Layers with each 10 Nodes and PM10 as output. 10 of 16 Energies 2021, 14, 7271 Input Layer ∈ℝ⁴ Hidden Layer ∈ℝ¹⁰ Hidden Layer ∈ℝ¹⁰ Output Layer ∈ℝ¹ CT CO2 RH V PM10 Figure 6. Feedforward Neural Network (FFNN) schematic with 4 inputs, 2 hidden layers @10 nodes. Hidden Layer ∈ℝ¹⁰ Hidden Layer ∈ℝ¹⁰ Hidden Layer ∈ℝ¹⁰ Input Layer ∈ℝ⁴ Output Layer ∈ℝ¹ Figure 6. Feedforward Neural Network (FFNN) schematic with 4 inputs, 2 hidden layers @10 nodes. Figure 6. 3.4. Data Model Feedforward Neural Network (FFNN) schematic with 4 inputs, 2 hidden layers @10 no The schematic of the Feedforward Neural Network Model for classiﬁcation (Proba- bilistic Neural Network) using the same inputs, two pattern (hidden) layers and PM10 as output is shown in Figure 7. The number of nodes in the pattern layers are equal to the number of training points in the dataset. In the ﬁrst pattern layer, the training data set uses 40% of the data, while the second layer utilizes 28% of the data set. Input Layer ∈ℝ⁴ Pattern Layer Patern Layer Summation Layer Output Layer ∈ℝ¹ CT CO2 RH V Σ Σ PM10 40 % of data 28 % of data Figure 7. Probabilistic Neural Network (PNN) schematic with 4 inputs and 2 pattern layers. Input Layer ∈ℝ⁴ Pattern Layer Patern Layer Figure 7. Probabilistic Neural Network (PNN) schematic with 4 inputs and 2 pattern layers. Last, Long Short-Term Memory (LSTM), a deep learning algorithm, is used in this study, the data from the aforementioned parameters are used as input variables, 1 dense layer with 25 LSTM Blocks and the particulate matter PM10 as output variables. Figure 8 shows the schematic of the LSTM model for this study. 11 of 16 Energies 2021, 14, 7271 Input Layer ∈ℝ⁴ LSTM Layer ∈ℝ25 Dense Layer ∈ℝ25 Output Layer ∈ℝ¹ n =25 CT CO2 RH V PM10 n =25 Figure 8. Long Short-Term Memory (LSTM) schematic with 4 inputs and 1 dense layer. PM10 Output Layer ∈ℝ¹ Figure 8. Long Short-Term Memory (LSTM) schematic with 4 inputs and 1 dense layer. The models were trained with different amount of data starting from 20% to 80%. The models did not overﬁt until 40%, as the results continue to get better with the increase in the amount of training data. Models showed little betterment with increase in training data above 40%. The computational time increases drastically with the increase in training data above 40% for LSTM and FFNN, so 40% training data was the ideal choice to compare the models. After building and training the model with the deﬁned amount of data, the rest of the measured data was used for validating (testing) the model. The model is tested with different kinds of statistical methods such as R2 criteria for best ﬁt, Minimum Absolute Error (MAE), Minimum Square Error (MSE) and Root Mean Square Error (RMSE). 3.4. Data Model Visualization as well as statistical evaluation have been carried out to compare the actual data and the predicted data. By comparing different models with statistical as well as visual criteria, the best model for the application is chosen. 3.4.1. Model Result of Linear Regression, Feedforward Neural Network and Probabilistic Neural Network From the correlation study, it was sufﬁcient to say that a linear relationship exists up to some extent between the input variables and desired output. Building a linear regression model using the chosen variables from the correlation study seems to be a logical way to go. After training the parameters by using the training data, which consist of 40% of the data set, the model was tested using the rest of the data set. g Figure 9 shows a direct comparison of the prediction from the linear regression model and the test data from the measured data set. An evident discrepancy exist can be seen from the comparison. Visually, Figure 10 shows a more satisfying result, which means the feedforward neural network model is a better ﬁt of a model compared to linear regression. However, both results are not sufﬁcient to use because of their inaccuracy. Both models use regression methods, in which the output layer consists of only one neuron and it gives a continuous output value. Supervised learning classiﬁcation methods also serve as alternative; they are used in building the model using probabilistic Neural Network. The result, thus the comparison to the test data is shown in the Figure 11, which displays a more satisfying result in compare to Figures 9 and 10. However, merely relying on visual analysis, a scientiﬁc analysis would not be valid. 12 of 16 12 of 16 Energies 2021, 14, 7271 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Normalized PM10 26.Nov.2018 27.Nov.2018 28.Nov.2018 29.Nov.2018 30.Nov.2018 Linear Regression Actual Predicted Figure 9. Comparison between the Linear Regression prediction and the test data. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Normalized PM10 26.Nov.2018 27.Nov.2018 28.Nov.2018 29.Nov.2018 30.Nov.2018 Feed Forward Neural Network Actual Predicted Figure 10. Comparison between the Feedforward Neural Network prediction (FFNN) and the test data. 3.4. Data Model Comparison between the Long short-term memory (LSTM) prediction and the test data. Figure 12. Comparison between the Long short-term memory (LSTM) prediction and the test data. 3.4. Data Model 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Normalized PM10 Probabilistic Neural Network 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Normalized PM10 26.Nov.2018 27.Nov.2018 28.Nov.2018 29.Nov.2018 30.Nov.2018 Linear Regression Actual Predicted 0 0.05 0.1 0.15 0.2 Nor 26.Nov.2018 27.Nov.2018 28.Nov.2018 29.Nov.2018 30.Nov.2018 Actual Predicted Figure 9. Comparison between the Linear Regression prediction and the test data. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Normalized PM10 26.Nov.2018 27.Nov.2018 28.Nov.2018 29.Nov.2018 30.Nov.2018 Feed Forward Neural Network Actual Predicted Figure 10. Comparison between the Feedforward Neural Network prediction (FFNN) and the test data. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Normalized PM10 26.Nov.2018 27.Nov.2018 28.Nov.2018 29.Nov.2018 30.Nov.2018 Probabilistic Neural Network Actual Predicted Figure 11. Comparison between the Probabilistic Neural Network (PNN) prediction and the test data. Figure 9. Comparison between the Linear Regression prediction and the test data. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Normalized PM10 26.Nov.2018 27.Nov.2018 28.Nov.2018 29.Nov.2018 30.Nov.2018 Feed Forward Neural Network Actual Predicted Feed Forward Neural Network Figure 10. Comparison between the Feedforward Neural Network prediction (FFNN) and the test data. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Normalized PM10 26.Nov.2018 27.Nov.2018 28.Nov.2018 29.Nov.2018 30.Nov.2018 Probabilistic Neural Network Actual Predicted Figure 11. Comparison between the Probabilistic Neural Network (PNN) prediction and the test data. Figure 11. Comparison between the Probabilistic Neural Network (PNN) prediction and the test data. Energies 2021, 14, 7271 13 of 16 For example, the results shown in Figure 12, which show the visual representation of LSTM model output with the actual measured data, cannot show the difference or beneﬁt of LSTM model compared to PNN model in Figure 11. Therefore, one needs to compare the results using statistical methods as well. For example, the results shown in Figure 12, which show the visual representation of LSTM model output with the actual measured data, cannot show the difference or beneﬁt of LSTM model compared to PNN model in Figure 11. Therefore, one needs to compare the results using statistical methods as well. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Normalized PM10 26.Nov.2018 27.Nov.2018 28.Nov.2018 29.Nov.2018 30.Nov.2018 LSTM Actual Predicted 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Normalized PM10 26.Nov.2018 27.Nov.2018 28.Nov.2018 29.Nov.2018 30.Nov.2018 LSTM Actual Predicted Figure 12. 3.4.2. Statistical Analysis of the Models In order to do statistical checks, the models are tested using different kinds of methods, such as R2-criteria for best ﬁt, Minimum Absolute Error (MAE), Minimum Square Error (MSE) and Root Mean Square Error (RMSE) and using their results the best ﬁt model can be chosen. Table 6 shows that the LSTM Model has the lowest error in all error checks and the best squared correlation. In contrast to the linear regression model that shows a poor ﬁt on the R2-Corellation and the lowest performance on the error tests, the LSTM Model shows good results, which is well matched with the visual comparison of Figures 4 and 5, and thus supports the proof that LSTM is capable to be the best ﬁt model for this study. Table 6. Results of the statistical checks of the models with Current (CT), CO2, Relative Humidity (RH) and Air Velocity (V) as inputs. Table 6. Results of the statistical checks of the models with Current (CT), CO2, Relative Humidity (RH) and Air Velocity (V) as inputs. Model R2 MAE MSE RMSE Linear Regression 0.008 0.03563 0.00377 0.05689 FNN 0.211 0.02336 0.00301 0.05126 PNN 0.453 0.01901 0.00267 0.04956 LSTM 0.821 0.01020 0.00122 0.0257 In order to get a better result, one could use a cumulative training method. A cumula- tive learning method uses aggregation of data as it grows with time. Consequently, it uses knowledge acquired on prior training to improve learning performance on subsequent training. On the contrary, a static training method uses discrete data, which means for each new training time period the algorithms are reset and fed with new data. Fixed training data are applied to a machine learning algorithm, and it does not use any knowledge from prior training. Therefore, cumulative method reuses learned knowledge to constrain new training, whereas static method depends entirely upon new training data as external inputs [27]. Thor [28] gave further description and deﬁnition of cumulative learning in the context of machine learning in detail. Figure 13 shows the comparison of static and cumulative training methods and the statistic result from this study according to R2 criteria. In the static training, the algorithm was trained for ﬁve days each month with new data. On the other hand, cumulative Energies 2021, 14, 7271 14 of 16 14 of 16 training uses aggregated data of the current and previous months. 3.4.2. Statistical Analysis of the Models The cumulative training method showed a gradually increasing R2 score reaching 0.81 R2, while the static method does not show a constant behavior and staying under 0.78 R2. This shows, that in training the LSTM Model, the cumulative training method is preferred. Feb March April May June Training period (5 days in each month) 0.74 0.75 0.76 0.77 0.78 0.79 0.80 0.81 0.82 R2 score Cumulative Static Figure 13. Long Short-Term Memory (LSTM) model analysis. Figure 13. Long Short-Term Memory (LSTM) model analysis. 4. Conclusions and Future Scope Because PM10 and PM2.5 Energies 2021, 14, 7271 15 of 16 15 of 16 show a strong increasing correlation with each other, PM2.5 could be excluded as an output variable. This will save a generous amount of computing time and cost. As supervised machine learning and deep learning algorithms are used in this study, the data from the aforementioned parameters are used as input variables and the particulate matter as output variables. The data were then split for training and testing of the models. The models are then tested statistically using different methods, such as R2 criteria for best ﬁt, Minimum Absolute Error (MAE), Minimum Square Error (MSE) and Root Mean Square Error (RMSE). This shows that LSTM dominated the test with 0.821 discrepancy on R2, 0.01020 on MAE, 0.00122 on MSE and 0.0257 on RMSE. As a comparison, the Linear Regression Model showed a poor result with 0.008 discrepancy on R2, 0.03563 on MAE, 0.00377 on MSE and 0.05689 on RMSE. The results were also examined using graphical method. The LSTM or Long Short-Term Memory showed the best result and therefore is best suited for the HVAC control concept. In the future it is possible to enhance the HVAC control using the LSTM model. It could be possible to install several micro controllers in the factory hall, that send the acquired data that is not expensive to measure data, such as relative humidity, CO2, electricity current and air velocity. Data can then be used to train the LSTM model continuously. The output of the models can send computed ﬁne dust values in real time to the HVAC system, which would be used as control parameter, besides typical control parameters, such as desired temperature, relative humidity and CO2 especially on factory scale, the cost savings would be signiﬁcant and are very interesting for factory operators and planners. Author Contributions: Y.W.: Conceptualization, Methodology, Validation, Investigation, Data Cura- tion, Writing—Original draft, Visualization, Project administration. M.V.: Conceptualization, Method- ology, Validation, Investigation, Data Curation, Writing—Reviewing and Editing. R.K.: Methodology, Software, Visualization, Investigation, Data Curation. C.F.: Supervision. C.H.: Supervision. All authors have read and agreed to the published version of the manuscript. 4. Conclusions and Future Scope Funding: The research regarding the presented use case in this paper was funded by the Federal Ministry for Economic Affairs (BMWi) by means of the 7th Energy Research Programme of the German Federal Government under grant number 03ET1660A (3DEMO—Safe and energy efﬁcient factories through 3D emission monitoring). Institutional Review Board Statement: Not applicable. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Informed Consent Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Restrictions apply to the availability of these data. Data was obtained from Volkswagen AG and are available from the authors with the permission of Volkswagen AG. Data Availability Statement: Restrictions apply to the availability of these data. Data was obtained from Volkswagen AG and are available from the authors with the permission of Volkswagen AG. Conﬂicts of Interest: The authors declare no conﬂicts of interest. The results, opinions and conclu- sions expressed in this paper are not necessarily those of Volkswagen AG. Conﬂicts of Interest: The authors declare no conﬂicts of interest. The results, opinions and conclu- sions expressed in this paper are not necessarily those of Volkswagen AG. 4. Conclusions and Future Scope This paper studied proof of concept for the Heating, Ventilation and Air-Conditioning (HVAC) system for a welding area. The proof of concept was modeled using four different machine learning algorithms and their performances were compared. The algorithms im- plemented were linear regression, feedforward neural network for regression, probabilistic neural network (Bayesian neural networks) for classiﬁcation and Long Short-Term Memory (deep learning algorithm). Long Short-Term Memory showed the best result and potential for the control system. y A complete cyber-physical HVAC system in a body shop use case for the substitution of cost-intensive ﬁne dust sensors was presented. The CPPS is then applied and validated in a real-world environment in one of the welding areas of Volkswagen’s production hall in Wrzesnia, Poland. A setup was conﬁgured and established based on the regulatory threshold to collect experimental data during production hours, in order to have a solid validation ground. The data were collected in 2 phases (in summer and winter), each extended through 5 production days. The concerned data set consist of air temperature,; relative humidity; air velocity; electric current used for welding; carbon dioxide; inhalable coarse particles; PM10, which are dust particles with a diameter of 10 micrometers (10 µm) or less and ﬁne particles; and PM2.5, which has a diameter of 2.5 µm or less. These were measured in the welding area during manual welding corrections, welding of ﬂoor elements (galvanized steel) with the welding torch (MAG method), sanding and cleaning of welded seam. As a double-prooﬁng measure, 8-hour gravimetric measurements were taken, commissioned by the Volkswagen work safety department, in order to ensure that the reductions of air exchange rate kept the exposition thresholds under the regulated limits. Gravimetric measurement (respectively analysis) describes a method to quantify a substance or chemical constituent in a mixture based on its mass. In our case, the relevant substance is ﬁne dust and the mixture is air. The gravimetric measurements were executed according to PN EN 689:2018. It showed that the reduction of air exchange rate does not have a linear correlation with the change in the amount of ﬁne dust in the working area. The correlation study according to Spearman’s rho and Pearson showed that only four variables have direct correlation with the outputs (PM10 and PM2.5): relative humidity, CO2 Concentration, electric current and air velocity in the area. References 2020. Available online: https://www.foedisch.de/app/download/12654990627/de.FDS%2015_Produktinformation. pdf?t=1596630054 (accessed on 29 October 2021). p 11. Födisch Umweltmesstechnik, A.G. Feinstaubsensor Ofﬁzielles Angebot; Zwenkauer Straße 159, 04420 Markranstädt, Deutschland, 2018. Unpublished internal company document. 12. Al-Fuqaha, A.; Guizani, M.; Mohammadi, M.; Aledhari, M.; Ayyash, M. Internet of Things: A Survey on Enabling Technologies, Protocols and Applications. IEEE Commun. Surv. Tutorials 2015, 17, 2347–2376. [CrossRef] 13. Bajrami, X.; Murturi, I. An efﬁcient approach to monitoring environmental conditions using a wireless sensor network and NodeMCU. Elektrotech. Inftech. 2018, 135, 294–301. [CrossRef] 14. Li, L.; Zhang, R.; Sun, J.; He, Q.; Kong, L.; Liu, X. Monitoring and prediction of dust concentration in deep-learning algorithm. J. Environ. Health Sci. Eng. 2021, 19, 401–414. doi:10.1007/s40201-021-00613-0 14. Li, L.; Zhang, R.; Sun, J.; He, Q.; Kong, L.; Liu, X. Monitoring and prediction of dust concentration in an open-pit mine using a deep-learning algorithm. J. Environ. Health Sci. Eng. 2021, 19, 401–414. doi:10.1007/s40201-021-00613-0. [CrossRef] [PubMed] 15 Umweltbundesamt Hintergrundpapier zum Thema Feinstaubbelastung in Deutschland Wörlitzer Platz 1 06844 Dessau Roßlau 15. Umweltbundesamt, Hintergrundpapier zum Thema Feinstaubbelastung in Deutschland, Wörlitzer Platz 1, 06844 Dessau-Roßlau. 2009. Available online: https://www.umweltbundesamt.de/sites/default/files/medien/publikation/long/3565.pdf (accessed on 29 October 2021). 16. European Environment Agency, Sources of Air Pollution in Europe, Published 13 Jun 2014; Last modiﬁed 23 Nov 2020. Available online: https://www.eea.europa.eu/ds_resolveuid/INF-13-en (accessed on 29 October 2021). 17. Madanchi, N.; Leiden, A.; Winter, M.; Asbach, C.; Lindermann, J.; Herrmann, C.; Thiede, S. Cutting ﬂuid emissions in grinding processes: Inﬂuence of process parameters on particle size and mass concentration. Int. J. Adv. Manuf. Technol. 2019, 101, 773–783. [CrossRef] 18. European Environment Agency, 6.4. Suspended particulates (TSP/SPM), Last Modiﬁed 23 Nov 2020. Available online: https: //www.eea.europa.eu/publications/2-9167-057-X/page021.html (accessed on 29 October 2021). 19. Xing, Y.F.; Xu, Y.H.; Shi, M.H.; Lian, Y.X. The impact of PM2.5 on the human respiratory system. J. Tho [CrossRef] [PubMed] u, Y.H.; Shi, M.H.; Lian, Y.X. The impact of PM2.5 on the human respiratory system. J. Thorac. Dis. 2016, 8, E ubMed] 20. Scott, R.M. Chemical Hazards in the Workplace; CRC Press: Boca Raton, FL, USA, 2020; ISBN 9781000716344. Available online: https://books.google.de/books?id=CkQPEAAAQBAJ (accessed on 29 October 2021). p g g 21. Riccelli, M.G.; Goldoni, M.; Poli, D.; Mozzoni, P.; Cavallo, D.; Corradi, M. Welding Fumes, a Risk Factor for Lung Diseases. Int. J. Environ. Res. Public Health 2020, 17, 2552. [CrossRef] [PubMed] 22. Spiegel-Ciobanu V.E. Aktuelles aus Gesetzgebung und Regelwerk zu Schadstoffen beim Schweißen. References 1. Hyoung, S.K.; Ju, Y.L.; Sang, S.C.; Hyun, K.; Jun, H.P.; Ji, Y.S.; Bo, H.K.; Sang, D.N. Smart manufacturing: Past research, present ﬁndings, and future directions. Int. J. Precis. Eng. Manuf.-Green Technol. 2016, 3, 111–128. 2. Thiede, S. Smart Environmental Sustainability of Cyber Physical Production Systems. Procedia CIRP 2018, 69, 644–649; ISSN 22128271. [CrossRef] 3. IEA’s World Energy Balances and Statistics. Total Final Consumption by Source (TFC), World, Data Retrieved from IEA World Energy Balances and Statistics. 2020. Available online: https://www.iea.org/data-and-statistics?country=WORLD&fuel= Energy%20consumption&indicator=TFCbySource (accessed on 29 October 2021). 3. IEA’s World Energy Balances and Statistics. Total Final Consumption by Source (TFC), World, Data Retrieved from IEA World Energy Balances and Statistics. 2020. Available online: https://www.iea.org/data-and-statistics?country=WORLD&fuel= Energy%20consumption&indicator=TFCbySource (accessed on 29 October 2021). gy p y 4. IEA’s World Energy Balances and Statistics. Share of Total Final Consumption (TFC) by Sector, World, Data Retrieved from IEA World Energy Balances and Statistics. 2020. Available online: https://www.iea.org/data-and-statistics?country=WORLD&fuel= Energy%20consumption&indicator=TFCbySector (accessed on 29 October 2021). 4. IEA’s World Energy Balances and Statistics. Share of Total Final Consumption (TFC) by Sector, World, Data Retrieved from IEA World Energy Balances and Statistics. 2020. Available online: https://www.iea.org/data-and-statistics?country=WORLD&fuel= Energy%20consumption&indicator=TFCbySector (accessed on 29 October 2021). 5. IEA’s World Energy Balances and Statistics. Total Final Consumption (TFC) by Sector, World, Data Retrieved from IEA World Energy Balances and Statistics. 2020. Available online: https://www.iea.org/data-and-statistics?country=WORLD&fuel= Energy%20consumption&indicator=TFCShareBySector (accessed on 29 October 2021). 5. IEA’s World Energy Balances and Statistics. Total Final Consumption (TFC) by Sector, World, Data Retrieved from IEA World Energy Balances and Statistics. 2020. Available online: https://www.iea.org/data-and-statistics?country=WORLD&fuel= Energy%20consumption&indicator=TFCShareBySector (accessed on 29 October 2021). Energies 2021, 14, 7271 16 of 16 7. Wang, S.K. Handbook of Air Conditioning and Refrigeration; Engineering Handbook; McGraw-Hill Education: New York, NY, USA; Vulkan-Verlag GmbH: Essen, Germany; DIV Deutscher Industrieverlag GmbH: München, Germany, 2015; ISBN 97800706816752001. Available online: https://books.google.de/books?id=0tVSAAAAMAAJ (accessed on 29 October 2021). p g g J ( ) 8. Recknagel, H. Heizung + Klimatechnik mit Neuem Hauptkapitel “Energiekonzepte”; Vulkan-Verlag GmbH, DIV Deutscher Indus- trieverlag GmbH: Essen, München, 2011. Available online: http://slubdd.de/katalog?TN_libero_mab216556880 (accessed on 29 October 2021). ) 9. Bundesanstalt für Arbeitsschutz und Arbeitsmedizin (BAuA). TRGS 900, Technische Regeln für Gefahrstoffe—Arbeitsplatzgrenzwerte in Bundesarbeitsblatt (BArBl.) Heft 1/2006, S. 41–55. Bundesministerium für Arbeit und Soziales (BMAS). Januar 2006, Bundesanstalt für Arbeitsschutz und Arbeitsmedizin (BAuA), Friedrich-Henkel-Weg 1–25. D-44149 Dortmund. Available online: https://www.baua.de/DE/Angebote/Rechtstexte-und-Technische-Regeln/Regelwerk/TRGS/pdf/TRGS-900.pdf?__ blob=publicationFile (accessed on 29 October 2021). 10. Födisch Umweltmesstechnik, A.G. Feinstaubsensor FDS 15 Produktinformation, Zwenkauer Straße 159, 04420 Markranstädt, Deutschland. References In Schneiden und Schweißen, Ausgabe 04/2016; BGHM—Berufsgenossenschaft Holz und Metal: Mainz, Germany, 2016. Available online: https://www.esta. com/ﬁleadmin/mount/ESTA_DE/Download/Wissenswertes/BGHM-Regelwerk-Schadstoffe-Schweissen.pdf (accessed on 29 October 2021). 23. Deutsche Institut für Normung e.V. (Hrsg.). DIN EN 689, Exposition am Arbeitsplatz—Messung der Exposition Durch Einatmung Chemischer Arbeitsstoffe—Strategie zur Überprüfung der Einhaltung von Arbeitsplatzgrenzwerten; Workplace Exposure—Measurement of Exposure by Inhalation to Chemical Agents—Strategy for Testing Compliance with Occupational Exposure Limit Values; Beuth Verlag GmbH: Berlin, Germany, 2020. y 24. Polish Committee for Standardization. PN-EN 689, Exposure at Work Stations—Measurements of Inhalation Exposure to Chemical Agents—Strategy for Testing Compliance with Limit Values; Polish Committee for Standardization:Warsaw, Poland; 2019. 25. Bishop, C.M. Pattern Recognition and Machine Learning, Information Science and Statistics; Softcover Published in 2016; Springer: New York, NY, USA, 2006. Available online: https://cds.cern.ch/record/998831 (accessed on 29 October 2021). 26. Spearman Rank Correlation Coefﬁcient. In The Concise Encyclopedia of Statistics; Springer: New York, NY, USA, 2008; doi:10.1007/978-0-387-32833-1_379. [CrossRef] 27. Michelucci, P.; Oblinger, D.; Sammut, C.; Webb, G.I. (Eds.) Cumulative Learning, Encyclopedia of Machine Learning and Data Mining; Springer: Boston, MA, USA, 2017; pp. 306–314. ISBN 978-1-4899-7687-1. [CrossRef] 28. Thórisson, K.; Bieger, J.; Li, X.; Wang, P. Cumulative Learning. In International Conferenc Cham, Switzerland, 2019; Volume 7, pp. 198–208. ISBN 978-3-030-27004-9. [CrossRef] 28. Thórisson, K.; Bieger, J.; Li, X.; Wang, P. Cumulative Learning. In International Conference on Artiﬁcial General Intelligence; Springer: Cham, Switzerland, 2019; Volume 7, pp. 198–208. ISBN 978-3-030-27004-9. [CrossRef]
https://openalex.org/W4226200309	http://ejournal.unp.ac.id/students/index.php/pek/article/download/12353/4862	Indonesian	null	Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise	Jurnal Ecogen	2,021	cc-by	3,160	available at http://ejournal.unp.ac.id/students/index.php/pek/index available at http://ejournal.unp.ac.id/students/index.php/pek/index ONLINE ISSN 2654-8429 Vol. 4 No. 4, 2021 Page 551-558 Vol. 4 No. 4, 2021 Page 551-558 Published by Jurusan Pendidikan Ekonomi Fakultas Ekonomi Universitas Negeri Padang, Indonesia Published by Jurusan Pendidikan Ekonomi Fakultas Ekonomi Universitas Negeri Padang, Indonesia Keywords : Experiential Marketing and Visitor Satisfaction Keywords : Experiential Marketing and Visitor Satisfaction This is an open access article distributed under the Creative Commons 4.0 Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ©20121 by author. 551 Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise 51 Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise Tri Sugiarti Ramadhan 1, Nanik Wahyuningtiyas 2 Jurusan Manajemen, Fakultas Ekonomi & Bisnis, Universitas Islam Malang trisugiarti@unisma.ac.id Tri Sugiarti Ramadhan 1, Nanik Wahyuningtiyas 2 Jurusan Manajemen, Fakultas Ekonomi & Bisnis, Universitas Islam Malang trisugiarti@unisma.ac.id Abstract : The purpose of this study was to determine the effect of the Dimension Experiential Marketing (Y) on visitor satisfaction at Malang Night Paradise. The population taken in this study were visitors who came to visit Malang Night Paradise Tourism. This research includes visitor satisfaction variable as dependent variable and dimension experiential marketing variable as independent variable consisting of Sense(X1), Feel(X2), Think(X3), Act(X4) and Relate(X5). in this study using accidental sampling technique as many as 96 respondents. Simultaneously, the Experiential Marketing variable has a significant effect on visitor satisfaction of Malang Night Paradise. Then partially there are two variables that do not have a significant effect, namely the taste variable and the related variable. Meanwhile, the other three variables, namely feel (X2), think (X3), and act (X4) have a significant effect on the visitor satisfaction variable (Y) at Malang Night Paradise. With an R2 value of 63.7%, while 36.3% is influenced by other factors that can affect visitor satisfaction. Keywords : Experiential Marketing and Visitor Satisfaction Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise Tri Sugiarti Ramadhan 1, Nanik Wahyuningtiyas 2 Jurusan Manajemen, Fakultas Ekonomi & Bisnis, Universitas Islam Malang trisugiarti@unisma.ac.id PENDAHULUAN Negara Indonesia merupakan negara yang sangat indah dan memiliki Sumber Daya Alam (SDA) yang kaya dan sangat melimpah. Melimpahnya SDA ini merupakan potensi yang dapat dilihat untuk pembangunan kepariwisataan menjadi salah satu sektor unggulan untuk meningkatkan penerimaan devisa. Indonesia memiliki visi kepariwisataan yang tertuang dalam Depbudpar (2000) adalah “Pariwisata Menumbuh kembangkan Kesejahteraan dan Perdamaian” dari Visi ini didapatkan Misi yang dimaksudkan agar pengembangan dan pendayagunaan berbagai potensi di industri pariwisata dapat memberikan sumbangsih yang akan berpengaruh bagi Negara Indonesia. Parwiwisata menjadi sektor ekonomi yang penting di Indonesia. pariwisata menjadi penyumbang devisa negara terbesar ke empat setelah industri agrikultural, manufaktur otomotif dan pertambangan. Kekayaan alam dan budaya merupakan komponen penting dalam pariwisata di Indonesia. Indonesia memiliki ribuan pulau yang terbentang dari Sabang sampai Merauke serta memiliki ciri khas dan budaya yang berbeda di setiap daerahnya. Hal ini menjadi keunggulan 551 Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise Indonesia untuk menarik minat wisatawan berkunjung ke Indonesia. Jumlah pengunjung mancanegara yang datang ke Indonesia setiap tahunnya terus mengalami peningkatan yakni dari 10.406.759 di tahun 2015 menjadi 12.023.971 di tahun 2016 atau tumbuh sebesar 15,54% (sumber: Ditjen Imigrasi dan BPS (diolah kembali oleh Asdep Litbangjakpar Kemenpar). Pertumbuhan ekonomi disektor pariwisata seperti sekarang menjadi kesempatan yang dilihat bagi para pelaku kegiatan bisnis. Sekarang banyak tempat wisata baru yang bermunculan salah satunya adalah wilayah Kota Malang. Kota Malang saat mulai menjadi pusat perhatian bagi wisatawan domestik sebagai salah satu tujuan untuk berwisata. Besarnya minat wisatawan untuk datang berlibur di Kota Malang mengakibatkan persaingan bisnis yang kuat antar perusahaan penyedia pelayanan wisata. Dal hal ini pelaku bisnis harus mampu menerapkan konsep marketing yang dapat memberikan pengalaman yang unik dan postif kepada pengunjung atau yang biasa disebut dengan Experiential Marketing. Experiential Marketing merupakan konsep pemasaran yang berfokus pada emosional dari konsumen, pada konsep ini pemasar membuat konsep pemasaran dengan menciptakan pengalaman yang bersifat positif dan berkesan menarik bagi konsumen. Menurut Schmitt dalam Kartajaya (2006:228), “experiential marketing dapat dihadirkan melalui lima unsur yaitu panca indera (sense), perasaan (feel), cara berpikir (think), kebiasaan (act) dan pertalian atau relasi (relate)”. Perusahaan sekarang ini dituntut untuk bisa menghadirkan pengalaman yang menarik melalui dimensi Experiential Marketing (sense, feel, think, act, dan relate) tersebut dengan maksimal. ini akan berpengaruh pada kondisi emosional/persepsi konsumen terhadap produk atu jasa yang ditawarkan oleh perusahaan. Schiffman dalam Wicaksono (2005:2) menyatakan bahwa “pelanggan melakukan keputusan berdasarkan persepsi yang dimiliki ketimbang kenyataan”. Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise PENDAHULUAN Hal ini menunjukan bahwa citra kualitas yang baik tidak berdasarkan prsepsi perusahaan, melainkan berdasarkan persepsi pelanggan (Tjiptono, 2002:61). Evaluasi pelanggan terhadap jasa merupakan perbandingan antara harapan pelanggan dengan tafsiran atau persepsi mengenai jasa yang senyatanya diterima”. Perbandingan tersebut akan menimbulkan perasaan puas dan tidak puas pelanggan terhadap produk atau jasa yang yang ditawarkan oleh perusahaan, sehingga dapat disimpulkan bahwa presepsi memiliki kaitan yang cukup erat terhadap tingkat kepuasan konsumen. (Schiffman dalam Wicaksono (2005:3). Wisata di kota Malang memang sangat populer, selain karena lokasinya yang strategis sebagai kawasan wisata dengan suhu udara yang sejuk. Objek wisata Malang Night Paradise merupakan dan tempat hiburan keluarga dikota Malang yang menawarkan 2 taman hiburan utama yang berisi taman lampion dan taman dinosaurus. tempat wisata ini terletlak disatu lokasi dengan Hawai Water Park yaitu berada di JL. Graha Kencana Raya Karanglo No.66, Karanglo Balearjosari, kec. Blimbing, Malang. Jawa Timur. Malang Night Pradise baru dibuka pada tanggal 19 juni 2017. Meskipun tergolong baru, tetapi pengunjung yang mendatangi wisata ini terbilang lumayan banyak Keberadaan Experiential Marketing sangat bermanfaat dalam menciptakan Visitor satisfaction karena perusahaan bisa lebih menyentuh emosi dari setiap pelanggan atas jasa yang sudah dirasakan. ini sejalan dengan penelitian yang dilakukan oleh Jannah (2014), dalam penelitiannya berjudul “Pengaruh Strategi Experiential Marketing 552 Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise Terhadap Kepuaan pengunjung Museum Sepuluh Nopember Surabaya”, hasil penelitian menunjukkan variabel sense, feel, think, act, dan relate secara simultan berpengaruh terhadap kepuasan pengunjung, kemudian secara parsial hanya variabel sense dan relate berpengaruh signifikan terhadap kepuasan pengunjung. Dimensi Experiential marketing merupakan suatu konsep pemasaran yang berfokus pada penciptaan pengalaman secara emosional kepada konsumen dengan tujuan untuk membentuk rasa loyal konsumen terhadap produk/jasa yang ditawarkan oleh perusahaan. dengan memberikan pengalaman yang unik dan postif terhadap konsumen perusahaan berharap dapat memenuhi harapan konsumen atau bahkan melebihinya, sehingga tercipta Visitor satisfaction. Apabila penerapan Experiential Marketing ini tidak dilakukan, maka akan kesulitan untuk mengukur Visitor satisfaction yang berkunjung pada tempat wisata Malang Night Paradise. Tujuan penelitian ini adalah untuk melihat pengaruh sense, feel, think, act dan relate terhadap Visitor Satisfation Malang Night Paradise secara simultan dan parsial. 553 Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise METODE PENELITIAN Jenis penelitian yang digunakan dalam penelitian ini adalah explanatory. Menurut Singarimbun dan Effendi (2008: 66) explanatory adalah penelitian yang menjabarkan hubungan kausal antara variabel-variabel penelitian melalui pengujian hipotesis yang telah dirumuskan sebelumny. Dalam penelitian Ini Dimensi Experiential Marketing Sebagai variabel bebas (X) dan Kepuasan Pengunjung sebagai veriabel terikat (Y). Populasi dalam penelitian ini adalah pengunjung yang telah berkunjung ke Malang Night Paradise dengan jumlah pengunjung di bulan Oktober 2020 sebanyak 2.476 orang dan Sampel dalam penelitian ini adalah sebagian dari pengunjung Malang Night Paradise. Teknik penentuan sampel ini mengacu pada pendapat Slovin. Untuk Sampel yang di gunakan sebagai responden dalam penelitian ini sebanyak 96 responden. Teknik Sampling Insidental (accidental sampling) menjadi teknik yang dipilih dalam pengambilan sampel pada penelitian ini. 554 Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise Uji Hipotesis Hasil uji regresi linier berganda digunakan untuk mengetahui besarnya pengaruh variabel experiential marketing (sense, feel, think, act, relate) terhadap variabel Kepuasan Pengunjung yang disajikan dalam tabel berikut: Hasil uji regresi linier berganda digunakan untuk mengetahui besarnya pengaruh variabel experiential marketing (sense, feel, think, act, relate) terhadap variabel Kepuasan Pengunjung yang disajikan dalam tabel berikut: Tabel 1. Uji Analisis Regresi Variabe l B Std. Error SC. Beta t hitung Sig. Ket X1 → Y , 021 .047 .029 .443 .659 Tidak Signifikan X2 → Y ,265 .042 .424 6.363 .000 Signifikan X3 → Y ,370 .042 .571 8.786 .000 Signifikan X4 → Y ,124 .041 .226 3.452 .001 Signifikan X5 → Y ,045 .040 .073 1.146 .255 Tidak Signifikan Constanta : -1.234 Fhitung : 31.559 R : 0,798 Sig F : 0,000 R2 : 0,617 Sumber: Data diolah 2021 Tabel 1. Uji Analisis Regresi Sumber: Data diolah 2021 Dari hasil analisis diperoleh hasil bahwa besarnya pengaruh Dimensi Experiential Marketing (X) terhadap Kepuasan Pengunjung (Y) ditunjukkan dengan nilai R2 sebesar 0,617 (61,7%). Dapat diketahui variabel bebas yaitu dimensi Experiential Marketing yang meliputi variable sense, feel, tink, act, relate secara bersama-sama dapat menjelaskan bahwa variabel terikat yaitu Visitor Satisfaction sebesar 63,7%, sedangkan sisanya sebesar 36,3% dijelaskan oleh variabel lain yang dalam penelitian ini tidak diteliti. Hasil uji hipotesis sebagai berikut: Hasil uji hipotesis sebagai berikut: Hasil uji hipotesis sebagai berikut: H1: Variabel sense (X1), feel (X2), think (X3), act (X4), dan relate (X5) berpengaruh secara simultan (terikat) terhadap Variabel Visitor Satisfaction (Y) Malang Night Paradise H2: Variabel sense (X1), feel (X2), think (X3), act (X4), dan relate (X5) berpengaruh secara Parsial (bebas) terhadap Variabel Visitor Satisfaction (Y) Malang Night Paradise H2: Variabel sense (X1), feel (X2), think (X3), act (X4), dan relate (X5) berpengaruh secara Parsi (bebas) terhadap Variabel Visitor Satisfaction (Y) Malang Night Paradise Uji Asumsi Klasik Uji asumsi klasik dalam penelitian ini adalah Uji Normalitas, Uji Multikolinearitas dan Uji Heteroskedastisitas. Uji normaitas didasarkan nilai One-Sample Kolmogorov-Smirnov, diketahui bahwa data dalam penelitian ini berdistribusi normal, yang ditunjukkan dengan nilai sig > dari 0,05 (0,428>0,05). Uji Multikoliniearitas diketahui nilai Variance Inflaction Faktor (VIF) pada masing-masing variabel memiliki nilai < 10 dan nilai tolerance > 0,1. Jadi dapat diketahui tidak terjadi multikolinearitas antara variabel bebas Experiential Marketing (X) terhadap variabel terikat kepuasan pengunjung (Y). Sedangkan untuk Uji Heteroskedastisitas dalam penelitian ini menunjukkan bahwa plot menyebar secara merata di atas dan di bawah angka 0 pada sumbu Y tanpa membentuk sebuah pola tertentu. Dengan demikian maka dapat disimpulkan bahwa variabel pada pengujian penelitian ini bebas dari gejala heterokedastisitas. Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise Pembahasan Berdasarkan hasil analisis dapat diketahui bahwa dari kelima Dimensi Experiential Marketing yang meliputi variable (sense, feel, think, act, relate) terdapat dua dimensi yang tidak berpengaruh secara signifikan terhadap variable Visitor Satisfaction (Y) Malang Night Paradise, yaitu dimensi sense (X1), dan relate (X5) karena kedua Dimensi Experiential Marketing tersebut memiliki t-hitung < t-tabel yaitu sebesar 0,443 < 1,662 Hal ini membuktikan bahwa variable bebas (X1) atau variabel sense tidak berpengaruh signifikan terhadap variable terikat (Y ) atau Visitor Satisfaction. Karena kebanyakan pengunjung tidak menjadikan kesejukan udara, kebersihan area wisata, dan desain eksterior yang menarik sebagai tolok ukur kepuasan mereka. Sehingga data yang diperoleh dalam penelitian ini tidak dapat membuktikan keterkaitan antara sense (X1) dengan Kepuasan Pengunjung (Y). dan variabel relate (X5) sebesar 1,146 < 1,662 Ini membuktikan bahwa secara bebasa (X5) atau variabel relate (X5) tidak berpengaruh signifikan terhadap (Y) atau variabel Visitor Satisfaction. Karena jarang terjadi keluhan dari pengujung, serta pelayanan yang dirasakan pengunjung tidak begitu istimewa. Ini berarti data yang dikumpulkan dalam penelitian ini tidak dapat membuktikan keterkaitan antara variable relate (X5) dengan variable Visitor Satisfaction (Y). Sementara itu ketiga Dimensi Experiential Marketing yang meliputi variable Feel (X2), Think (X3), Act (X4) berpengaruh secara signifikan terhadap variable Visitor Satisfaction (Y) pada wisata Malang Night Paradise. Kemudian secara simultan variabel Dimensi Experiential Marketing yang meliputi variable (sense, feel, think, act, relate) berpengaruh signifikan terhadap variable Visitor Satisfaction (Y) pada wisata Malang Night Paradise. Penelitian ini sejalan dengan penelitian yang dilakukan oleh Jatmiko dan Andharini (2012), dalam penelitiannya yang berjudul “Analisis Experiential Marketing dan Loyalitas Pelanggan Jasa Wisata (Studi Pada Taman Rekreasi Sengkaling Malang), secara parsial feel dan sense berpengaruh postif signifikan terhadap loyalitas pelanggan, sedang think, act, dan relate berpengaruh positif tidak signifikan terhadap loyalitas pelanggan., namun penelitian ini bertentangan dengan penelitian yang dilakukan oleh Jannah (2014), dalam penelitiannya yang berjudul “Pengaruh Strategi Experiential Marketing Terhadap Kepuaan pengunjung Museum Sepuluh Nopember Surabaya”, Secara simultan terdapat pengaruh signifikan, namun secara parsial terdapat tiga variabel yang tidak signifikan yaitu Feel, Think, dan Act. 555 Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise 5 Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Ni ht P di SIMPULAN Berdasarkan hasil penelitian dan pembahasan yang telah dilakukan maka dapat ditarik kesimpulan bahwa variabel bebas (parsial) Sense dan relate atau (X1) dan (X5) tidak berpengaruh secara signifikan terhadap variebel Visitor Satisfaction (Y) pada Wisata Malang Night Paradise. Namun, pada variabel bebas (parsial) lainnya yang meliputi variabel feel, think dan act atau (X2), (X3), (X4) berpengaruh secara signifikan terhadap variabel Visitor Satisfaction (Y) pada Wisata Malang Night Paradise. Dan secara terikat (simultan) seluruh variabel Dimensi Experiential Marketing yang meliputi variabel (sense, feel, think, act, relate) berpengaruh signifikan terhadap variabel Visitor Satisfaction (Y) pada wisata Malang Night Paradise. Pada penelitian terdapat keterbatasan terkait jawaban yang diberikan oleh sampel dari kuisioner yang dibagikan tidak menunjukkan keadaan sebenarnya, serta terbatasnya jumlah responden dalam penelitian ini yaitu sebesar 96 responden. Penelitian ini sudah dilaksanakan sesuai dengan prosedur ilmiah, namun masih memiliki keterbatasan seperti hal-hal yang mempengaruhi Visitor Satisfaction dalam penelitian ini hanya berasal dari Dimensi Experiential Marketing saja yang menghasilkan nilai R2 sebesar 63,7%, sedangkan sebesar 36,3% dipengaruhi oleh faktor lain yang dapat mempengaruhi Visitor Satisfaction (Y) Berdasarkan hasil kesimpulan dari hasil penelitian, maka diajukan beberapa saran kepada pihak wisata Malang Night Paradise diharapkan mempertahankan serta memperhatikan strategi Experiential Marketing, khususnya dalam hal pengalaman yang melibatkan panca idera (sense) dan juga hubungan dengan pengunjung (relate) agar dapat lebih meningkatkan Visitor Satisfaction pada Wisata Malang Night Paradise. peneliti berikutnyanya diharapkan untuk menyempurnakan atau melenngkapi penelitian ini dengan menambahkan beberapa variabel lain yang mempengaruhi Visitor Satisfaction pada Wisata Malang Night Paradise serta menambah jumlah dari responden agar diperoleh hasil yang lebih akurat dan signifikan. DAFTAR PUSTAKA Algifari. 2000. Analisis Regresi: Teori, Kasus dan Solusi Edisi 2. Yogyakarta: BPFE UGM Alma, Dan Buchari. 2007. Manajemen Pemasaran dan Pemasaran Jasa. Bandung : Alfabeta. Andreani, Dan Fransisca. 2007.”Experiential Marketing (Sebuah Pendekatan Pemasaran)”. Jurnal Manajemen Pemasaran, Vol.2 No.1 April hal.1 – 8. Arikunto, Suharsimi. 2010. Prosedur Penelitian: Suatu Pendekatan Praktek. Jakarta: PT Rineka Cipta. Bitner, M. J. Dan Zeithaml, V. A. 2003. Service Marketing (3rd ed.). New Delhi: Tata McGraw Hill. Dewi, Ratih Dkk. 2013. Pengaruh Experiential Marketing Terhadap Kepuasan Pelanggan dan Dampaknya Terhadap Loyalitas Pelanggan (Survwei Pada Pelanggan Tempat Wisaa Jawa Timur Park 1 Kota Wisata Batu). Fakultas Ilmu Administrasi. Universitas Brawijaya Malang Ghozali, Dan Imam. 2007. Analisis Multivariate Dengan Program SPSS. Semarang: BP Universitas Diponegoro. .2009. Aplikasi Analisis Multivarian Dengan Program SPSS. Semarang : Badan Penerbit Universitas Diponegoro. Gujarati, Dan Damodar. 2003. Ekonometri Dasar, Terjemahan: Sumamo Zain, Jakarta: Erlangga Grundey, Dan Daniora. 2008. Experiential Marketing vs Traditional Marketing. Dalam Romanian Economic Journal. Vol.11 No.3. Hal 133-151. Handal, Nehemia S. 2010. Analisis Pengaruh Experiential Marketing Terhadap Loyalitas. Jurnal Manajemen Universitas Diponegoro. Hasan, A. 2008. Marketing. Yogyakarta: Media Utama. Indriantoro, Nur dan Supomo. 2014. Metodologi Penelitian Bisnis untuk Akuntasi dan Manajemen. Yogyakarta: BPFE Jannah, Dan Miftahul. 2014. Pengaruh Experiential Marketing Terhadap Kepuasan Pengunjung Museum sepuluh Nopember Surabaya. Fakultas Ekonomi. Universitas Trunojoyo Madura. J.Setiadi 2010. Perilaku Konsumen, KENCANA prenada media group : Jakarta Jatmiko, Dwi dan Andharini. 2012. Analisi Experiential Marketing dan Loyalitas Pelanggan Jasa Wisata (Studi pada Taman Rekreasi Sengkaling Malang). (Jurnal) Fakultas Ekonomidan Bisnis. Universitas Muhammadiyah Malang. 6 Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise Kartajaya, Hermawan. 2004. Hermawan Kartajaya on brand Seri 9 Elemen Marketing. Bandung: P Mizan Pustaka. Kartajaya, Hermawan. 2004. Hermawan Kartajaya on brand Seri 9 Elemen Marketing. Bandung: PT Mizan Pustaka. . 2006. Hermawan Kartajaya On Marketing. Jakarta: PT Gramedia. Kotler, Philip. 2000. Manajemen Pemasaran DiIndonesia. Jakarta : Salemba Empat. Kotler. P. 2009. Manajemen Pemasaran. : Analisa, Perencanaan, Implikasi dan Kontrol, Jilid I. Jakarta :PT Prenhallindo. Kotler, Phillip. 2002. Manajemen Pemasaran 1. Milenium ed. Jakarta: PT Mowen Kotler, Philip. 2005. Marketing management: Analysis, Planing, Implementation, Control. Ed. 8, New jersey: Prentice Hall, Inc. Kotler, Philip and Keller. 2006. Manajemen Pemasaran. Jakarta: PT. Indeks. dan Armstrong, Gary. 2008. Prinsip-prinsip Pemasaran, Edisi 12. Jakarta: Erlangga. Kustini. 2007, Penerapan Experiential Marketing. Jurnal Riset Ekonomi dan Bisnis Vol. 7, No. 2. UPN “Veteran”Jawa Timur. Lulluangi, Irene Rosa. 2014. DAFTAR PUSTAKA Analisis Pengaruh Experiential Marketing Terhadap Kepuasan Konsumen (Skripsi). Semarang: Universitas Diponegoro. Lupiyoadi, Rambat. 2001. Manajemen Pemasaran Jasa Teori dan Praktik. Jakarta: Salemba Empat. . 2013. Manajemen Pemasaran Jasa. Jakarta : Salemba Empat. Priansa. 2017. Komunikasi Peamasaran Terpadu. Bandung: Pustaka Setia Payne, Andrian. 2001. Pemasaran Jasa. Cetak Kedua. Yogyakarta. Andi And Pearson Education. Priyatno, Dan Duwi. 2010. Tehnik Mudah Dan Cepat Melakukan Analisis Data Penelitian Dengan Sp Dan T J b Uji P d d Yogyakarta: Gaya Media Priyatno, Dan Duwi. 2010. Tehnik Mudah Dan Cepat Melakukan Analisis Data Penelitian Dengan Spss Dan Tanya Jawab Ujian Pendadaran. Yogyakarta: Gaya Media. y , p g p Dan Tanya Jawab Ujian Pendadaran. Yogyakarta: Gaya Media. Tanya Jawab Ujian Pendadaran. Yogyakarta: Gaya Media. . 2012. Cara Kilat Belajar Analisis Data Dengan Spss 2.0. Yogyakarta: Andi. Priyono, Agus. 2015. Analisis Data dengan SPSS. BPFE UNISMA Priyono, Agus. 2015. Analisis Data dengan SPSS. BPFE UNISMA Raharso, Sri. 2005. Pengaruh Customer Delight terhadap Behavioral Intentions Battrey. Jurnal, Usahawan No. 5, mei 2005 Rini, Endang Sulistya. 2009. Menciptakan Pengalaman Konsumen Dengan Experienxial Marketing. Jurnal Manajemen Bisnis. Vol II, No 1, h. 15-20. Ruslan, dan Rosady. 2007. Manajemen Public Relations & Media Komunikasi: Presepsi dan Aplikasi. Jakarta: PT Raja Grafindo Persada. a roni 2017. Pengaruh Dimensi Experiential Marketing Terhadap Kepuasan Pengunjung (Studi Kasus Pada Taman Rekreasi Tlogomas Malang. Jurnal Ilmiah Riset Manajemen e-Jrm Vol. 6 No. 1 557 Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise Schmitt, B. 1999. Experiential marketing: How to get customers to sense, feel, think, act and relate to your company and brands. The Free Press, New York, USA. Singarimbun, Masri dan Sofian Effendi. 2008. Metode Penelitian Survei. Jakrta: LP3ES Indonesia. Singgih, Dan Susanto. 2002. SPSS Versi 11.5 . Cetakan Kedua. Jakarta: Gramedia. Soemantri, Ating dan Ali M, Sambas. 2006. Aplikasi Statistika dalam Penelitian. Bandung: CV Pustaka Setia. Sugiyono. 2004. Metode Penelitian Binis. Bandung. Alfabeta. Sugiyono. 2009. Metode Penelitian Bisnis , Edisi ke-14. Bandung: Alfabeta. Tjiptono, Fandi. 2002. Manajemen Jasa. Yogyakarta : Andy. . 2005. Pemasaran Jasa. Edisi 1. Malang. Bayu Media Publissing . 2006. Pemasaran Jasa. Malang : Bayumedia Publishing. . 2008. Strategi Pemasaran. Edisi Ketiga. Yogyakarta. Andi Offset . 2014. Pemasaran Jasa: Prinsip Penerapan Dan Penelitian. Yoggyakarta. Andi Offset. Umar, Dan Husen. 2012. Riset Sdm Dalam Organisasi. Jakarta: Gramedia Pustaka Utama. Utami, Chistina Widya. 2006. Manajemen Ritel. Jakarta: Salemba Empat. Wicaksono, Arry. 2005. 558 Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise DAFTAR PUSTAKA Pengaruh Persepsi Kualitas Layanan Terhadap Kepuasan Klien dan Dampaknya Pada Preferensi Rekomendasi Klien. Jurnal Ekonomi Perusahaan, IBII Vol.12- No.3. Wood, Ivonne. 2009. Layanan Pelanggan: Cara Praktis, Murah dan Inspiratif Memuaskan Pelanggan Anda. Yogyakarta: Graha Ilmu. Yazid. 1999. Pemasaran Jasa. Fakultas Ekonomi. UII 58 Dimensi Experiential Marketing Terhadap Visitor Satisfaction Pada Wisata Malang Night Paradise
https://openalex.org/W3033384051	https://www.frontiersin.org/articles/10.3389/fimmu.2020.01124/pdf	English	null	Interleukins in Thyroid Cancer: From Basic Researches to Applications in Clinical Practice	Frontiers in immunology	2,020	cc-by	11,972	REVIEW published: 12 June 2020 doi: 10.3389/ﬁmmu.2020.01124 REVIEW published: 12 June 2020 doi: 10.3389/ﬁmmu.2020.01124 Edited by: Edited by: Avery Dexter Posey Jr., University of Pennsylvania, United States Avery Dexter Posey Jr., University of Pennsylvania, United States Reviewed by: Evelien Smits, University of Antwerp, Belgium Connie Jackaman, Curtin University, Australia Correspondence: Zhong-Ling Qiu qiuzhongling123@163.com Quan-Yong Luo lqyn@sh163.net Reviewed by: Evelien Smits, University of Antwerp, Belgium Connie Jackaman, Curtin University, Australia Keywords: interleukins, thyroid cancer, inﬂammation, tumor microenvironment, immunotherapy Correspondence: Zhong-Ling Qiu qiuzhongling123@163.com Quan-Yong Luo lqyn@sh163.net Interleukins in Thyroid Cancer: From Basic Researches to Applications in Clinical Practice Chuang Xi †, Guo-Qiang Zhang †, Zhen-Kui Sun, Hong-Jun Song, Chen-Tian Shen, Xiao-Yue Chen, Jian-Wen Sun, Zhong-Ling Qiu* and Quan-Yong Luo* Department of Nuclear Medicine, Shanghai Jiao Tong University Afﬁliated Sixth People’s Hospital, Shanghai, China Inﬂammation is crucial to tumorigenesis and progression of many cancers. Inﬂammatory molecules in tumor microenvironment exert pro- or anti-tumor effects. Among them, interleukin, mainly produced by CD3+ and CD4+ T lymphocytes, is a class of small molecule proteins which play an important role in intercellular communication. Numerous studies have conﬁrmed that interleukins are closely related to thyroid cancer. Interleukins regulate the proliferation and migration of thyroid cancer cells and they have prospects in discriminating benign and malignant thyroid diseases, predicting the risk of tumorigenesis, evaluating the prognosis and monitoring the recurrence of thyroid cancer. Besides, the effective application of interleukins in treatment of thyroid cancer has been conﬁrmed by some cell and animal researches. The present review will introduce the potential mechanisms of interleukins in thyroid cancer and focus on the applications of interleukins in clinical practice of thyroid cancer, which will help update understanding of the progress of interleukins researches in thyroid cancer. Inﬂammation is crucial to tumorigenesis and progression of many cancers. Inﬂammatory molecules in tumor microenvironment exert pro- or anti-tumor effects. Among them, interleukin, mainly produced by CD3+ and CD4+ T lymphocytes, is a class of small molecule proteins which play an important role in intercellular communication. Numerous studies have conﬁrmed that interleukins are closely related to thyroid cancer. Interleukins regulate the proliferation and migration of thyroid cancer cells and they have prospects in discriminating benign and malignant thyroid diseases, predicting the risk of tumorigenesis, evaluating the prognosis and monitoring the recurrence of thyroid cancer. Besides, the effective application of interleukins in treatment of thyroid cancer has been conﬁrmed by some cell and animal researches. The present review will introduce the potential mechanisms of interleukins in thyroid cancer and focus on the applications of interleukins in clinical practice of thyroid cancer, which will help update understanding of the progress of interleukins researches in thyroid cancer. INTRODUCTION Thyroid cancer is the most common endocrine malignancy with increasing incidence rate over the past decades (1). It happens as a result of hereditary susceptibility and environment factors such as iodine excess, radiation exposure, obesity (2). According to pathology types, thyroid cancer is divided into diﬀerentiated thyroid carcinoma (DTC), anaplastic thyroid carcinoma (ATC), and medullary thyroid carcinoma (MTC). DTC, including papillary thyroid carcinoma (PTC) and follicular thyroid carcinoma (FTC), accounts for 90% of all thyroid cancers and has a relatively good prognosis. More than 80% of patients with DTC can achieve excellent response to current treatment model, such as surgery, radioiodine (RAI) therapy, and TSH suppressive therapy (3). Although ATC only accounts for 1–2% of thyroid cancer, it is responsible for 14–50% of all thyroid cancer-related deaths due to the lack of eﬀective treatment (4). Molecular targeted therapy is the most promising emerging treatment for ATC and the involved drugs are multiple receptor tyrosine kinase inhibitors (5). MTC, derived from thyroid C cells, accounts for about 5–10% of thyroid cancer and the current treatment of MTC is limited to surgery (6). †These authors have contributed equally to this work †These authors have contributed equally to this work Specialty section: This article was submitted to Cancer Immunity and Immunotherapy, a section of the journal Frontiers in Immunology Specialty section: This article was submitted to Cancer Immunity and Immunotherapy, a section of the journal Frontiers in Immunology Received: 08 February 2020 Accepted: 07 May 2020 Published: 12 June 2020 REGULATING TUMOR CELL PROLIFERATION the immune system, regulation of transcription factors, and control of inﬂammation. The role of interleukins in cancer was ﬁrst described by Vose (7), and in the following decades many studies have conﬁrmed that interleukins, from IL-1 to IL-38, play signiﬁcant roles in many cancers, such as breast cancer, hepatoma, thyroid cancer etc. (8, 9). Tumor cell proliferation is an important step in tumor development. Several studies have demonstrated that interleukins could regulate the proliferation of thyroid cancer cells. To date, numerous studies have conﬁrmed that interleukins play signiﬁcant roles in the diagnosis and treatment of thyroid cancer. This review will present the eﬀects of interleukins in thyroid cancer and the clinical applications in the diagnosis and treatment of thyroid cancer in order to help update understanding of the progress of interleukins researches in thyroid cancer. IL-1 includes two activator cytokines IL-1α and IL-1β, as well as an inhibitory cytokine, the IL-1 receptor antagonist (IL- 1ra). IL-1α and IL-1β bind to the same receptor, the type 1 IL-1 receptor (IL-1R), and activate the downstream signaling cascades, ultimately promoting the immune and inﬂammatory responses (12). The role of IL-1 in cancer has been well- demonstrated (13) and it is well-demonstrated that IL-1 could regulate the proliferation of thyroid cancer through diﬀerent mechanisms. Due to diﬀerent thyroid cancer cell lines used in diﬀerent studies, the results are contradictory. IL-1α could promote the proliferation of PTC cell line NIM1 via stimulation of Ca2+ channels (14). IL-1 could also suppress the proliferation of thyroid cancer cells. IL-1 inhibits the growth of the thyroid cancer cell line NPA, which was in part associated with the suppression of c-myc (15). IL-1β exerts strong antitumor eﬀects on PTC (16, 17) and FTC cell lines (18) through suppressing proliferation and invasiveness. Furthermore, IL- 1β did not have an anti-proliferative eﬀect on ATC cell lines, which indicates that PTC cancer cells escaping from antitumor eﬀect of IL-1β may be a step toward anaplasia change, resulting in more aggressiveness of thyroid cancer (17). However, the mechanisms of this process are not clear and further studies are needed. Citation: Xi C, Zhang G-Q, Sun Z-K, Song H-J, Shen C-T, Chen X-Y, Sun J-W, Qiu Z-L and Luo Q-Y (2020) Interleukins in Thyroid Cancer: From Basic Researches to Applications in Clinical Practice. Front. Immunol. 11:1124. doi: 10.3389/ﬁmmu.2020.01124 Tumor microenvironment is closely related to the occurrence and development of cancer. It consists of immune cells, stroma cells, cytokines, and chemokines, which exert pro- or anti- tumor eﬀects. Interleukins are small protein signaling molecules that belong to the superfamily of cytokines and are mainly produced by T lymphocytes, monocytes, macrophages, and endothelial cells. The main functions of interleukins include the facilitation of communication between cells of June 2020 \| Volume 11 \| Article 1124 Frontiers in Immunology \| www.frontiersin.org Xi et al. Xi et al. Interleukins and Thyroid Cancer THE EFFECTS OF INTERLEUKINS IN THYROID CANCER Growing evidence suggests that imbalance of pro-inﬂammatory and anti-inﬂammatory cytokines is correlated to the pathogenesis of thyroid cancer. Inﬂammatory molecules in tumor microenvironment exert two main eﬀects. One hand, they sustain features of the malignant phenotype of tumors, such as proliferation and invasiveness (10). Moreover, they recruit inﬂammatory and immune cells, and induce the remodeling of the tumor stroma and stimulate angiogenesis. Thus, inﬂammatory molecules could further promote tumor progression. In addition, the recruitment of immune cells into tumor sites could result in the immune escape of cancer cells, because cancer cells could induce the secretion of molecules that suppress immune responses and the recruitment of regulatory T cells (11). Interleukins are also crucial components of microenvironment of thyroid cancer and some studies conﬁrmed that interleukins play signiﬁcant roles in thyroid cancer through some potential mechanisms (Figure 1, Table 1). IL-22, produced by Th17 and Th22 cells, exerts its biological eﬀects through binding to IL-22 receptor and IL-10 receptor. IL- 22 triggers a variety of downstream signaling pathways including JAK/STAT3 and MAPK, resulting in cancer progression (19). In thyroid cancer, IL-22 induces miR-595 expression, which in FIGURE 1 \| The mechanism of interleukins in thyroid cancer. Interleukins (A) regulate the proliferation of thyroid cancer cells and promote the process of (B) Epithelial-to-Mesenchymal Transition (EMT) and (C) angiogenesis. Besides, they also regulate the abilities of thyroid cancer cells to (D) resist to cell apoptosis and (E) escape the immune system. Through these mechanisms, interleukins cloud play important roles in the tumorigenesis and development of thyroid cancer. FIGURE 1 \| The mechanism of interleukins in thyroid cancer. Interleukins (A) regulate the proliferation of thyroid cancer cells and promote the process of (B) Epithelial-to-Mesenchymal Transition (EMT) and (C) angiogenesis. Besides, they also regulate the abilities of thyroid cancer cells to (D) resist to cell apoptosis and (E) escape the immune system. Through these mechanisms, interleukins cloud play important roles in the tumorigenesis and development of thyroid cancer. June 2020 \| Volume 11 \| Article 1124 Frontiers in Immunology \| www.frontiersin.org 2 Ca2+ e PA. e d –Slug C cell Sample size Method ILs Findings — — IL-1α IL-1α promotes the proliferation of NIM1 cells by stimulating Ca2 inﬂux voltage-dependent Ca2+ channels. — — IL-1α IL-1β IL-1α and IL-1β inhibited NPA cells growth associated with the suppression of c-myc. ll lines nd NP O lines lines lines ines lines mples lines C cell cancer; P chemistry THE EFFECTS OF INTERLEUKINS IN THYROID CANCER — — IL-1β IL-1β inhibit the proliferation and invasiveness of TPC-1 and NPA — — IL-1β IL-1β inhibit growth of BHP and NP cells but not ARO cells. — — IL-1β IL-1β inhibit the proliferation of follicular cell line. — — IL-22 IL-22 induces miR-595 expression which in turn downregulate Sox17 expression and then results in increasing migration and invasion of PTC cells. PTC: 30 Cell culture, western blotting IHC IL-8 1. IL-8 is required for EMT in thyroid cancer through IL-8–Akt–Sl pathway. 2. IL-8 increases the stem features of thyroid cancer cells. ATC: 76 PTC: 100 IHCELISAPCR IL-11 Hypoxia-induced stimulation of IL-11 production results in ATC c invasion, migration, and EMT via the PI3K/Akt/GSK3β pathway. TC: 65 FTC: 50 ATC: 22 IHC PCR Western blot IL13Rα2 IL13RA2 is differentially regulated in PTC and is involved in cell migration via enhancing EMT. PTC: 40 IHCCell culture IL13Rα2 Knockdown of IL13Rα2 could decline the number of invading ce in PTC. PTC: 18 PCR IL-8 IL-8 is an important downstream mediator of NF-kB signaling in advanced thyroid cancer growth and progression. TC: 61, FTC:19, MTC:8 IHC IL-17 1. Expression of IL-17 proteins in DTC and MTC tissues is highe 2. High IL-17 expression was associated with recurrence and mortality in thyroid neoplasm. PTC: 89 PTC+HT: 49 PCR IHC IL-17A 1. IL-17A induced MHC class I expression and promotes inhibiti of tumor immune escape in PTC with HT. 2. The immune escape suppressed by IL-17A may be linked to PD-1/PD-L1 pathway. — — IL-4 IL-10 IL4 and IL-10 plays signiﬁcantly role in protecting thyroid cancer cells from apoptosis when complicated with Graves’ disease. PTC: 8 FTC: 8 UTC: 5 Immunostaining Western Blotting IL-4 IL-10 IL-4 and IL-10 protect thyroid cancer cells from cytotoxic effect antineoplastic drugs by induce the expression of Bcl-xL and Bcl PTC: 8 FTC: 8 UTC: 4 Immunostaining Western Blotting IL-4 IL-10 Autocrine of IL-4 and IL-10 in thyroid cancer results in resistance to CD95-mediated apoptosis. PTC: 45 PCR Flow cytometry IL13Rα2 1. IL13Rα2 is contributed to the tumorigenesis, cell progression a invasion of thyroid cancer. 2. IL13Rα2 may function as an oncogene during PTC carcinogenesis. , papillary thyroid cancer; FTC, follicular thyroid cancer; ATC, anaplastic thyroid cancer; MTC, Medullary thyroid cancer; PCR, Polymerase Chain Reaction; T, Hashimoto’s thyroiditis; EMT, epithelial-to-Mesenchymal Transition. er. REGULATING TUMOR IMMUNE ESCAPE It is an essential step in successful migration and metastasis of tumor cells. Some interleukins promote the EMT process of thyroid cancer and then enhance the aggressiveness of thyroid cancer. Human immune system is capable of recognizing and resisting cancer cells, however, by altering the host immune system, tumors can escape immune control and continue to progress (37). The tumor microenvironment provides conditions for tumors to escape the immune surveillance, and some interleukins play an important role in this process. IL-8, a pro-inﬂammatory chemokine, functions through binding to CXCR1 and CXCR2. Considering the characteristic expression of CXCR1 and CXCR2 on cancer cells, endothelial cells, and tumor-associated macrophages, the increased secretion of IL-8 from tumor cells has signiﬁcance to the tumor microenvironment (22). IL-8 has been repeatedly reported to be a tumor-promoting cytokine in several cancers, but rarely reported in thyroid cancer. Mast cells, which correlate to malignant features and invasiveness of thyroid cancer, are the main source of IL-8 in thyroid cancer. IL-8 is required for mast cells mediated EMT in thyroid cancer through the IL-8-Akt-Slug pathway (23). IL-10 is an anti-inﬂammatory and immunosuppressive cytokine that inﬂuence the course of cancer by promoting immune escape through inhibition of the antitumor activity of immune cells (38). IL-10 is expressed in thyroid cancer and inﬂuence the aggressiveness of it (39). The immunosuppressive eﬀect of IL-10 may be involved in the immune escape of thyroid cancer cells and promote the aggressiveness of thyroid cancer. However, studies are needed to explain the mechanisms of this process. IL-11 interactives with its receptor IL-11Rα and activates signaling pathways of targeted cells such as JAK/STAT, MAPK, Src-family kinases, and PI3K pathway (24). The IL-11 gene is a hypoxia-inducible gene whose expression is induced by hypoxia via HIF-1α (25). IL-11 promotes the invasion, migration and EMT of ATC cell via the PI3K/Akt/GSK3β pathway (26). Higher expression of IL-11 in ATC tissues than in PTC could explain the higher metastasis rates of ATC (26). The promotion of EMT induced by IL-11 could take part in this process. Tumor cells achieve immune escape by downregulating the expression of major histocompatibility complex (MHC) class I and loss of MHC class I expression is a frequent mechanism of tumor immune escape in PTC (40). However, after IL-17 treatment, the membrane expression of MHC class I in K1 and PTC-1 increased signiﬁcantly (41). THE EFFECTS OF INTERLEUKINS IN THYROID CANCER ell lines nd NP O lines lines lines lines lines mples s lines s s C cell cancer; P chemistry ll lines nd NP O lines lines lines ines lines mples lines Interleukins and Thyroid Cancer Xi et al. Xi et al. of IL-1β and TNF-α in the tumor microenvironment (34). The expression of IL-17 is observed in various tumor tissues and considered as the most important pro-angiogenic mediator (35). The expression level of IL-17 is higher in DTC and MTC than that in benign thyroid neoplasms (36), which suggests that IL-17 is closely correlated to the aggressiveness of thyroid cancer, and the tumor pro-angiogenesis of IL-17 could have roles in this process. However, no study has ever reported the pro-angiogenesis of IL-17 in thyroid cancer. turn downregulates Sox17 expression, thereby enhancing the migration and invasion of thyroid cancer (20). turn downregulates Sox17 expression, thereby enhancing the migration and invasion of thyroid cancer (20). REGULATING TUMOR IMMUNE ESCAPE Programmed cell death ligand 1 (PD-L1), expressed on the surface of tumor cells, binds to its receptor PD-1 on T cells membrane, inducing T cells anergy. The PD-1 expression of T cells reduced in the presence of IL- 17 (41). It is suggested that IL-17 inhibit tumor immune escape by upregulating MHC class I expression on tumor cells and suppressing PD-L1/PD-1 pathway. IL13Rα2 is a type II cytokine receptor with high binding aﬃnity to IL-13 (27) and it has an oncogenic role in many cancers (28, 29). In PTC, IL13Rα2-induced cell migration is associated with the upregulation of EMT markers such as N-cadherin, Vimentin and Snail, indicating that IL13Rα2 enhances thyroid cancer aggressiveness through promoting EMT process (30). A recent study found that the number of invading cells in PTC declined signiﬁcantly after IL13Rα2 knockdown, indicating that IL13Rα2 is involved in the invasion of PTC cells (31). However, the potential mechanism of how IL13Rα2 inﬂuence the EMT process of thyroid cancer is not clear. Some interleukins can inhibit the anti-tumor immune response, allowing tumor cells to escape recognition and attack by the immune system. Therefore, tumor cells can proliferate and metastasize to distant organs. Better understanding of the mechanisms of interleukins in immune escape will provide new targets for immunotherapy of thyroid cancer. PROMOTING EPITHELIAL-TO-MESENCHYMAL TRANSITION (EMT) EMT is a process in which epithelial cells lose adhesion properties and turn into a mesenchymal phenotype, allowing non-invasive tumor cells to attain the ability of invasion and metastasis (21). It is an essential step in successful migration and metastasis of tumor cells. Some interleukins promote the EMT process of thyroid cancer and then enhance the aggressiveness of thyroid cancer. EMT is a process in which epithelial cells lose adhesion properties and turn into a mesenchymal phenotype, allowing non-invasive tumor cells to attain the ability of invasion and metastasis (21). Frontiers in Immunology \| www.frontiersin.org INHIBITING CANCER CELL APOPTOSIS Besides, in PTC cell lines, transfection with siRNA targeting IL13Rα2 induces cell apoptosis by upregulation of caspase 3 and then results in inhibition of cell proliferation, which indicates that IL13Rα2 promotes the invasion and metastasis of tumors through inhibiting apoptosis (31). It can induce Th2 cell proliferation and diﬀerentiation, and inhibit apoptosis of B and T cells. A variety of malignancies, such as melanoma and breast cancer, express IL-4 receptor (IL-4Rα), and IL-4 has antiproliferative and/or proapoptotic eﬀects in these cancer cells. On the contrary, IL-4 weakly stimulates the proliferation of thyroid cancer and protects it from apoptosis. The pro-tumor eﬀect of IL-4 is associated with the up-regulation of anti-apoptotic molecule Bcl-2 and the weak down-regulation of the pro-apoptotic molecule Bax (43). Besides, autocrine production of IL-4 and IL-10 induces the over-expression of Bcl-xL and Bcl-2, two anti-apoptotic proteins, which subsequently protect thyroid cancer cells from the cytotoxic eﬀects of antineoplastic drugs (44). Both PTC and FTC cells express CD95 and its ligand CD95L, which mediate cell apoptosis, however, expression of CD95 and CD95L does not aﬀect tumorigenesis and progression in thyroid cancer (45). Therefore, there is a molecular mechanism that restrains the CD95-mediated apoptosis signaling pathway in thyroid cancer and autocrine production of IL-4 and IL-10 may have signiﬁcant roles in this mechanism, because IL-4 and IL-10 promote resistance to CD95-mediated apoptosis via the activating the Jak/Stat pathway and up-regulating cFLIPL and PED (46). Besides, in PTC cell lines, transfection with siRNA targeting IL13Rα2 induces cell apoptosis by upregulation of caspase 3 and then results in inhibition of cell proliferation, which indicates that IL13Rα2 promotes the invasion and metastasis of tumors through inhibiting apoptosis (31). treatment of thyroid cancer, further studies focused on potential mechanisms in are still needed. treatment of thyroid cancer, further studies focused on potential mechanisms in are still needed. INHIBITING CANCER CELL APOPTOSIS To support the high proliferation of cancer cells, tumors need to rapidly develop a new vascular network. Angiogenesis, the formation of new blood vessels, is one of crucial steps in tumor progression. NF-κB is a key regulator of angiogenesis in thyroid cancer (32), and IL-8 may be a signiﬁcant downstream eﬀector of NF-κB signaling pathway in the progression of advanced thyroid cancer (33). IL-17 is involved in the pathogenesis of inﬂammatory responses and is known to induce the production Apoptosis, also called programmed cell death, is ﬁnely regulated at the gene level to orderly remove damaged cells (42). Its alteration is not only responsible for tumor progression but also for tumor resistance to therapies. Autocrine production of interleukins in thyroid cancer results in upregulation of anti-apoptotic proteins, which contributes to tumor progression. IL-4 is a pleiotropic cytokine produced by Th2 cells and exerts regulatory eﬀect on the immune response. June 2020 \| Volume 11 \| Article 1124 Frontiers in Immunology \| www.frontiersin.org 4 Xi et al. Interleukins and Thyroid Cancer It can induce Th2 cell proliferation and diﬀerentiation, and inhibit apoptosis of B and T cells. A variety of malignancies, such as melanoma and breast cancer, express IL-4 receptor (IL-4Rα), and IL-4 has antiproliferative and/or proapoptotic eﬀects in these cancer cells. On the contrary, IL-4 weakly stimulates the proliferation of thyroid cancer and protects it from apoptosis. The pro-tumor eﬀect of IL-4 is associated with the up-regulation of anti-apoptotic molecule Bcl-2 and the weak down-regulation of the pro-apoptotic molecule Bax (43). Besides, autocrine production of IL-4 and IL-10 induces the over-expression of Bcl-xL and Bcl-2, two anti-apoptotic proteins, which subsequently protect thyroid cancer cells from the cytotoxic eﬀects of antineoplastic drugs (44). Both PTC and FTC cells express CD95 and its ligand CD95L, which mediate cell apoptosis, however, expression of CD95 and CD95L does not aﬀect tumorigenesis and progression in thyroid cancer (45). Therefore, there is a molecular mechanism that restrains the CD95-mediated apoptosis signaling pathway in thyroid cancer and autocrine production of IL-4 and IL-10 may have signiﬁcant roles in this mechanism, because IL-4 and IL-10 promote resistance to CD95-mediated apoptosis via the activating the Jak/Stat pathway and up-regulating cFLIPL and PED (46). Serum Interleukin Level Serum level of other interleukins could also be used in discriminating benign and malignant thyroid diseases. Serum levels of IL-7, IL-10, and IL-13 are higher in thyroid diseases than healthy controls (49). However, a recent study (51) reported that serum concentrations of IL-10 could not discriminate benign and malignant groups from healthy controls. Besides, Serum concentration of IL-35 in thyroid cancer is lower than that in thyroid adenoma, which indicates IL-35 could be used in discriminating thyroid cancer and adenoma (54). Serum biomarkers are generally used for tumor screening due to their eﬀectiveness and convenience, and speciﬁc serum tumor biomarkers are important for tumor diagnosis. With the development of tumor immunity, some studies conﬁrm the possibility of serum cytokine level for tumor diagnosis. It has been reported that the serum level of some interleukins could serve as potential biomarkers in the diagnosis of thyroid cancer, especially in discriminating benign and malignant thyroid diseases (Table 2). However, the results of diﬀerent studies are contradictory. Interleukins are potential biomarkers in discriminating benign and malignant thyroid diseases, however, the clinical application value of single type of interleukin still have some limitations. In order to solve this problem, some researchers have begun to focus on the combined detection of diﬀerent types of interleukins or the combined detection of interleukins and other biochemical indicators. It is reported that the combination of IL-13 and IL-8 is highly eﬃcient in identify thyroid diseases (AUC 0.90) (49). Another study evidenced that a panel of four serum biomarkers (IL-8, HGF, MIG, and IL-12) might assist in discriminating thyroid cancer and benign thyroid diseases (AUC 0.81) (55). IL-1 consists of IL-1α, IL-1β, and IL1 receptor antagonist (IL1ra). Compared with healthy controls, IL-1β was found to be underexpressed in the serum of patients with PTC, and it was considered to be a valuable factor in discriminating atrophic thyroiditis and thyroid cancer (47). IL1ra inhibits the pro- inﬂammatory eﬀects of IL-1 through competitively binding to the IL-1 receptor (61). Compared with the control group, the serum level of IL1ra is higher in patients with FTC or ATC, but not in patients with PTC or MTC (48), which indicates that serum IL- 1ra level could be used as a biomarker for FTC or ATC. However, this eﬀect reaches signiﬁcance only in women. It could due to the fact that production of IL-1ra in monocytes from female patients is increased (62). CLINICAL UTILITIES OF INTERLEUKINS IN THYROID CANCER Thyroid cancer is the most common malignant tumor of endocrine system with increasing incidence over recent decades (1). How to improve the diagnosis and treatment quality of thyroid cancer has been paid more attention by clinicians. In clinical practice, the diagnosis, treatment, prognosis evaluation, and disease recurrence monitoring of thyroid cancer are challenging. Either in process of initial diagnosis or postoperative pathological classiﬁcation, discriminating benign and malignant thyroid nodules is not easy. In order to achieve the purpose of early detection, early diagnosis and early treatment, determination of the high-risk population is signiﬁcant in clinical practice. Evaluation of aggressiveness of thyroid cancer can provide more information for accurate risk stratiﬁcation of patients. In addition, most thyroid cancer progress slowly and patients has long survival time after treatments, thus the prognosis evaluation and follow-up after treatment are particularly important. These problems in clinical practice suggest more eﬀective biochemical indicators are needed to make up for the imperfections of current clinical diagnosis and treatment of thyroid cancer in order to improve the level of clinical diagnosis and treatment and make patients attain more beneﬁts. This section will concentrate on the possible applications of interleukins in the clinical practice of thyroid cancer in order to provide more help for clinicians (Figure 2). Although the mechanisms of some interleukins in thyroid cancer have been described by some studies, how other interleukins conﬁrmed to be correlated to thyroid cancer in clinical researches aﬀect the progression of thyroid cancer is still not clear so far. Therefore, in order to provide more help for FIGURE 2 \| Clinical applications of interleukins in thyroid cancer. IGURE 2 \| Clinical applications of interleukins in thyroid cancer. IGURE 2 \| Clinical applications of interleukins in thyroid cancer. June 2020 \| Volume 11 \| Article 1124 Frontiers in Immunology \| www.frontiersin.org Frontiers in Immunology \| www.frontiersin.org 5 Interleukins and Thyroid Cancer Xi et al. Xi et al. DISCRIMINATING BENIGN AND MALIGNANT THYROID DISEASES serum expression level of thyroid cancer is higher than that of thyroid adenoma, suggesting that it could be used as a potential biomarker to distinguish thyroid cancer from adenoma (54). Serum Interleukin Level Serum interleukin level is correlated to thyroid cancer and could be used to discriminate benign and malignant thyroid diseases. However, due to the diﬀerent measurement methods and sample size, the results of diﬀerent studies are not similar and even opposite. Besides, there are still some limitations in the eﬀect of single type of interleukin. At present, few studies focused on the combined detection of diﬀerent types of interleukins or the combined detection of interleukins and other biochemical indicators, which also limits the application of serum interleukins in clinical practice (49, 55). Therefore, in order to improve the clinical application value of interleukins, it is necessary to use the same detection methods for multi-center, large sample size research. In addition, studies on the combined detection are also needed. IL-6 is a pleiotropic cytokine which could induce pro- and anti-inﬂammatory eﬀects under speciﬁc conditions. In tumor microenvironment, IL-6 interacts with its receptor and associates with the target cell membrane glycoprotein 130, inducing pro- inﬂammation cytokines production to support the chronic inﬂammation. In addition, the IL-6 pathway could transmit positive signals to tumor cells, promoting the proliferation, invasion and anti-apoptosis of cancer. The published researches reported that IL-6 expression is higher in patients with benign and malignant thyroid neoplasms than healthy controls, and it is associated with tumor aggressiveness and poor survival (49). Besides, serum level of IL-6 is high in presurgical serum samples of PTC patients and returns to normal level following surgery (50). However, a recent study reported that IL-6 could not discriminate benign and malignant groups from healthy controls (51). This diﬀerence may due to diﬀerent interleukins measurement methods and sample size. Frontiers in Immunology \| www.frontiersin.org Expression of Interleukins in Thyroid Tissues Eﬀective biomarkers are needed to improve the accuracy of pathological diagnosis of thyroid cancer. Even though serum interleukins could be used to discriminate benign and malignant thyroid diseases, they might not accurately reﬂect their actual expression level in thyroid tissues. Thus, studies on expression level of interleukins in thyroid tissues are still needed. Previous studies showed that IL-6 expression was signiﬁcantly down- regulated in undiﬀerentiated thyroid cancer tissues (56). Besides, PTC tissues had the highest level of IL-6 expression while FTC and ATC issues were consistently negative for IL-6 expression (57). Compared with benign or normal thyroid tissues, higher expressions of IL-13 (58) in PTC tissues, higher expressions of IL- 17 (59) and IL-23 (36) in DTC and MTC tissues were observed. Besides, the expression of IL-32 (60) in thyroid cancer issues is also higher in benign thyroid tissues. The ﬁndings of serum IL-8 level in thyroid cancer are contradictory in diﬀerent studies. Some studies found that the serum level of IL-8 is lower in thyroid disease than healthy controls (49), while others found that the level of IL-8 expression in serum samples of PTC patients is higher than healthy controls (50). Besides, a recent study (51) reported that compared with the benign thyroid disease groups, higher concentration of IL-8 in thyroid cancer was observed. The measurement methods and sample size in these studies are diﬀerent, which indicates that a multicenter large sample study is needed to determine the serum expression level of IL-8 in thyroid cancer. Serum level of IL-17 increases signiﬁcantly in patients with thyroid tumors compared with healthy controls (52, 53). And its June 2020 \| Volume 11 \| Article 1124 6 ases. amples Sample size Method ILs Findings m samples PTC: 15 Multiplex technology IL-1β Serum IL-1β levels was under expressed in PTC group compared to healthy control group and other thyroid diseases. m samples PTC:21 FTC:8 MTC:12 ATC: 11 ELISA IL-1ra Serum IL-1ra level is associated with the development of ATC and FTC. m samples TC: 20 Multiplex technology (IL)6, 7, 10, 13 Serum level of IL-6, IL-7, IL-10, and IL-13 was higher in thyroid disease, while IL-8 was lower than healthy controls. m samples PTC: 31 ELISA IL-6, IL-8 Serum IL-6 and IL-8 were higher in presurgical thyroid samples and returned to normal following surgery. m samples DTC: 200 ELISA (IL)2, 6R, 8, 12 1. Expression of Interleukins in Thyroid Tissues Serum concentration of IL-2, IL-6R, IL-8, and IL-12 might assist in the characterization of thyroid nodules’ malignancy and tumor aggressiveness. 2. Concentration of serum IL-8 was higher in the malignant group compared with controls. 3. IL-6 and IL-10 were not able to discriminate the benign and malignant groups from controls. 4. Serum concentrations of IL-2, 2R and 10 were associated with thyroglobulin levels. m samples PTC: 29 FTC: 11 ELISA IL-17 Serum IL-17 level was signiﬁcantly increased in patients with DTC. m samples TC: 11 ELISA IL-17 Concentration of serum IL-17 was signiﬁcantly higher in patients with thyroid tumors. m samples PTC: 42 ELISA IL-17 Serum IL-17 level was higher in thyroid cancer than that in thyroid adenoma. m samples TC: 23 Multiplex technology IL-8 and IL-12 A panel of four serum biomarkers (IL-8, HGF, MIG, and IL-12) might assist in the discriminating thyroid cancer and benign thyroid diseases. oid samples TC: 99 IHC IL-6 Down-regulation of serum IL-6 may be a biomarker of UTC. oid samples TC: 130 IHC IL-6 IL-6 is negatively correlated to aggressiveness of thyroid cancer. oid samples PTC: 13 IHC IL-13 PTC expressed higher IL-13 protein and PTC had more IL-13 genetic changes compared to benign nodules. oid samples PTC: 60 PCR IHC IL-17 IL-17 is correlated to tumor TNM stage, capsule invasion, and lymph node metastasis of thyroid cancer. oid samples PTC: 61 FTC: 19 MTC: 8 IHC IL-17 1. Expression of IL-17 proteins in DTC and MTC tissues are higher than healthy controls. 2. High IL-17 expression was associated with recurrence and mortality in thyroid neoplasm. es Thyroid samples TC: 139 HC: 138 IHC PCR IL-32 IL-32 protein was expressed higher in thyroid cancer tissues. cancer; PTC, papillary thyroid cancer; FTC, follicular thyroid cancer; UTC, undifferentiated thyroid cancer; ATC, anaplastic thyroid cancer; MTC, Medullary thyroid cancer; nosorbent assay. Interleukins and Thyroid Cancer Xi et al. higher aggressiveness need more active surgical resection and radioiodine treatment. However, it is not always easy to distinguish these patients from others in initial diagnosis. At present, some studies have explored the possibility of interleukins in predicting the aggressiveness of thyroid cancer and have obtained promising results. The expression of interleukins in thyroid cancer tissues is signiﬁcantly diﬀerent compared with benign or normal thyroid tissues. Evaluating the Risk of Tumorigenesis of Thyroid Cancer Lymph node metastasis is the most common type of thyroid cancer metastasis. Most patients are detected with multiple lymph node metastases at the time of initial diagnosis. It is reported that interleukins are signiﬁcantly associated with lymph node metastasis of thyroid cancer. The expression level of IL-8 is higher in thyroid cancer tissues with lymph node metastases than that without lymph node metastases (75). Gene polymorphism evidenced that IL-8 may contribute to DTC lymph node metastasis (76), and IL-1β may cause PTC lymph node metastasis (71). However, gene polymorphism IL-18 (69) and IL-27 (73) are negatively associated with lymph node metastasis in patients with PTC. y Heritable factors are crucial to the occurrence and development of cancers. Several single nucleotide polymorphisms (SNPs) found in cytokine genes aﬀect the expression or function of proteins which have been evaluated for their roles in cancer predisposition (63). Considering that genetically inherited predisposition is the initiating factor to thyroid cancer occurrence, some studies have focused on gene polymorphisms of interleukins and found some favorable results. The interleukins of whose SNPs are associated with the increased risk of PTC include IL-1α (64), IL6 (65, 66), IL10 (67, 68), IL-18 (69), and IL-27 (70). IL-1β (64, 71) and IL-21 (72) gene polymorphisms were related to the decreased risk of tumorigenesis in thyroid cancer. However, another study found that IL-27 gene polymorphism was not a risk factor of tumorigenesis but a risk factor lymph node metastasis in PTC (73). Patient selection bias may account for this diﬀerence. Moreover, the study focused on the serum interleukins has shown that high serum IL-10 level was positively associated with an increased risk of thyroid cancer, but it was signiﬁcant only in women (74). The correlation between interleukins and aggressiveness of thyroid cancer has been conﬁrmed, indicating that interleukins might be used to predict the aggressiveness of thyroid cancer and provide some information for clinicians to make suitable treatment decisions. Distant metastasis is the main cause of disease-speciﬁc death in thyroid cancer patients (77). None of the above studies found the correlation between interleukins and distant metastasis of thyroid cancer. Considering the small proportion of patients with distant metastasis in these studies, the results may have some limitations. Therefore, further studies should focus on the patients with distant metastasis, and determine whether interleukins can be used to predict the risk of distant metastasis. Evaluating the Risk of Tumorigenesis of Thyroid Cancer Multiple genetic studies have demonstrated the association between interleukins and the risk of thyroid cancer. Therefore, interleukin gene testing of high-risk populations, including patients with family history of thyroid cancer or radiation exposure, could help to assess the risk of thyroid cancer more accurately. Due to the diﬀerences of heritable factors among diﬀerent populations, genetic studies are needed in diﬀerent population to determine the clinical value of interleukins in the risk of thyroid cancer. In addition, interleukins in human body are aﬀected by many other diseases, such as thyroiditis, immune diseases, or other malignancy. However, there are no reports on the tumor risk assessment of patients suﬀering from thyroid cancer combined with other diseases. Furthermore, it is unclear how these genetic polymorphisms aﬀect the function and production of interleukins. Expression of Interleukins in Thyroid Tissues Therefore, interleukins could be potential biomarkers in the pathological diagnosis of thyroid cancer, and a simple immunohistochemical analysis in thyroid tissues could help pathologists discriminate benign and malignant thyroid disease accurately. However, similar to serum interleukins, further studies on the expression of diﬀerent interleukins and other biomarkers in thyroid cancer are needed to improve their clinical value. High levels of serum IL-2, IL-10, and IL-12 are correlated to the aggressive tumor characteristics in patients with DTC (51). Another study also reported that higher levels of positive expression of IL-10 in thyroid cancer tissues were signiﬁcantly correlated to extrathyroidal invasion and larger tumor size (39). Recently, high expression of IL-13RA2 was also observed to be correlated with advanced tumor stage in PTC tissues (30). Prognosis Evaluation and Recurrence Surveillance Considering the utility in predicting recurrence and mortality of thyroid cancer, postoperative risk estimation is recommended to guide radioiodine therapy and follow-up strategies (78). Molecule proﬁle enriches the risk estimate system and is considered a prognostic factor of thyroid cancer. More and more studies reported that interleukins are associated with survival of thyroid cancer patients and are considered as potential predictors of prognosis. Elevated serum IL-6 level is signiﬁcantly associated with poor overall survival in PTC patients (79). Besides, higher expression of IL-10 in cancer tissues (39) and IL-17 in serum (36) are related to shorter recurrence-free survival of thyroid cancer patients. Frontiers in Immunology \| www.frontiersin.org Predicting the Aggressiveness of Thyroid Cancer Clinicopathological factors such as tumor size, extrathyroid extension, lymph node metastasis, and distant metastasis are common indicators of the aggressiveness of thyroid cancer. Aggressiveness is a signiﬁcant factor for tumor risk stratiﬁcation in clinical practice. Thyroid cancer patients with Although the prognosis of DTC patients is relatively favorable, the recurrence rate after initial treatment reaches up to 8–23% (80). Therefore, recurrence monitoring during follow-up is of June 2020 \| Volume 11 \| Article 1124 Frontiers in Immunology \| www.frontiersin.org 8 Interleukins and Thyroid Cancer Xi et al. gene transfer and immunotherapy provide promising results for gene therapy for thyroid cancer. An adenoviral vector expressing thymidine kinase of herpes simplex virus (HSV-TK) and IL-2 (AdCMVTKhIL2) was constructed, and about 63% of MTC tumors were destroyed after intratumoral injection of itAdCMVTKhIL2 and the antitumor eﬀect of AdCMVTKhIL2 is superior than each single vector (86). Besides, a retroviral vector expressing HSV-TK and IL-2 could completely eradicate ATC tumors and reduce DTC tumor size by more than 80% (87). To further optimize this therapeutic approach, new vector was constructed by replacing the viral enhancer with the enhancer sequence of the human Tg gene. This new vector allows selective transgene expression and cell killing in DTC cells but not in ATC cells (88). In addition, an in vitro and vivo study suggested that combination of IL-12 and lanreotide (LAN), a somatostatin analog, could suppress and kill MTC cells and improve quality of life of MTC patients (89). great signiﬁcance to provide timely and accurate information to patients for intensive treatment. Serum thyroglobulin (Tg) level after thyroidectomy and iodine ablation is the most eﬀective indicator of disease recurrence during follow-up, with high sensitivity and accuracy. Interleukins such as IL-2 and IL-10 were correlated to tumor aggressiveness and to serum Tg level, which suggests the involvement of interleukins would improve the eﬃciency of Tg evaluation system (51). For DTC patients with autoimmune disease or post-trauma immune system response, the evaluation eﬃciency of Tg is greatly reduced due to interference of thyroglobulin antibody (TgAb). Thus, eﬀective indicators are needed to monitor tumor recurrence in these patients. In PTC patients with or without Hashimoto’s thyroiditis, serum IL-4 and IL-10 levels were reported to be higher in cases with persistent or recurrent disease than those without persistent or recurrent disease, suggesting that they could select patients who need close monitoring and intensive treatment (81). Predicting the Aggressiveness of Thyroid Cancer p RET mutations are responsible for the course of human cancers (90). RET stimulation with its ligand GDNF induced IL-8 production in thyroid cancer (97), In MTC animal model, the decrease of serum IL-8 level is induced by RET inhibitor Sunitinib (91). IL-8 promotes the proliferation, survival, invasion, and angiogenesis of tumor. Therefore, inhibiting activation of RETof thyroid cancer, which results in the reduction of IL-8 secretion, may be an eﬀective strategy for thyroid cancer treatment. And the decreased expression of IL-8 could be used to evaluate the curative eﬀect of RET- inhibited treatment. It is gradually accepted that cancer stem cells (CSC) promote tumor growth, metastasis, recurrence and drug-resistance. CSC is more abundant in ATC sample than DTC (98). In human PTC specimens, a signiﬁcant correlation between Mast cell (MC) density and stemness features was observed. It is reported that MC-dependent IL-8–Akt–Slug pathway that sustains EMT/stemness of thyroid cancer cells (23). Thus, the IL- 8-CXCR1/2 axis might be used as a targeted therapy in advanced thyroid cancer. Taken together, interleukins could be potential biomarkers used in prognosis evaluation and disease surveillance of thyroid cancer. The risk stratiﬁcation information based on interleukin levels and clinicopathologic features could guide follow-up management decisions of patients with thyroid cancer. However, many other diseases also aﬀect the production and function of interleukins in human body. Therefore, more studies are needed on thyroid cancer patients combined with other diseases to further conﬁrm the potential utility of interleukins in prognosis evaluation and recurrence surveillance of thyroid cancer. CURRENT SITUATION OF INTERLEUKINS IN TREATMENT OF THYROID CANCER Eighty percentage of DTC patients respond well to the current combined treatment model of surgery, RAI therapy, and TSH suppression. However, ∼10% of DTC patients have distant metastasis at the time of diagnosis or develop distant metastasis during follow-up, and these patients usually have poor response to treatments (3). Besides, some DTC patients gradually progress to be refractory to RAI (RR-DTC) during treatment. There is no eﬀective treatment for aggressive pathological types of thyroid cancer, such as ATC. Therefore, new therapy model for these patients are needed. Over recent decades, it has been suggested that some interleukins have good prospects in the treatment of thyroid cancer (Table 3). Due to its most signiﬁcant antitumor activity among all cytokines, IL-12 is one of the most promising interleukins in clinical application. Both IL-12 gene therapy and recombinant protein therapy inhibit thyroid cancer growth and prolonged survival (99). However, systemic administration of recombinant IL-12 caused dose-dependent toxicity in animals. In order to alleviate this situation, some studies have investigated the local expression of the IL-12 induced by transduction of IL-12 expressing vectors into the tumor tissue. Injection of AdTCPmIL-12 (92) and AdCMVmIL-12 (93) into MTC tumors has antitumor eﬀect to both primary and distant lesions, and long-term antitumor immune was established. The combination of AdTCPtk and AdTCPmIL-12 has stronger antitumor eﬀect to MTC than each single vector in vivo study (94). In addition, AdCMVIL-12 has the same antitumor eﬀect in FTC animal model (95). Besides, a single-chain IL-12 fusion protein has antitumor eﬀect in ATC animal model, however lone-time tumor immune was not observed in this study (96). IL-2 exerts its antitumor eﬀect by activating cytotoxic T lymphocytes, and natural killer (NK) cells. Genetic immunotherapy of IL-2 has a promising prospect in the application of the treatment of thyroid cancer, especially for ATC and MTC (82, 83). In animal models of MTC, about 42.9% of cases were cured by directly intratumoral injection of Replication-defective adenovirus expressing IL-2 (AdCMVmIL2) and most of the cured rats developed systemic immunity (84). Several studies have evaluated the possibility of IL-2 combined with other treatments to enhance the antitumor eﬀects. The tumor growth can be inhibited and even eliminated after injection of IL-2 or IL-4. Moreover, IL-2 and IL-4 have a synergistic inhibitory eﬀect (85). Frontiers in Immunology \| www.frontiersin.org CURRENT SITUATION OF INTERLEUKINS IN TREATMENT OF THYROID CANCER The combination of suicide IL13Rα2 could serve as a target of therapeutic intervention of some malignant tumor and current trials mainly involve glioblastoma multiforme (100). In thyroid cancer, IL13Rα2 was June 2020 \| Volume 11 \| Article 1124 Frontiers in Immunology \| www.frontiersin.org 9 Interleukins and Thyroid Cancer Xi et al. TABLE 3 \| Current situation of treatments targeting on interleukins in thyroid cancer. Authors Year Samples ILs Zhang et al. (84) 1999 Wag/Rij rats IL-2 Cressent. et al.(85) 1995 WagRij rats IL-2 Zhang and DeGroot (86) 2001 Wag/Rij rats IL-2 Barzon et al. (87) 2003 nude mice IL-2 Barzon et al. (88) 2002 nude mice IL-2 Vitale et al. (89) 2013 TT cells MTC patients IL-2 Iwahashi et al. (90) 2002 TPC-1 and TT cell line IL-8 Broutin et al. (91) 2011 TT cell line Animal model IL-8 Yamazaki et al. (92) 2002 Animal model IL-12 Zhang and DeGroot (93) 2000 Animal model IL-12 Yamazaki et al. (94) 2004 WAG/Rij rats IL-12 Zhang and DeGroot (95) 2003 Rats model IL-12 Shi et al. (96) 2003 Nude mice IL-12 Gu (31) 2018 TPC-1 and ARO cell line IL13Rα2 ILs, interleukins; MTC, medullary thyroid cancer. TABLE 3 \| Current situation of treatments targeting on interleukins in thyroid cancer. Authors Year Samples ILs Findings Zhang et al. (84) 1999 Wag/Rij rats IL-2 1. AdCMVmIL2 has antitumor effects and could establish tumor immune in MTC animal model. 2. AdCMVmIL2 has low toxicity Cressent. et al.(85) 1995 WagRij rats IL-2 1. Injection of IL-2 or IL-4 inhibit the growth of tumor in MTC animal model. 2. IL-2 and IL-4 were synergistic in their inhibitory effects Zhang and DeGroot (86) 2001 Wag/Rij rats IL-2 1. AdCMVTKhIL2 destroyed 63% of tumors in MTC animal model. 2. The antitumor effect of AdCMVTKhIL2 is superior than each single vector Barzon et al. (87) 2003 nude mice IL-2 A retroviral vector expressing HSV-TK and IL-2 completely eradicate tumors of ATC and reduce more than 80% tumor size of DTC Barzon et al. (88) 2002 nude mice IL-2 A new targeted vector of which viral enhancer replaced by human Tg gene enhancer has antitumor effect to DTC Vitale et al. (89) 2013 TT cells MTC patients IL-2 Combination of IL-12 and LAN could suppress MTC cells and improve quality of life in MTC patients Iwahashi et al. CURRENT SITUATION OF INTERLEUKINS IN TREATMENT OF THYROID CANCER (90) 2002 TPC-1 and TT cell line IL-8 RET induces IL-8 production from both PTC and MTC cells through many signal pathways Broutin et al. (91) 2011 TT cell line Animal model IL-8 Sunitinib decreases the serum level of IL-8 in mice model Yamazaki et al. (92) 2002 Animal model IL-12 1. AdTCPmIL-12 has antitumor effects on tumor in MTC animal model. 2. AdTCPmIL-12 induces tumor bearing rats to establish long-time tumor immune Zhang and DeGroot (93) 2000 Animal model IL-12 1. AdCMVmIL-12 has antitumor effects on tumor in MTC animal model. 2. AdCMVmIL-12 induced tumor bearing rats to establish long-time tumor immune Yamazaki et al. (94) 2004 WAG/Rij rats IL-12 The combination of AdTCPtk and AdTCPmIL-12 has stronger antitumor effects on MTC than each single vector in MTC animal model Zhang and DeGroot (95) 2003 Rats model IL-12 AdCMVIL-12 has antitumor effects in FTC animal model Shi et al. (96) 2003 Nude mice IL-12 1. A single-chain IL-12 fusion protein has antitumor effects in ATC animal model. 2. Long-time tumor immune was not observed in this study Gu (31) 2018 TPC-1 and ARO cell line IL13Rα2 The number of invading cells declined signiﬁcantly after the knockdown of IL IL13Rα2 ILs interleukins; MTC medullary thyroid cancer TABLE 3 \| Current situation of treatments targeting on interleukins in thyroid cancer. cancer in order to ﬁnd more potential targets for the treatment of thyroid cancer. Besides, the interleukins, which has been proved to be eﬀective in basic researches, should be studied in combination with other drugs. observed to promote aggressive behavior of tumor through promoting EMT (30). The number of invading cells declined signiﬁcantly after the knockdown of IL IL13Rα2, indicating the possibility of IL13Rα2 used as a novel target in the treatment of thyroid cancer (31). However, no study has reported IL13Rα2 could be used in treatment of thyroid cancer. Frontiers in Immunology \| www.frontiersin.org ILs, interleukins; MTC, medullary thyroid cancer. REFERENCES Current approaches and perspectives in the therapy of medullary thyroid carcinoma. Cancer. (2001) 91:1797–808. doi: 10.1002/ 1097-0142(20010501)91:9<1797::aid-cncr1199>3.0.co;2-p 7. Vose BM. Expansion of autorecognitive cytotoxic eﬀectors in human cancer by T cell growth factor (Interleukin 2)1. Arch Geschwulstforsch. (1981) 51:317–26. 20. Mei Z, Zhou L, Zhu Y, Jie K, Fan D, Chen J, et al. Interleukin- 22 promotes papillary thyroid cancer cell migration and invasion through microRNA-595/Sox17 axis. Tumour Biol. (2016) 37:11753– 62. doi: 10.1007/s13277-016-5030-1 8. Fasoulakis Z, Kolios G, Papamanolis V, Kontomanolis EN. Interleukins associated with breast cancer. Cureus. (2018) 10:e3549. doi: 10.7759/cureus.3549 21. Yang J, Weinberg RA. Epithelial-mesenchymal transition: at the crossroads of development and tumor metastasis. Dev Cell. (2008) 14:818–29. doi: 10.1016/j.devcel.2008.05.009 9. Yang H, Xuefeng Y, Jianhua X. Systematic review of the roles of interleukins in hepatocellular carcinoma. Clin Chim Acta. (2020) 506:33– 43. doi: 10.1016/j.cca.2020.03.001 22. Waugh DJ, Wilson C. The interleukin-8 pathway in cancer. Clin Cancer Res. (2008) 14:6735–41. doi: 10.1158/1078-0432.CCR- 07-4843 10. Melillo RM, Castellone MD, Guarino V, De Falco V, Ciraﬁci AM, Salvatore G, et al. The RET/PTC-RAS-BRAF linear signaling cascade mediates the motile and mitogenic phenotype of thyroid cancer cells. J Clin Invest. (2005) 115:1068–81. doi: 10.1172/JCI22758 23. Visciano C, Liotti F, Prevete N, Cali G, Franco R, Collina F, et al. Mast cells induce epithelial-to-mesenchymal transition and stem cell features in human thyroid cancer cells through an IL-8-Akt-Slug pathway. Oncogene. (2015) 34:5175–86. doi: 10.1038/onc.2014.441 11. Ameri P, Ferone D. Diﬀuse endocrine system, neuroendocrine tumors and immunity: what’s new? Neuroendocrinology. (2012) 95:267–76. doi: 10.1159/000334612 24. Taga T, Kishimoto T. Gp130 and the interleukin-6 family of cytokines. Annu Rev Immunol. (1997) 15:797–819. doi: 10.1146/annurev.immunol.15.1.797 12. Weber A, Wasiliew P, Kracht M. Interleukin-1 (IL-1) pathway. Sci Signal. (2010) 3:cm1. doi: 10.1126/scisignal.3105cm1 25. Onnis B, Fer N, Rapisarda A, Perez VS, Melillo G. Autocrine production of IL-11 mediates tumorigenicity in hypoxic cancer cells. J Clin Invest. (2013) 123:1615–29. doi: 10.1172/JCI59623 13. Baker KJ, Houston A, Brint E. IL-1 family members in cancer; two sides to every story. Front Immunol. (2019) 10:1197. doi: 10.3389/ﬁmmu.2019. 01197 26. Zhong Z, Hu Z, Jiang Y, Sun R, Chen X, Chu H, et al. Interleukin-11 promotes epithelial-mesenchymal transition in anaplastic thyroid carcinoma cells through PI3K/Akt/GSK3beta signaling pathway activation. Oncotarget. (2016) 7:59652–63. doi: 10.18632/oncotarget.10831 14. Inokuchi N, Zeki K, Morimoto I, Nakano Y, Fujihira T, Yamashita U, et al. CONCLUSION Interleukins have good prospects in the treatment of thyroid cancer. Studies mentioned above have reported that some interleukins could be used in the treatment of thyroid cancer and evaluation of curative eﬀect. However, researches on the treatment of thyroid cancer by interleukin are limited to cell and animal researches. This may be because the mechanism of interleukins in thyroid cancer is not fully understood. Studies should further explore the mechanisms of interleukins in thyroid Basic researches have conﬁrmed that interleukins have signiﬁcant roles in thyroid cancer through diﬀerent potential mechanisms. Few interleukins exhibit anti-tumor roles in thyroid cancer, while most show pro-tumor eﬀects. Interleukins have promising prospect in clinical practice of thyroid cancer. They enhance the accuracy of diagnosis of thyroid cancer and provide novel June 2020 \| Volume 11 \| Article 1124 Frontiers in Immunology \| www.frontiersin.org 10 Xi et al. Xi et al. Interleukins and Thyroid Cancer ACKNOWLEDGMENTS We thank Liwen Bianji, Edanz Editing China (www.liwenbianji. cn/ac), for editing the English text of this manuscript. FUNDING This review was supported by the National Natural Science Foundation of China, under grant no. 81771865, and by the Shanghai Key Discipline of Medical Imaging, under grant no. 2017ZZ02005. AUTHOR CONTRIBUTIONS treatments approaches. Besides, interleukins may also be used as biomarkers of surveillance for recurrence and serve as prognostic factors in thyroid cancer. However, there still some limitations on the clinical applications of interleukins in thyroid cancer. Firstly, only a few interleukins in thyroid cancer have been described, and potential mechanisms of other interleukins are not very clear. Secondly, there are some diﬀerences in the results of the clinical study on the utility of interleukins in thyroid cancer, which may be due to the diﬀerent methods and diﬀerent sample size. In addition, interleukins level in human is aﬀected by many diseases, and to date there is no studies focusing on the accuracy and clinical value of interleukins in discriminating thyroid tumors combined with other diseases. Finally, the use of interleukins in the treatment of thyroid cancer is limited to cell and animal researches. However, what certain is that with future researches, interleukins will provide more help for clinicians and patients. treatments approaches. Besides, interleukins may also be used as biomarkers of surveillance for recurrence and serve as prognostic factors in thyroid cancer. However, there still some limitations on the clinical applications of interleukins in thyroid cancer. Firstly, only a few interleukins in thyroid cancer have been described, and potential mechanisms of other interleukins are not very clear. Secondly, there are some diﬀerences in the results of the clinical study on the utility of interleukins in thyroid cancer, which may be due to the diﬀerent methods and diﬀerent sample size. In addition, interleukins level in human is aﬀected by many diseases, and to date there is no studies focusing on the accuracy and clinical value of interleukins in discriminating thyroid tumors combined with other diseases. Finally, the use of interleukins in the treatment of thyroid cancer is limited to cell and animal researches. However, what certain is that with future researches, interleukins will provide more help for clinicians and patients. All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication. REFERENCES thyroid carcinoma cell line, NIM 1. Jpn J Cancer Res. (1995) 86:670–6. doi: 10.1111/j.1349-7006.1995.tb02451.x thyroid carcinoma cell line, NIM 1. Jpn J Cancer Res. (1995) 86:670–6. doi: 10.1111/j.1349-7006.1995.tb02451.x 1. Harmer C. Foreword for special issue on thyroid cancer 2017. Clin Oncol. (2017) 29:275. doi: 10.1016/j.clon.2017.01.008 1. Harmer C. Foreword for special issue on thyroid cancer 2017. Clin Oncol. (2017) 29:275. doi: 10.1016/j.clon.2017.01.008 15. Kimura H, Yamashita S, Namba H, Tominaga T, Tsuruta M, Yokoyama N, et al. Interleukin-1 inhibits human thyroid carcinoma cell growth. J Clin Endocrinol Metab. (1992) 75:596–602. doi: 10.1210/jcem.75.2. 1322431 2. Nagataki S, Nystrom E. Epidemiology and primary prevention of thyroid cancer. Thyroid. (2002) 12:889–96. doi: 10.1089/10507250276 1016511 16. Yip I, Pang XP, Berg L, Hershman JM. Antitumor actions of interferon-gamma and interleukin-1 beta on human papillary thyroid carcinoma cell lines. J Clin Endocrinol Metab. (1995) 80:1664–9. doi: 10.1210/jcem.80.5.7745015 3. Mazzaferri EL, Jhiang SM. Long-term impact of initial surgical and medical therapy on papillary and follicular thyroid cancer. Am J Med. (1994) 97:418– 28. doi: 10.1016/0002-9343(94)90321-2 17. Ohta K, Pang XP, Berg L, Hershman JM. Antitumor actions of cytokines on new human papillary thyroid carcinoma cell lines. J Clin Endocrinol Metab. (1996) 81:2607–12. doi: 10.1210/jcem.81.7.86 75585 4. Ito K, Hanamura T, Murayama K, Okada T, Watanabe T, Harada M, et al. Multimodality therapeutic outcomes in anaplastic thyroid carcinoma: improved survival in subgroups of patients with localized primary tumors. Head Neck. (2012) 34:230–7. doi: 10.1002/hed.21721 5. Molinaro E, Romei C, Biagini A, Sabini E, Agate L, Mazzeo S, et al. Anaplastic thyroid carcinoma: from clinicopathology to genetics and advanced therapies. Nat Rev Endocrinol. (2017) 13:644–60. doi: 10.1038/nrendo.2017.76 18. Lin JD, Chao TC, Weng HF, Lin KD. The roles of cytokines and retinoic acid in the regulation of human thyroid cancer cell growth. Cytokine. (1998) 10:536–9. doi: 10.1006/cyto.1997.0323 19. Kotenko SV, Izotova LS, Mirochnitchenko OV, Esterova E, Dickensheets H, Donnelly RP, et al. Identiﬁcation of the functional interleukin-22 (IL-22) receptor complex: the IL-10R2 chain (IL-10Rbeta) is a common chain of both the IL-10 and IL-22 (IL-10-related T cell-derived inducible factor, IL-TIF) receptor complexes. J Biol Chem. (2001) 276:2725– 32. doi: 10.1074/jbc.M007837200 6. Vitale G, Caraglia M, Ciccarelli A, Lupoli G, Abbruzzese A, Tagliaferri P, et al. Current approaches and perspectives in the therapy of medullary thyroid carcinoma. Cancer. (2001) 91:1797–808. doi: 10.1002/ 1097-0142(20010501)91:9<1797::aid-cncr1199>3.0.co;2-p 6. Vitale G, Caraglia M, Ciccarelli A, Lupoli G, Abbruzzese A, Tagliaferri P, et al. REFERENCES Stimulatory eﬀect of interleukin-1 alpha on proliferation through a Ca2+/calmodulin-dependent pathway of a human June 2020 \| Volume 11 \| Article 1124 Frontiers in Immunology \| www.frontiersin.org 11 Xi et al. Interleukins and Thyroid Cancer 27. Arima K, Sato K, Tanaka G, Kanaji S, Izuhara K. Characterization of the interaction between interleukin-13 and interleukin-13 receptors. J Biol Chem. (2005) 280:24915–22. doi: 10.1074/jbc.M502571200 45. Mitsiades N, Poulaki V, Tseleni-Balafouta S, Koutras DA, Stamenkovic I. Thyroid carcinoma cells are resistant to FAS-mediated apoptosis but sensitive to tumor necrosis factor-related apoptosis-inducing ligand. Cancer Res. (2000) 60:4122–9. doi: 10.1097/00002820-200008000-00010 28. Mintz A, Gibo DM, Slagle-Webb B, Christensen ND, Debinski W. IL- 13Ralpha2 is a glioma-restricted receptor for interleukin-13. Neoplasia. (2002) 4:388–99. doi: 10.1038/sj.neo.7900234 46. Todaro M, Zerilli M, Ricci-Vitiani L, Bini M, Perez Alea M, Maria Florena A, et al. Autocrine production of interleukin-4 and interleukin-10 is required for survival and growth of thyroid cancer cells. Cancer Res. (2006) 66:1491– 9. doi: 10.1158/0008-5472.CAN-05-2514 29. Fujisawa T, Joshi BH, Puri RK. IL-13 regulates cancer invasion and metastasis through IL-13Ralpha2 via ERK/AP-1 pathway in mouse model of human ovarian cancer. Int J Cancer. (2012) 131:344–56. doi: 10.1002/ ijc.26366 47. Kammoun-Krichen M, Bougacha-Elleuch N, Mnif M, Bougacha F, Charﬀedine I, Rebuﬀat S, et al. IL-1beta a potential factor for discriminating between thyroid carcinoma and atrophic thyroiditis. Eur Cytokine Netw. (2012) 23:101–6. doi: 10.1684/ecn.2012.0312 30. Chong ST, Tan KM, Kok CYL, Guan SP, Lai SH, Lim C, et al. IL13RA2 is diﬀerentially regulated in papillary thyroid carcinoma versus follicular thyroid carcinoma. J Clin Endocrinol Metab. (2019) 104:5573– 84. doi: 10.1210/jc.2019-00040 48. Niedzwiecki S, Stepien T, Kuzdak K, Stepien H, Krupinski R, Seehofer D, et al. Serum levels of interleukin-1 receptor antagonist (IL-1ra) in thyroid cancer patients. Langenbecks Arch Surg. (2008) 393:275– 80. doi: 10.1007/s00423-007-0251-9 31. Gu M. IL13Ralpha2 siRNA inhibited cell proliferation, induced cell apoptosis, and suppressed cell invasion in papillary thyroid carcinoma cells. Onco Targets Ther. (2018) 11:1345–52. doi: 10.2147/OTT. S153703 49. Provatopoulou X, Georgiadou D, Sergentanis TN, Kalogera E, Spyridakis J, Gounaris A, et al. Interleukins as markers of inﬂammation in malignant and benign thyroid disease. Inﬂamm Res. (2014) 63:667– 74. doi: 10.1007/s00011-014-0739-z 32. Bauerle KT, Schweppe RE, Haugen BR. Inhibition of nuclear factor-kappa B diﬀerentially aﬀects thyroid cancer cell growth, apoptosis, and invasion. Mol Cancer. (2010) 9:117. doi: 10.1186/1476-4598-9-117 50. Beksac K, Sonmez C, Cetin B, Kismali G, Sel T, Tuncer Y, et al. REFERENCES Evaluation of proinﬂammatory cytokine and neopterin levels in women with papillary thyroid carcinoma. Int J Biol Markers. (2016) 31:e446– 50. doi: 10.5301/jbm.5000214 33. Bauerle KT, Schweppe RE, Lund G, Kotnis G, Deep G, Agarwal R, et al. Nuclear factor kappaB-dependent regulation of angiogenesis, and metastasis in an in vivo model of thyroid cancer is associated with secreted interleukin-8. J Clin Endocrinol Metab. (2014) 99:E1436–44. doi: 10.1210/ jc.2013-3636 51. Martins MB, Marcello MA, Batista FA, Peres KC, Meneghetti M, Ward MAL, et al. Serum interleukin measurement may help identify thyroid cancer patients with active disease. Clin Biochem. (2017) 52:1– 7. doi: 10.1016/j.clinbiochem.2017.10.003 34. Ouyang W, Kolls JK, Zheng Y. The biological functions of T helper 17 cell eﬀector cytokines in inﬂammation. Immunity. (2008) 28:454– 67. doi: 10.1016/j.immuni.2008.03.004 52. Jiang G, Ma S, Wei Y, Wu Y, Yu X, Liu H. The prevalence and distribution of Th17 and Tc17 cells in patients with thyroid tumor. Immunol Lett. (2014) 162(1 Pt A):68–73. doi: 10.1016/j.imlet.2014.07.005 j 35. Numasaki M, Lotze MT, Sasaki H. Interleukin-17 augments tumor necrosis factor-alpha-induced elaboration of proangiogenic factors from ﬁbroblasts. Immunol Lett. (2004) 93:39–43. doi: 10.1016/j.imlet.2004.01.014 53. Zhang L, Chen J, Xu C, Qi L, Ren Y. Eﬀects of iodine-131 radiotherapy on Th17/Tc17 and Treg/Th17 cells of patients with diﬀerentiated thyroid carcinoma. Exp Ther Med. (2018) 15:2661–6. doi: 10.3892/etm. 2017.5663 36. Carvalho DFG, Zanetti BR, Miranda L, Hassumi-Fukasawa MK, Miranda- Camargo F, Crispim JCO, et al. High IL-17 expression is associated with an unfavorable prognosis in thyroid cancer. Oncol Lett. (2017) 13:1925– 31. doi: 10.3892/ol.2017.5638 54. Lu Y, Yuan Y. Serum level of interleukin-17 and interleukin-35 as a biomarker for diagnosis of thyroid cancer. J Cancer Res Ther. (2015) 11(Suppl. 2):C209–11. doi: 10.4103/0973-1482.168187 37. Schreiber RD, Old LJ, Smyth MJ. Cancer immunoediting: integrating immunity’s roles in cancer suppression and promotion. Science. (2011) 331:1565–70. doi: 10.1126/science.1203486 55. Linkov F, Ferris RL, Yurkovetsky Z, Marrangoni A, Velikokhatnaya L, Gooding W, et al. Multiplex analysis of cytokines as biomarkers that diﬀerentiate benign and malignant thyroid diseases. Proteomics Clin Appl. (2008) 2:1575–85. doi: 10.1002/prca.200780095 38. Moore KW, de Waal Malefyt R, Coﬀman RL, O’Garra A. Interleukin- 10 and the interleukin-10 receptor. Annu Rev Immunol. (2001) 19:683– 765. doi: 10.1146/annurev.immunol.19.1.683 56. Basolo F, Fiore L, Pollina L, Fontanini G, Conaldi PG, Toniolo A. Reduced expression of interleukin 6 in undiﬀerentiated thyroid carcinoma: in vitro and in vivo studies. Clin Cancer Res. (1998) 4:381–7. 39. REFERENCES Cunha LL, Morari EC, Nonogaki S, Marcello MA, Soares FA, Vassallo J, et al. Interleukin 10 expression is related to aggressiveness and poor prognosis of patients with thyroid cancer. Cancer Immunol Immunother. (2017) 66:141– 8. doi: 10.1007/s00262-016-1924-4 57. Ruggeri RM, Villari D, Simone A, ScarﬁR, Attard M, Orlandi F, et al. Co-expression of interleukin-6 (IL-6) and interleukin-6 receptor (IL-6R) in thyroid nodules is associated with co-expression of CD30 ligand/CD30 receptor. J Endocrinol Invest. (2002) 25:959–66. doi: 10.1007/bf033 44068 40. Angell TE, Lechner MG, Jang JK, LoPresti JS, Epstein AL. MHC class I loss is a frequent mechanism of immune escape in papillary thyroid cancer that is reversed by interferon and selumetinib treatment in vitro. Clin Cancer Res. (2014) 20:6034–44. doi: 10.1158/1078-0432.CCR- 14-0879 58. Zhao Z, Wei Q, Zhao Y, Sun F, Jin X, Cui B, et al. Genetic copy number alterations and IL-13 expression diﬀerences in papillary thyroid cancers and benign nodules. Endocrine. (2009) 36:155–60. doi: 10.1007/s12020-009-9206-y 41. Han LT, Hu JQ, Ma B, Wen D, Zhang TT, Lu ZW, et al. IL-17A increases MHC class I expression and promotes T cell activation in papillary thyroid cancer patients with coexistent Hashimoto’s thyroiditis. Diagn Pathol. (2019) 14:52. doi: 10.1186/s13000-019-0832-2 59. Jiang XL, Zhang H, Chen YL, Peng L. [Expression of microRNA- 221 and IL-17 in papillary thyroid carcinoma and correlation with clinicopathologic features]. Zhonghua Bing Li Xue Za Zhi. (2017) 46:160– 5. doi: 10.3760/cma.j.issn.0529-5807.2017.03.004 42. Pistritto G, Trisciuoglio D, Ceci C, Garuﬁ A, D’Orazi G. Apoptosis as anticancer mechanism: function and dysfunction of its modulators and targeted therapeutic strategies. Aging. (2016) 8:603–19. doi: 10.18632/aging.100934 60. Plantinga TS, Costantini I, Heinhuis B, Huijbers A, Semango G, Kusters B, et al. A promoter polymorphism in human interleukin-32 modulates its expression and inﬂuences the risk and the outcome of epithelial cell-derived thyroid carcinoma. Carcinogenesis. (2013) 34:1529– 35. doi: 10.1093/carcin/bgt092 43. Vella V, Mineo R, Frasca F, Mazzon E, Pandini G, Vigneri R, et al. Interleukin-4 stimulates papillary thyroid cancer cell survival: implications in patients with thyroid cancer and concomitant Graves’ disease. J Clin Endocrinol Metab. (2004) 89:2880–9. doi: 10.1210/jc.2003- 031639 61. Garlanda C, Dinarello CA, Mantovani A. The interleukin-1 family: back to the future. Immunity. (2013) 39:1003–18. doi: 10.1016/j.immuni.2013.11.010 44. Stassi G, Todaro M, Zerilli M, Ricci-Vitiani L, Di Liberto D, Patti M, et al. Thyroid cancer resistance to chemotherapeutic drugs via autocrine production of interleukin-4 and interleukin-10. Cancer Res. (2003) 63:6784– 90. REFERENCES An adenoviral vector expressing functional heterogeneous proteins herpes simplex viral thymidine kinase and human interleukin-2 has enhanced in vivo antitumor activity against medullary thyroid carcinoma. Endocr Relat Cancer. (2001) 8:315–25. doi: 10.1677/erc.0.0080315 69. Chung JH, Lee YC, Eun YG, Chung JH, Kim SK, Chon S, et al. Single nucleotide polymorphism of interleukin-18 and interleukin-18 receptor and the risk of papillary thyroid cancer. Exp Clin Endocrinol Diabetes. (2015) 123:598–603. doi: 10.1055/s-0035-1559780 87. Barzon L, Bonaguro R, Castagliuolo I, Chilosi M, Franchin E, Del Vecchio C, et al. Gene therapy of thyroid cancer via retrovirally-driven combined expression of human interleukin-2 and herpes simplex virus thymidine kinase. Eur J Endocrinol. (2003) 148:73–80. doi: 10.1530/eje.0. 1480073 70. Nie X, Yuan F, Chen P, Pu Y, Zhu J, Wang Y, et al. Association between IL-27 gene polymorphisms and risk of papillary thyroid carcinoma. Biomark Med. (2016) 11:141–9. doi: 10.2217/bmm-2016-0283 71. Ban JY, Kim MK, Park SW, Kwon KH. Interleukin-1 beta polymorphisms are associated with lymph node metastasis in Korean patients with papillary thyroid carcinoma. Immunol Invest. (2012) 41:888–905. doi: 10.3109/08820139.2012.724751 88. Barzon L, Bonaguro R, Castagliuolo I, Chilosi M, Gnatta E, Parolin C, et al. Transcriptionally targeted retroviral vector for combined suicide and immunomodulating gene therapy of thyroid cancer. J Clin Endocrinol Metab. (2002) 87:5304–11. doi: 10.1210/jc.2002-020975 72. Xiao M, Hu S, Tang J, Zhang L, Jiang H. Interleukin (IL)-21 promoter polymorphism increases the risk of thyroid cancer in Chinese population. Gene. (2014) 537:15–9. doi: 10.1016/j.gene.2013.12.050 89. Vitale G, Lupoli G, Guarrasi R, Colao A, Dicitore A, Gaudenzi G, et al. Interleukin-2 and lanreotide in the treatment of medullary thyroid cancer: in vitro and in vivo studies. J Clin Endocrinol Metab. (2013) 98:E1567– 74. doi: 10.1210/jc.2013-1443 73. Zhang S, Gao X, Wang Y, Jia J, Zhang Q, Ji Z. Interleukin 27−964A > G genetic polymorphism and serum IL-27p28 levels in Chinese patients with papillary thyroid cancer. Tumour Biol. (2015) 36:8207– 11. doi: 10.1007/s13277-015-3570-4 90. Jhiang SM. The RET proto-oncogene in human cancers. Oncogene. (2000) 19:5590–7. doi: 10.1038/sj.onc.1203857 74. Dossus L, Franceschi S, Biessy C, Navionis AS, Travis RC, Weiderpass E, et al. Adipokines and inﬂammation markers and risk of diﬀerentiated thyroid carcinoma: The EPIC study. Int J Cancer. (2018) 142:1332– 42. doi: 10.1002/ijc.31172 91. Broutin S, Ameur N, Lacroix L, Robert T, Petit B, Oumata N, et al. Identiﬁcation of soluble candidate biomarkers of therapeutic response to sunitinib in medullary thyroid carcinoma in preclinical models. REFERENCES doi: 10.1016/S0165-4608(02)00795-1 62. Witkin SS, Gerber S, Ledger WJ. Inﬂuence of interleukin-1 receptor antagonist gene polymorphism on disease. Clin Infect Dis. (2002) 34:204– 9. doi: 10.1086/338261 June 2020 \| Volume 11 \| Article 1124 Frontiers in Immunology \| www.frontiersin.org 12 Xi et al. Interleukins and Thyroid Cancer 81. Stanciu AE, Serdarevic N, Hurduc AE, Stanciu MM. IL-4, IL- 10 and high sensitivity-CRP as potential serum biomarkers of persistent/recurrent disease in papillary thyroid carcinoma with/without Hashimoto’s thyroiditis. Scand J Clin Lab Invest. (2015) 75:539–48. doi: 10.3109/00365513.2015.1057895 63. Balasubramanian SP, Azmy IA, Higham SE, Wilson AG, Cross SS, Cox A, et al. Interleukin gene polymorphisms and breast cancer: a case control study and systematic literature review. BMC Cancer. (2006) 6:188. doi: 10.1186/1471-2407-6-188 64. Li H, Duan N, Zhang Q, Shao Y. IL1A & IL1B genetic polymorphisms are risk factors for thyroid cancer in a Chinese Han population. Int Immunopharmacol. (2019) 76:105869. doi: 10.1016/j.intimp.2019.105869 82. Drosten M, Putzer BM. Gene therapeutic approaches for medullary thyroid carcinoma treatment. J Mol Med. (2003) 81:411–9. doi: 10.1007/s00109-003-0455-6 65. Ozgen AG, Karadeniz M, Erdogan M, Berdeli A, Saygili F, Yilmaz C. The (- 174) G/C polymorphism in the interleukin-6 gene is associated with risk of papillary thyroid carcinoma in Turkish patients. J Endocrinol Invest. (2009) 32:491–4. doi: 10.1007/BF03346494 83. Spitzweg C, Morris JC. Gene therapy for thyroid cancer: current status and future prospects. Thyroid. (2004) 14:424– 34. doi: 10.1089/105072504323150732 66. Li H, Dai H, Li H, Li B, Shao Y. Polymorphisms of the highly expressed IL-6 gene in the papillary thyroid cancer susceptibility among Chinese. Curr Mol Med. (2019) 19:443–51. doi: 10.2174/1566524019666190426142432 84. Zhang R, Straus FH, DeGroot LJ. Eﬀective genetic therapy of established medullary thyroid carcinomas with murine interleukin-2: dissemination and cytotoxicity studies in a rat tumor model. Endocrinology. (1999) 140:2152– 8. doi: 10.1210/endo.140.5.6719 67. Erdogan M, Karadeniz M, Ozbek M, Ozgen AG, Berdeli A. Interleukin-10 gene polymorphism in patients with papillary thyroid cancer in Turkish population. J Endocrinol Invest. (2008) 31:750–4. doi: 10.1007/bf03349252 85. Cressent M, Pidoux E, Cohen R, Modigliani E, Roth C. Interleukin- 2 and interleukin-4 display potent antitumour activity on rat medullary thyroid carcinoma cells. Eur J Cancer. (1995) 31A:2379–2384. doi: 10.1016/0959-8049(95)00445-9 68. Cil E, Kumral A, Kanmaz-Ozer M, Vural P, Dogru-Abbasoglu S, Altuntas Y, et al. Interleukin-10-1082 gene polymorphism is associated with papillary thyroid cancer. Mol Biol Rep. (2014) 41:3091–7. doi: 10.1007/s11033-014-3169-7 86. Zhang R, DeGroot LJ. REFERENCES Clin Cancer. Res. (2011) 17:2044–54. doi: 10.1158/1078-0432.CCR-10-2041 75. Liotti F, Collina F, Pone E, La Sala L, Franco R, Prevete N, et al. Interleukin- 8, but not the related chemokine CXCL1, sustains an autocrine circuit necessary for the properties and functions of thyroid cancer. Stem Cells. (2017) 35:135–46. doi: 10.1002/stem.2492 92. Yamazaki M, Zhang R, Straus FH, Messina M, Robinson BG, Hashizume K, et al. Eﬀective gene therapy for medullary thyroid carcinoma using recombinant adenovirus inducing tumor-speciﬁc expression of interleukin- 12. Gene Ther. (2002) 9:64–74. doi: 10.1038/sj.gt.3301617 76. Kilic I, Guldiken S, Sipahi T, Palabiyik O, Akker M, Celik O, et al. Investigation of VEGF and IL-8 gene polymorphisms in patients with diﬀerentiated thyroid cancer. Clin Lab. (2016) 62:2319–25. doi: 10.7754/Clin.Lab.2016.160403 93. Zhang R, DeGroot LJ. Genetic immunotherapy of established tumours with adenoviral vectors transducing murine interleukin-12 (mIL12) subunits in a rat medullary thyroid carcinoma model. Clin Endocrinol. (2000) 52:687– 94. doi: 10.1046/j.1365-2265.2000.01003.x 94. Yamazaki M, Straus FH, Messina M, Robinson BG, Takeda T, Hashizume K, et al. Adenovirus-mediated tumor-speciﬁc combined gene therapy using Herpes simplex virus thymidine/ganciclovir system and murine interleukin- 12 induces eﬀective antitumor activity against medullary thyroid carcinoma. Cancer Gene Ther. (2004) 11:8–15. doi: 10.1038/sj.cgt.7700636 77. Durante C, Haddy N, Baudin E, Leboulleux S, Hartl D, Travagli JP, et al. Long-term outcome of 444 patients with distant metastases from papillary and follicular thyroid carcinoma: beneﬁts and limits of radioiodine therapy. J Clin Endocrinol Metab. (2006) 91:2892–9. doi: 10.1210/jc.2005-2838 78. Haugen BR, Alexander EK, Bible KC, Doherty GM, Mandel SJ, Nikiforov YE, et al. 2015 American thyroid association management guidelines for adult patients with thyroid nodules and diﬀerentiated thyroid cancer: the American thyroid association guidelines task force on thyroid nodules and diﬀerentiated thyroid cancer. Thyroid. (2016) 26:1– 133. doi: 10.1089/thy.2015.0020 Cancer Gene Ther. (2004) 11:8–15. doi: 10.1038/sj.cgt.7700636 95. Zhang R, DeGroot LJ. Gene therapy of a rat follicular thyroid carcinoma model with adenoviral vectors transducing murine interleukin-12. Endocrinology. (2003) 144:1393–8. doi: 10.1210/en.2002-221013 96. Shi Y, Parhar RS, Zou M, Baitei E, Kessie G, Farid NR, et al. Gene therapy of anaplastic thyroid carcinoma with a single- chain interleukin-12 fusion protein. Hum Gene Ther. (2003) 14:1741–51. doi: 10.1089/104303403322611755 79. Kobawala TP, Trivedi TI, Gajjar KK, Patel DH, Patel GH, Ghosh NR. Signiﬁcance of interleukin-6 in papillary thyroid carcinoma. J Thyroid Res. (2016) 2016:6178921. doi: 10.1155/2016/6178921 97. Iwahashi N, Murakami H, Nimura Y, Takahashi M. 100. Brown CE, Badie B, Barish ME, Weng L, Ostberg JR, Chang WC, et al. Bioactivity and safety of IL13Ralpha2-redirected chimeric antigen receptor CD8+ T cells in patients with recurrent glioblastoma. Clin Cancer Res. (2015) 21:4062–72. doi: 10.1158/1078-0432.CCR-15-0428 99. Parhar RS, Zou M, Al-Mohanna FA, Baitei EY, Assiri AM, Meyer BF, et al. IL-12 immunotherapy of Braf(V600E)-induced papillary thyroid cancer in a mouse model. Lab Invest. (2015) 96:89–97. doi: 10.1038/labinvest. 2015.126 98. Todaro M, Iovino F, Eterno V, Cammareri P, Gambara G, Espina V, et al. Tumorigenic and metastatic activity of human thyroid cancer stem cells. Cancer Res. (2010) 70:8874–85. doi: 10.1158/0008-5472.CAN-10-1994 Frontiers in Immunology \| www.frontiersin.org June 2020 \| Volume 11 \| Article 1124 REFERENCES Activation of RET tyrosine kinase regulates interleukin-8 production by multiple signaling pathways. Biochem Biophys Res Commun. (2002) 294:642– 9. doi: 10.1016/S0006-291X(02)00528-4 80. Brito JP, Morris JC, Montori VM. Thyroid cancer: zealous imaging has increased detection and treatment of low risk tumours. BMJ. (2013) 347:f4706. doi: 10.1136/bmj.f4706 June 2020 \| Volume 11 \| Article 1124 Frontiers in Immunology \| www.frontiersin.org 13 Xi et al. Interleukins and Thyroid Cancer 98. Todaro M, Iovino F, Eterno V, Cammareri P, Gambara G, Espina V, et al. Tumorigenic and metastatic activity of human thyroid cancer stem cells. Cancer Res. (2010) 70:8874–85. doi: 10.1158/0008-5472.CAN-10-1994 Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. 99. Parhar RS, Zou M, Al-Mohanna FA, Baitei EY, Assiri AM, Meyer BF, et al. IL-12 immunotherapy of Braf(V600E)-induced papillary thyroid cancer in a mouse model. Lab Invest. (2015) 96:89–97. doi: 10.1038/labinvest. 2015.126 Copyright © 2020 Xi, Zhang, Sun, Song, Shen, Chen, Sun, Qiu and Luo. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. June 2020 \| Volume 11 \| Article 1124 Frontiers in Immunology \| www.frontiersin.org 14
https://openalex.org/W4388821612	https://www.researchsquare.com/article/rs-3566658/latest.pdf	English	null	Targeted knockdown of PGAM5 in synovial macrophages efficiently alleviates osteoarthritis	Research Square (Research Square)	2,023	cc-by	10,146	Zhouyi Duan Department of Orthopedic Surgery, Xin Hua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (SJTUSM), Shanghai 200092, China;National Facility for Translational Medicine Bingjun Zhang Department of Orthopedic Surgery, Xin Hua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (SJTUSM), Shanghai 200092, China;National Facility for Translational Medicine Bingjun Zhang xiaoling zhang xiaoling zhang Xin Hua Hospital Affiliated to Shanghai Jiao Tong University School of MedicineInstitute of health science https://orcid.org/0000-0002-0134-1347 Xin Hua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine https://orcid.org/0000-0001-9151-5146 Ruihan Hao Xin Hua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (SJTUSM) Jia Lv Shanghai Frontiers Science Center of Genome Editing and Cell Therapy, Shanghai Key Laboratory of Regulatory Biology, School of Life Sciences, East China Normal University Jie Yuan The Second Hospital of Shanxi Medical University Ruihan Hao Xin Hua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (SJTUSM) Jia Lv Shanghai Frontiers Science Center of Genome Editing and Cell Therapy, Shanghai Key Laboratory of Regulatory Biology, School of Life Sciences, East China Normal University Jie Yuan The Second Hospital of Shanxi Medical University Bingjun Zhang Department of Orthopedic Surgery, Xin Hua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (SJTUSM), Shanghai 200092, China;National Facility for Translational Medicine Liming Dai Department of Orthopedic Surgery, Xin Hua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (SJTUSM), Shanghai 200092, China;National Facility for Translational Medicine Liming Dai Department of Orthopedic Surgery, Xin Hua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (SJTUSM), Shanghai 200092, China;National Facility for Translational Medicine Yiyun Cheng Page 1/26 Page 1/26 Shanghai Frontiers Science Center of Genome Editing and Cell Therapy, Shanghai Key Laboratory of Regulatory Biology, School of Life Sciences, East China Normal University, Shanghai 200241, P. R. Chi Wei Lu CAS Key Laboratory of Tissue Microenvironment and Tumor, Shanghai Institute of Nutrition and Health University of Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, China Article Keywords: osteoarthritis, macrophage polarization, PGAM5, β-catenin, macrophage-targeted therapy Posted Date: November 20th, 2023 DOI: https://doi.org/10.21203/rs.3.rs-3566658/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Additional Declarations: (Not answered) Version of Record: A version of this preprint was published at Bone Research on March 4th, 2024. See th published version at https://doi.org/10.1038/s41413-024-00318-8. Shanghai Frontiers Science Center of Genome Editing and Cell Therapy, Shanghai Key Laboratory of Regulatory Biology, School of Life Sciences, East China Normal University, Shanghai 200241, P. R. Chi Wei Lu Wei Lu CAS Key Laboratory of Tissue Microenvironment and Tumor, Shanghai Institute of Nutrition and Heal University of Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, China Article Keywords: osteoarthritis, macrophage polarization, PGAM5, β-catenin, macrophage-targeted therapy Posted Date: November 20th, 2023 DOI: https://doi.org/10.21203/rs.3.rs-3566658/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Additional Declarations: (Not answered) Version of Record: A version of this preprint was published at Bone Research on March 4th, 2024. See t CAS Key Laboratory of Tissue Microenvironment and Tumor, Shanghai Institute of Nutrition and Health, University of Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, China Article Keywords: osteoarthritis, macrophage polarization, PGAM5, β-catenin, macrophage-targeted therapy Posted Date: November 20th, 2023 DOI: https://doi.org/10.21203/rs.3.rs-3566658/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Additional Declarations: (Not answered) Version of Record: A version of this preprint was published at Bone Research on March 4th, 2024. See the published version at https://doi.org/10.1038/s41413-024-00318-8. Article Additional Declarations: (Not answered) Version of Record: A version of this preprint was published at Bone Research on March 4th, 2024. See the published version at https://doi.org/10.1038/s41413-024-00318-8. Page 2/26 Abstract Osteoarthritis (OA) is a common degenerative disease worldwide and new therapeutics that target inflammation and the crosstalk between immunocytes and chondrocytes are being developed to prevent and treat OA. These attempts involve repolarizing pro-inflammatory M1 macrophages into the anti- inflammatory M2 phenotype in synovium. In this study, we found that phosphoglycerate mutase 5 (PGAM5) significantly increased in macrophages in OA synovium compared to controls based on histology of human samples and single-cell RNA sequencing results of mice models. To address the role of PGAM5 in macrophages in OA, we found conditional knockout of PGAM5 in macrophages greatly alleviated OA symptoms in vivo and promoted anabolic metabolism of chondrocytes in vitro. Mechanistically, we found that PGAM5 enhanced M1 polarization via AKT-mTOR/p38/ERK pathways, whereas inhibited M2 polarization via STAT6-PPARγ pathway in murine bone marrow-derived macrophages. Furthermore, we found that PGAM5 directly dephosphorylated Dishevelled Segment Polarity Protein 2 (DVL2) which resulted in the inhibition of β-catenin and repolarization of M2 macrophages into M1 macrophages. Conditional knockout of both PGAM5 and β-catenin in macrophages significantly exacerbated osteoarthritis compared to PGAM5-deficient mice. Motivated by these findings, we successfully designed mannose modified fluoropolymers combined with siPGAM5 to inhibit PGAM5 specifically in synovial macrophages via intra-articular injection, which possessed desired targeting abilities of synovial macrophages and greatly attenuated murine osteoarthritis. Collectively, these findings defined a key role for PGAM5 in orchestrating macrophage polarization and provides insights into novel macrophage-targeted strategy for treating OA. Introduction Osteoarthritis (OA) is the most common joint degenerative and age-related disorder and is characterized by cartilage degradation, ectopic osteophyte formation, subchondral bone remodeling, and synovial inflammation.[1, 2] Current standard of care does not provide satisfactory relief for many patients due to the complex pathophysiology of OA. [3] One of the primary pathologies of osteoarthritis is chronic and low-grade inflammation mainly caused by joint damage and debris, which triggers the innate immune response in the early stage of OA and finally results in infiltration of immune cells and synovial hyperplasia.[4] In OA synovium, M1 and M2 macrophages compete with each other in various pathologic conditions, which is critical for the homeostasis of OA.[5] M1 macrophages accumulated more in the synovial membrane of experimental OA than M2 macrophages and further exacerbated the progression of OA, while M2 macrophages could induce cartilage synthesis and inhibit chondrocyte apoptosis.[5–7] Thus, exploring the mechanisms underlying macrophage polarization and remodeling synovial macrophages is emerging as a strategy for OA intervention. Phosphoglycerate mutase 5 (PGAM5) is a mitochondrial serine/threonine phosphatase located in the mitochondrial membrane,[8] which acts as a critical regulator of mitochondrial metabolism and dynamics and controls a series of functions of cells. PGAM5 is vital in programmed cell necrosis by dephosphorylating Drp1, leading to mitochondrial fragmentation.[9] PGAM5 also regulates mitophagy by Page 3/26 Page 3/26 recruiting the E3 ubiquitin ligase PARKIN or dephosphorylating FUNDC1 to modulate the degradation of mitochondria.[10] In addition, PGAM5 modulates cellular senescence by regulating mitochondrial dynamics[11] and was also shown to enhance inflammasome activation in macrophages, which has a critical role in processing of pro-IL-1β in bone marrow-derived macrophages (BMDMs).[12] However, the role of PGAM5 in regulating synovial macrophages in OA has not been reported before. We herein found macrophage PGAM5 significantly increased in macrophages in OA synovium based on histology of human samples and single-cell RNA sequencing results of mice models. Furthermore, we found that conditional knockout of PGAM5 in macrophages alleviated murine OA symptoms via repolarizing M1 macrophages into M2 macrophages in synovium. Mechanistically, we found PGAM5 enhanced M1 polarization via AKT-mTOR/P38/ERK signaling pathways, whereas inhibited M2 polarization via STAT6-PPARγ signaling pathway. Besides, we identified PGAM5 inhibited β-catenin via dephosphorylation of DVL2 to regulate macrophage polarization and conditional knockout of both PGAM5 and β-catenin in macrophages led to increased OA symptoms compared to PGAM5-deficient mice. Introduction In order to precisely modulate PGAM5 in synovial macrophages of OA, we constructed nanoparticles (NPs) composed of mannose modified fluoropolymers and siPGAM5 to inhibit PGAM5 specifically in macrophages, which significantly reduced the OA symptoms via repolarization of M1 macrophages into M2 macrophages(Fig. 1). Our results, therefore, demonstrated that PGAM5 plays a critical role in regulating macrophage polarization which has excellent potential for clinical OA treatment. PGAM5 expression increased in macrophages of the OA synovium To investigate the factors that modulate the progression of osteoarthritis, we examined human synovium from OA patients and normal controls (patients with femoral fracture) and observed enhanced expression of inducible nitric oxide synthase (iNOS) and reduced expression of CD206 in the OA synovium compared to controls (Fig. 2a). Immunohistochemistry showed that PGAM5 was mainly detected in the human OA synovium (Fig. 2a). To address the role of PGAM5 in OA macrophages, we further observed increased expression of PGAM5 in macrophages in human OA synovium compared to controls, as confirmed by double-positive immunostaining for PGAM5 and CD68, a marker of macrophages (Fig. 2b), indicating the potential role of PGAM5 in modulating macrophages in the OA synovium. In addition, we confirmed that the protein and RNA levels of PGAM5 increased in human OA synovium (Fig. 2c). Based on the results in human samples, we further performed destabilization of the medial meniscus (DMM) surgery in 8-week- old male wild - type (WT) mice. The knee joints of mice were collected and sectioned for histological examination 28 days after DMM surgery, safranin O staining and immunohistochemistry of matrix metalloproteinase 13 (MMP13) and aggrecan (ACAN) in the knee joints indicated successful joint osteoarthritis induction in the DMM group (Fig. 2d). As expected, we also observed enhanced iNOS and reduced CD206 expression in the OA synovium of mice, and PGAM5 was upregulated in the murine Page 4/26 Page 4/26 synovium of DMM model compared to the sham group (Fig. 2d). We further investigated the databa NCBI and reanalyzed the single-cell RNA sequencing for murine synovium in an OA model induced b anterior cruciate ligament rupture (ACLR)[13] and found that the expression of PGAM5 increased in t synovial cells of OA mice compared to the sham group, specifically in the macrophage cluster which showed enhanced expression of IL1b, IL6, and reduced expression of Pparg and IL10 in the surgical group (Fig. 2e, f), indicating that PGAM5 might be a potential regulatory factor of OA macrophages i synovium. PGAM5 regulated osteoarthritis in mice by modulating macrophage polarization To address the role of macrophage PGAM5 in osteoarthritis, we generated Pgam5 fl/fl -lyz2-Cre (Pgam5 cKO) mice, by crossing Pgam5 fl/fl mice with transgenic mice that carried lysozyme (Lyz2) proximal promoter - mediated Cre recombinase, which specifically ablates PGAM5 from macrophages. We then employed DMM surgery in 8-week-old male Pgam5 fl/fl and Pgam5 cKO mice for 28 days. We found that Pgam5 cKO mice exhibited relieved OA symptoms compared to Pgam5 fl/fl mice, indicated by enhanced safranin O staining, increased expression of ACAN, and decreased expression of MMP13 in cartilage (Fig. 3a). We further observed decreased M1 and increased M2 macrophages in the synovium of Pgam5 cKO mice compared to Pgam5 fl/fl mice, as confirmed by immunofluorescence of iNOS and CD206 (Fig. 3a, b). RNA-seq of bone marrow-derived macrophages (BMDMs) from Pgam5 cKO mice and Pgam5 fl/fl mice was performed to evaluate the modulatory role of PGAM5 deletion on macrophages at the transcriptome level. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis showed that various macrophage polarization-associated pathways, including the Toll-like receptor signaling pathway, PI3K-Akt signaling pathway, MAPK signaling pathway, and arginine biosynthesis[14] in BMDMs of Pgam5 cKO mice were significantly different from that in BMDMs of Pgam5 fl/fl mice, suggesting that PGAM5 potentially modulated macrophage polarization (Fig. 3c). mRNA levels of M1 markers, iNOS and CD80, and M2 markers, arginase 1 (Arg1) and CD206, were detected in peritoneal macrophages of Pgam5 cKO and Pgam5 fl/fl mice, which indicated that Pgam5 cKO mice showed fewer proinflammatory phenotypes than Pgam5 fl/fl mice (Fig. 3d). These results were validated by flow cytometry of peritoneal macrophages, which indicated fewer CD86+ (M1) and more CD206+ (M2) cells in Pgam5 cKO macrophages (Fig. 3e). To further examine the potential function of PGAM5 in macrophage polarization, we examined whether the expression of PGAM5 changed during the induction of macrophage polarization. BMDMs of WT mice were collected and induced to the M1 polarized state by lipopolysaccharide (LPS) and interferon-γ (IFNγ) or induced to the M2 state by stimulation with IL4. M1 induction significantly reduced PGAM5 mRNA and protein levels within 24 h. However, M2 induction increased PGAM5 mRNA and protein levels in macrophages within 24 h(Fig. 3f, g). The results above indicated that PGAM5 might be involved in regulating macrophage polarization to modulate OA. PGAM5 enhanced M1 polarization and inhibited M2 polarization of BMDMs PGAM5 enhanced M1 polarization and inhibited M2 polarization of BMDMs The protein levels of ACAN and MMP3 and the mRNA levels of COL2A1 and SOX9 in chondrocytes indicated that PGAM5 deletion in macrophages limited the proinflammatory phenotypes of chondrocytes (Fig. 4e, f). Next, we aimed to investigate whether PGAM5 plays an essential role in M2 macrophage polarization. Pgam5 cKO and Pgam5 fl/fl BMDMs were isolated and stimulated with IL4 for 24 hr for M2 induction. After IL-4 stimulation, Pgam5 cKO BMDMs expressed enhanced mRNA levels of Arg1, CD206, PPARγ, and IL10, compared to Pgam5 fl/fl BMDMs (Fig. 4g). The number of CD206 + cells was significantly higher in Pgam5 cKO BMDMs than in Pgam5 fl/fl BMDMs after M2 induction, as determined by flow cytometry (Fig. 4h). The protein level of CD206 significantly increased in Pgam5−cKO BMDMs compared to Pgam5 fl/fl macrophages after M2 induction (Fig. 4i). In conclusion, Pgam5 led to an increased M1 response and a decreased M2 response in macrophages in vitro. PGAM5 enhanced M1 polarization and inhibited M2 polarization of BMDMs Page 5/26 Page 5/26 To further illustrate the role of PGAM5 in macrophage polarization, we isolated BMDMs from Pgam5 cKO and Pgam5 fl/fl mice and induced them to an M1-polarized state by LPS plus IFN-γ for 24 hr in vitro. PGAM5 deletion in BMDMs significantly suppressed the mRNA levels of proinflammatory genes, including iNOS, IL1α, IL1β, IL6, and IL12, compared with that in Pgam5 fl/fl macrophages (Fig. 4a). The number of CD86+ cells was significantly lower in Pgam5 cKO BMDMs than in Pgam5 fl/fl BMDMs after M1 induction, as determined by flow cytometry (Fig. 4b). Enzyme-linked immunosorbent assay (ELISA) of the supernatant of BMDMs treated with LPS plus IFN-γ also verified that PGAM5 deletion led to decreased secretion of proinflammatory cytokines in macrophages (Fig. 4c). In addition, the protein level of iNOS was significantly lower in Pgam5 cKO BMDMs than in Pgam5 fl/fl macrophages after M1 induction (Fig. 4d). To clarify the influence of PGAM5-deficient macrophages on cartilage, BMDMs from Pgam5 cKO and Pgam5 fl/fl mice were cocultured with chondrocytes after M1 induction in vitro. The protein levels of ACAN and MMP3 and the mRNA levels of COL2A1 and SOX9 in chondrocytes indicated that PGAM5 deletion in macrophages limited the proinflammatory phenotypes of chondrocytes (Fig. 4e, f). To further illustrate the role of PGAM5 in macrophage polarization, we isolated BMDMs from Pgam5 cKO and Pgam5 fl/fl mice and induced them to an M1-polarized state by LPS plus IFN-γ for 24 hr in vitro. PGAM5 deletion in BMDMs significantly suppressed the mRNA levels of proinflammatory genes, including iNOS, IL1α, IL1β, IL6, and IL12, compared with that in Pgam5 fl/fl macrophages (Fig. 4a). The number of CD86+ cells was significantly lower in Pgam5 cKO BMDMs than in Pgam5 fl/fl BMDMs after M1 induction, as determined by flow cytometry (Fig. 4b). Enzyme-linked immunosorbent assay (ELISA) of the supernatant of BMDMs treated with LPS plus IFN-γ also verified that PGAM5 deletion led to decreased secretion of proinflammatory cytokines in macrophages (Fig. 4c). In addition, the protein level of iNOS was significantly lower in Pgam5 cKO BMDMs than in Pgam5 fl/fl macrophages after M1 induction (Fig. 4d). To clarify the influence of PGAM5-deficient macrophages on cartilage, BMDMs from Pgam5 cKO and Pgam5 fl/fl mice were cocultured with chondrocytes after M1 induction in vitro. PGAM5 induced M1 polarization via the AKT-mTOR/P38/ERK signaling pathway, whereas inhibited M2 polarization through STAT6-PPARγ signaling pathway In previous studies, a series of signaling pathways were activated by induction of M1 polarization, such as the AKT-mTOR and MAPK signaling pathways.[12] We further examined whether PGAM5 promotes M1 polarization through these specific signals. Treated with LPS and IFNγ, BMDMs from Pgam5 cKO mice significantly showed lower protein levels of p-p38 and p-ERK than Pgam5 fl/fl macrophages and nearly identical expression of total p38 and total ERK, indicating that PGAM5 activates the p38 and ERK signaling pathways to enhance M1 polarization since MAPK signaling has been proven to increase M1 polarization in previous studies[15] (Fig. 5a). However, we did not find different expression pattern of p- JNK in Pgam5 cKO and Pgam5 fl/fl BMDMs (Fig. 5a). p-AKT and p-mTOR significantly decreased in Pgam5 cKO macrophages compared with Pgam5 fl/fl macrophages stimulated by LPS and IFNγ (Fig. 5a). To explore the possible role of AKT and mTOR in mediating the intensive role of PGAM5 in macrophage inflammation, we employed the specific mTOR activator MHY1485[16] to activate the mTOR pathway. Activating mTOR rescued the decreased levels of iNOS, IL1α, IL1β and IL-6 in PGAM5-deficient macrophages after LPS and IFNγ stimulation, as examined by qPCR (Fig. 5b), indicating that mTOR Page 6/26 Page 6/26 activity is involved in the regulation of M1 macrophage polarization by PGAM5. The protein level of iNOS was also enhanced in the presence of MHY1485 in Pgam5−cKO macrophages (Fig. 5c). The activation of mTOR, also presented by phosphorylation of mTOR, increased in BMDMs when treated with doses of 0, 1, 5, and 10 µM of MHY1485 (Fig. 5d), and the mRNA levels of iNOS, IL1α, IL1β, and IL6 increased in Pgam5 cKO macrophages in a dose-dependent manner (Fig. 5e), indicating that PGAM5 promoted the M1 phenotype by activating the mTOR signaling pathway. activity is involved in the regulation of M1 macrophage polarization by PGAM5. The protein level of iNOS was also enhanced in the presence of MHY1485 in Pgam5−cKO macrophages (Fig. 5c). The activation of mTOR, also presented by phosphorylation of mTOR, increased in BMDMs when treated with doses of 0, 1, 5, and 10 µM of MHY1485 (Fig. 5d), and the mRNA levels of iNOS, IL1α, IL1β, and IL6 increased in Pgam5 cKO macrophages in a dose-dependent manner (Fig. 5e), indicating that PGAM5 promoted the M1 phenotype by activating the mTOR signaling pathway. PGAM5 induced M1 polarization via the AKT-mTOR/P38/ERK signaling pathway, whereas inhibited M2 polarization through STAT6-PPARγ signaling pathway To explore the mechanisms by which PGAM5 modulates M2 polarization, we focused on the STAT6- PPARγ signaling pathway, which regulates M2 polarization.[17] After stimulation with IL4 for 24 hr, the protein levels of PPARγ and p-STAT6 enhanced in Pgam5−cKO BMDMs compared to Pgam5 fl/fl BMDMs (Fig. 5f), indicating that the STAT6-PPARγ signaling pathway might be involved in PGAM5-mediated M2 polarization. Next, inhibition of PPARγ by a specific inhibitor, T0070907,[18] markedly reduced the protein level of CD206 in a dose-dependent manner (Fig. 5g). Furthermore, T0070907 decreased the M2 polarization response in Pgam5 cKO macrophages, as indicated by significantly decreased mRNA levels of Arg1, chitinase-like 3 (Ym1), and CD206 (Fig. 5h). Besides, the protein level of CD206 was also markedly reduced by treatment with T0070907 in Pgam5−cKO macrophages stimulated by IL4 (Fig. 5i), suggesting that M2 polarization was regulated by PGAM5 through the STAT6-PPARγ signaling pathway. To identify the role of STAT6 in the PGAM5 modulation of M2 polarization, specific inhibition of STAT6 by AS1517499[19] was added to Pgam5 cKO and Pgam5 fl/fl macrophages at a dose of 10 µM. As a result, the protein expression of PPARγ and CD206 significantly reduced, indicating that STAT6 functions as a regulator of PPARγ (Fig. 5j). In conclusion, PGAM5 regulates M2 polarization via the STAT6-PPARγ signaling pathway. PGAM5 regulated macrophage polarization by targeting the β-catenin pathway via dephosphorylation of DVL2. Although we have identified the related signaling pathways that regulate PGAM5-mediated macrophage polarization, the direct target of PGAM5 in regulating macrophage polarization has not been verified. PGAM5 was reported to inhibit the Wnt/β-catenin signaling pathway on the mitochondrial membrane in human cells and Xenopus embryogenesis.[20] Moreover, the β-catenin signaling pathway is closely correlated with macrophage activation and polarization.[21–23] Nevertheless, whether PGAM5 modulates β-catenin in synovial macrophage polarization and the related targets have not been reported. Thus, we first examined the potential regulation of PGAM5 on β-catenin in macrophage polarization. Pgam5 cKO and Pgam5fl/fl BMDMs were isolated and stimulated to induce M1 or M2 polarization. After 24 hr of stimulation, Pgam5 cKO BMDMs showed increased phosphorylation of β-catenin at Ser675 and decreased phosphorylation at Thr41/Ser45 in both the M1- and M2-polarized states, and increased protein levels of β-catenin were detected in the M1-polarized state compared to Pgam5fl/fl BMDMs, indicating that PGAM5 triggered inhibition of β-catenin in both M1 and M2 polarization (Fig. 6a). To better confirm the target of PGAM5 on β-catenin in macrophage polarization, we focused on whether PGAM5 could dephosphorylate Dishevelled Segment Polarity Protein 2 (DVL2), which is a inhibitory Page 7/26 Page 7/26 regulator in the upstream of β-catenin signaling pathway. Interestingly, the phosphorylation sites of DVL2, S143 and T224, were both increased in Pgam5 cKO macrophages when induced to M1 and M2-polarized states compared to Pgam5fl/fl BMDMs (Fig. 6a), indicating potential interaction of PGAM5 and DVL2. Thus, we detected whether PGAM5 could directly bind to DVL2 and found that PGAM5 could be coimmunoprecipitated with DVL2 in both M1 and M2 macrophages (Fig. 6b), indicating PGAM5 directly dephosphorylated DVL2 via binding to it. We then focused on whether β-catenin was involved in PGAM5- modulated macrophage polarization. Specific inhibition of β-catenin by ICG-001[24] increased the protein levels of p-p38, p-ERK and iNOS. ICG-001 also reduced the expression of p-STAT6, PPARγ, and CD206, indicating that β-catenin functions as a regulator of macrophage polarization (Fig. 6c, d). In addition, as predicted, the expression of β-catenin increased in the OA synovium of Pgam5 cKO mice compared to Pgam5fl/fl mice (Fig. 6e). Thus, these results indicated that PGAM5 modulated macrophage polarization by inhibiting the β-catenin pathway via directly targeting DVL2. PGAM5 regulated macrophage polarization by targeting the β-catenin pathway via dephosphorylation of DVL2. Take F7 for example, the mannose grafted polymer conjugated with 10 F7 ligands and 15 F7 ligands were termed MFP7-1 and MFP7-2, respectively. The siRNA delivery efficacy of the synthesized nanoparticles (NPs) was first screened on Raw264.7 cells. MFP9-2/siPGAM5 complexes exhibited the highest gene knockdown efficiency, which was higher than that of the Lipofectamine 2000 (Lipo)/siPGAM5 complexes (Fig. 7b). The MFP9-2 transported the FAM-labelled siRNA into the CD68 + synovial macrophages rather than chondrocytes efficiently after intraarticular injection (Fig. 7c), which is beneficial for achieving macrophage-targeted RNAi. To validate the therapeutic efficacy of the MFP9-2 based siRNA delivery system, MFP9-2/siPGAM5 complex was injected into the joint of WT mice twice weekly in early stage of OA established by DMM surgery and the joints were collected 28 days after DMM surgery. Sham group, DMM group with no injection and DMM group with injection of MFP9-2 combined with siNC (MFP9-2/siNC) were established as controls (Fig. 7d). Intraarticular injection of MFP9-2/siPGAM5 greatly relieved the OA symptoms compared to DMM group and DMM with injection of MFP9-2/siNC group, indicated by increased safranine O staining area, decreased expression of MMP13 and enhanced level of ACAN in cartilage (Fig. 7d). Besides, the amount of iNOS positive cells in synovium treated with MFP9-2/siPGAM5 greatly decreased, while CD206 positive cells increased compared to DMM group and DMM with injection of MFP9-2/siNC group (Fig. 7e, g), indicating the availability of macrophage modulation in OA synovium by MFP9-2/siPGAM5. To further confirm the modulatory function of MFP9-2/siPGAM5, we detected whether the relief of OA symptoms was the result of PGAM5 inhibition in synovial macrophage by MFP9-2/siPGAM5. We found that injection of MFP9-2/siPGAM5 greatly decreased the level of PGAM5 in CD68 positive macrophages in synovium compared to injection of MFP9-2/siNC with FAM (Fig. 7f, g), indicating successful targeting and inhibiting of PGAM5 in synovial macrophages, which could further achieve better OA outcomes. via an amine-isocyanate reaction, and an average number of 1 mannose was conjugated on each polymer calculated by 1H NMR. Take F7 for example, the mannose grafted polymer conjugated with 10 F7 ligands and 15 F7 ligands were termed MFP7-1 and MFP7-2, respectively. The siRNA delivery efficacy of the synthesized nanoparticles (NPs) was first screened on Raw264.7 cells. MFP9-2/siPGAM5 complexes exhibited the highest gene knockdown efficiency, which was higher than that of the Lipofectamine 2000 (Lipo)/siPGAM5 complexes (Fig. 7b). PGAM5 regulated macrophage polarization by targeting the β-catenin pathway via dephosphorylation of DVL2. The MFP9-2 transported the FAM-labelled siRNA into the CD68 + synovial macrophages rather than chondrocytes efficiently after intraarticular injection (Fig. 7c), which is beneficial for achieving macrophage-targeted RNAi. To validate the therapeutic efficacy of the MFP9-2 based siRNA delivery system, MFP9-2/siPGAM5 complex was injected into the joint of WT mice twice weekly in early stage of OA established by DMM surgery and the joints were collected 28 days after DMM surgery. Sham group, DMM group with no injection and DMM group with injection of MFP9-2 combined with siNC (MFP9-2/siNC) were established as controls (Fig. 7d). Intraarticular injection of MFP9-2/siPGAM5 greatly relieved the OA symptoms compared to DMM group and DMM with injection of MFP9-2/siNC group, indicated by increased safranine O staining area, decreased expression of MMP13 and enhanced level of ACAN in cartilage (Fig. 7d). Besides, the amount of iNOS positive cells in synovium treated with MFP9-2/siPGAM5 greatly decreased, while CD206 positive cells increased compared to DMM group and DMM with injection of MFP9-2/siNC group (Fig. 7e, g), indicating the availability of macrophage modulation in OA synovium by MFP9-2/siPGAM5. To further confirm the modulatory function of MFP9-2/siPGAM5, we detected whether the relief of OA symptoms was the result of PGAM5 inhibition in synovial macrophage by MFP9-2/siPGAM5. We found that injection of MFP9-2/siPGAM5 greatly decreased the level of PGAM5 in CD68 positive macrophages in synovium compared to injection of MFP9-2/siNC with FAM (Fig. 7f, g), indicating successful targeting and inhibiting of PGAM5 in synovial macrophages, which could further achieve better OA outcomes. To conclude, PGAM5 serves as a novel factor of regulating macrophage polarization in osteoarthritis via dephosphorylating DVL2, resulting in increased activity of GSK3β and degradation of β-catenin,[27] which disables the translocation of β-catenin into nucleus to bind to promotors for downstream signaling pathways, further contributes to increased M1 and decreased M2 phenotypes via specific signals. To better treat OA via early intervention of macrophage PGAM5, specifically inhibition of PGAM5 in macrophages was achieved by intraarticular injection of MFP9-2/siPGAM5, which could significantly target synovial macrophages and reduce the expression of PGAM5 in macrophages, resulting in the relief of OA symptoms. Together, we have clarified the modulatory role of PGAM5 in OA macrophage and designed a functional macrophage-targeted therapy, which might contribute to early and precise immunological interventions in OA in clinic. PGAM5 regulated macrophage polarization by targeting the β-catenin pathway via dephosphorylation of DVL2. To verify the role of β-catenin in PGAM5-mediated macrophage polarization in synovium, we generated mice in which both PGAM5 and β-catenin were ablated in macrophages, herein referred to as DKO mice, by crossing β-cateninfl/fl mice with Pgam5 cKO mice. Knee joints of DKO and Pgam5 cKO male mice were collected 28 days after DMM surgery for further investigations. DKO mice significantly exacerbated OA symptoms compared to Pgam5 cKO mice, as determined by safranin O staining (Fig. 6f). Expression of ACAN decreased while expression of MMP13 increased in the knee joint of DKO mice compared to Pgam5 cKO mice, indicating that the relieved OA symptom by Pgam5 deficiency in macrophages was partly mediated by the enhanced activity of β-catenin (Fig. 6f). In addition, iNOS-positive cells were significantly increased and CD206 + cells were significantly decreased in the synovium of DKO mice compared to Pgam5 cKO mice (Fig. 6f, g), suggesting that PGAM5 regulates synovial macrophage polarization by inhibiting the β-catenin signaling pathway, further aggravating the progression of OA. Peritoneal macrophages in DKO mice showed increased mRNA levels of iNOS and CD80 and decreased mRNA levels of CD206 (Fig. 6h). In addition, flow cytometry of peritoneal macrophages also indicated that DKO mice showed more proinflammatory phenotypes than Pgam5 cKO mice (Fig. 6i). In conclusion, PGAM5 regulates macrophage polarization by inhibiting the β-catenin signaling pathway, and inhibition of β-catenin extensively reversed the alleviation of OA symptoms in Pgam5 cKO mice. Targeted knockdown of PGAM5 in synovial macrophages by MFP9-2/siPGAM5 relieved OA symptoms.Based on the mechanisms of PGAM5 in regulating macrophage polarization, we aimed to establish targeted deletion of PGAM5 in synovial macrophages to treat OA in early stage. RNA interference (RNAi) is a powerful technique to treat various diseases via specific gene silence.[25] However, targeted delivery of siRNA into synovial macrophages is challenging owing to the complicated synovial fluid composition and extracellular matrix in the joint microenvironment.[26] Here, we developed a series of mannose modified fluoropolymers for macrophage-targeted siRNA delivery to treat OA via inhibition of PGAM5. ε-PLL was conjugated with fluoroalkanes (F7-F17, Fig. 7a) via amine-epoxide reactions, and different feeding ratios were chosen to get PLL modified with average numbers of about 10 and 15 fluoroalkanes, respectively. The obtained fluoropolymers were further grafted with mannose Page 8/26 Page 8/26 via an amine-isocyanate reaction, and an average number of 1 mannose was conjugated on each polymer calculated by 1H NMR. Discussion Early intervention of OA have been focused to gain better clinical outcomes, which requires for identification of molecular biomarkers and validation of molecular targets for novel targeted therapies in the early-stage OA.[28] In the OA pathophysiology, synovial macrophages exhibit distinct functions.[29] Page 9/26 Macrophages showed distinct phenotypes of M1 and M2 macrophages. M1 macrophages are also termed proinflammatory macrophages, which are activated by LPS produced by microbes or IFN-γ and characterized by the production of inflammatory cytokines such as IL-1, IL-6, IL-12, TNF-α and iNOS, which promote the Th1 response.[14, 17, 30] In contrast, M2 macrophages showed anti-inflammatory functions by promoting the Th2 response and enhanced tissue remodelling. M2 macrophages can be alternatively activated by IL-4, IL-13, glucocorticoids, IL-10, and immunoglobulin complexes/Toll-like receptor ligands and are characterized by enhanced expression of multiple cell surface markers, including mannose receptor Mrc1 (also known as CD206), CD9, and CD36. They also produce various cytokines, like arginase (Arg), IL-10, IL-1 receptor antagonist (IL-1ra), and the type II IL-1 decoy receptor.[30–32] Since investigations have shown that synovial macrophage polarization is closely associated with OA, [33–35] precise therapies for macrophage to decrease synovitis and attenuate OA progression need to be clearly proposed. In this study, we found that PGAM5 significantly increased in macrophages in OA synovium compared to controls based on histology of human samples and single-cell RNA sequencing results of mice models. Accordingly, we constructed DMM model in transgenic mice with macrophage-specific deletion of PGAM5 to identify the modulatory role of PGAM5 in synovial macrophages. We observed alleviated OA symptoms in Pgam5 cKO mice compared to controls due to repolarization of M1 macrophages into M2 macrophages in synovium. Additionally, Pgam5 cKO BMDMs tilted polarization toward M2 macrophages, which promoted anabolic metabolism of chondrocytes in vitro. To clarify the underlying mechanisms of macrophage polarization regulated by PGAM5 in OA, we examined macrophage-related signalings in Pgam5 cKO BMDMs. In previous studies, Akt-mTOR signaling pathway is significantly activated in the process of M1 polarization by stimulation of LPS and IFNγ, in which Akt activation could lead to the inactivation of tuberous sclerosis complex (TSC) 1/2, a critical factor for M2 polarization and attenuation of M1 responses,[31] further activating mTORC1 and increasing the M1 phenotype.[36] Thus, we examined Akt-mTOR signaling in Pgam5 cKO BMDMs during M1 induction and found that the decreased M1 response of Pgam5 cKO macrophages was partly due to reduced activation of the Akt-mTOR signaling pathway. Discussion However, Jin Fan et al. also revealed that the activation of Akt pathway could enhance M2 macrophage polarization,[37] indicating the different mechanisms of Akt signaling in modulating macrophage polarization. To clarify the intracellular pathways involved in PGAM5-mediated M1 polarization, we used MHY1485, a specific mTOR agonist,[16] to activate mTOR signaling in Pgam5 cKO macrophages. As a result, MHY1485 significantly reversed the decreased M1 phenotype of Pgam5 cKO macrophages in a dose-dependent manner, indicating PGAM5 induced M1 polarization via activating Akt- mTOR signaling. Furthermore, we detected whether PGAM5 also activated MAPK-related signaling to modulate M1 polarization, since p38/stat3, ERK/MEK, and JNK signals are proved activated by M1- correlated stimuli such as LPS, serving as signal transducers to enhance the M1 response.[31, 38–41] We then proved that PGAM5 deletion in macrophages could inactivate p38 and ERK signaling when treated with LPS and IFNγ, suggesting that p38/ERK MAPK signaling pathways participated in PGAM5-induced M1 polarization. Similarly, we revealed that PGAM5 inhibited M2 polarization by STAT6-PPARγ by usage In this study, we found that PGAM5 significantly increased in macrophages in OA synovium compared to controls based on histology of human samples and single-cell RNA sequencing results of mice models. Accordingly, we constructed DMM model in transgenic mice with macrophage-specific deletion of PGAM5 to identify the modulatory role of PGAM5 in synovial macrophages. We observed alleviated OA symptoms in Pgam5 cKO mice compared to controls due to repolarization of M1 macrophages into M2 macrophages in synovium. Additionally, Pgam5 cKO BMDMs tilted polarization toward M2 macrophages, which promoted anabolic metabolism of chondrocytes in vitro. To clarify the underlying mechanisms of macrophage polarization regulated by PGAM5 in OA, we examined macrophage-related signalings in Pgam5 cKO BMDMs. In previous studies, Akt-mTOR signaling pathway is significantly activated in the process of M1 polarization by stimulation of LPS and IFNγ, in which Akt activation could lead to the inactivation of tuberous sclerosis complex (TSC) 1/2, a critical factor for M2 polarization and attenuation of M1 responses,[31] further activating mTORC1 and increasing the M1 phenotype.[36] Thus, we examined Akt-mTOR signaling in Pgam5 cKO BMDMs during M1 induction and found that the decreased M1 response of Pgam5 cKO macrophages was partly due to reduced activation of the Akt-mTOR signaling pathway. However, Jin Fan et al. also revealed that the activation of Akt pathway could enhance M2 macrophage polarization,[37] indicating the different mechanisms of Akt signaling in modulating macrophage polarization. Discussion Furthermore, Pgam5/catenin DKO mice showed increased OA symptoms compared to Pgam5 cKO mice, indicating that PGAM5 promoted M1 polarization and inhibited M2 polarization of synovial macrophages by downregulating β-catenin both in vitro and in vivo. In previous studies, PGAM5 was shown to play a controversial role of modulating β-catenin signaling pathway. DVL2 serves as a substrate of PGAM5 and could be dephosphorylated by PGAM5 directly on the mitochondrial membrane due to its phosphatase activity, which activates GSK3β, an antagonist of the Wnt-β-catenin signaling pathway, by phosphorylating β-catenin on Thr41 and inhibiting its activity,[20, 42, 43] indicating that PGAM5 functions as an inhibitor of the β-catenin signaling pathway. Nevertheless, some studies also proved that PGAM5 could directly bind to β-catenin and activate the Wnt-β-catenin signaling pathway.[44, 45] To further verify the association between PGAM5 and β-catenin signaling during synovial macrophage polarization, we further examined the expression patterns of potential targets of PGAM5. Our observations showed that the levels of β-catenin and p-DVL2 were enhanced in Pgam5 cKO macrophages during M1 and M2 induction, suggesting that PGAM5 might directly dephosphorylate DVL2 to inhibit the β-catenin pathway in macrophages. Next, co-IP of PGAM5 and DVL2 indicated that PGAM5 directly binds to DVL2 and leads to inhibition of the β-catenin pathway. Motivated by these findings, we aimed to deliver siRNA targeting PGAM5 in synovial macrophages to treat OA. Few studies have achieved macrophage-targeted delivery of siRNA in OA synovium, since siRNA is a large hydrophilic molecule which is difficult to penetrate across the semipermeable cell membrane, resulting in the low efficacy of intracellular delivery. [46] Besides, the negatively charged siRNA could be easily degraded in the cytoplasm leading to short residence of siRNA.[46] Accordingly, we constructed mannose modified fluoropolymers combined with siPGAM5 and treated mice via intraarticular injection. The fluoropolymers modified by perfluorinated lipids(PFL) have excellent advances in gene delivery,[47] since the strong hydrophobicity of fluoropolymers ensures efficient cell membrane penetration and endosomal escape of siRNA, and cationic polymers by fluorination protect nucleic acids from enzyme degradation. As a result, we detected stable residency of FAM-modified siPGAM5 in synovial macrophages up to a week. Discussion To clarify the intracellular pathways involved in PGAM5-mediated M1 polarization, we used MHY1485, a specific mTOR agonist,[16] to activate mTOR signaling in Pgam5 cKO macrophages. As a result, MHY1485 significantly reversed the decreased M1 phenotype of Pgam5 cKO macrophages in a dose-dependent manner, indicating PGAM5 induced M1 polarization via activating Akt- mTOR signaling. Furthermore, we detected whether PGAM5 also activated MAPK-related signaling to modulate M1 polarization, since p38/stat3, ERK/MEK, and JNK signals are proved activated by M1- correlated stimuli such as LPS, serving as signal transducers to enhance the M1 response.[31, 38–41] We then proved that PGAM5 deletion in macrophages could inactivate p38 and ERK signaling when treated with LPS and IFNγ, suggesting that p38/ERK MAPK signaling pathways participated in PGAM5-induced M1 polarization. Similarly, we revealed that PGAM5 inhibited M2 polarization by STAT6-PPARγ by usage Page 10/26 of STAT6 inhibitor (AS1517499) and PPARγ inhibitor (T0070907). These results verified that PGAM5 is a critical regulator of macrophage polarization. In previous studies, PGAM5 was shown to play a controversial role of modulating β-catenin signaling pathway. DVL2 serves as a substrate of PGAM5 and could be dephosphorylated by PGAM5 directly on the mitochondrial membrane due to its phosphatase activity, which activates GSK3β, an antagonist of the Wnt-β-catenin signaling pathway, by phosphorylating β-catenin on Thr41 and inhibiting its activity,[20, 42, 43] indicating that PGAM5 functions as an inhibitor of the β-catenin signaling pathway. Nevertheless, some studies also proved that PGAM5 could directly bind to β-catenin and activate the Wnt-β-catenin signaling pathway.[44, 45] To further verify the association between PGAM5 and β-catenin signaling during synovial macrophage polarization, we further examined the expression patterns of potential targets of PGAM5. Our observations showed that the levels of β-catenin and p-DVL2 were enhanced in Pgam5 cKO macrophages during M1 and M2 induction, suggesting that PGAM5 might directly dephosphorylate DVL2 to inhibit the β-catenin pathway in macrophages. Next, co-IP of PGAM5 and DVL2 indicated that PGAM5 directly binds to DVL2 and leads to inhibition of the β-catenin pathway. Accordingly, inhibition of the β-catenin pathway by ICG-001, a specific inhibitor of the β-catenin pathway, [24] significantly enhanced the M1 phenotype by activating the p38/ERK signaling pathway and reduced the M2 response by inhibiting the STAT6-PPARγ pathway in vitro. Animals Myeloid cell-specific PGAM5 conditional knockout mice (Pgam5 cKO mice) were obtained by crossing PGAM5loxp/loxp mice with mice expressing Cre recombinase under the control of the Lysozyme promoter (Lyz2). Myeloid cell-specific PGAM5 and β-catenin conditional knockout mice (DKO mice) were obtained by crossing β-cateninloxp/loxp mice with Pgam5 cKO mice. Lyz2Cre-negative, PGAM5loxp/loxp and β-cateninloxp/loxp littermates served as the controls. Eight-week-old mice (male) were usually used for the in vitro experiments. PGAM5loxp/loxp and Lyz2Cre mice were generous gifts from Prof. Wei Lu.[8] β-cateninloxp/loxp mice were kindly provided by Prof. Xuefeng Wu.[49] All mice were maintained in a specific pathogen-free facility. All experimental manipulations were undertaken in accordance with the Institutional Guidelines for the Care and Use of Laboratory Animals, Institute of Zoology (Shanghai, China). Human synovium Normal human synovium was obtained from bone fracture patients with no history of arthritic diseases (n = 3). Human OA synovium was obtained from patients undergoing total knee replacement surgery (n = 5). Human samples were obtained from Shanghai Xinhua Hospital and Shanghai Dongfang Hospital. All patients gave informed consent to use their clinical information for scientific research. The study was approved by the Ethics Committee of the Shanghai Xinhua Hospital and Shanghai Dongfang Hospital. Discussion Furthermore, fluoropolymers could be easily modified by mannose for targeting synovial macrophages, owing to the widely expressed mannose receptors on the surface of macrophages.[48] Amazingly, we found that nearly a hundred percent of synovial macrophages were targeted and PGAM5 in macrophages was inhibited via the intraarticular injection of siPGAM5 NPs, leading to repolarization of M1 macrophages into M2 macrophages which greatly alleviated OA. What’s more, the cytotoxicity of fluoropolymers is less than other cationic polymers due to less nitrogen to phosphorus ratio,[46] which is a promising therapeutic for targeting synovial macrophages in clinical intervention. Page 11/26 In summary, we demonstrated that PGAM5 directly interacted with DVL2 to modulate the β-catenin pathway, further amplifying the M1 response via the Akt-mTOR, p38, and ERK signaling and inhibiting M2 polarization mediated by downregulating the STAT6-PPARγ pathway. Accordingly, targeted knockdown of PGAM5 in macrophages by macrophage-targeted siPGAM5 NPs greatly relieved OA symptoms in mice. The critical function of PGAM5 in OA might provide novel clinical targets and the related macrophage- targeted therapy will offer new therapeutic strategies of precise immunomodulation in OA for better outcomes. Reagents Anti-mCD86-PE/Cy5 and anti-mCD206-FITC were purchased from Cell Signaling Technology. Bacterial lipopolysaccharide (LPS; E. coli 055: B5) was purchased from Sigma‒Aldrich. Recombinant mouse IL-4 and IFN-γ were purchased from PeproTech (Rocky Hill, NJ). An agonist of mTOR (MHY1985), inhibitor of PPARγ (T0070907) and inhibitor of STAT6 (AS1517499) were purchased from Selleck. Primary antibodies against ERK, p-ERK1/2 (Thr202/Tyr204), p-JNK (Thr183/Tyr185), P38, p-p38 (Thr180/Tyr182), Akt, p-Akt (Thr308), p-Akt (Ser473), STAT6, p-STAT6 (Tyr641), p-β-catenin (Ser675, Thr41/Ser45) and β- tubulin were purchased from Cell Signaling Technology. Primary antibodies against PGAM5, PPARγ, MMP13, β-catenin, DVL2 and p-DVL2 (S143, T224) were purchased from ABclonal. ACAN antibody was purchased from ABclonal (A11691; Wuhan, China). All of these antibodies were diluted at 1:1,000 in 5% Page 12/26 Page 12/26 bovine serum albumin (BSA). ELISA was performed using the Mouse IL-12p70 ELISA Kit (EK212/3–96, Lianke), IL-6 Mouse Uncoated ELISA Kit (88-7064-88, Thermo Fisher), and Mouse IL-1 beta Uncoated ELISA (88-7013-88, Thermo Fisher). Cell preparation Bone marrow cells were cultured with Dulbecco's modified Eagle medium (DMEM) containing 10% (v/v) FBS and 10 ng ml − 1 mouse M-CSF or 10 ng ml − 1 mouse GM-CSF for 12 days to obtain bone marrow- derived macrophages. The nonadherent cells were removed by washing with PBS. The inflammatory response of macrophages was induced by LPS (100 ng ml − 1) for 24 h. M2 macrophages were induced by IL-4 (1,000 U ml − 1) treatment for 24 h. Quantitative PCR analysis Total RNA was extracted by transfer to TRIzol reagent (Invitrogen, Waltham, MA) from 6-well plates and then homogenized at high speed on ice. DNase I (Sigma‒Aldrich, St. Louis, MO) was added to the extracted mRNA to remove genomic DNA. The quantification of mRNA was performed and calculated using a Nanodrop 2000 (Thermo Fisher Scientific). mRNA was reversely transcribed into complementary DNA (cDNA) in each experimental and control group using the PrimeScript RT Master Mix Kit (Takara Bio Inc., Dalian, China). Then, cDNA was tested by PCR via the SYBR Premix Ex Taq Kit (RR420a; Takara, Tokyo, Japan). To normalize the mRNA expression, the level of the housekeeping gene GAPDH served as a control. Quantitative PCR (qPCR) primers for the genes and forwards (F) and reverse (R) primer sequences were as follows: Immunohistochemistry and pathological staining ACAN, MMP13, β-catenin, iNOS and CD206 in pathological sections were examined by immunohistochemistry. Paraffin sections of joints were dewaxed, rehydrated, pretreated with pepsin at 37°C for 30 minutes, and then incubated with 3% H2O2 in methanol solution. After rinsing with PBS, the slices were blocked with BSA at room temperature for 1 hour and then incubated overnight with primary antibodies at 4°C. Then, the slices were incubated with the secondary antibody provided in the HRP polymer anti-rabbit IHC kit (Kit 5005; MaxVision, Shenzhen, China) for 15 minutes and stained with substrate from the DAB Plus kit (DAB-2031) for 10 minutes. Histological images of the knee joint were obtained after staining with safranine O and fast green. We performed immunofluorescence using a multicoloured immunofluorescence kit (abs50012, absin) and observed the cells using a microscope from Zeiss. Page 13/26 hPGAM5-F: TCGTCCATTCGTCTATGACGC hPGAM5-R: GGCTTCCAATGAGACACGG hGAPDH-F: GGAGCGAGATCCCTCCAAAAT hGAPDH-R: GGCTGTTGTCATACTTCTCATGG mCOL2A1-F: ACGAGGCAGACAGTACCTTG mCOL2A1-R: CAGCCCTGGTTGGGATCAAT mMMP13-F: TTGGCTTAGAGGTGACTGGC mMMP13-R: CCACATCAGGCACTCCACAT mSOX9-F: TCAGCAAGACTCTGGGCAAG mSOX9-R: TCCGTTCTTCACCGACTTCC mMMP3-F: CCACTCCCTGGGACTCTAC hPGAM5-F: TCGTCCATTCGTCTATGACGC hPGAM5-R: GGCTTCCAATGAGACACGG hGAPDH-F: GGAGCGAGATCCCTCCAAAAT hGAPDH-R: GGCTGTTGTCATACTTCTCATGG mCOL2A1-F: ACGAGGCAGACAGTACCTTG mCOL2A1-R: CAGCCCTGGTTGGGATCAAT mMMP13-F: TTGGCTTAGAGGTGACTGGC mMMP13-R: CCACATCAGGCACTCCACAT mSOX9-F: TCAGCAAGACTCTGGGCAAG mSOX9-R: TCCGTTCTTCACCGACTTCC mMMP3-F: CCACTCCCTGGGACTCTAC Page 13/26 mMMP3-R: TGAGAGAGATGGAAACGGGAC mPGAM5-F: CCCTGCAAGAAGACTGTGGT mPGAM5-R: GTCAGCGGGGGCTAAATCTT mGAPDH-F: TGACCTCAACTACATGGTCTACA mGAPDH-R: CTTCCCATTCTCGGCCTTG mCD80-F: TCAGTTGATGCAGGATACACCA mCD80-R: AAAGACGAATCAGCAGCACAA mIL1β-F: GCAACTGTTCCTGAACTCAACT mIL1β-R: ATCTTTTGGGGTCCGTCAACT mTNF-F: CCCTCACACTCAGATCATCTTCT mTNF-R: GCTACGACGTGGGCTACAG mIL10-F: CTTACTGACTGGCATGAGGATCA mIL10-R: GCAGCTCTAGGAGCATGTGG mIL4-F: GGTCTCAACCCCCAGCTAGT mIL4-R: GCCGATGATCTCTCTCAAGTGAT mIL6-F: CTGCAAGAGACTTCCATCCAG mIL6-R: AGTGGTATAGACAGGTCTGTTGG mIL1A-F: AGTATCAGCAACGTCAAGCAA mIL1A-R: TCCAGATCATGGGTTATGGACTG mIl1b-F: GAAATGCCACCTTTTGACAGTG mIl1b-R: TGGATGCTCTCATCAGGACAG mIL12-F: TGGTTTGCCATCGTTTTGCTG mIL12-R: ACAGGTGAGGTTCACTGTTTCT mPPARγ-F: GGAAGACCACTCGCATTCCTT Page 14/26 Page 14/26 Synthesis of mannose modified fluoropolymers. ε-PLL (Mw: 4224 Da) was mixed with epoxides bearing fluoroalkanes at different molar ratios in anhydrous methanol, and stirred at 60 oC for 48 h. The crude products were dialyzed against methanol and distilled water, then lyophilized to obtain the fluoropolymers. The average number of fluoroalkanes modified on each polymer was tested by a well-established ninhydrin assay[50] (Supplementary Figure S1). The obtained fluoropolymers were further reacted with d-mannopyranosylphenyl isothiocyanate in dimethylsulfoxide at a molar ratio of 1:1 at room temperature for 24 h, then lyophilized to obtain the mannose modified fluoropolymers. The average number of mannoses modified on each polymer was characterized and calculated by 1H-NMR (Bruker, 500 MHz). Western blot assay and coimmunoprecipitation Macrophages were cultured in DMEM with 10% FCS in 6-well plates. Cells were treated with IL-4 (100 nM) or LPS (100 ng ml − 1) for the indicated times. After stimulation, cells were washed once in cold PBS and lysed in RIPA buffer (50 mM Tris–HCl pH 7.4, 1% NP-40, 0.25% Na-deoxycholate, 150 mM NaCl, 1 mM EDTA pH 7.4) with protease and phosphatase inhibitor cocktails (Sigma) for 10 min on a rocker at 4°C. Protein concentration was determined using a BCA assay. Protein samples were analysed by SDS polyacrylamide gel electrophoresis (SDS–PAGE) and transferred onto PVDF membranes (Millipore, CA)60. Each polyvinylidene fluoride membrane was blocked with TBST (100 mM Tris–HCl pH 7.5, 150 mM NaCl, 0.05% Tween 20) with 5% BSA for 1 h and then incubated with primary antibodies overnight on a shaker at 4°C. The appropriate HRP-coupled secondary antibody was then added and detected through chemiluminescence (Millipore). GAPDH and β-tubulin were used as protein loading controls. Coimmunoprecipitation was performed with indicated antibodies and G/A beads following information of co-IP kit (absin, abs955.) Coimmunoprecipitation was performed with indicated antibodies and G/A beads following information of co-IP kit (absin, abs955.) Flow cytometry Flow cytometry Page 15/26 Primary BMDMs were induced under specific conditions and labeled with the following antibodies for identification of M1 or M2 macrophages: PE/Cyanine7 anti-mouse F4/80 Antibody (Biolegend, 123113) PE/Cyanine5 anti-mouse CD86 Antibody (Biolegend, 105015), FITC anti-mouse CD206 (MMR) Antibody (Biolegend, 141703). Related isotypes for control are PE/Cyanine7 Rat IgG2a, κ Isotype Ctrl Antibody (Biolegend, 400521), PE/Cyanine5 Rat IgG2a, κ Isotype Ctrl Antibody (Biolegend, 400509), FITC Rat IgG2a, κ Isotype Ctrl Antibody (Biolegend, 400505). Cells were suspended and incubated with antibodies for 30 minutes, and analyzed via Guava Easycyte 12HT (Luminex). Primary BMDMs were induced under specific conditions and labeled with the following antibodies for identification of M1 or M2 macrophages: PE/Cyanine7 anti-mouse F4/80 Antibody (Biolegend, 123113) PE/Cyanine5 anti-mouse CD86 Antibody (Biolegend, 105015), FITC anti-mouse CD206 (MMR) Antibody (Biolegend, 141703). Related isotypes for control are PE/Cyanine7 Rat IgG2a, κ Isotype Ctrl Antibody (Biolegend, 400521), PE/Cyanine5 Rat IgG2a, κ Isotype Ctrl Antibody (Biolegend, 400509), FITC Rat IgG2a, κ Isotype Ctrl Antibody (Biolegend, 400505). Cells were suspended and incubated with antibodies for 30 minutes, and analyzed via Guava Easycyte 12HT (Luminex). siRNAs siRNAs targeting mouse PGAM5 with FAM labeled were purchased from Sangon Biotech. Scrambled siRNA with FAM labeled was used as a control. Raw264.7 cells were transfected with siRNA using Lipofectamine 2000 (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) in serum-free conditional medium for examination. Sequence of siRNA targeting PGAM5 is: sense(5'-3’): (FAM)CUGGAGAAGACGAGUUGACAUTT, antisense (5'-3’): AUGUCAACUCGUCUUCUCCAGTT. Experiments were repeated for at least three times independently. ( ) (FAM)CUGGAGAAGACGAGUUGACAUTT, antisense (5'-3’): AUGUCAACUCGUCUUCUCCAGTT. Experiments were repeated for at least three times independently. Statistical analysis All data are presented as the mean ± s.d. Two-way ANOVA was used for comparisons among multiple groups with SPSS 16.0 software. Student's unpaired t test for comparison of means was used to compare two groups. Log-rank tests were used for mouse survival assays. A P value less than 0.05 was considered to be statistically significant. Data availability All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Data will be made available upon reasonable request. All data needed to evaluate the conclusions in the paper are present in the paper and/or the All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Data will be made available upon reasonable request. Conflict of interests The authors declare no competing interests. Author contributions Y.L., J.L., Y.C., W.L., and X.Z. designed research; Y.L., R.H., J.L., J.Y., X.W., D.M., C.R., Z.D., B.Z., and L.D.performed research; Y.L., J.L., Y.C., W.L., X.Z. analyzed data; and Y.L., J.L., W.L. and X.Z. drafted the manuscript. Acknowledgements This work was supported by grants from National Natural Science Foundation of China (81830078, 82071868, 32370892), Science and Technology Commission of Shanghai Municipality (23141901200), Health Commission of Shanghai Municipality (2022JC029), Biomaterials and Regenerative Medicine Page 16/26 Institute Cooperative Research Project, Shanghai Jiaotong University School of Medicine (2022LHA11) and Shanghai Key Laboratory of Orthopedic Implant（No.KFKT202206）. Institute Cooperative Research Project, Shanghai Jiaotong University School of Medicine (2022LHA11) and Shanghai Key Laboratory of Orthopedic Implant（No.KFKT202206）. Institute Cooperative Research Project, Shanghai Jiaotong University School of Medicine (2022LHA11) and Shanghai Key Laboratory of Orthopedic Implant（No.KFKT202206）. References 1. Coryell, P.R., B.O. Diekman, and R.F. Loeser, Mechanisms and therapeutic implications of cellular senescence in osteoarthritis. Nat. Rev. Rheumatol., 2021. 17(1): p. 47–57. 1. Coryell, P.R., B.O. Diekman, and R.F. Loeser, Mechanisms and therapeutic implications of cellular senescence in osteoarthritis. Nat. Rev. Rheumatol., 2021. 17(1): p. 47–57. 2. Loeser, R.F., J.A. Collins, and B.O. Diekman, Ageing and the pathogenesis of osteoarthritis. Nat. Rev. Rheumatol., 2016. 12(7): p. 412–20. 2. Loeser, R.F., J.A. Collins, and B.O. Diekman, Ageing and the pathogenesis of osteoarthritis. Nat. Rev. Rheumatol., 2016. 12(7): p. 412–20. 3. Conaghan, P.G., et al., Therapeutic options for targeting inflammatory osteoarthritis pain. Nat Rev Rheumatol, 2019. 15(6): p. 355-363. 3. Conaghan, P.G., et al., Therapeutic options for targeting inflammatory osteoarthritis pain. Nat Rev Rheumatol, 2019. 15(6): p. 355-363. 4. Robinson, W.H., et al., Low-grade inflammation as a key mediator of the pathogenesis of osteoarthritis. Nat. Rev. Rheumatol., 2016. 12(10): p. 580–92. 4. Robinson, W.H., et al., Low-grade inflammation as a key mediator of the pathogenesis of osteoarthritis. Nat. Rev. Rheumatol., 2016. 12(10): p. 580–92. 5. Zhang, H., et al., Synovial macrophage M1 polarisation exacerbates experimental osteoarthritis partially through R-spondin-2. Ann. Rheum. Dis., 2018. 77(10): p. 1524–1534. 5. Zhang, H., et al., Synovial macrophage M1 polarisation exacerbates experimental osteoarthritis partially through R-spondin-2. Ann. Rheum. Dis., 2018. 77(10): p. 1524–1534. 6. Dai, M., et al., Cartilage repair in degenerative osteoarthritis mediated by squid type II collagen via immunomodulating activation of M2 macrophages, inhibiting apoptosis and hypertrophy of chondrocytes. Biomaterials, 2018. 180: p. 91-103. 6. Dai, M., et al., Cartilage repair in degenerative osteoarthritis mediated by squid type II collagen via immunomodulating activation of M2 macrophages, inhibiting apoptosis and hypertrophy of chondrocytes. Biomaterials, 2018. 180: p. 91-103. 7. Ma, Y., et al., Artificial M2 macrophages for disease-modifying osteoarthritis therapeutics. Biomaterials, 2021. 274: p. 120865. 7. Ma, Y., et al., Artificial M2 macrophages for disease-modifying osteoarthritis therapeutics. Biomaterials, 2021. 274: p. 120865. 8. Lu, W., et al., Genetic deficiency of the mitochondrial protein PGAM5 causes a Parkinson's-like movement disorder. Nat. Commun., 2014. 5: p. 4930. 8. Lu, W., et al., Genetic deficiency of the mitochondrial protein PGAM5 causes a Parkinson's-like movement disorder. Nat. Commun., 2014. 5: p. 4930. 9. Wang, Z., et al., The mitochondrial phosphatase PGAM5 functions at the convergence point of multiple necrotic death pathways. Cell, 2012. 148(1-2): p. 228–43. 9. References Wang, Z., et al., The mitochondrial phosphatase PGAM5 functions at the convergence point of multiple necrotic death pathways. Cell, 2012. 148(1-2): p. 228–43. 10. Ma, K., et al., Dynamic PGAM5 multimers dephosphorylate BCL-xL or FUNDC1 to regulate mitochondrial and cellular fate. Cell Death Differ., 2020. 27(3): p. 1036–1051. 10. Ma, K., et al., Dynamic PGAM5 multimers dephosphorylate BCL-xL or FUNDC1 to regulate mitochondrial and cellular fate. Cell Death Differ., 2020. 27(3): p. 1036–1051. Page 17/26 Page 17/26 11. Yu, B., et al., Mitochondrial phosphatase PGAM5 modulates cellular senescence by regulating mitochondrial dynamics. Nat. Commun., 2020. 11(1): p. 2549. 12. Moriwaki, K., et al., The mitochondrial phosphatase PGAM5 is dispensable for necroptosis but promotes inflammasome activation in macrophages. J. Immunol., 2016. 196(1): p. 407–15. 13. Knights, A.J., et al., Synovial fibroblasts assume distinct functional identities and secrete R-spondin 2 in osteoarthritis. Ann. Rheum. Dis., 2023. 82(2): p. 272–282. 14. Sica, A. and A. Mantovani, Macrophage plasticity and polarization: in vivo veritas. J. Clin. Invest., 2012. 122(3): p. 787–95. 15. Schaefer, L., et al., The matrix component biglycan is proinflammatory and signals through Toll-like receptors 4 and 2 in macrophages. J. Clin. Invest., 2005. 115(8): p. 2223–33. 16. Yang, H., et al., MTORC1 coordinates the autophagy and apoptosis signaling in articular chondrocytes in osteoarthritic temporomandibular joint. Autophagy, 2020. 16(2): p. 271-288. 17. Olefsky, J.M. and C.K. Glass, Macrophages, inflammation, and insulin resistance. Annu. Rev. Physiol., 2010. 72: p. 219–46. 18. Guo, B., et al., Antagonism of PPAR-γ signaling expands human hematopoietic stem and progenitor cells by enhancing glycolysis. Nat. Med., 2018. 24(3): p. 360–367. 18. Guo, B., et al., Antagonism of PPAR-γ signaling expands human hematopoietic stem and progenitor cells by enhancing glycolysis. Nat. Med., 2018. 24(3): p. 360–367. 19. Li, J., et al., STAT6 contributes to renal fibrosis by modulating PPARα-mediated tubular fatty acid oxidation. Cell Death Dis., 2022. 13(1): p. 66. 19. Li, J., et al., STAT6 contributes to renal fibrosis by modulating PPARα-mediated tubular fatty acid oxidation. Cell Death Dis., 2022. 13(1): p. 66. 20. Rauschenberger, V., et al., The phosphatase Pgam5 antagonizes Wnt/β-Catenin signaling in embryonic anterior-posterior axis patterning. Development, 2017. 144(12): p. 2234–2247. 21. Feng, Y., et al., Wnt/β-catenin-promoted macrophage alternative activation contributes to kidney fibrosis. J. Am. Soc. Nephrol., 2018. 29(1): p. 182–193. 22. Tian, X., et al., Long noncoding RNA LINC00662 promotes M2 macrophage polarization and hepatocellular carcinoma progression via activating Wnt/β-catenin signaling. Mol. References Oncol., 2020. 14(2): p. 462–483. 22. Tian, X., et al., Long noncoding RNA LINC00662 promotes M2 macrophage polarization and hepatocellular carcinoma progression via activating Wnt/β-catenin signaling. Mol. Oncol., 2020. 14(2): p. 462–483. 23. Oguma, K., et al., Activated macrophages promote Wnt signalling through tumour necrosis factor- alpha in gastric tumour cells. EMBO J., 2008. 27(12): p. 1671–81. 23. Oguma, K., et al., Activated macrophages promote Wnt signalling through tumour necrosis factor- alpha in gastric tumour cells. EMBO J., 2008. 27(12): p. 1671–81. 24. Lu, Y., et al., lncRNA MIR100HG-derived miR-100 and miR-125b mediate cetuximab resistance via Wnt/β-catenin signaling. Nat. Med., 2017. 23(11): p. 1331–1341. 24. Lu, Y., et al., lncRNA MIR100HG-derived miR-100 and miR-125b mediate cetuximab resistance via Wnt/β-catenin signaling. Nat. Med., 2017. 23(11): p. 1331–1341. 25. Fire, A., et al., Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature, 1998. 391(6669): p. 806-11. 25. Fire, A., et al., Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature, 1998. 391(6669): p. 806-11. 26. Evans, C.H., Catering to chondrocytes. Sci Transl Med, 2018. 10(469). 26. Evans, C.H., Catering to chondrocytes. Sci Transl Med, 2018. 10(469). 26. Evans, C.H., Catering to chondrocytes. Sci Transl Med, 2018. 10(469). 27. Kazi, A., et al., GSK3 suppression upregulates β-catenin and c-Myc to abrogate KRas-dependent tumors. Nat Commun, 2018. 9(1): p. 5154. 27. Kazi, A., et al., GSK3 suppression upregulates β-catenin and c-Myc to abrogate KRas-dependent tumors. Nat Commun, 2018. 9(1): p. 5154. 28. Mahmoudian, A., et al., Early-stage symptomatic osteoarthritis of the knee - time for action. Nat Rev Rheumatol, 2021. 17(10): p. 621-632. 28. Mahmoudian, A., et al., Early-stage symptomatic osteoarthritis of the knee - time for action. Nat Rev Rheumatol, 2021. 17(10): p. 621-632. Page 18/26 29. Wynn, T.A. and K.M. Vannella, Macrophages in Tissue Repair, Regeneration, and Fibrosis. Immunity, 2016. 44(3): p. 450-462. 30. Mantovani, A., et al., Macrophage polarization: tumor-associated macrophages as a paradigm for polarized M2 mononuclear phagocytes. Trends Immunol., 2002. 23(11): p. 549–55. 30. Mantovani, A., et al., Macrophage polarization: tumor-associated macrophages as a paradigm for polarized M2 mononuclear phagocytes. Trends Immunol., 2002. 23(11): p. 549–55. 31. Zhu, L., et al., TSC1 controls macrophage polarization to prevent inflammatory disease. Nat. Commun., 2014. 5: p. 4696. 32. Orecchioni, M., et al., Macrophage polarization: different gene signatures in M1(LPS+) vs. classically and M2(LPS-) vs. alternatively activated macrophages. Front. Immunol., 2019. 10: p. 1084. 33. References Lv, Z., et al., TRPV1 alleviates osteoarthritis by inhibiting M1 macrophage polarization via Ca(2+)/CaMKII/Nrf2 signaling pathway. Cell Death Dis, 2021. 12(6): p. 504. 34. Zhang, H., et al., Synovial macrophage M1 polarisation exacerbates experimental osteoarthritis partially through R-spondin-2. Ann Rheum Dis, 2018. 77(10): p. 1524-1534. 35. Zhang, Y., et al., Reprogramming of Mitochondrial Respiratory Chain Complex by Targeting SIRT3- COX4I2 Axis Attenuates Osteoarthritis Progression. Adv Sci (Weinh), 2023. 10(10): p. e2206144. 36. Vergadi, E., et al., Akt signaling pathway in macrophage activation and M1/M2 polarization. J. Immunol., 2017. 198(3): p. 1006–1014. 37. Zhao, S.J., et al., Macrophage MSR1 promotes BMSC osteogenic differentiation and M2-like polarization by activating PI3K/AKT/GSK3β/β-catenin pathway. Theranostics, 2020. 10(1): p. 17-35. 37. Zhao, S.J., et al., Macrophage MSR1 promotes BMSC osteogenic differentiation and M2-like polarization by activating PI3K/AKT/GSK3β/β-catenin pathway. Theranostics, 2020. 10(1): p. 17-35. 38. Bode, J.G., C. Ehlting, and D. Häussinger, The macrophage response towards LPS and its control through the p38(MAPK)-STAT3 axis. Cell Signal., 2012. 24(6): p. 1185–94. 38. Bode, J.G., C. Ehlting, and D. Häussinger, The macrophage response towards LPS and its control through the p38(MAPK)-STAT3 axis. Cell Signal., 2012. 24(6): p. 1185–94. 39. Wang, L., et al., Gut microbial bile acid metabolite skews macrophage polarization and contributes to high-fat diet-induced colonic inflammation. Gut Microbes, 2020. 12(1): p. 1-20. 39. Wang, L., et al., Gut microbial bile acid metabolite skews macrophage polarization and contributes to high-fat diet-induced colonic inflammation. Gut Microbes, 2020. 12(1): p. 1-20. 40. Chen, J.Y., et al., CTRP9 induces macrophages polarization into M1 phenotype through activating JNK pathway and enhances VSMCs apoptosis in macrophages and VSMCs co-culture system. Exp. Cell Res., 2020. 395(1): p. 112194. 41. Zhang, H., D. Cai, and X. Bai, Macrophages regulate the progression of osteoarthritis. Osteoarthr. Cartil., 2020. 28(5): p. 555–561. 42. Gao, C. and Y.G. Chen, Dishevelled: the hub of Wnt signaling. Cell Signal., 2010. 22(5): p. 717–27. 42. Gao, C. and Y.G. Chen, Dishevelled: the hub of Wnt signaling. Cell Signal. 43. Wu, D. and W. Pan, GSK3: a multifaceted kinase in Wnt signaling. Trends Biochem. Sci., 2010. 35(3): p. 161–8. 44. Denk, D., et al., Expansion of T memory stem cells with superior anti-tumor immunity by Urolithin A- induced mitophagy. Immunity, 2022. 55(11): p. 2059-2073.e8. 44. Denk, D., et al., Expansion of T memory stem cells with superior anti-tumor immunity by Urolithin A- induced mitophagy. Immunity, 2022. 55(11): p. 2059-2073.e8. 45. References Bernkopf, D.B., et al., Pgam5 released from damaged mitochondria induces mitochondrial biogenesis via Wnt signaling. J Cell Biol, 2018. 217(4): p. 1383-1394. 45. Bernkopf, D.B., et al., Pgam5 released from damaged mitochondria induces mitochondrial biogenesis via Wnt signaling. J Cell Biol, 2018. 217(4): p. 1383-1394. 46. Lv, J., et al., Fluorination Promotes the Cytosolic Delivery of Genes, Proteins, and Peptides. Acc Chem Res, 2022. 55(5): p. 722-733. Page 19/26 47. Lv, J. and Y. Cheng, Fluoropolymers in biomedical applications: state-of-the-art and future perspectives. Chem Soc Rev, 2021. 50(9): p. 5435-5467. 48. Cummings, R.D., The mannose receptor ligands and the macrophage glycome. Curr Opin Struct Biol, 2022. 75: p. 102394. 48. Cummings, R.D., The mannose receptor ligands and the macrophage glycome. Curr Opin Struct Bio 2022. 75: p. 102394. 49. Cao, W., et al., Tumor suppressor adenomatous polyposis coli sustains dendritic cell tolerance through IL-10 in a β-catenin-dependent manner. J. Immunol., 2023. 210(10): p. 1589–1597. 49. Cao, W., et al., Tumor suppressor adenomatous polyposis coli sustains dendritic cell tolerance through IL-10 in a β-catenin-dependent manner. J. Immunol., 2023. 210(10): p. 1589–1597. 50. Wang, M., et al., A fluorinated dendrimer achieves excellent gene transfection efficacy at extremely low nitrogen to phosphorus ratios. Nat Commun, 2014. 5: p. 3053. 47. Lv, J. and Y. Cheng, Fluoropolymers in biomedical applications: state-of-the-art and future perspectives. Chem Soc Rev, 2021. 50(9): p. 5435-5467. Figures Figures Figure 1 Legend not included with this version g Figure 1 Legend not included with this version Figure 1 Figure 1 Legend not included with this version Legend not included with this version Page 20/26 Figure 2 Legend not included with this version Figure 2 Legend not included with this version Legend not included with this version Page 21/26 Figure 3 Figure 3 Legend not included with this version Legend not included with this version Page 22/26 Page 22/26 Figure 4 Figure 4 Legend not included with this version Page 23/26 Legend not included with this version Legend not included with this version Page 24/26 Figure 6 Legend not included with this version Figure 6 Figure 6 Legend not included with this version Legend not included with this version Page 25/26 Figure 7 Legend not included with this version Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. Fig.S1.tif Fig.S1.tif Page 26/26 Page 26/26
https://openalex.org/W1972977299	https://europepmc.org/articles/pmc2905333?pdf=render	English	null	RNF122: A novel ubiquitin ligase associated with calcium-modulating cyclophilin ligand	BMC cell biology	2,010	cc-by	6,779	RESEARCH ARTICLE Open Access © 2010 Peng et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Background: RNF122 is a recently discovered RING finger protein that is associated with HEK293T cell viability and is overexpressed in anaplastic thyroid cancer cells. RNF122 owns a RING finger domain in C terminus and transmembrane domain in N terminus. However, the biological mechanism underlying RNF122 action remains unknown. Results: In this study, we characterized RNF122 both biochemically and intracellularly in order to gain an understanding of its biological role. RNF122 was identified as a new ubiquitin ligase that can ubiquitinate itself and undergoes degradation in a RING finger-dependent manner. From a yeast two-hybrid screen, we identified calcium- modulating cyclophilin ligand (CAML) as an RNF122-interacting protein. To examine the interaction between CAML and RNF122, we performed co-immunoprecipitation and colocalization experiments using intact cells. What is more, we found that CAML is not a substrate of ubiquitin ligase RNF122, but that, instead, it stabilizes RNF122. Conclusions: RNF122 can be characterized as a C3H2C3-type RING finger-containing E3 ubiquitin ligase localized to the ER. RNF122 promotes its own degradation in a RING finger-and proteasome-dependent manner. RNF122 interacts with CAML, and its E3 ubiquitin ligase activity was noted to be dependent on the RING finger domain. RING finger proteins contain a RING finger domain, which was first identified as being encoded by the Really Interesting New Gene in the early 1990 s [5]. The RING finger domain contains 8 metal-binding residues that coordinate 2 zinc atoms in an interleaved pattern to facil- itate correct folding, which is necessary for the biological actions of these proteins. Many RING finger proteins have been identified as ubiquitin ligases and are known to play an important role in various physiological processes. For example, MDM2, a representative RING finger pro- tein, is a ubiquitin ligase of p53 [6]. Further, Cbl is known to play an important role in the ligand-induced ubiquit- ination of epidermal growth factor receptor (EGFR) via a mechanism that involves interaction between the RING finger domain and UbcH7 [7]. The tumor necrosis factor (TNF) receptor associated factors (TRAFs) contain a RING finger domain at the N-terminal region; TNFs play an important role in both adaptive and innate immunity [8]. However, there are still hundreds of RING finger pro- teins whose functions have yet to be characterized. Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 Research article RNF122: A novel ubiquitin ligase associated with calcium-modulating cyclophilin ligand Zhi Peng1,4, Taiping Shi1,2,3 and Dalong Ma1,2,3 Correspondence: taiping_shi@yahoo.com.cn 1 Chinese National Human Genome Center, #3-707 North YongChang Road BDA, Beijing 100176, PR China Full list of author information is available at the end of the article Background The ubiquitin-proteasome system is involved in protein degradation and many biological processes such as tran- scription, cell cycle progression, antigen processing, cel- lular defense, signaling, and apoptosis [1,2]. In the ubiquitin-proteasome pathway, a ubiquitin-activating enzyme (E1) activates ubiquitin (Ub) by attaching it to a substrate via a thiol-ester linkage and then transferring the complex formed to the active-site cysteine of a ubiq- uitin carrier protein (E2). Formation of isopeptide bonds between the C terminus of Ub and the lysines on the sub- strate is catalyzed by a ubiquitin ligase (E3), which binds the substrate and catalyzes the transfer of Ub from a spe- cific E2 to the substrate. The formation of a chain of Ub molecules on the substrate generally targets it for degra- dation by the 26 S proteasome [3]. Comparative genome analysis has revealed the presence of a few genes encod- ing E1, others encoding E2, and hundreds encoding E3 ligases [4]. A cell-based screening performed in our laboratory revealed that the RNF122 gene is associated with cell via- bility [9]. RNF122 is expressed in several normal tissues and in tumor tissues and cell lines; RNF122 has been Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 Page 2 of 9 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 RING finger that catalyses its own degradation in a pro- teasome-dependent manner. localized to the endoplasmic reticulum (ER) and the Golgi apparatus [10]. Comparative genomic hybridiza- tion (CGH) has revealed that RNF122 is overexpressed in anaplastic thyroid cancer cells [11]. However, RNF122 expression is not invariably detected in mammalian and Escherichia coli cells. The presence of the RING finger domain in RNF122 implies that RNF122 may be involved in the ubiquitylation pathway; however, to date, there has been no study that has sought to biochemically character- ize of RNF122. Moreover, the mechanism of apoptosis induction mediated by RNF122 remains unclear. Given the importance of the biological functions of the RING finger proteins and the overexpression of RNF122 in ana- plastic thyroid cancer cells, a functional characterization of this gene is highly warranted. The present study pro- vides evidence that RNF122 is a new uncharacterized ubiquitin ligase. Further, we demonstrate that RNF122 interacts with CAML. We constructed a full-length GST-RNF122 fusion pro- tein, but we were unable to demonstrate successful expression of this protein in bacteria (data not shown), which is consistent with the results obtained by Yu [11]. Background The transmembrane sites, usually represented by a hydrophobic sequence, are not expressed in E. coli. Hence, we constructed two truncated mutants of GST- RNF122, namely, GST-RNF122ΔTM and GST- mutRNF122ΔTM, in which the transmembrane (TM) domain was deleted. As illustrated in Figure 1C, the expression of both GST-RNF122ΔTM and GST- mutRNF122ΔTM was inducible. These proteins were purified successfully by affinity chromatography and con- firmed by western blotting using an anti-GST antibody. In vivo and in vitro ubiquitination of RNF122 q To further examine the ubiquitination of RNF122, we performed in vivo ubiquitination studies. We coex- pressed RNF122-myc with HA-ubiquitin in HEK293T cells, recovered the immunoprecipitates using anti-myc antibody, and performed immunoblotting using the anti- HA antibody. As shown in Figure 2A, a smeared band, which indicates ubiquitination, appeared in the blots of lysates of cells cotransfected with RNF122-myc and HA- ubiquitin, but this was not detected in the vector control group. Hence, we confirmed that RNF122 undergoes ubiquitination in vivo. Further, we performed the same experiment using an RNF122 mutated in the RING finger domain and found that this mutant (RNF122C92A-myc) was also ubiquitylated in vivo, as evident from the high- molecular-weight smeared band detected on blots of the lysates of the cells coexpressing RNF122C92A-myc and ubiquitin (Figure 2A). However, the degree of ubiquitina- tion of this mutant protein was significantly lower than that of RNF122-myc; a finding that is consistent with the proteasome degradation of the wild-type and mutant RNF122. Characterization of RNF122 expression in mammalian cells and E. coli BMC Cell Biology 2010, 11:41 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 Page 3 of 9 g gy http://www.biomedcentral.com/1471-2121/11/41 Figure 1 Characterization of RNF122 expression in mammalian and E. coli cells. (A). HEK293T and HeLa cells were transiently transfected with RNF122-myc. At 24 h post-transfection, the cells were treated with 2.5 μg/mL MG132 or 50 μg/mL tunicamycin and harvested at the indicated time. The cell lysates were subjected to SDS-PAGE and analyzed by western blotting using anti-myc and anti-β-actin antibodies. (B). HEK293T cells were transiently transfected with RNF122-myc or with RNF122C92A-myc or RNF122C95A-myc mutants. At 24 h post-transfection, the cells were treated with 50 μg/mL MG132, harvested after 6 h, and then analyzed by western blotting. (C). GST-RNF122ΔTM or GST-RNF122-ΔTM mutant proteins were expressed in E. coli and purified by affinity chromatography. The proteins were verified by SDS-PAGE and western blotting using an anti-GST primary antibody. a: pre-induction; b: post-induction; c: effluent obtained after affinity chromatography; d: purified protein. Figure 1 Characterization of RNF122 expression in mammalian and E. coli cells. (A). HEK293T and HeLa cells were transiently transfected with RNF122-myc. At 24 h post-transfection, the cells were treated with 2.5 μg/mL MG132 or 50 μg/mL tunicamycin and harvested at the indicated time. The cell lysates were subjected to SDS-PAGE and analyzed by western blotting using anti-myc and anti-β-actin antibodies. (B). HEK293T cells were transiently transfected with RNF122-myc or with RNF122C92A-myc or RNF122C95A-myc mutants. At 24 h post-transfection, the cells were treated with 50 μg/mL MG132, harvested after 6 h, and then analyzed by western blotting. (C). GST-RNF122ΔTM or GST-RNF122-ΔTM mutant proteins were expressed in E. coli and purified by affinity chromatography. The proteins were verified by SDS-PAGE and western blotting using an anti-GST primary antibody. a: pre-induction; b: post-induction; c: effluent obtained after affinity chromatography; d: purified protein. results also suggested that CAML interacts with RNF122 through the RING finger domain. Hence, the RING fin- ger protein functions as a ubiquitin ligase that ubiquitiny- lates the substrate and targets it to the ubiquitin proteasome system. We accordingly performed experi- ments to determine whether CAML is a substrate of RNF122 and whether it can be degraded by the ubiquitin proteasome system. As shown in Figure 3B, however, CAML levels in MG132-treated cells were not elevated, suggesting that CAML may not be involved in the protea- some-dependent degradation pathway. Characterization of RNF122 expression in mammalian cells and E. coli Furthermore, CAML protein levels remained unchanged even when CAML-FLAG and RNF122-myc or mutant RNF122 were coexpressed (Figure 3C). Overall, we conclude that CAML interacts with RNF122 but is not a substrate for its ubiquitylating activity. Interestingly, RNF122 expres- sion was occasionally detected in cells that overexpressed CAML. As illustrated in Figure 3D, RNF122 was stabi- lized when CAML was overexpressed. Hence, it can be speculated that the interaction between CAML and RNF122 inhibits the ubiquitination of RNF122, and thereby stabilizes RNF122. assay using RNF122 containing a single-site mutations (Cys92Ala) in the RING finger domain (GST- RNF122C92A-ΔTM). As shown in Figure 2C, the muta- tion compromised the autoubiquitylation activity of UbcH5a, UbcH5b, UbcH5c, Ubc6, and Ubc13. These results demonstrate that RNF122 is an E3 ligase that is selective for UbcH5a, UbcH5b, UbcH5c, Ubc6, and Ubc13. assay using RNF122 containing a single-site mutations (Cys92Ala) in the RING finger domain (GST- RNF122C92A-ΔTM). As shown in Figure 2C, the muta- tion compromised the autoubiquitylation activity of UbcH5a, UbcH5b, UbcH5c, Ubc6, and Ubc13. These results demonstrate that RNF122 is an E3 ligase that is selective for UbcH5a, UbcH5b, UbcH5c, Ubc6, and Ubc13. Characterization of RNF122 expression in mammalian cells and E. coli Characterization of RNF122 expression in mammalian cells and E. coli Yu found that epitope-tagged RNF122 was not success- fully expressed in mammalian and bacterial cells [11]. In the present study, we performed similar experiments and obtained the same results. However, northern blotting, RT-PCR, and subcellular localization studies have dem- onstrated the existence of RNF122 [10]. Many RING fin- ger proteins are ubiquitin ligases that ubiquitinate themselves to facilitate their degradation by the ubiquitin proteasome. We accordingly examined the expression of RNF122 in HeLa and HEK293T cells after treatment with one of the ubiquitin proteasome inhibitors, namely, MG132. As illustrated in Figure 1A, RNF122 was detected in the MG132-treated cells. This result demon- strates that RNF122 is a substrate of the ubiquitin protea- some system. To verify the importance of the RING finger domain in the stability of RNF122, a mutant RNF122 gene was constructed using PCR-based muta- genesis and transfected into HEK293T cells. Subse- quently, the total protein was extracted and analyzed by western blotting. The mutant RNF122 was detected in cells not treated with MG132; further, its expression was not significantly affected by MG132 treatment (Figure 1B). These results suggest that the RING finger domain is critical for the degradation of RNF122. Figure 1A and 1B shows 2 bands of approximately 18 kD and 26 kD, indi- cating that RNF122 may be glycosylated; this result is consistent with the fact that the RNF122 sequence con- tains some potential N-linked glycosylation sites (NxT/S). Moreover, treatment of the cells with tunicamycin, which inhibits N-linked glycosylation, caused a shift in the pro- tein bands (Figure 1A). These results indicate that RNF122 is an N-linked glycosylated protein containing a The presence of the RING finger domain in RNF122 confers the properties of a ubiquitin ligase to this protein. To investigate whether RNF122 has ubiquitin ligase activ- ity, we established an in vitro autoubiquitination assay, which was performed according to the manufacturer's instructions. Briefly, E1, E2, and bacterially expressed GST-RNF122ΔTM were added to a reaction mixture containing ubiquitin and ATP. We treated the RNF122 fusion proteins with different E2 enzymes to determine which E2 supports RNF122-mediated ubiquitination. As is evident from Figure 2B, polyubiquitination was detected only in cells treated with the following E2-conju- gating enzymes: UbcH5a, UbcH5b, UbcH5c, Ubc6, and Ubc13. To clarify the role of the RING finger domain in ubiquitination, we performed an auto ubiquitination Peng et al. RNF122 interacts with CAML Next, the yeast two-hybrid system was used to identify the potential substrates of RNF122. The experiments were performed by Ruixing Corp. (Shanghai, China). The results revealed that RNF122 interacts with CAML, a protein associated with calcium signaling and T cell acti- vation. To confirm the interaction between RNF122 and CAML, we co-expressed these proteins in HEK293T cells and immunoprecipitated RNF122-myc with the anti-myc antibody. Western blot analysis of the immunoprecipitate obtained using the anti-FLAG antibody revealed that RNF122 but not its mutant form was coprecipitated with FLAG-tagged CAML (CAML-FLAG), demonstrating their association in mammalian cells (Figure 3A). The Page 4 of 9 Peng et al. BMC Cell Biology 2010, 11:41 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 g gy http://www.biomedcentral.com/1471-2121/11/41 Figure 2 In vivo and in vitro ubiquitylation of RNF122. (A) HEK293T cells were seeded into 60-mm plates and transfected with 5 μg of Ub-HA and 5 μg of pcDB (empty) or wild-type or mutant RNF122-myc expression plasmid. At 48 h post-transfection, cell lysates in RIPA buffer were immunopre- cipitated with anti-myc antibody agarose beads. The immunoprecipitates were analyzed on an immunoblot for HA. A smeared band was observed in the RNF122-myc group. (B) E2-selective ubiquitin ligase activity of RNF122. Bacterially expressed GST-RNF122ΔTM was incubated with ubiquitin, E1, and various E2 enzymes. Autoubiquitylation of RNF122 was detected by HRP-streptavidin immunoblotting. (C) RNF122 E3 ligase activity required an intact RING finger domain. The reactions were performed using ubiquitin, E1, E2 (UbcH5b or UbcH5c), and E3 (GST-RNF122ΔTM, or mutant GST- RNF122ΔTM), as indicated. Colocalization of RNF122 and CAML in HEK293T and HeLa cells In order to confirm the intracellular interaction between RNF122 and CAML, we constructed a plasmid, namely, pEGFP-N1-RNF122, encoding green fluorescent protein, and a CAML-FLAG-encoding plasmid and cotransfected these into HeLa cells or HEK293T cells. As shown in Fig- Colocalization of RNF122 and CAML in HEK293T and HeLa cells Colocalization of RNF122 and CAML in HEK293T and HeLa cells HEK293T cells were seeded into 60-mm plates and transfected with pcDB (empty), RNF122-myc or mut-RNF122-myc vectors in conjunction with CAML-FLAG. After 48 h, the cells were treated with 2.5 μg/mL MG132 for 6 h, lysed in PBS containing 1% Triton X-100, and immunoprecipitated with anti-myc antibody agarose beads. (B) CAML is not involved in the proteasome-dependent degradation pathway. HEK293T cells were treated with 2.5 μM MG132 or vehicle (DMSO) for 6 h. The cells were then harvested and the expression of CAML was analyzed by western blotting using an anti-CAML antibody. The membrane was subsequently reprobed with an anti-β-actin antibody. (C) HEK293T cells were transiently cotransfected with CAML-FLAG and pcDB, RNF122-myc, RNF122C92A-myc, or RNF122C95A-myc. At 24 h post-transfection, the cells were treated with 50 μg/mL CHX and harvested at the indicated times. The cell lysates were subjected to SDS-PAGE and analyzed by western blotting by using antisera against FLAG and β-actin. (D) The level of RNF122 was found to increase in a proteasome-independent manner when it was coexpressed with CAML. HEK293T cells were transfected with empty (pcDB) or CAML expression vector along with RNF122-myc, and subsequently treated with 2.5 μg/mL MG132 or DMSO at 24 h post-transfection. After 24 h, the cell lysates were suspended in PBS containing 1% Triton X-100 and a protease inhibitor cocktail, and then immunoblotted with an anti-myc antibody. ous finding that RNF122 is localized to the ER. These data confirmed that RNF122 and CAML interact to form a complex that is functional in the cytoplasm. teins destined for proteasome-dependent degradation, although there have been no reports confirming their E3 activities toward these proteins. In our previous study, we performed northern blotting and RT-PCR analysis, and confirmed that RNF122 is widely expressed in several tissues and cell lines and that it is localized to the ER and the Golgi apparatus. A cell- based screening test revealed that RNF122 can downreg- ulate the expression of the prolactin (PRL) gene. Further studies showed that RNF122 was associated with the via- bility of HEK293T cells. However, the exact biological function of RNF122 is still unknown. In the present study, we characterized the biological role of RNF122. RNF122 contains a RING domain, which suggests that it catalyzes its own degradation. Furthermore, through in vivo and in vitro ubiquitination assays, RNF122 was identified as a new ubiquitin ligase. The TM domain is critical to the localization of RNF122 RNF122 contains a putative TM domain near the N ter- minus (amino acid residues 37-59). To investigate whether this domain is required for the localization of RNF122 to the ER, we generated a truncated mutant of RNF122 in which 60 N-terminal amino acids were deleted. The expression vectors pcDB/RNF122-ΔTM and RNF122-ΔTM-GFP were constructed and used to evalu- ate the functions and subcellular localization of the trun- cated RNF122. The cytoplasmic expression pattern of the RNF122 mutant was noted to be diffuse as compared to the expression pattern of the wild-type RNF122 (Figure 4B), indicating that the N-terminal TM domain of RNF122 is required for its specific localization. These findings suggest that the putative N-terminal TM domain is necessary for the localization of RNF122. CAML is a ubiquitous protein containing 296 amino acids. It contains 3 putative TM domains near its C ter- minus and has been shown to be a resident protein of the ER [12,13]. There are several proteins that can interact with CAML, including CAML interactor (TACI) [14], epidermal growth factor receptor (EGFR) [15], p56lck [16], Kaposi's sarcoma virus-associated mitochondrial K7 pro- Colocalization of RNF122 and CAML in HEK293T and HeLa cells ure 4A, both pEGFP-N1-RNF122 and CAML-FLAG were observed to be localized in the cytoplasm and were highly concentrated in spots around the nucleus; moreover, the distribution of CAML was similar to that of RNF122. Fur- ther experiments on the localization of CAML yielded the same result, namely, that CAML is partially localized to the ER. This observation is consistent with our previ- In order to confirm the intracellular interaction between RNF122 and CAML, we constructed a plasmid, namely, pEGFP-N1-RNF122, encoding green fluorescent protein, and a CAML-FLAG-encoding plasmid and cotransfected these into HeLa cells or HEK293T cells. As shown in Fig- Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 Peng et al. BMC Cell Biology 2010, 11:41 Page 5 of 9 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 g gy http://www.biomedcentral.com/1471-2121/11/41 Figure 3 Interaction of RNF122 with CAML. (A) In mammalian cells, RNF122 interacted with CAML. HEK293T cells were seeded into 60-mm plates and transfected with pcDB (empty), RNF122-myc or mut-RNF122-myc vectors in conjunction with CAML-FLAG. After 48 h, the cells were treated with 2.5 μg/mL MG132 for 6 h, lysed in PBS containing 1% Triton X-100, and immunoprecipitated with anti-myc antibody agarose beads. (B) CAML is not involved in the proteasome-dependent degradation pathway. HEK293T cells were treated with 2.5 μM MG132 or vehicle (DMSO) for 6 h. The cells were then harvested and the expression of CAML was analyzed by western blotting using an anti-CAML antibody. The membrane was subsequently reprobed with an anti-β-actin antibody. (C) HEK293T cells were transiently cotransfected with CAML-FLAG and pcDB, RNF122-myc, RNF122C92A-myc, or RNF122C95A-myc. At 24 h post-transfection, the cells were treated with 50 μg/mL CHX and harvested at the indicated times. The cell lysates were subjected to SDS-PAGE and analyzed by western blotting by using antisera against FLAG and β-actin. (D) The level of RNF122 was found to increase in a proteasome-independent manner when it was coexpressed with CAML. HEK293T cells were transfected with empty (pcDB) or CAML expression vector along with RNF122-myc, and subsequently treated with 2.5 μg/mL MG132 or DMSO at 24 h post-transfection. After 24 h, the cell lysates were suspended in PBS containing 1% Triton X-100 and a protease inhibitor cocktail, and then immunoblotted with an anti-myc antibody. Figure 3 Interaction of RNF122 with CAML. (A) In mammalian cells, RNF122 interacted with CAML. Colocalization of RNF122 and CAML in HEK293T and HeLa cells We also found that the TM domain of RNF122 is critical for its localization to the ER. Discussion The importance of the RING finger domain in ubiquit- ination is supported by the fact that several RING finger proteins are associated with Ubcs and/or specific pro- Page 6 of 9 Peng et al. BMC Cell Biology 2010, 11:41 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 tein [17], type-1 angiotensin II receptor-associated pro- tein (ATRAP) [18], and γ2 subunit-containing gamma- i b t i id t A t (GABA R) [19] C the CAML levels. Nonetheless, it seems unlikely that CAML is a substrate for the ubiquitin ligase activity of RNF122 Thi iti i b d th b ti Figure 4 Subcellular localization of RNF122. (A) HeLa or HEK293T cells were grown on coverslips and transfected with RNF122-N1 and CAML-FLAG. To inhibit the degradation of RNF122-N1, the cells were treated with 2.5 μM MG132 for 6 h. The cells were then treated with anti-FLAG antibody and FITC-conjugated goat anti-mouse antibody, and observed under a confocal microscopy. The results showed that the cytoplasmic distribution of CAML entirely overlapped that of RNF122. (B) HEK293T cells were transfected with RNF122-N1 and RNF122ΔTM-N1 and observed under a fluores- cence microscope. Figure 4 Subcellular localization of RNF122. (A) HeLa or HEK293T cells were grown on coverslips and transfected with RNF122-N1 and CAML-FLAG. To inhibit the degradation of RNF122-N1, the cells were treated with 2.5 μM MG132 for 6 h. The cells were then treated with anti-FLAG antibody and FITC conjugated goat anti mouse antibody and observed under a confocal microscopy The results showed that the cytoplasmic distribution of Figure 4 Subcellular localization of RNF122. (A) HeLa or HEK293T cells were grown on coverslips and transfected with RNF122-N1 and CAML-FLAG. To inhibit the degradation of RNF122-N1, the cells were treated with 2.5 μM MG132 for 6 h. The cells were then treated with anti-FLAG antibody and FITC-conjugated goat anti-mouse antibody, and observed under a confocal microscopy. The results showed that the cytoplasmic distribution of CAML entirely overlapped that of RNF122. (B) HEK293T cells were transfected with RNF122-N1 and RNF122ΔTM-N1 and observed under a fluores- cence microscope. the CAML levels. Nonetheless, it seems unlikely that CAML is a substrate for the ubiquitin ligase activity of RNF122. This supposition is based on the observation that CAML levels were not reduced by the coexpression of RNF122 and were not increased in MG132-treated cells. tein [17], type-1 angiotensin II receptor-associated pro- tein (ATRAP) [18], and γ2 subunit-containing gamma- aminobutyric acid type A receptor (GABAAR) [19]. Con- sistent with the findings of the previous studies on CAML, our findings showed that RNF122 interacts with CAML. The coimmunoprecipitation and colocalization experiments verified the interaction between RNF122 and CAML in mammalian cells and, at the same time, showed that overexpression of RNF122 does not reduce We have validated the interaction between CAML and RNF122 in vivo when the proteins are overexpressed, which may not necessarily represent the endogenous Page 7 of 9 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 condition. However, we have failed to testify endogenous interaction because of the limited RNF122 expression and its poor antibody. In the stability study, RNF122 can be stabilized when CAML was overexpressed, which give a clue that RNF122 and CAML are functional related. It has been concluded that CAML has been proposed to regulate trafficking of EGFR and the GABAA receptor. The stalibility of EGFR and GABAA receptor can be mod- ulated through c-Cbl and Plic-1 respectively, and they both are involved in ubiquitin-proteasome system [20,21]. So we speculate that RNF122 may regulate EGFR and GABAA receptor stability through ubiquitin-protea- some system. We are still focus on the biological role of RNF122 in the next investigation. The results may give us a new direction to clarify the biological function of RNF122. S-transferase) were obtained using PCR fragments encoding residues 61-155 of human RNF122 or RNF122- C92A-myc. The PCR primers contained XhoI and BamHI sites. Cell lines and reagents Human embryonic kidney cell line HEK293T and human cervix carcinoma cell line HeLa were obtained from the American Type Culture Collection (Manassas, VA). HEK293T and HeLa cells were cultured (37°C, 5% CO2 humidified atmosphere) in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum (FBS) (Hyclone, Logan, UT) and supplemented with 2 mM L-glutamine (Invitrogen, Carls- bad, CA). Monoclonal mouse antibodies against β-actin, c-myc, and FLAG were purchased from Sigma (Sigma- Aldrich, St Louis, MO); horseradish peroxidase (HRP)- streptavidin and HRP-myc antibodies, from Upstate (USA); MG132, from Sigma (Sigma-Aldrich, St Louis, MO); tunicamycin, from Alexis (USA); cycloheximide, from Calbiochem (USA); and a ubiquitin-conjugating enzyme kit (mammalian) from Biomol (USA). The resulting PCR fragment was inserted into the pGEX-4T-1 vector (GE Healthcare, USA). All the inser- tions were confirmed by DNA sequencing. DNA transfection was performed using VigoFect (Vig- orous, China), a non-liposomal cationic formula, accord- ing to the manufacturer's instructions. Prokaryotic expression E. coli strain BL-21(DE3) (Novagen, Madison, WI) were cultured overnight at 37°C and induced with 0.4 mM iso- propyl-1-thio-β-D-galactopyranoside for 1 h at room temperature (RT). The bacterial pellets were resuspended in a solution containing 50 mM Tris (pH 7.4), 1 mM eth- ylenediaminetetraacetic acid (EDTA), 1% Triton X-100, 5 mM dithiothreitol (DTT), and 2 mM phenylmethylsulfo- nyl fluoride (PMSF) (sonication buffer), and lysed by probe sonication using 4 mL of sonication buffer per 100 mL of bacterial culture. The sonicate was clarified by cen- trifugation at 4°C for 15 min at 18,000 × g, divided into aliquots, and stored at -70°C. To estimate the level of GST fusion proteins expressed, the sonicates were incubated with glutathione-Sepharose (GS) beads, washed, and sub- jected to sodium dodecyl sulfate-polyacrylamide gel elec- trophoresis (SDS-PAGE); subsequently, the gel was stained with Coomassie Brilliant Blue R-250. Previous study show that CAML can interact with EGFR in ligand-dependent way, which gave us a clue that RNF122 may involved in EGFR pathway. We have also found that RNF122 effect on cell viability may relate to ERK pathway (data not shown). However, the precise molecular mechanisms such as the interaction between EGFR and RNF122 still need to be further explored. Conclusions In conclusion, RNF122 can be characterized as a C3H2C3-type RING finger-containing E3 ubiquitin ligase localized to the ER. RNF122 promotes its own degrada- tion in RING finger-and proteasome-dependent manner. RNF122 interacts with CAML, and its E3 ubiquitin ligase activity was noted to be dependent on the RING finger domain. Further studies are required to clarify the molec- ular mechanism, such as substrate binding, that underlie RNF122 action. Constructs and transfection Full-length cDNA of RNF122 (GenBank accession no. NM_024787.2) was obtained and pEGFP-N1-RNF122 was constructed according to previously described meth- ods [10](Wang, Shi et al. 2006). RNF122-myc was ampli- fied and then ligated into pcDNA.3.1/myc-His(-)B (Invitrogen, USA). The first or the second cysteine resi- due of the RING finger was substituted with an alanine residue by polymerase chain reaction (PCR)-based site- directed mutagenesis, which resulted in the formation of products designated as RNF122C92A-myc or RNF122C95A-myc, respectively. We get RNF122-ΔTM- GFP and CAML-FLAG by inserting RNF122-ΔTM and CAML into the pEGFP-N1 and pFLAG-CMV2 vectors (Clontech, USA), respectively, using a PCR-based method. The bacterially expressed GST-RNF122ΔTM or GST-mutRNF122ΔTM fusion proteins (GST; glutathione Yeast Two-Hybrid Screening The HEK293T or HeLa cells were cotransfected with pEGFP-N1-RNF122 and CAML22-N1 CCAML-FLAG, grown on coverslips, fixed using 3.7% paraformaldehyde at RT for 30 min, and then permeabilized with 0.1% Tri- ton X-100 at RT for a further 10 min. The cover slips were washed 3 times with PBS, treated with a blocking buffer (5% bovine serum albumin in PBS) for 30 min, and incu- bated with an anti-FLAG primary antibody for 2 h at RT. The cells were then washed 3 times (10 min each) in PBS and incubated with a secondary antibody for 1 h at 37°C. Rhodamine-conjugated goat anti-mouse immunoglobu- lin G (IgG; ZhongShan Biotechnology) was used as the secondary antibody. Yeast Two-Hybrid Screening was performed at Shanghai Genomics (Shanghai, China). The two-hybrid screening system has been previously described [22]. Briefly, the library consisted of 1500 known genes associated with cell apoptosis, cell proliferation, and cell cycles. Each ORF was amplified by PCR using Pfu DNA polymerase and cloned into pGBK-RC, a Gal4 DNA-binding domain- based bait vector, and pGAD-RC, a Gal4 activation domain-based prey vector, following the MATCH- MAKER GAL4 Two-Hybrid System 3 and Libraries User Manual PT3247-1 (PR94575) protocol (Clontech, Moun- tain View, CA). Plasmids with inserts of expected sizes were confirmed by colony PCR followed by agarose gel electrophoresis. RNF122 bait vector and prey vectors were cotransfected in yeast Y190 and spread into SD/-T- L-H. Formed colonies were picked out, cracked in lique- facient nitrogen, and subsequently utilized in colony lift filter assays. Authors' contributions d h d ZP participated in the design of the study, carried out the function research, performed the validation assays, and was pivotal in drafting the manuscript. DM and TS supervised the study and were involved in the conceptualization and writing. All authors read and approved the final manuscript. In vitro ubiquitination In vitro ubiquitination was performed according to the instructions provided with the ubiquitination kit (BioMol, USA). Briefly, the assays were carried out at 30°C in a 50-μL reaction mixture containing 25 mM 4-(2- hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 100 U/mL of isopentenylpyrophosphate (IPP) (Sigma- Aldrich, St Louis, MO), 50 mM DTT, 50 mM EDTA, 1 mM Mg-ATP (pH 7.5), 100 nM E1 enzyme, 1 μM E2 enzyme, 10 μM GST-RNF122-ΔTM (or mutant GST- RNF122-ΔTM) fusion protein, and 2.5 μM biotin-labeled ubiquitin. After 30-60 min, the reactions were terminated by the addition of a non-reducing gel loading buffer, and subsequently the protein components were separated by SDS-PAGE and analyzed by western blotting using the HRP-streptavidin antibody. In vivo ubiquitination In order to analyze ubiquitination, we transfected the cells with a myc-tagged RNF122 vector and hemaggluti- nin (HA)-tagged ubiquitin, as described previously. Next, we harvested the cells and incubated them with 1 volume of 2% SDS in TBS (10 mM Tris-HCl, pH 8.0) at 95°C for 10 min in order to achieve lysis. Subsequently, we added 9 volumes of 1% Triton X-100 and 2 mM EDTA in TBS to the cell lysates, and incubated the lysates on ice for 1 h. The protein concentrations of the lysates were deter- mined by the bicinchoninic acid (BCA) assay. For immu- noprecipitation, 1 mg of protein was incubated with the anti-myc antibody at 4°C overnight; subsequently, this mixture was incubated with protein G beads for 2 h. The Immunoprecipitation and immunoblotting The samples were then eluted using 60 μL of 2 × SDS loading buffer (Sigma-Aldrich, St Louis, MO) and analyzed by western blotting using anti-myc or anti-FLAG antibodies. Subsequently, the bands were detected using Ig-HRP-conjugated antibodies and an enhanced chemiluminescence (ECL) detection system (GE Healthcare, USA). beads were washed twice with NaCl (1 M) in TBS supple- mented with NP-40 (1%), β-mercaptoethanol (0.05%), and EDTA (1 mM). The proteins were loaded onto a 12% SDS-PAGE gel and analyzed by immunoblotting using the aforementioned antibodies and an ECL detection kit (GE Healthcare, USA). Immunoprecipitation and immunoblotting HEK293T cells were transiently transfected with the epitope-tagged constructs using VigoFect, as described above. Forty-eight hours after transfection, the cells were washed 3 times in phosphate-buffered saline (PBS), har- vested by scraping, and centrifuged (5 min, 500 × g). The pelleted cells were homogenized in a cell lysis buffer [50 Page 8 of 9 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 mmol/L Tris-HCl (pH 7.4), 150 mmol/L NaCl, 5 mmol/L EDTA-Na2, 1% NP-40; 1 μg/μL pepstain, 1 μg/μL aproti- nin, 1 μg/μL leupeptin, 1 mmol/L DTT, and 0.0174 μg/μL PMSF]. The cell lysates thus obtained were centrifuged (20 min, 14,000 × g, 4°C), and the resulting supernatants were combined with 12.5 μL (packed gel) of either anti-c- Myc or anti-FLAG M2 affinity agarose (Sigma-Aldrich, St Louis, MO), mixed, and incubated in 4°C overnight. The immunoadsorbents were recovered by centrifugation (5 min, 700 × g) and washed 3 times with a high-salt buffer [500 mmol/L Tris-HCl (pH 7.4), 150 mmol/L NaCl, 5 mmol/L EDTA-Na2, 1% NP-40, 1 μg/μL pepstain, 1 μg/μL aprotinin, 1 μg/μL leupeptin, 1 mmol/L DTT, and 0.0174 μg/μL PMSF], followed by re-suspension in cell lysis cen- trifugation (5 min, 700 × g), and the same procedure was repeated using PBS. The samples were then eluted using 60 μL of 2 × SDS loading buffer (Sigma-Aldrich, St Louis, MO) and analyzed by western blotting using anti-myc or anti-FLAG antibodies. Subsequently, the bands were detected using Ig-HRP-conjugated antibodies and an enhanced chemiluminescence (ECL) detection system (GE Healthcare, USA). mmol/L Tris-HCl (pH 7.4), 150 mmol/L NaCl, 5 mmol/L EDTA-Na2, 1% NP-40; 1 μg/μL pepstain, 1 μg/μL aproti- nin, 1 μg/μL leupeptin, 1 mmol/L DTT, and 0.0174 μg/μL PMSF]. The cell lysates thus obtained were centrifuged (20 min, 14,000 × g, 4°C), and the resulting supernatants were combined with 12.5 μL (packed gel) of either anti-c- Myc or anti-FLAG M2 affinity agarose (Sigma-Aldrich, St Louis, MO), mixed, and incubated in 4°C overnight. The immunoadsorbents were recovered by centrifugation (5 min, 700 × g) and washed 3 times with a high-salt buffer [500 mmol/L Tris-HCl (pH 7.4), 150 mmol/L NaCl, 5 mmol/L EDTA-Na2, 1% NP-40, 1 μg/μL pepstain, 1 μg/μL aprotinin, 1 μg/μL leupeptin, 1 mmol/L DTT, and 0.0174 μg/μL PMSF], followed by re-suspension in cell lysis cen- trifugation (5 min, 700 × g), and the same procedure was repeated using PBS. 1. Haglund K, Dikic I: Ubiquitylation and cell signaling. EMBO J 2005, 24(19):3353-3359. 1. Haglund K, Dikic I: Ubiquitylation and cell signaling. EMBO J 2005, 24(19):3353-3359. 2. Nalepa G, Rolfe M, Harper JW: Drug discovery in the ubiquitin- proteasome system. Nat Rev Drug Discov 2006, 5(7):596-613. 3. Nandi D, Tahiliani P, Kumar A, Chandu D: The ubiquitin-proteasome system. J Biosci 2006, 31(1):137-155. 4. Semple CA, RIKEN GER Group, GSL Menbers: The comparative proteomics of ubiquitination in mouse. Genome Res 2003, 13(6B):1389-1394. 5. Freemont PS, Hanson IM, Trowsdale J: A novel cysteine-rich sequence motif. Cell 1991, 64(3):483-484. 6. Honda R, Tanaka H, Yasuda H: Oncoprotein MDM2 is a ubiquitin ligase E3 for tumor suppressor p53. FEBS Lett 1997, 420(1):25-27. 7. Yokouchi M, Kondo T, Houghton A, Bartkiewicz M, Horne WC, Zhang H, Yoshimura A, Baron R: Ligand-induced ubiquitination of the epidermal growth factor receptor involves the interaction of the c-Cbl RING finger and UbcH7. J Biol Chem 1999, 274(44):31707-31712. 8. Chung JY, Park YC, Ye H, Wu H: All TRAFs are not created equal: common and distinct molecular mechanisms of TRAF-mediated signal transduction. J Cell Sci 2002, 115(Pt 4):679-688. 9. Wang L, Gao X, Gao P, Deng W, Yu P, Ma J, Guo J, Wang X, Cheng H, Zhang C, Yu C, Ma X, Lv B, Lu Y, Shi T, Ma D: Cell-based screening and validation of human novel genes associated with cell viability. J Biomol Screen 2006, 11(4):369-376. 10. Wang L, Shi T, Wang L, Yu CF, Zeng LE, Wang J, Wang L: Cloning, expression and subcellular localization of a novel human gene- RNF122. Beijing Da Xue Xue Bao 2006, 38(3):239-243. 11. Yu W, Imoto I, Inoue J, Onda M, Emi M, Inazawa J: A novel amplification target, DUSP26, promotes anaplastic thyroid cancer cell growth by inhibiting p38 MAPK activity. Oncogene 2007, 26(8):1178-1187. 11. Yu W, Imoto I, Inoue J, Onda M, Emi M, Inazawa J: A novel amplification target, DUSP26, promotes anaplastic thyroid cancer cell growth by inhibiting p38 MAPK activity. Oncogene 2007, 26(8):1178-1187. 12. Bram RJ, Crabtree GR: Calcium signalling in T cells stimulated by a cyclophilin B-binding protein. Nature 1994, 371(6495):355-358. g p y g 12. Bram RJ, Crabtree GR: Calcium signalling in T cells stimulated by a cyclophilin B-binding protein. Nature 1994, 371(6495):355-358. y 13. Holloway MP, Bram RJ: Co-localization of calcium-modulating cyclophilin ligand with intracellular calcium pools. J Biol Chem 1998, 273(26):16346-16350. 14. Author Details 1Chinese National Human Genome Center, #3-707 North YongChang Road BDA, Beijing 100176, PR China, 2Laboratory of Medical Immunology, School of Basic Medical Science, Peking University Health Science Center, 38# Xueyuan Road, Beijing, 100191, PR China, 3Peking University Center for Human Disease Genomics, 38# Xueyuan Road, Beijing, 100191, PR China and 4Department of Medical Oncology, Yijishan Hospital, Wannan Medical College, Wuhu, Anhui, 241001, PR China 22. Ma X, Zhao H, Shan J, Long F, Chen Y, Chen Y, Zhang Y, Han X, Ma D: PDCD10 interacts with Ste20-related kinase MST4 to promote cell growth and transformation via modulation of the ERK pathway. Mol Biol Cell 2007, 18(6):1965-1978. Received: 2 December 2009 Accepted: 17 June 2010 Published: 17 June 2010 This article is available from: http://www biomedcentral com/1471 2121/11/41 © 2010 Peng et al; licensee BioMed Central Ltd This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons org/licenses/by/2 0) which permits unrestricted use distribution and repro BMC Cell Biology 2010 11:41 doi: 10.1186/1471-2121-11-41 Cite this article as: Peng et al., RNF122: A novel ubiquitin ligase associated with calcium-modulating cyclophilin ligand BMC Cell Biology 2010, 11:41 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 Page 9 of 9 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 Peng et al. BMC Cell Biology 2010, 11:41 http://www.biomedcentral.com/1471-2121/11/41 21. Yokouchi M, Kondo T, Houghton A, Bartkiewicz M, Horne WC, Zhang H, Yoshimura A, Baron R: Ligand-induced ubiquitination of the epidermal growth factor receptor involves the interaction of the c-Cbl RING finger and UbcH7. J Biol Chem 1999, 274(44):31707-31712. 22. Ma X, Zhao H, Shan J, Long F, Chen Y, Chen Y, Zhang Y, Han X, Ma D: PDCD10 interacts with Ste20-related kinase MST4 to promote cell growth and transformation via modulation of the ERK pathway. Mol Biol Cell 2007, 18(6):1965-1978. doi: 10.1186/1471-2121-11-41 Cite this article as: Peng et al., RNF122: A novel ubiquitin ligase associated with calcium-modulating cyclophilin ligand BMC Cell Biology 2010, 11:41 21. Yokouchi M, Kondo T, Houghton A, Bartkiewicz M, Horne WC, Zhang H, Yoshimura A, Baron R: Ligand-induced ubiquitination of the epidermal growth factor receptor involves the interaction of the c-Cbl RING finger and UbcH7. J Biol Chem 1999, 274(44):31707-31712. 21. Yokouchi M, Kondo T, Houghton A, Bartkiewicz M, Horne WC, Zhang H, Yoshimura A, Baron R: Ligand-induced ubiquitination of the epidermal growth factor receptor involves the interaction of the c-Cbl RING finger and UbcH7. J Biol Chem 1999, 274(44):31707-31712. Acknowledgements This study was supported by grants from the National High-tech R&D Program (863 Program) of China (no. 2006AA02A305), the National Natural Science Foundation of China (no. 30600547), and the key national S&T Program--- "Major New Drug Development" (no. 2009ZX09503-004) 1. Haglund K, Dikic I: Ubiquitylation and cell signaling. EMBO J 2005, 24(19):3353-3359. von Bulow GU, Bram RJ: NF-AT activation induced by a CAML- interacting member of the tumor necrosis factor receptor superfamily. Science 1997, 278(5335):138-141. 15. Tran DD, Russell HR, Sutor SL, van Deursen J, Bram RJ: CAML is required for efficient EGF receptor recycling. Dev Cell 2003, 5(2):245-256. 16. Tran DD, Edgar CE, Heckman KL, Sutor SL, Huntoon CJ, van Deursen J, McKean DL, Bram RJ: CAML is a p56Lck-interacting protein that is required for thymocyte development. Immunity 2005, 23(2):139-152. 17. Feng P, Park J, Lee BS, Lee SH, Bram RJ, Jung JU: Kaposi's sarcoma- associated herpesvirus mitochondrial K7 protein targets a cellular calcium-modulating cyclophilin ligand to modulate intracellular calcium concentration and inhibit apoptosis. J Virol 2002, 76(22):11491-11504. 18. Guo S, Lopez-Ilasaca M, Dzau VJ: Identification of calcium-modulating cyclophilin ligand (CAML) as transducer of angiotensin II-mediated nuclear factor of activated T cells (NFAT) activation. J Biol Chem 2005, 280(13):12536-12541. 19. Yuan X, Yao J, Norris D, Tran DD, Bram RJ, Chen G, Luscher B: Calcium- modulating cyclophilin ligand regulates membrane trafficking of postsynaptic GABA(A) receptors. Mol Cell Neurosci 2008, 38(2):277-289. 20. Bedford FK, Kittler JT, Muller E, Thomas P, Uren JM, Merlo D, Wisden W, Triller A, Smart TG, Moss SJ: GABAA receptor cell surface number and subunit stability are regulated by the ubiquitin-like protein Plic-1. Nat Neurosci 2001, 4(9):908-916.
https://openalex.org/W4360616637	https://www.frontiersin.org/articles/10.3389/fonc.2023.1142022/pdf	English	null	Diagnostic performance of transperineal prostate targeted biopsy alone according to the PI-RADS score based on bi-parametric magnetic resonance imaging	Frontiers in oncology	2,023	cc-by	6,212	Diagnostic performance of transperineal prostate targeted biopsy alone according to the PI-RADS score based on bi-parametric magnetic resonance imaging OPEN ACCESS EDITED BY Ran Xu, Second Xiangya Hospital, Central South University, China REVIEWED BY Biagio Barone, University of Naples Federico II, Italy Jianbo Li, Case Western Reserve University, United States CORRESPONDENCE Sung Gu Kang kkangsung7@korea.ac.kr SPECIALTY SECTION This article was submitted to Genitourinary Oncology, a section of the journal Frontiers in Oncology RECEIVED 11 January 2023 ACCEPTED 10 March 2023 PUBLISHED 23 March 2023 CITATION Noh TI, Shim JS, Kang SH, Cheon J and Kang SG (2023) Diagnostic performance of transperineal prostate targeted biopsy alone according to the PI-RADS score based on bi-parametric magnetic resonance imaging. Front. Oncol. 13:1142022. doi: 10.3389/fonc.2023.1142022 EDITED BY Ran Xu, Second Xiangya Hospital, Central South University, China REVIEWED BY Biagio Barone, University of Naples Federico II, Italy Jianbo Li, Case Western Reserve University, United States Tae Il Noh, Ji Sung Shim, Seok Ho Kang, Jun Cheon and Sung Gu Kang Tae Il Noh, Ji Sung Shim, Seok Ho Kang, Jun Cheon and Sung Gu Kang* Department of Urology, Anam Hospital, Korea University College of Medicine, Seoul, Republic of Korea Purpose: To compare the diagnostic performance of transperineal targeted biopsy (TB) or systematic biopsy (SB) alone based on combined TB+SB and radical prostatectomy (RP) specimen for detecting prostate cancer (PCa) according to the prostate imaging reporting and data system (PI-RADS) score. Noh TI, Shim JS, Kang SH, Cheon J and Kang SG (2023) Diagnostic performance of transperineal prostate targeted biopsy alone according to the PI-RADS score based on bi-parametric magnetic resonance imaging. Materials and methods: This study included 1077 men who underwent transperineal bi-parametric (bp) magnetic resonance imaging (MRI)–ultrasound (US) fusion TB+SB (bpMRI-US FTSB) between April 2019 and March 2022. To compare the performance of each modality (TB, SB, and combined TB+SB) with the RP specimen (as the standard) for detecting PCa and clinically signiﬁcant PCa (csPCa), receiver operating characteristic (ROC) curves were plotted. COPYRIGHT © 2023 Noh, Shim, Kang, Cheon and Kang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Results: PCa was detected in 581 of 1077 men (53.9%) using bpMRI-US FTSB. TYPE Original Research PUBLISHED 23 March 2023 DOI 10.3389/fonc.2023.1142022 Diagnostic performance of transperineal prostate targeted biopsy alone according to the PI-RADS score based on bi-parametric magnetic resonance imaging Study design Prostate cancer (PCa) diagnosis relies on prostate-speciﬁc antigen (PSA) and prostate biopsy, and transrectal ultrasonography-guided systematic biopsy (TRUSB) has been considered the standard diagnostic pathway in men with a clinical suspicion of PCa (1). We analyzed the medical records of 1077 men, between April 2019 and March 2022, who were clinically suspected for PCa with an elevated prostate-speciﬁc antigen (PSA) level (≥4.0 ng/mL), and/or abnormal ﬁndings on digital rectal examination (DRE). All enrolled patients underwent bi-parametric MRI (bpMRI) prior to the prostate biopsy, and regions of interest (ROIs) on MRI were established according to the PI-RADS version 2.0. Subsequent transperineal bpMRI-US FTSB and RP were performed (Figure 1). However, TRUSB has led to missed diagnosis in >30% of patients with PCa and has poor discriminative power in diagnosing cancerous tissue (2, 3). In this regard, to improve the discriminative power and diagnostic accuracy of prostate biopsy, visualization of PCa through magnetic resonance imaging (MRI) has been attempted. Accordingly, the prostate imaging reporting and data system (PI-RADS) was developed to maximize the standardized utilization of MRI for detecting PCa, which led to increased usage of MRI as a guide for targeted biopsy (TB) (4). Studies have suggested that MRI-TB can provide additional value in diagnosis of PCa for clinically signiﬁcant PCa (csPCa) categorized as International Society for Urological Pathology (ISUP) grade ≥2 (5). Additionally, MRI-TB based on PI-RADS signiﬁcantly outperforms systematic biopsy (SB) for detection of csPCa with the probability of sparing the potential redundancy of SB (6–8). FIGURE 1 Flowchart of study design. ADC, apparent diffusion coefﬁcient; T2WI, T2-weighted images; DWI, diffusion weighted images; TB, targeted biopsy; SB, systematic biopsy; US, ultrasound. However, MRI was missing PCa in 20% of index tumor and 79% of non-index tumor (9). Therefore, the performance of MRI- TB alone may be not good enough to omit systematic biopsy (SB) in every man with a clinical suspicion for PCa (10). TB is the standard pathway in most cancers, nevertheless the current guidelines for detecting PCa have recommended SB and additional TB with a suspicious lesion in MRI (11). However, SB may be associated with over-diagnose the clinically insigniﬁcant PCa and result in overtreatment and impose the risk of adverse events, complications, and comes with consequence of medical burden (12, 13). Abbreviations: ADC, apparent diffusion coefﬁcient; AUC, area under the ROC curve; bpMRI, bi-parametric MRI; bpMRI–US FTSB, bi-parametric MRI-US fusion (transperineal) targeted and systematic biopsy; csPCa, clinically signiﬁcant prostate cancer; DCE, dynamic contrast enhanced; DWI, diffusion weighted images; FTSB, (transperineal) fusion targeted and systematic biopsy; GA, general anesthesia; MRI, magnetic resonance imaging; mpMRI, multi-parametric MRI; US, ultrasound; PCa, prostate cancer; PI-RADS, prostate Imaging Reporting and Data Systems; ROC, receiver-operating characteristic; ROI, regions of interest; SB, (template) systematic biopsy; TRUS, transrectal ultrasound; TRUSB, transrectal ultrasound guided systematic biopsy; T2WI, T2-weighted images; TB, targeted biopsy; US, ultrasound. Diagnostic performance of transperineal prostate targeted biopsy alone according to the PI-RADS score based on bi-parametric magnetic resonance imaging CsPCa was detected in 383 of 1077 men (35.6%), 17 of 285 (6.0%) with PI-RADS 0 to 2, 35 of 277 (12.6%) with PI-RADS 3, 134 of 274 (48.9%) with PI-RADS 4, and 197 of 241 (81.7%) with PI-RADS 5, respectively. The additional diagnostic value of TB vs. SB compared to combined TB+SB for diagnosing csPCa were 4.3% vs. 3.2% (p=0.844), 20.4% vs 5.1% (p<0.001), and 20.3% vs. 0.7% (p<0.001) with PI-RADS 3, 4, and 5, respectively. TB alone showed no signiﬁcant difference in diagnostic performance for csPCa with combined TB+SB based on RP specimens in patients with PI-RADS 5 (p=0.732). Conclusion: A need for addition of SB to TB in patients with PI-RADS 3 and 4 lesions, however, TB alone may be performed without affecting the management of patients with PI-RADS 5. magnetic resonance imaging, transperineal biopsy, prostate cancer, PI-RADS, target biopsy 01 01 Frontiers in Oncology frontiersin.org Noh et al. 10.3389/fonc.2023.1142022 MRI acquisition protocol of Urological Pathology (ISUP) grade groups (GG). Clinically insigniﬁcant PCa was deﬁned as an ISUP GG1. Clinically signiﬁcant PCa was deﬁned as > ISUP GG2 (Figure 2B). The bpMRI, contrast-free protocol, was performed using a 3.0-T scanner (Magnetom Skyra and Prisma, Siemens Healthineers, Erlangen, Germany or Achieva, Philips Healthcare, Best, Netherlands) with a multichannel phased-array external surface coil. T2-weighted images (T2WI) and diffusion-weighted images (DWI) were obtained, whereas dynamic contrast-enhanced (DCE) images were omitted. ROIs on the bpMRI were marked by three dedicated uroradiologists based on PI-RADS version 2.0 (Figure 2A). Study end points The endpoint was to compare the impact of TB or SB alone according to PI-RADS scores, referring to the standard of combined TB+SB and RP specimens. Prostate biopsy protocol We have previously reported a protocol for transperineal bpMRI-US FTSB (16). In brief, the elastic image registration type of the MRI-US fusion technique using a mechanical position encoder and robotic articulated arm system (Biojet, USA) was used and TB and SB were performed by urologists during the same session. Further, we considered suspicious lesions as ROI (PI- RADS ≥3) for TB, and 3-4 cores of TB and sequential 22-cores of SB were performed using a prostate mapping template (modiﬁed Barzell-template). The ROI for the TB was not intentionally avoided. Each core was labelled separately and subjected to histopathology. The number of biopsy cores was decided depending on the prostate size. The prostate biopsy results were reported by three uropathologists based on the International Society RP and histopathologic examination protocol Localized PCa with PI-RADS 3-5, sequentially underwent robot-assisted RP (RARP) using da Vinci Si, Xi, or SP system (Intuitive Surgical, Sunnyvale, CA, USA) by two surgeons. For histopathological examination, whole-mount histopathology slides were used, and each prostate was sectioned in the axial plane from the basal to the apex at approximately 4-5 mm intervals (Figure 2C). Study design Notably, in PI-RADS 5, MRI-TB have shown good performance with high predictive rates for csPCa that suggests TB alone might also be valuable in diagnosing csPCa (14–16). The purpose of this study was to compare the diagnostic performance of TB or SB alone according to the PI-RADS scores with combined TB+SB based on the standard transperineal bi- parametric magnetic resolution imaging-ultrasound fusion TB+SB (bpMRI-US FTSB) and radical prostatectomy (RP) specimen. 02 Frontiers in Oncology frontiersin.org Noh et al. 10.3389/fonc.2023.1142022 Diagnostic performance of bpMRI-US FTSB In patients with PI-RADS 3 to 5, csPCa (ISUP ≥GG2) was detected in 43.1%, 31.4%, and 46.2% cases via TB, SB, and combined TB+SB, respectively. Accordingly, the diagnosis rate of TB, SB, and combined TB+SB for diagnosing csPCa were 9.4%, 8.3%, and 12.6% in patients with PI-RADS 3, 43.8%, 28.5%, and 48.9% in patients with PI-RADS 4, and 81.0%, 61.4%, and 81.7% in patients with PI-RADS 5, respectively (Figure 3B). The additional diagnostic value for csPCa detection of TB vs. SB alone compared to combined TB+SB was 14.8% vs. 3.1%(p<0.001) in patients with PI- RADS 3-5; PI-RADS 3: 4.3% vs. 3.2% (p=0.844), PI-RADS 4: 20.4% vs. 5.1% (p<0.001), and PI-RADS 5: 20.3% vs. 0.7% (p<0.001), respectively (Table 3). Further, TB alone showed no signiﬁcant difference in diagnostic performance for csPCa to combined TB+SB in patients with PI-RADS 5; PI-RADS 3: area under the curve PCa (GG1) was detected in 581 of 1077 men (53.9%) by bpMRI- US FTSB. Accordingly, it was detected in 58 of 285 cases (35.6%) with PI-RADS 0-2, in 91 of 277 cases (32.9%) with PI-RADS 3, in 209 of 274 cases (76.3%) with PI-RADS 4, and in 220 of 241 cases (91.3%) with PI-RADS 5 (Figure 3A). Further, csPCa (≥GG2) was detected in 383 of 1077 men (35.6%). Accordingly, it was detected in 17 of 285 men (6.0%) with PI-RADS 0-2, in 35 of 277 men (12.6%) with PI-RADS 3, in 134 of 274 men (48.9%) with PI-RADS 4, and in 197 of 241 men (81.7%) with PI-RADS 5 (Figure 3B). The distribution of ISUP grade groups is shown in Table 2. Patients with csPCa (GG2≥2) had higher median PSA, PSAD, and lower prostate volume than those with GG1 pathology; PSA(IQR) [66.0 vs. 72.0, TABLE 1 Demographics of men according to PI-RADS distribution. Statistical analysis To quantify and compare the performance of each modality (TB, SB, and combined TB+SB) in detecting PCa and csPCa, FIGURE 2 Protocols of study. (A) Bi-parametric magnetic resonance imaging (bpMRI) (B) Transperineal bpMRI-Ultrasound fusion targeted and systematic biopsy (C) Whole-mount radical prostatectomy specimen ISUP, International Society for Urological Pathology; GG, grade group; MRI, Magnetic resonance imaging; TB, targeted biopsy; SB, systematic biopsy. FIGURE 2 Protocols of study. (A) Bi-parametric magnetic resonance imaging (bpMRI) (B) Transperineal bpMRI-Ultrasound fusion targeted and systematic biopsy (C) Whole-mount radical prostatectomy specimen ISUP, International Society for Urological Pathology; GG, grade group; MRI, Magnetic resonance imaging; TB, targeted biopsy; SB, systematic biopsy. 03 Frontiers in Oncology frontiersin.org Noh et al. 10.3389/fonc.2023.1142022 p= 0.038], PSAD (0.14 vs. 0.35, p=0.011), and lower prostate volume (41.2 vs. 30.3, p=0.047) than those with GG1 pathology (Supplementary Table 1). receiver operating characteristic (ROC) curve analyses were performed considering combined TB+SB and RP specimens as standards. Accordingly, the results were summarized using Delong’s test as the areas under the ROC curves (AUCs) and 95% CI. All statistical analyses were performed using IBM SPSS version 26.0 (IBM Corp., Armonk, NY, USA). The level of statistical signiﬁcance was considered P<0.05. Patient demographics Combined TB+SB showed superior diagnostic performance for TB or SB alone in patients with PI-RADS 3 and 4 (p <0.001). However, TB alone showed no signiﬁcant difference in diagnostic performance for csPCa with combined TB+SB in patients with PI- RADS 5; PI-RADS 3: area under the curve (AUC) [95% conﬁdence interval (CI)], 0.882 [0.838–0.918], p<0.001; PI-RADS 4: AUC, 0.964 [0.935–0.983], p<0.001; PI-RADS 5: AUC, 0.986 [0.961– 0.997], p=0.078 (Table 3). In total, 1077 men were included in the analysis. The median (interquartile range (IQR)) age was 69.0 (62.0-75.0) years. The median (IQR) PSA and PSA density (PSAD) were 6.66 (4.57-11.57) ng/mL and 0.18 (0.11-0.35) ng/mL2. The demographics of the study population are reported in Table 1. Ethics statement In patients with PI-RADS 3 to 5, TB, SB, and Combined TB+SB were able to detect PCa in 61.0%, 54.0%, and 66.0% of cases, respectively. Accordingly, the diagnosis rate of TB, SB, and combined TB+SB for diagnosing PCa were 24.9%, 26.4%, and 32.9% in patients with PI-RADS 3, 70.8%, 63.5%, and 76.3% in patients with PI-RADS 4, and 91.3%, 75.1%, and 92.5% in patients with PI-RADS 5, respectively (Figure 3A). The additional diagnostic value for PCa detection of TB vs. SB compared to combined TB+SB were 12.0% vs. 5.0% (p<0.001) in patients with PI-RADS 3-5; PI-RADS 3: 6.5% vs. 8.0% (p=0.535), PI- RADS 4: 12.8% vs. 5.5% (p<0.001), and PI-RADS 5: 17.4% vs. 1.2% (p<0.0001), respectively (Table 3). This study was conducted in accordance with the Declaration of Helsinki and current ethical guidelines. The study was reviewed and approved by the Ethics Committee and Institutional Review Board of Korea University Anam Hospital (IRB No. 2018AN0339). PI-RADS, prostate imaging-reporting and data systems; IQR, interquartile range; PSA, prostate-speciﬁc antigen; DRE, digital rectal exam. frontiersin.org PI-RADS, prostate imaging-reporting and data systems; IQR, interquartile range; PSA, prostate-speciﬁc antigen; DRE, digital rectal exam. Diagnostic performance of bpMRI-US FTSB All PI-RADS 0-2 PI-RADS 3 PI-RADS 4 PI-RADS 5 Distribution of PI-RADS, n (%) 1077 285 (26.5) 277 (25.7) 274 (25.4) 241 (22.4) Median Age (IQR) 69.0 (62.0-75.0) 61.0 (56.0-68.0) 66.0 (61.0-72.0) 72.0 (64.8-77.0) 72.0 (68.0-78.0) Median PSA, ng/mL (IQR) 6.66 (4.57-11.57) 5.27 (4.14-6.73) 5.65 (4.28-8.64) 6.88 (4.89-10.87) 13.3 (7.03-34.3) Median prostate volume, cm3 (IQR) 36.3 (26.4-50.1) 38.9 (27.7-54.0) 39.4 (30.2-51.1) 34.9 (25.3-46.4) 32.1 (24.2-44.4) Median PSA density (IQR) 0.18 (0.11-0.35) 0.13 (0.08-0.20) 0.15 (0.10-0.22) 0.19 (0.13-0.35) 0.45 (0.22-1.03) Median free/total PSA ratio (IQR) 0.15 (0.10-0.21) 0.17 (0.12-0.24) 0.17 (0.12-0.23) 0.13 (0.10-0.19) 0.12 (0.08-0.17) DRE nodule, n (%) 122 (11.3) 15 (5.3) 27 (9.7) 42 (15.3) 38 (15.8) PI RADS prostate imaging reporting and data systems; IQR interquartile range; PSA prostate speciﬁc antigen; DRE digital rectal exam 04 frontiersin.org Noh et al. 10.3389/fonc.2023.1142022 FIGURE 3 Diagnostic performance of TB, SB, TB+SB in patients with PI-RADS 3 to 5 (A) Detection rate for prostate cancer (B) Detection rate for clinically signiﬁcant prostate cancer PI-RADS, prostate imaging-reporting and data systems; TB, targeted biopsy; SB, systematic biopsy. FIGURE 3 Diagnostic performance of TB, SB, TB+SB in patients with PI-RADS 3 to 5 (A) Detection rate for prostate cancer (B) Detection rate for clinically signiﬁcant prostate cancer PI-RADS, prostate imaging-reporting and data systems; TB, targeted biopsy; SB, systematic biopsy. (AUC) [95% conﬁdence interval (CI)], 0.893 [0.851–0.927], p=0.0088; PI-RADS 4: AUC, 0.961 [0.931–0.981], p=0.002; PI- RADS 5: AUC, 0.989 [0.971–0.998], p=0.093 (Table 3). (3.7%) and 3 of 108 (2.8%) with PI-RADS 5, respectively. Further, downgrading of GG ≥2 to GG1 occurred in only one in 289 (0.3%) (Table 4). The combined TB+SB showed superior diagnostic performance compared to TB alone for diagnosing csPCa when compared to the standard of RP specimen; TB alone vs TB+SB, AUC (95% CI); PI- RADS 3-5: 0.824 (0.777-0.864) vs. 0.860 (0.809-0.911), p=0.034; PI- RADS 3: 0.663 (0.524-0.802) vs. 0.722 (0.593-0.852), p=0.016; PI- RADS 4: 0.817 (0.730-0.904) vs. 0.844 (0.766-0.921), p=0.049. TB alone showed no signiﬁcant difference in diagnostic performance for csPCa to combined TB+SB in patients with PI-RADS 5; TB alone vs. combined TB+SB, AUC (95% CI), 0.951(0.909-0.994) vs. 0.961(0.924-0.998), p=0.732 (Table 4). Comparison of diagnostic performances referring to RP specimen The RARP was performed in 289 of 483 diagnosed with PCa with PI-RADS 3-5; 59 of 91 (64.8%) with PI-RADS 3, 122 of 209 (58.4%) with PI-RADS 4, and 108 of 220 (49.1%) with PI-RADS 5, respectively (Table 4). Accordingly, TB alone and combined TB+SB showed 45.7% and 33.2% of any upgrading in RP specimens with PI-RADS 3-5; 76.2% and 37.3% with PI-RADS 3, 50.4% and 39.3% with PI-RADS 4, 28.7% and 24.1% with PI-RADS 5, respectively; and upgrading of GG1 to GG ≥2 occurred in 59 of 265 (22.3%) and 59 of 289 (20.4%) cases with PI- RADS 3-5; 17 of 42 (40.5%) and 21of 59 (35.6%) with PI-RADS 3, 38 of 115 (33.0%) and 35 of 122 (28.7%) with PI-RADS 4, and 4 of 108 * ISUP grade groups (GG):1 = Gleason 6 (or less), 2 = Gleason 7(3 + 4), 3 = Gleason 7(4 + 3), 4 = Gleason 8(4 + 4 or 3 + 5 or 5 + 3), and 5 = Gleason 9 or 10. csPCa: ≥ISUP GG2. MRI-US, magnetic resonance imaging-ultrasonography; ISUP, International Society for Urological Pathology; TB, targeted biopsy; SB, systematic biopsy; PCa, prostate cancer; PI-RADS, prostate imaging reporting and data system. Frontiers in Oncology Discussion In recent years with signiﬁcant improvements in the accuracy of MRI after implementation of the PI-RADS, the use of prebiopsy TABLE 2 Diagnostic performance of transperineal MRI-US fusion TB and SB. All PI-RADS 0-2 PI-RADS 3 PI-RADS 4 PI-RADS 5 1077 285 277 274 241 SB TB SB TB+SB TB SB TB+SB TB SB TB+SB PCa, n (%) 581 (53.9) 58 (20.4) 69 (24.9) 73 (26.4) 91 (32.9) 194 (70.8) 174 (63.5) 209 (76.3) 220 (91.3) 181 (75.1) 223 (92.5) csPCa, n (%) 383 (35.6) 17 (6.0) 26 (9.4) 23 (8.3) 35 (12.6) 120 (43.8) 78 (28.5) 134 (48.9) 195 (81.0) 148 (61.4) 197 (81.7) ISUP, n (%) 1 198 (34.1) 41 (14.4) 43 (15.5) 50 (18.1) 56 (20.2) 74 (27.0) 96 (35.0) 75 (27.4) 25 (10.4) 33 (13.7) 26 (10.8) 2 119 (20.5) 10 (3.5) 15 (5.4) 17 (6.1) 22 (7.9) 43 (15.7) 25 (9.1) 44 (16.1) 46 (19.1) 42 (17.4) 43 (17.8) 3 37 (6.4) 5 (1.8) 3 (1.1) 1 (0.4) 2 (0.7) 11 (4.0) 7 (2.6) 15 (5.5) 28 (11.6) 12 (5.0) 15 (6.2) 4 180 (30.9) 1 (0.4) 8 (2.9) 5 (1.8) 11 (4.0) 60 (21.9) 41 (15.0) 66 (24.1) 91 (37.8) 66 (27.4) 102 (42.3) 5 47 (8.1) 1 (0.4) 0 (0.0) 0 (0.0) 0 (0.0) 6 (2.2) 5 (1.8) 9 (3.3) 30 (12.4) 28 (11.6) 37 (15.4) ISUP grade groups (GG):1 = Gleason 6 (or less), 2 = Gleason 7(3 + 4), 3 = Gleason 7(4 + 3), 4 = Gleason 8(4 + 4 or 3 + 5 or 5 + 3), and 5 = Gleason 9 or 10. csPCa: ≥ISUP GG2. TABLE 2 Diagnostic performance of transperineal MRI-US fusion TB and SB. 05 Frontiers in Oncology frontiersin.org 10.3389/fonc.2023.1142022 Noh et al. TABLE 3 Diagnostic performance of TB or SB alone according to PI-RADS compared to combined TB and SB. Discussion All (PI-RADS 3-5) PI-RADS 3 PI-RADS 4 PI-RADS 5 792 277 274 241 TB SB TB +SB TB SB TB +SB TB SB TB +SB TB SB TB +SB PCa, n (%) 483 (61.0) 428 (54.0) 523 (66.0) 69 (24.9) 73 (26.4) 91 (32.9) 194 (70.8) 174 (63.5) 209 (76.3) 220 (91.3) 181 (75.1) 223 (92.5) Additional value of TB 12.0% 6.5 % 12.8 % 17.4 % Additional value of SB 5.0% 8.0 % 5.5 % 1.2 % AUC (CI 95%) Reference to TB+SB 0.932 (0.915- 0.947) 0.914 (0.895- 0.931) 0.882 (0.838- 0.918) 0.904 (0.863- 0.937) 0.964 (0.935- 0.983) 0.916 (0.877- 0.946) 0.986 (0.961- 0.997) 0.906 (0.862- 0.940) p value (vs. TB+SB) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.078 <0.001 csPCa (≥GG2) , n (%) 341 (43.1) 249 (31.4) 366 (46.2) 26 (9.4) 23 (8.3) 35 (12.6) 120 (43.8) 78 (28.5) 134 (48.9) 195 (81.0) 148 (61.4) 197 (81.7) Additional value of TB 14.8% 4.3% 20.4% 20.3% Additional value of SB 3.1% 3.2% 5.1% 0.7% AUC (CI 95%) Reference to TB+SB 0.957 (0.942- 0.968) 0.881 (0.895- 0.901) 0.893 (0.851- 0.927) 0.883 (0.839- 0.918) 0.961 (0.931- 0.981) 0.841 (0.792- 0.882) 0.989 (0.971- 0.998) 0.867 (0.817- 0.907) p value (vs. TB+SB) <0.001 <0.001 0.009 0.004 0.0021 <0.001 0.093 <0.001 ISUP grade groups (GG):1 = Gleason 6 (or less), 2 = Gleason 7(3+4), 3 = Gleason 7(4+3), 4 = Gleason 8(4+4 or 3+5 or 5+3), and 5 = Gleason 9 or 10. csPCa: ≥ISUP GG2 AUC, area under the curve; CI, conﬁdence interval; ISUP, International Society for Urological Pathology; TB, targeted biopsy; SB, systematic biopsy; PCa, prostate cancer; csPCa, clinically signiﬁcant prostate cancer; PI-RADS, prostate imaging reporting and data system. * ISUP grade groups (GG):1 = Gleason 6 (or less), 2 = Gleason 7(3+4), 3 = Gleason 7(4+3), 4 = Gleason 8(4+4 or 3+5 or 5+3), and 5 = Gleason 9 or 10. csPCa: ≥ISUP GG2 AUC, area under the curve; CI, conﬁdence interval; ISUP, International Society for Urological Pathology; TB, targeted biopsy; SB, systematic biopsy; PCa, prostate cancer; csPCa, clinically signiﬁcant prostate cancer; PI-RADS, prostate imaging reporting and data system. calculators (RCs) have been suggested, thereby may be reducing the number of unnecessary biopsies (22). MRI for PCa diagnosis has increased (4, 6, 17). Discussion Furthermore, numerous studies have demonstrated that MRI-TB could offer improved diagnostic value for csPCa with pooled sensitivity and speciﬁcity of 0.80 (95%CI: 0.69-0.87) and 0.94 (95%CI: 0.90-0.97) (5). However, addition of TB to SB increases the number of csPCa (≥ISUP GG2) by 6.7-7.6%, while added value of SB to TB is 4.3- 5.2% for csPCa (5, 14, 18). Further, MRI was missing PCa in 20% of index tumor and 79% of non-index tumor (9). Therefore, due to the additional diagnostic value of SB and the risk of missing csPCa with TB alone, combined TB + SB has been suggested for dignosing PCa (5, 10, 11). Notably, MRI-TB showed good performance and was highly predictive for diagnosing csPCa in cases with PI-RADS 5 (77-85%) (7, 14, 16). In a study comparing the concordances between PI- RADS and histologic reports of the RP specimen, the PI-RADS≥3 was further associated with csPCa in 92.4% of cases, with 100% association in cases with a PI-RADS 5 score (23). High performance of MRI-TB and low additional diagnostic value (2-4%) of SB for detection of csPCa in patients with PI-RADS 5 that suggests the probability of sparing the potential redundancy of SB in PI-RADS 5 (12, 24, 25). However, it should be noted that obtaining more prostate cores accompanies with a greater risk of complications, such as prostatitis, sepsis events, visits to the emergency room, rectal bleeding, hematuria, and pain (7, 19, 20). MRI-TB alone with fewer core biopsies per patient might lead to fewer complications. The net beneﬁt of adding SB to TB for prostate biopsy optimization according to PI-RADS score should be weighed against accuracy for csPCa detection and additional burden such as overdiagnosis of indolent PCa, resulting in overtreatment and complications from increased numbers of biopsies. For predicting csPCa, several predictors and their combination such as clinical parameters including PSAD and PI-RADS score have been suggested (21). In addition, for risk assessment to determine the need for biopsy, risk For MRI-TB, mpMRI have shown a high sensitivity and negative predictive value (NPV) of 93.0% and 89.0% for csPCa (6). However, it is time-consuming (~ 40 min) to acquire T2- weighted imaging (T2WI) and diffusion-weighted imaging (DWI), and dynamic contrast-enhanced (DCE) imaging requires intravenous administration of contrast media. Several studies have demonstrated comparable diagnostic performance of bpMRI (contrast-free protocol) to mpMRI (26). Frontiers in Oncology TB 11 (94 58 (50 30 (26 44 (38 38 (33 71 (61 0 (0. 0.8 (0.730- 0.04 0. csPCa: cancer; cs 11 (94 58 (50 30 (26 44 (38 38 (33 71 (61 0 (0.0 0.8 (0.730- 0.04 csPCa: ncer; cs 30 (26. 44 (38. 38 (33. 71 (61. 0 (0.0 0.8 .730- 0.04 PCa: er; cs . csP ance (0.7 0. csP cance my, n T 1 (94 5 (50 3 (26 4 (38 3 (33 7 (61 0 (0 0.8 (0.730 0.0 csPCa ncer; c 0.04 PCa: er; cs 5 (50 3 (26 4 (38 3 (33 7 (61 0 (0. 0.8 (0.730- 0.0 0. csPCa: cancer; cs (0.7 0. csP ancer omy, ( ( ( ( ( ( 0 (0.73 0 csPC ancer; Data availability statement In the current study, we compared the impact of TB, SB, and combined TB+SB according to the PI-RADS score. Accordingly, the SB had only additional diagnostic values of 1.2% and 0.7% for detection of PCa and csPCa in patients with PI-RADS 5. Further, TB alone showed no signiﬁcant difference of diagnostic performance with combined TB+SB for csPCa. Similarly, in a study conducted on 112 patients with PI-RADS 5 on MRI and subsequentially 78 of RP, TB alone could diagnose PCa with very high probability (97%) in patients with PSAD >0.15ng/ml2 (12). Accordingly, if SB was omitted, none of the PCa cases and only 4% of csPCa cases would be missed. Thus, the authors suggest that SB might be omitted for cases with PI-RADS 5 and PSAD >0.15ng/ml2. The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author. Conﬂict of interest The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. The limitations of this study are its retrospective nature and accompanying bias. The other limitation is that this study was performed in a single tertiary center with transperineal prostate biopsy and bpMRI, and transrectal prostate biopsy with mpMRI, which is the common practice, was not considered. This may raise concerns toward extrapolating a general trend from our results. Nevertheless, this study can support that performing TB alone in patients with PI-RADS 5 lesions, might mitigate the medical burden by SB omission. Discussion In a systematic review and meta-analysis of the diagnostic accuracy of bpMRI and mpMRI for PCa detection, pooled sensitivity and speciﬁcity did not show signiﬁcant difference and the AUCs were similar; 0.87 and 0.90 for mpMRI and 06 Frontiers in Oncology frontiersin.org ording to radical prostatectomy specimen by PI-RADS scores. y, n (%) PI-RADS 4 PI-RADS 5 122 108 TB SB TB+SB TB SB TB+SB 115 (94.2) 104 (85.2) 122 (100.0) 108 (100.0) 84 (77.8) 108 (100.0) 58 (50.4) 79 (76.0) 48 (39.3) 31 (28.7) 54 (64.3) 26 (24.1) 30 (26.1) 24 (23.1) 37 (30.3) 37 (34.3) 28 (33.3) 47 (43.5) 44 (38.3) 62 (59.6) 45 (36.9) 9 (8.3) 23 (27.9) 8 (7.4) 38 (33.0) 54 (51.9) 35 (28.7) 4 (3.7) 19 (22.6) 3 (2.8) 71 (61.7) 42 (40.4) 77 (63.1) 99 (91.7) 66 (78.6) 100 (92.6) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0.817 730-0.904) 0.688 (0.559-0.816) 0.844 (0.766-0.921) 0.951 (0.909-0.994) 0.820 (0.711-0.929) 0.961 (0.924-0.998) 0.049 <0.001 0.732 <0.001 PCa: ≥ISUP GG2. er; csPCa, clinically signiﬁcant prostate cancer; PI-RADS, Prostate Imaging Reporting and Data System. Noh et al. 10.3389/fonc.2023.1142022 performed for diagnosing csPCa in patients with PI-RADS 5, without changing the subsequent clinical management. bpMRI (27). In this regard, bpMRI is more rapid (~15 min) due to exclusion of DCE, and safer from potential side effects of contrast media than mpMRI while retaining a sufﬁcient diagnostic value (16). Publisher’s note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Author contributions TIN: protocol/project development, data acquisition, data analysis and interpretation, drafting of the manuscript. JSS: Protocol/project development, supervision. SHK: Protocol/project development, Data acquisition, supervision. JC: Protocol/project development, supervision. SGK: Protocol/project development, Data acquisition, supervision. All authors contributed to the article and approved the submitted version. Ethics statement This study was approved by the Ethics Committee and the Institutional Review Board of KUMC (IRB No. 2018AN0339). Since the upgrading grade group of RP specimens from prostate biopsy has been reported, the omission of SB may lead to misclassiﬁcation of PCa; TB (30.9%) and TB+SB (14.4%) of the upgraded grade group (10). These inconsistencies between biopsy and specimen of prostate, upgrading and misclassiﬁcation of PCa, are the inherent limitations of prostate needle biopsy (28). Nevertheless, in this study, upgrading from GG1 to ≥GG2, which has a potential risk of changing subsequent clinical management, showed difference in only one patient; TB alone vs. combined TB +SB, 4 of 108 (3.7%) vs. 3 of 108 (2.8%). Similarly, in another study, MRI-TB alone in PI-RADS 5 cases had meager upgrade rate (3.4%) (29). Further, addition of SB to TB in PI-RADS 5 cases altered only 3.1% of the highest grade group of PCa patients, all of whom had already been categorized as GG≥2 based on TB, and SB did not change subsequent clinical management (24). Current study supports the need for SB in patients with PI-RADS 3 and 4 lesions. However, minimal additional diagnostic values of SB and comparable diagnostic performance of MRI-TB suggest that SB potentially can be omitted in patients with PI-RADS 5. References 1. Mottet N, Bellmunt J, Bolla M, Briers E, Cumberbatch MG, De Santis M, et al. EAU-ESTRO-SIOG guidelines on prostate cancer. part 1: Screening, diagnosis, and local treatment with curative intent. Eur Urol (2017) 71:618. doi: 10.1016/ j.eururo.2016.08.003 template systematic prostate biopsy based on bi-parametric magnetic resonance imaging. Cancer Res Treat (2020) 52:714. doi: 10.4143/crt.2019.716 17. Oberlin DT, Casalino DD, Miller FH, Meeks JJ. Dramatic increase in the utilization of multiparametric magnetic resonance imaging for detection and management of prostate cancer. Abdom Radiol (NY) (2017) 42:1255. doi: 10.1007/s00261-016-0975-5 17. Oberlin DT, Casalino DD, Miller FH, Meeks JJ. Dramatic increase in the utilization of multiparametric magnetic resonance imaging for detection and management of prostate cancer. Abdom Radiol (NY) (2017) 42:1255. doi: 10.1007/s00261-016-0975-5 2. Das CJ, Razik A, Sharma S, Verma S. Prostate biopsy: when and how to perform. Clin Radiol (2019) 74:853. doi: 10.1016/j.crad.2019.03.016 18. van der Leest M, Cornel E, Israël B, Hendriks R, Padhani AR, Hoogenboom M, et al. Head-to-head comparison of transrectal ultrasound-guided prostate biopsy versus multiparametric prostate resonance imaging with subsequent magnetic resonance-guided biopsy in biopsy-naïve men with elevated prostate-speciﬁc antigen: A Large prospective multicenter clinical study. Eur Urol (2019) 75:570. doi: 10.1016/j.eururo.2018.11.023 3. Epstein JI, Feng Z, Trock BJ, Pierorazio PM. Upgrading and downgrading of prostate cancer from biopsy to radical prostatectomy: incidence and predictive factors using the modiﬁed Gleason grading system and factoring in tertiary grades. Eur Urol (2012) 61:1019. doi: 10.1016/j.eururo.2012.01.050 4. Weinreb JC, Barentsz JO, Choyke PL, Cornud F, Haider MA, Macura KJ, et al. PI- RADS prostate imaging - reporting and data system: 2015, version 2. Eur Urol (2016) 69:16. doi: 10.1016/j.eururo.2015.08.052 19. Klotz L, Chin J, Black PC, Finelli A, Anidjar M, Bladou F, et al. Comparison of multiparametric magnetic resonance imaging-targeted biopsy with systematic transrectal ultrasonography biopsy for biopsy-naive men at risk for prostate cancer: A phase 3 randomized clinical trial. JAMA Oncol (2021) 7:534. doi: 10.1001/jamaoncol.2020.7589 5. Drost FH, Osses DF, Nieboer D, Steyerberg EW, Bangma CH, Roobol MJ, et al. Prostate MRI, with or without MRI-targeted biopsy, and systematic biopsy for detecting prostate cancer. Cochrane Database Syst Rev (2019) 4:Cd012663. doi: 10.1002/14651858.CD012663.pub2 20. Ghani KR, Dundas D, Patel U. Bleeding after transrectal ultrasonography- guided prostate biopsy: a study of 7-day morbidity after a six-, eight-and 12-core biopsy protocol. BJU Int (2004) 94:1014. doi: 10.1111/j.1464-410X.2004.05096.x 6. Conclusion The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fonc.2023.1142022/ full#supplementary-material The current study suggests a need for addition of SB to TB in patients with PI-RADS 3 and 4 lesions, and TB alone may be 08 Frontiers in Oncology frontiersin.org 10.3389/fonc.2023.1142022 Noh et al. References Ahmed HU, El-Shater Bosaily A, Brown LC, Gabe R, Kaplan R, Parmar MK, et al. Diagnostic accuracy of multi-parametric MRI and TRUS biopsy in prostate cancer (PROMIS): a paired validating conﬁrmatory study. Lancet (2017) 389:815. doi: 10.1016/S0140-6736(16)32401-1 21. Massanova M, Vere R, Robertson S, Crocetto F, Barone B, Dutto L, et al. Clinical and prostate multiparametric magnetic resonance imaging ﬁndings as predictors of general and clinically signiﬁcant prostate cancer risk: A retrospective single-center study. Curr Urol (2023). doi: 10.1097/cu9.0000000000000173 22. Alberts AR, Roobol MJ, Verbeek JFM, Schoots IG, Chiu PK, Osses DF, et al. Prediction of high-grade prostate cancer following multiparametric magnetic resonance imaging: Improving the Rotterdam European randomized study of screening for prostate cancer risk calculators. Eur Urol (2019) 75:310–8. doi: 10.1016/j.eururo.2018.07.031 7. Kasivisvanathan V, Rannikko AS, Borghi M, Panebianco V, Mynderse LA, Vaarala MH, et al. MRI-Targeted or standard biopsy for prostate-cancer diagnosis. N Engl J Med (2018) 378:1767. doi: 10.1056/NEJMoa1801993 8. Siddiqui MM, Rais-Bahrami S, Turkbey B, George AK, Rothwax J, Shakir N, et al. Comparison of MR/ultrasound fusion-guided biopsy with ultrasound-guided biopsy for the diagnosis of prostate cancer. JAMA (2015) 313:390. doi: 10.1001/ jama.2014.17942 23. Rapisarda S, Bada M, Crocetto F, Barone B, Arcaniolo D, Polara A, et al. The role of multiparametric resonance and biopsy in prostate cancer detection: comparison with deﬁnitive histological report after laparoscopic/robotic radical prostatectomy. Abdom Radiol (NY) (2020) 45:4178–84. doi: 10.1007/s00261-020-02798-8 9. Le JD, Tan N, Shkolyar E, Lu DY, Kwan L, Marks LS, et al. Multifocality and prostate cancer detection by multiparametric magnetic resonance imaging: correlation with whole-mount histopathology. Eur Urol (2015) 67:569. doi: 10.1016/ j.eururo.2014.08.079 24. Drobish JN, Bevill MD, Tracy CR, Sexton SM, Rajput M, Metz CM, et al. Do patients with a PI-RADS 5 lesion identiﬁed on magnetic resonance imaging require systematic biopsy in addition to targeted biopsy? Urol Oncol (2021) 39:235 e1. doi: 10.1016/j.urolonc.2020.12.015 10. Ahdoot M, Wilbur AR, Reese SE, Lebastchi AH, Mehralivand S, Gomella PT, et al. MRI-Targeted, systematic, and combined biopsy for prostate cancer diagnosis. N Engl J Med (2020) 382:917. doi: 10.1056/NEJMoa1910038 25. Kilic M, Vural M, Coskun B, Acar Ö, Saglican Y, Akpek , et al. Accuracy of sampling PI-RADS 4-5 index lesions alone by MRI-guided in-bore biopsy in biopsy- naive patients undergoing radical prostatectomy. Eur Urol Focus (2020) 6:249. doi: 10.1016/j.euf.2019.04.010 11. References Mottet N, van den Bergh RCN, Briers E, Van den Broeck T, Cumberbatch MG, De Santis M, et al. EAU-EANM-ESTRO-ESUR-SIOG guidelines on prostate cancer- 2020 update. part 1: Screening, diagnosis, and local treatment with curative intent. Eur Urol (2021) 79:243. doi: 10.1016/j.eururo.2020.09.042 26. Bass EJ, Pantovic A, Connor M, Gabe R, Padhani AR, Rockall A, et al. A systematic review and meta-analysis of the diagnostic accuracy of biparametric prostate MRI for prostate cancer in men at risk. Prostate Cancer Prostatic Dis (2021) 24:596. doi: 10.1038/s41391-020-00298-w 12. Tafuri A, Iwata A, Shakir A, Iwata T, Gupta C, Sali A, et al. Systematic biopsy of the prostate can be omitted in men with PI-RADS™5 and prostate speciﬁc antigen density greater than 15. J Urol (2021) 206:289. doi: 10.1097/JU.0000000000001766 12. Tafuri A, Iwata A, Shakir A, Iwata T, Gupta C, Sali A, et al. Systematic biopsy of the prostate can be omitted in men with PI-RADS™5 and prostate speciﬁc antigen density greater than 15. J Urol (2021) 206:289. doi: 10.1097/JU.0000000000001766 27. Alabousi M, Salameh JP, Gusenbauer K, Samoilov L, Jafri A, Yu H, et al. Biparametric vs multiparametric prostate magnetic resonance imaging for the detection of prostate cancer in treatment-naïve patients: a diagnostic test accuracy systematic review and meta-analysis. BJU Int (2019) 124:209. doi: 10.1111/bju.14759 13. Weiner AB, Manjunath A, Kirsh GM, Scott JA, Concepcion RD, Verniero J, et al. The cost of prostate biopsies and their complications: a summary of data on all medicare fee-for-service patients over 2 years. Urol Pract (2020) 7:145. doi: 10.1097/ UPJ.0000000000000072 28. Noh TI, Shim JS, Kang SG, Cheon J, Lee JG, Lee JH, et al. Concordance between biparametric MRI, transperineal targeted plus systematic MRI-ultrasound fusion prostate biopsy, and radical prostatectomy pathology. Sci Rep (2022) 12:6964. doi: 10.1038/s41598-022-10672-4 14. Rouvière O, Puech P, Renard-Penna R, Claudon M, Roy C, Mège-Lechevallier F, et al. Use of prostate systematic and targeted biopsy on the basis of multiparametric MRI in biopsy-naive patients (MRI-FIRST): a prospective, multicentre, paired diagnostic study. Lancet Oncol (2019) 20:100. doi: 10.1016/S1470-2045(18)30569-2 15. Filson CP, Natarajan S, Margolis DJ, Huang J, Lieu P, Dorey FJ, et al. Prostate cancer detection with magnetic resonance-ultrasound fusion biopsy: The role of systematic and targeted biopsies. Cancer (2016) 122:884. doi: 10.1002/cncr.29874 15. Filson CP, Natarajan S, Margolis DJ, Huang J, Lieu P, Dorey FJ, et al. Prostate cancer detection with magnetic resonance-ultrasound fusion biopsy: The role of systematic and targeted biopsies. References Cancer (2016) 122:884. doi: 10.1002/cncr.29874 29. Arabi A, Deebajah M, Yaguchi G, Pantelic M, Williamson S, Gupta N, et al. Systematic biopsy does not contribute to disease upgrading in patients undergoing targeted biopsy for PI-RADS 5 lesions identiﬁed on magnetic resonance imaging in the course of active surveillance for prostate cancer. Urology (2019) 134:168. doi: 10.1016/ j.urology.2019.08.035 16. Noh TI, Tae JH, Kim HK, Shim JS, Kang SG, Sung DJ, et al. Diagnostic accuracy and value of magnetic resonance imaging-ultrasound fusion transperineal targeted and 16. Noh TI, Tae JH, Kim HK, Shim JS, Kang SG, Sung DJ, et al. Diagnostic accuracy and value of magnetic resonance imaging-ultrasound fusion transperineal targeted and 09 Frontiers in Oncology frontiersin.org
https://openalex.org/W3211107075	http://cds.cern.ch/record/2764483/files/fermilab-conf-21-063-scd.pdf	English	null	Porting CMS Heterogeneous Pixel Reconstruction to Kokkos	null	2,021	cc-by	5,883	Taylor Childers1,, Matti J. Kortelainen2,∗, Martin Kwok2, Alexei Strelchenko2, and Yunsong Wang3 Taylor Childers1,, Matti J. Kortelainen2,∗, Martin Kwok2, Alexei Strelchenko2, and Yunsong Wang3 1Argonne National Laboratory, Lemont, IL, USA 1Argonne National Laboratory, Lemont, IL, USA 1Argonne National Laboratory, Lemont, IL, USA 2Fermi National Accelerator Laboratory, Batavia, IL, USA 3Lawrence Berkeley National Laboratory, Berkeley, CA, USA 2Fermi National Accelerator Laboratory, Batavia, IL, USA 2Fermi National Accelerator Laboratory, Batavia, IL, USA 3Lawrence Berkeley National Laboratory, Berkeley, CA, USA 3Lawrence Berkeley National Laboratory, Berkeley, CA, USA Abstract. Programming for a diverse set of compute accelerators in addition to the CPU is a challenge. Maintaining separate source code for each archi- tecture would require lots of eﬀort, and development of new algorithms would be daunting if it had to be repeated many times. Fortunately there are sev- eral portability technologies on the market such as Alpaka, Kokkos, and SYCL. These technologies aim to improve the developer productivity by making it pos- sible to use the same source code for many diﬀerent architectures. In this paper we use heterogeneous pixel reconstruction code from the CMS experiment at the CERNL LHC as a realistic use case of a GPU-targeting HEP reconstruction software, and report experience from prototyping a portable version of it us- ing Kokkos. The development was done in a standalone program that attempts to model many of the complexities of a HEP data processing framework such as CMSSW. We also compare the achieved event processing throughput to the original CUDA code and a CPU version of it. This manuscript has been authored by Fermi Research Alliance, LLC under Contract No. DE-AC02-07CH11359 with the U.S. Department of Energy, Office of Science, Office of High Energy Physics. FERMILAB-CONF-21-063-SCD FERMILAB-CONF-21-063-SCD ∗e-mail: matti@fnal.gov 1 Introduction 1 Graphics processing units (GPUs) are being used in scientiﬁc computing because of their cost 2 and power eﬃciency in solving data-parallel problems. Currently each GPU vendor provides 3 their own APIs and programming models, that also diﬀer from the programming of the CPU. 4 There are, however, similarities in these GPU programming models, and in many cases the 5 code for very core pieces of algorithms can be shared between the CPU and the GPUs, but 6 the surrounding code arranging the data and calling the algorithms has to diﬀer. In multi- 7 million line code bases that have many custom algorithms and have to be maintained for tens 8 of years, such duplication of code would require signiﬁcant development and maintenance 9 eﬀort, and be error prone to maintain. 10 Over several years, many technologies for fully portable code between CPUs and com- 1 pute accelerators have emerged to ease the development and maintenance burden of het- 2 erogeneous applications. These technologies include C++ libraries, such as Alpaka [1– 3 3], Kokkos [4], and RAJA [5, 6]; SYCL [7] that can be implemented as libraries such as 4 triSYCL [8] and hipSYCL [9] or as speciﬁc compilers such as ComputeCpp [10] by Code- 5 play and DPC++ [11] by Intel; compiler pragma based solutions such as OpenMP [12] and 6 ∗e-mail: matti@fnal.gov OpenACC [13]; or as standard C++ itself via parallel STL where the compiler is solely re- 17 sponsible for generating necessary code for the oﬄoading. 18 OpenACC [13]; or as standard C++ itself via parallel STL where the compiler is solely re- 17 sponsible for generating necessary code for the oﬄoading. 18 In this work we explore the applicability of Kokkos for portability across CPU and GPUs 19 using the Patatrack heterogeneous pixel reconstruction workﬂow [14] from the CMS exper- 20 iment [15] at the CERN LHC [16] as a use case for a set of realistic HEP reconstruction 21 algorithms that are able to eﬀectively utilize a GPU. The work was done in the context of 22 the DOE HEP Center for Computational Excellent (HEP-CCE). We look into not only the 23 porting of the algorithms, but also the implications of integrating such an approach into a 24 HEP data processing software. 25 We mimic the setup of the CMS data processing software, CMSSW [17]. 1 Introduction 1 CMSSW is 26 multi-threaded [18–20] using the Intel Threading Building Blocks (TBB) [21], and the cur- 27 rent plan for direct same-node compute accelerators is to build code for all supported accel- 28 erators in the same release build, express all possibilities in the conﬁguration, and decide at 29 runtime what code exactly to run based on hardware availability [22, 23]. We are looking for 30 a single-source solution that would provide portability at least across CPU and GPUs, would 31 be relatively easy to program with by HEP physicists, would provide adequate performance 32 on all relevant platforms, and would require the least amount of change in the CMSSW build- 33 ing and data processing model. It is unlikely that all these goals would be met by a single 34 technology, and therefore it is necessary to learn the details in all these aspects to ﬁnd the 35 best compromise. 36 This paper is organized as follows. The technical aspects of the Patatrack pixel recon- 37 struction are described in Section 2. A brief introduction of Kokkos is given in Section 3. 38 The experience of porting the original CUDA application into Kokkos is reported in Sec- 39 tion 4. In Kokkos’ nomenclature a place that runs code is called an execution space. We 40 have tested Serial, Threads, CUDA, and HIP execution spaces of Kokkos, and we focus on 41 several aspects in how Kokkos would ﬁt into a framework like CMSSW. We have measured 42 the event processing throughput of the Kokkos version’s CPU and CUDA execution spaces, 43 and compare those to direct CPU and CUDA implementations in Section 5. Conclusions are 44 given in Section 6. 45 2 Patatrack Heterogeneous Pixel Reconstruction 46 The Patatrack pixel reconstruction pioneered oﬄoading algorithms to NVIDIA GPUs with di- 47 rect CUDA programming within CMSSW. The oﬄoaded chain of reconstruction algorithms 48 takes the raw data of the CMS pixel detector as an input, along with the beamspot parameters 49 and necessary calibration data, and produces pixel tracks and vertices. CMSSW schedules 50 algorithms as units that are called modules. The pixel reconstruction algorithms are orga- 51 nized in ﬁve modules, depicted in Figure 1, that communicate the intermediate data in the 52 GPU memory through the CMSSW event data. The BeamSpot module only transfers the 53 beamspot data into the GPU. The Clusters module transfers the raw data to the GPU, un- 54 packs them, calibrates the individual pixels, and clusters the pixels on each detector module. 55 The RecHits module estimates the 3D position of each cluster forming hits. The Tracks mod- 56 ule forms n-tuplets from the hits, and ﬁts the hit n-tuplets to obtain track parameters. The 57 Vertices module forms vertices from the pixel tracks. There are further modules to optionally 58 transfer the tracks and vertices to the CPU, and to convert the Structure-of-Array (SoA) data 59 structures to the data formats used by downstream algorithms in CMSSW, but those are not 60 considered in this work and therefore not shown in Figure 1. 61 In order to explore code portability technologies, the CUDA code of the Patatrack pixel 62 reconstruction was extracted from CMSSW into a standalone program [24]. The separation 63 from CMSSW gives us freedom to modify the compilers, build rules, external libraries, and 64 BeamSpot Clusters RecHits Tracks Vertices Figure 1. Directed acyclic graph of the framework modules in the Patatrack pixel reconstruction. The arrows denote the data dependencies of the modules, e.g. RecHits module depends on data produced by BeamSpot and Clusters modules. The Clusters module (red rectangle) is the only one that transfer data from the device to the host and uses External Worker synchronization mechanism, while the other modules (blue oval) do not. BeamSpot Clusters RecHits Tracks Vertices Clusters Figure 1. Directed acyclic graph of the framework modules in the Patatrack pixel reconstruction. The arrows denote the data dependencies of the modules, e.g. RecHits module depends on data produced by BeamSpot and Clusters modules. 2 Patatrack Heterogeneous Pixel Reconstruction 46 The Clusters module (red rectangle) is the only one that transfer data from the device to the host and uses External Worker synchronization mechanism, while the other modules (blue oval) do not. code organization that would be more laborious to achieve in the full CMSSW software stack. The standalone program was crafted to mimic several aspects of CMSSW, including similar organization of code into shared libraries, plugin libraries that are loaded dynamically based on run-time information, and a simple framework that uses TBB for multi-threading. From the CMSSW framework concurrency features this simple framework includes only event loop based on TBB tasks, processing of multiple events concurrently, and processing of independent modules concurrently for the same event. There is only a single module type of each module having a separate instance for each concurrent event, and the External Worker concept [23] is included in order to use the CPU threads to do other work while the GPU is running the oﬄoaded work. The CMSSW tools to use CUDA runtime directly in the modules [23] are also included. The standalone setup includes a binary data ﬁle that contains raw pixel detector data from 76 1000 simulated top quark pair production events from CMS Open Data [25], with an average 77 of 50 superimposed pileup collisions with a center-of-mass energy of 13 TeV, using design 78 conditions corresponding to the 2018 CMS detector. All of the data, about 250 MB, are 79 read into the memory at the job startup to exclude I/O from the throughput measurement. 80 The necessary pixel detector conditions data are also stored in binary ﬁles, and read into the 81 memory at the start of the job. The data processing throughput is calculated by measuring 82 the time spent in the event processing, and dividing the number of processed events with 83 that time. For each event, the object holding the raw data for that event is copied once from 84 the aforementioned memory buﬀer to another object owned by the event data structure. The 85 event processing time includes the time taken by this copy operation. 86 3 Kokkos 87 Kokkos is a programming model and a C++ library for writing performance portable appli- 88 cations. At the time of writing the latest version of Kokkos is 3.3.1, and it supports several 89 execution spaces. An algorithm can be run serially on the host CPU via a host serial ex- 90 ecution space, or it can be parallelized with one of two host parallel execution spaces that 91 are OpenMP and (POSIX) Threads. An algorithm can also be oﬄoaded to compute accel- 92 erators with device parallel execution spaces. NVIDIA GPUs can be used with CUDA or 93 HPX execution spaces, and AMD GPUs can be used with HIP execution space. There are 94 also OpenMP-Target and SYCL 2020 execution spaces that can support various platforms 95 depending on the underlying toolchain. Currently all other device parallel execution spaces 96 than CUDA are experimental. In this work we have tested Serial, Threads, CUDA, and HIP 97 execution spaces. 98 // declarations of variables // declarations of variables constexpr uint32_t MaxNumModules; // declarations of variables constexpr uint32_t MaxNumModules; constexpr uint32_t maxHitsInModule(); Kokkos::View<uint32_t const, Kokkos::CudaSpace > cluStart; Kokkos::View<uint32_t const, Kokkos::CudaSpace > cluStart; Kokkos::View<uint32_t, Kokkos::CudaSpace > moduleStart; Kokkos::parallel_for( Kokkos::RangePolicy <Kokkos::Cuda>(0, MaxNumModules)), KOKKOS_LAMBDA(const int index) { moduleStart(index + 1) = std::min(maxHitsInModule(), cluStart(index)); }); Kokkos::parallel_for( Kokkos::RangePolicy <Kokkos::Cuda>(0, MaxNumModules)), KOKKOS_LAMBDA(const int index) { moduleStart(index + 1) = std::min(maxHitsInModule(), cluStart(index)); }); Figure 2. A simpliﬁed example of using RangePolicy policy with parallel_for. The initializa tion of the declared variables is omitted for brevity. In this example the execution and memory spac template argument are spelled out explicitly. If the compile-time defaults for those suﬃce, the explic Kokkos::parallel_for( Kokkos::RangePolicy <Kokkos::Cuda>(0, MaxNumModules)), moduleStart(index + 1) = std::min(maxHitsInModule(), cluStart(index)); Figure 2. A simpliﬁed example of using RangePolicy policy with parallel_for. The initializa- tion of the declared variables is omitted for brevity. In this example the execution and memory space template argument are spelled out explicitly. If the compile-time defaults for those suﬃce, the explicit template arguments can be left out. Corresponding CUDA program is shown in Figure 3. Kokkos makes use of a runtime library. The library can have the Serial, one host parallel, 99 and one device parallel execution space enabled at the same time, and this set is chosen at the 100 library build conﬁguration time. 1The league corresponds to grid in CUDA, and team corresponds to block. 4.1 Impact on building 128 The current plan to support compute accelerators in CMSSW software stack is to build code 129 for all supported accelerators, and choose the exact version to be run at runtime [22]. The 130 various constraints of the Kokkos runtime library, described in Section 3, make it challenging 131 to deploy in this manner. A single runtime library supporting only one device parallel exe- 132 cution space, and only one CUDA major architecture or CPU vector architecture, would, in 133 this plan, imply the need to build many versions of the runtime library. The correct version 134 would have to be loaded dynamically based on the available hardware. In this work we used 135 exactly one runtime library at a time. 136 Every source ﬁle that includes any Kokkos header must be built with a compiler that is 137 capable of compiling the code for all the enabled execution spaces, even if the source ﬁle 138 would not use any Kokkos functionality. For example, if the Kokkos runtime library was 139 built with CUDA execution space enabled, all source ﬁles including Kokkos headers must be 140 compiled with a CUDA capable compiler. 141 Kokkos provides an integration with the CMake build system. In this work, however, we 142 used CMake only to build the Kokkos runtime library itself, and we used a plain Makeﬁle to 143 build the application code. We did this because CMSSW uses the SCRAM build system [26], 144 and therefore we’d have to understand the exact build rules in order to implement those for 145 SCRAM. 146 The inability of nvcc to link device code from shared objects imposed severe constraints 147 on how the Kokkos runtime library had to be built. We were able to use the runtime library 148 built as a dynamic library with RangePolicy, but with the ﬁrst use of TeamPolicy that 149 approach lead to link errors from nvcc. The only build setup we managed to get to work was 150 to build the Kokkos runtime library as a static library without support for relocatable device 151 code, but with position-independent code for the host (-fPIC) to be able to link the static 152 library with dynamic libraries of the application. // declarations of used variables constexpr uint32_t MaxNumModules; constexpr uint32_t maxHitsInModule(); // declarations of used variables constexpr uint32_t MaxNumModules; constexpr uint32_t maxHitsInModule() // declarations of used variables constexpr uint32_t MaxNumModules; constexpr uint32_t maxHitsInModule() // declarations of used variables constexpr uint32_t MaxNumModules; constexpr uint32_t maxHitsInModule(); __global__ void fillHitsModuleStart(uint32_t const cluStart , uint32_t* moduleStart) { for(int i = threadIdx.x, iend = MaxNumModules; i < iend; i += blockDim.x) { moduleStart[i + 1] = std::min(maxHitsInModule(), cluStart[i]); } __global__ void fillHitsModuleStart(uint32_t const* cluStart , uint32_t* moduleStart) { for(int i = threadIdx.x, iend = MaxNumModules; i < iend; i += blockDim.x) { moduleStart[i + 1] = std::min(maxHitsInModule(), cluStart[i]); } uint32_t const* cluStart_; uint32_t* moduleStart_; fillHitsModuleStart <<<1, 1024>>>(cluStart_ , moduleStart_); Figure 3. CUDA version of the simpliﬁed example expressed in Kokkos in Figure 2. The initialization of the declared variables is omitted for brevity. 3 Kokkos 87 In addition, at least for CUDA execution space, one library 101 can support only GPUs that have the same major compute capability number. For example, 102 one library can support Volta (compute capability 7.0) and Turing (7.5) GPUs, but not Volta 103 and Pascal (6.0) GPUs. In the code the execution space to be used can be chosen at compile 104 time with template arguments. If the execution space is not speciﬁed explicitly, the most 105 advanced execution space available in the library is used, i.e. device parallel execution space 106 is preferred over host parallel execution space, which is preferred over the Serial execution 107 space. Currently Kokkos supports only one device per process. 108 Kokkos provides high-level interface for parallel operations. These include 109 parallel_for for a for-loop of independent iterations, parallel_scan for a preﬁx scan, 110 and parallel_reduce for a reduction. Parallel operations can be nested with some restric- 111 tions. The details of the iteration are controlled with a policy. A RangePolicy can be used 112 for a 1-dimensional range where all elements of the range can be processed independently. 113 An example of parallel_for with RangePolicy is shown in Figure 2 and a corresponding 114 CUDA version in Figure 3. An MDRangePolicy extendes the concept of the 1-dimensional 115 RangePolicy to many, up to 6, dimensions. A TeamPolicy introduces a league of teams 116 that consist of threads1. Threads in a team can use a common scratch memory space, and can 117 synchronize within the team with a barrier. In addition, Kokkos has some support for tasks 118 and graphs, that are not explored in this work. 119 As well as parallel operations, Kokkos provides a datastructure for multi-dimensional ar- 120 ray, Kokkos::View. It is reference counted and behaves like std::shared_ptr, and can 121 be passed to device functions by value. A major feature of the Kokkos::View is that its 122 memory layout can be controlled with template arguments, and the default layout depends on 123 the memory space. In addition, intents for the memory can be expressed with additional tem- 124 plate arguments, for example specifying random-access constant data enables seamless use 125 of CUDA texture caches. Data transfers between the host and the device are done explicitly. 126 4.2 Impact on code As mentioned in Section 3, the Kokkos execution space is chosen at compile time. A choice 160 done at runtime would be a much better ﬁt in the current plans of using compute accelerators 161 in CMSSW. We implemented the capability of choosing the execution space at runtime by 162 building each source ﬁle containing Kokkos code once for each execution space and using 163 namespaces to guarantee diﬀerent symbols for each execution space. 164 Conversion of CUDA kernel calls to Kokkos parallel operations was mostly straightfor- 165 ward. Kokkos provides a parallel scan and sort, and therefore we decided to use those instead 166 of trying to port the implementations of scan and radix sort device functions in the direct 167 CUDA version. The code uses team-wide scan, but before version 3.3, Kokkos provided only 168 league-wide scan. Before updating to Kokkos 3.3 we used the league-wide scan with two ad- 169 ditional kernels to post-process the league-wide result to be equivalent to a team-wide scan. 170 Kokkos’ parallel sort function can be called only from the host code, which meant that we 171 had to split all the CUDA kernels that called the device-side sort function into two kernels, 172 and call the Kokkos’ host-side sort function in between. Finding out the proper and eﬃcient 173 way to transform the CUDA code to use the Kokkos’ scan and sort APIs was the most time 174 consuming single eﬀort. 175 For hierarchical parallelism, or thread teams, we found that the number of threads in a 176 team is not exactly portable. The Serial execution space requires it to be exactly one, Threads 177 execution space can use at most the number of CPU threads, and CUDA execution space has 178 the same limitations as CUDA itself. This disparity can be largely mitigated by specifying 179 the number of threads as Kokkos::AUTO(), that leaves the decision of the number of threads 180 to Kokkos. 181 We found Kokkos::View to be useful by providing a uniﬁed interface for memory allo- 182 cation, and smart pointer semantics for managing the ownership of the memory block. Also 183 the ability to avoid an additional memory allocation in code that transfers data from host to 184 device for CPU-only execution spaces is a plus. 4.1 Impact on building 128 This setup implies that CUDA separate 153 compilation model can not be used, and therefore each source ﬁle must contain all device 154 code called from that ﬁle, either directly or via including other ﬁles. Also, CUDA dynamic 155 parallelism can not be used. 156 p With the HIP execuion space we were able to use a dynamic Kokkos runtime library, and 157 in fact were not able to get a static build to work with the HIP compiler. 158 5 Performance comparison 208 The performance tests were done on GPU nodes of the Cori supercomputer at the National 209 Energy Research Scientiﬁc Computing Center (NERSC). A Cori GPU node has two sockets 210 with Intel Xeon Gold 6148 ("Skylake") processors, each with 20 cores and 2 threads per core, 211 and eight NVIDIA V100 GPUs. For this work we used only one CPU socket, to avoid the 212 impact of non-uniform memory access (NUMA), and one GPU. In all tests the threads were 213 pinned to a single socket. Each job was run for approximately 5 minutes, processing the set 214 of 1000 individual events for an integer number of times, and repeated 8 times on random 215 nodes of the GPU cluster. The code was compiled with GCC 8.3.0, and nvcc from CUDA 216 11.1. 217 In order to minimize the impact of the CPU frequency scaling the CPU programs were 218 tested by running another program on the background with as many threads as needed to ﬁll 219 all the 40 hardware threads of the socket. Table 1 shows the throughput of the Kokkos ver- 220 sion with Serial and Threads execution spaces, and of the direct CPU version with 1 and 40 221 threads. The Kokkos version processes one event at a time, and with the Threads execution 222 space each Kokkos parallel operation is parallelized with the same number of threads. The 223 direct CPU version, on the other hand, is parallelized by processing multiple events concur- 224 rently, one event per thread. While comparing the multi-threaded throughput of these two 225 approaches is not exactly fair, it does show what can be achieved with a single process using 226 the diﬀerent approaches. 227 The results in Table 1 show that the intra-event parallelization scales poorly, whereas par- 228 allelizing over events gives much better throughput and scales well. We have not concluded 229 yet why the direct CPU version gives 1.5 times better throughput than the Kokkos version 230 with Serial execution space. 231 The programs using CUDA were tested without any background activity on the CPU. 232 Table 2 shows the throughput of the Kokkos version with CUDA execution space, and of 233 the direct CUDA version. 4.2 Impact on code The more advanced features like multiple 185 dimensions and the layout control are not needed in this code, where nearly all arrays have 186 only one dimension. The only exception is the track covariance matrix, but we did not try 187 to transform the Eigen-based implementation in the original CUDA into multidimensional 188 Kokkos::View. In this code a SoA abstraction would be much more useful than multi- 189 dimensional array, and we do not see how Kokkos::View would help in crafting SoA data 190 structures. 191 In the ﬁrst Kokkos version we found that about 80 % overall kernel runtime was spent 192 in Kokkos::View initialization. In this code the ﬁrst operation for all device memory is 193 a write either by a memory copy from the host memory, or by a computation done in a 194 kernel. Therefore all the initialization done by default is unnecessary, and avoiding that with 195 Kokkos::ViewAllocateWithoutInitializing argument to Kokkos::View constructor 196 improved the event processing throughput by almost a factor of 3. 197 At the time of writing, we have not been able to successfully run the full application with 198 the HIP execution space. A test application that uses the same build and dynamic library 199 infrastructure works well, but is not complex-enough to give meaningful insights into the 200 performance. 201 Furthermore, we have not yet managed to run the application with multiple concurrent 202 events with Serial or CUDA execution spaces. The Threads execution space explicitly pre- 203 vents calls from more than one thread, even if the calls would come at diﬀerent times. Despite 204 of the Threads execution space being uninteresting to be used in the context of CMSSW, we 205 have included it as a comparison point in the performance measurements in Section 5 to show 206 how a parallelization strategy diﬀerent from concurrent events would perform. 207 6 Conclusions 245 We have ported the Patatrack heterogeneous pixel reconstruction code from CUDA to 246 Kokkos. In our experience Kokkos provides an API that is at a higher level than CUDA, 247 and would be easier to develop new algorithms by physicists that are not necessarily experts 248 in programming. We have achieved almost full portability between CPU, CUDA, and HIP, 249 even if work still continues to understand runtime failures of the HIP execution space version 250 of the code. This analysis shows that Kokkos can give 70 % of native CUDA performance in 251 a simpliﬁed setup without either a memory pool or concurrent events. If similar performance 252 proportion can be achieved also in a more realistic setup, it may be worth using a portable 253 framework to reduce person power in maintaining a code base despite the loss of compute 254 performance. 255 Our impression is that Kokkos would work well for a project that compiles the code 256 separately for each target architecture, does not rely much on shared libraries, uses CMake as 257 the build system, and does not rely on concurrent work outside of Kokkos. CMSSW doing 258 all these in the opposite way implies that integrating the current version of Kokkos into the 259 current data processing model of CMSSW would be challenging to do without sacriﬁcing 260 application performance. It is not, however, clear to us at this time to what extent these 261 challenges are caused by design choices in Kokkos, or by the nature of the portability problem 262 itself. 263 More work is needed to complete the study with Kokkos. In addition, comparisons to 264 other portability technologies are planned within the HEP-CCE. 265 5 Performance comparison 208 In order to compare to the current state of the Kokkos version, the CUDA version was tested also with 1 concurrent event and disabling the use of the memory pool. The reported uncertainty corresponds to sample standard deviation of 8 trials. Test case Throughput (events/s) CUDA version, peak (9 concurrent events and CPU threads) 1840 ± 20 CUDA version, 1 concurrent event 720 ± 20 CUDA version, 1 concurrent event, memory pool disabled 159 ± 1 Kokkos version, CUDA execution space 115.7 ± 0.3 Table 2. Comparison of the event processing throughput between the Kokkos version of the program using CUDA execution space and the original CUDA version. In all cases the CPU threads were pinned to a single CPU socket, and used one NVIDIA V100 GPU. Each test ran about 5 minutes, and the machine was free from other activity. The CUDA version can process data from multiple events concurrently using many CPU threads and CUDA streams, and uses a memory pool to amortize the cost of raw CUDA memory allocations. The maximum throughput from a scan from 1 to 20 concurrent events is reported for the CUDA version. In order to compare to the current state of the Kokkos version, the CUDA version was tested also with 1 concurrent event and disabling the use of the memory pool. The reported uncertainty corresponds to sample standard deviation of 8 trials. Test case Throughput (events/s) CUDA version, peak (9 concurrent events and CPU threads) 1840 ± 20 CUDA version, 1 concurrent event 720 ± 20 CUDA version, 1 concurrent event, memory pool disabled 159 ± 1 Kokkos version, CUDA execution space 115.7 ± 0.3 The Kokkos version with the CUDA execution space reaches about 70 % of the through- 239 put of the direct CUDA version when run on a single concurrent event and disabling the use 240 of the memory pool. Proﬁling indicates that various overheads e.g. in the Kokkos::View 241 are the main cause for the performance diﬀerence. From Table 2 it is also clear that the kind 242 of data processing done in this application beneﬁts greatly from a memory pool, and from 243 processing multiple events concurrently. 244 5 Performance comparison 208 The direct CUDA version can process data from multiple events 234 concurrent with CUDA streams, and this approach helps to get 2.5 times higher throughput 235 from the V100 GPU than when processing one event at a time. With a single event in ﬂight, 236 the memory pool, based on the CachingDeviceAllocator of the CUB [27] library, helps 237 to increase the throughput by 4.5 times compared to using raw CUDA memory allocations. 238 Table 1. Comparison of the event processing throughput between the Kokkos version of the program using Serial and Threads execution spaces and the CPU version implemented from the original CUDA version through a simple translation header. In all cases all the threads were pinned to a single CPU socket (Intel Xeon Gold 6148) that has 20 cores and 2 threads per core. Each test ran about 5 minutes, and CPU-heavy threads from a background process were used to ﬁll all the 40 hardware threads of the socket. The work in the CPU version is parallelized by processing as many events concurrently as the number of threads the job uses without any intra-event parallelization, whereas in the Kokkos version there is only one event in ﬂight, and all parallelization is within the data of that event. For the Kokkos version with Threads execution space the maximum throughput from a scan from 1 to 20 threads is reported. The reported uncertainty corresponds to sample standard deviation of 8 trials. Test case Throughput (events/s) CPU version, 1 thread 13.5 ± 0.2 Kokkos version, Serial execution space 8.5 ± 0.2 CPU version, 40 threads 539 ± 9 Kokkos version, Threads execution space, peak (18 threads) 28 ± 1 Table 2. Comparison of the event processing throughput between the Kokkos version of the program using CUDA execution space and the original CUDA version. In all cases the CPU threads were pinned to a single CPU socket, and used one NVIDIA V100 GPU. Each test ran about 5 minutes, and the machine was free from other activity. The CUDA version can process data from multiple events concurrently using many CPU threads and CUDA streams, and uses a memory pool to amortize the cost of raw CUDA memory allocations. The maximum throughput from a scan from 1 to 20 concurrent events is reported for the CUDA version. Acknowledgements 266 is work was supported by the U.S. Department of Energy, Oﬃce of Science, Oﬃce of High Ph i Hi h E Ph i C t f C t ti l E ll (HEP CCE) t This work was supported by the U.S. Department of Energy, Oﬃce of Science, Oﬃce of High 267 Energy Physics, High Energy Physics Center for Computational Excellence (HEP-CCE) at 268 This work was supported by the U.S. Department of Energy, Oﬃce of Science, Oﬃc 267 Energy Physics, High Energy Physics Center for Computational Excellence (HEP 268 Argonne National Laboratory, Fermi National Accelerator Laboratory, and Lawrence Berke- 269 ley National Laboratory under B&R KA2401045. This research used resources of the Na- 270 tional Energy Research Scientiﬁc Computing Center (NERSC), a U.S. Department of Energy 271 Oﬃce of Science User Facility located at Lawrence Berkeley National Laboratory, operated 272 under Contract No. DE- AC02-05CH11231. 273 References 274 [1] B. Worpitz, Investigating performance portability of a highly scalable particle-in-cell 275 simulation code on various multi-core architectures (2015) 276 [2] E. Zenker, B. Worpitz, R. Widera, A. Huebl, G. Juckeland, A. Knüpfer, W.E. Nagel, 277 M. Bussmann, Alpaka - An Abstraction Library for Parallel Kernel Acceleration (IEEE 278 Computer Society, 2016), 1602.08477 279 [3] A. Matthes, R. Widera, E. Zenker, B. Worpitz, A. Huebl, M. Bussmann, Tuning and 280 optimization for a variety of many-core architectures without changing a single line of 281 implementation code using the Alpaka library (2017), 1706.10086 282 [4] H.C. Edwards, C.R. Trott, D. Sunderland, Journal of Parallel and Distributed Computing 283 74, 3202 (2014) 284 [5] D.A. Beckingsale, J. Burmark, R. Hornung, H. Jones, W. Killian, A.J. Kunen, O. Pearce, 285 P. Robinson, B.S. Ryujin, T.R.W. Scogland, RAJA: Portable Performance for Large- 286 Scale Scientiﬁc Applications (2019), IEEE/ACM International Workshop on Perfor- 287 mance, Portability and Productivity in HPC (P3HPC), p. 71 288 [6] RAJA Performance Portability Layer, https://github.com/LLNL/RAJA (2021), ac- 289 cessed: 2021-02-07 290 [7] The Khoronos SYCL Working Group, SYCL 2020 Speciﬁcation (revision 2) (2021) [8] trisycl, https://github.com/triSYCL/triSYCL (2021), accessed: 2021-02-07 [9] A. Alpay, V. Heuveline, SYCL beyond OpenCL: The Architecture, Current State and Fu- 293 ture Direction of HipSYCL, in Proceedings of the International Workshop on OpenCL 294 (Association for Computing Machinery, New York, NY, USA, 2020), IWOCL ’20, 295 https://doi.org/10.1145/3388333.3388658 296 [10] ComputeCpp, https://developer.codeplay.com/products/computecpp/ce/ 297 home (2021), accessed: 2021-02-07 298 [11] Intel oneAPI DPC++/C++ compiler, https://software.intel.com/content/ 299 www/us/en/develop/tools/oneapi/components/dpc-compiler.html (2021), 300 accessed: 2021-02-07 301 [12] OpenMP Architecture Review Board, OpenMP Application Programming Interface, 302 version 5.1 (2020) 303 [13] OpenACC-Stadnrad.org, The OpenACC Application Programming Interface, version 304 3.1 (2020) 305 [14] A. Bocci, V. Innocente, M. Kortelainen, F. Pantaleo, M. Rovere, Front. Big. Data 3, 306 601728 (2020), 2008.13461 307 [15] CMS Collaboration, JINST 3, S08004 (2008) 308 [16] L. Evans, P. Bryant, JINST 3, S08001 (2008) 309 [16] L. Evans, P. Bryant, JINST 3, S08001 (2008) 309 [17] C.D. Jones, M. Paterno, J. Kowalkowski, L. Sexton-Kennedy, W. Tanenbaum, The New 310 CMS Event Data Model and Framework, in Proceedings of International Conference 311 on Computing in High Energy and Nuclear Physics (CHEP06) (2006) 312 [18] C.D. Jones, E. Sexton-Kennedy, J. Phys.: Conf. Series 513, 022034 (2014) 313 [19] C.D. Jones, L. Contreras, P. Gartung, D. Hufnagel, L. Sexton-Kennedy, J. Phys.: Conf. 314 Series 664, 072026 (2015) 315 [20] C.D. References 274 Jones, J. Phys.: Conf. Series 898, 042008 (2017) 316 [21] oneAPI Threading Building Blocks, https://github.com/oneapi-src/oneTBB 317 (2021), accessed: 2021-02-07 318 [22] CMS Oﬄine Software and Computing, Evolution of the CMS computing model towards 319 phase-2 (2021), CMS-NOTE-2021-001, https://cds.cern.ch/record/2751565 320 [23] A. Bocci, D. Dagenhart, V. Innocente, C. Jones, M. Kortelainen, F. Pantaleo, M. Rovere, 321 EPJ Web Conf. 245, 05009 (2020) 322 [24] Standalone Patatrack pixel tracking, https://github.com/cms-patatrack/ 323 pixeltrack-standalone/ (2021), accessed: 2021-02-07 324 [25] CMS Collaboration, TTToHadronic_TuneCP5_13TeV-powheg-pythia8 in FEVT- 325 DEBUGHLT format for 2018 collision data. CERN Open Data Portal., 326 doi:10.7483/OPENDATA.CMS.GOB0.0LEW (2019) 327 [26] J.P. Wellisch, C. Williams, S. Ashby, SCRAM: Software conﬁguration and manage- 328 ment for the LHC Computing Grid project, in Proceedings of International Conference 329 on Computing in High Energy and Nuclear Physics (CHEP03) (2003), p. TUJP001, 330 cs/0306014 331 [27] CUB, https://nvlabs.github.io/cub/ (2021), accessed: 2021-02-07 332
https://openalex.org/W4200589512	https://dro.deakin.edu.au/ndownloader/files/36835590	English	null	Exotic herbivores dominate Australian high‐elevation grasslands	Conservation science and practice	2,021	cc-by	9,112	K E Y W O R D S K E Y W O R D S grazing, herbaceous, invasive species, landscape elements, mountain, nonindigenous, non- native, plant–herbivore interactions, subalpine, vegetation structure grazing, herbaceous, invasive species, landscape elements, mountain, nonindigenous, non- native, plant–herbivore interactions, subalpine, vegetation structure Correspondence Correspondence Renée Hartley, Fenner School of Environment and Society, Australian National University, 141 Linnaeus Way, Canberra, ACT 2601, Australia. Email: renee.hartley@anu.edu.au C O N T R I B U T E D P A P E R C O N T R I B U T E D P A P E R Renée Hartley1 \| Wade Blanchard1 \| Mellesa Schroder2 \| David B. Lindenmayer1 \| Chloe Sato1 \| Ben C. Scheele1 1Fenner School of Environment and Society, Australian National University, Canberra, Australia 2Southern Ranges Branch, National Parks and Wildlife Service NSW, Jindabyne, Australia 1Fenner School of Environment and Society, Australian National University, Canberra, Australia 2Southern Ranges Branch, National Parks and Wildlife Service NSW, Jindabyne, Australia 1Fenner School of Environment and Society, Australian National University, Canberra, Australia Funding information This research was supported by an Australian Government Research Training Program (RTP) Scholarship, the NSW Government Saving our Species program, a Royal Zoological Society of NSW Paddy Pallin Science Grant, and the National Environmental Science Program; none of which were involved in the study design, data collection, analysis or interpretation, writing, or decision to publish. Exotic herbivores dominate Australian high-elevation grasslands Renée Hartley1 \| Wade Blanchard1 \| Mellesa Schroder2 \| David B. Lindenmayer1 \| Chloe Sato1 \| Ben C. Scheele1 Abstract Invasive species are major drivers of ecosystem degradation globally. How invasive herbivore impacts differ from native herbivore impacts remains understudied. We examined the relationships between herbivore sign and veg- etation height, foliage density, cover of forbs, weeds, bare ground, and soil compaction across environmental and herbivore activity gradients in the main- land Australian Alps. We detected native and exotic herbivore sign at 32.8% and 94.0% of sites, respectively. Total herbivore activity was primarily attrib- uted to exotic herbivores and was associated with elevation and grassland type. Greater horse (exotic) activity was associated with lower vegetation height, lower foliage density, higher forb cover, and higher soil compaction. Greater rabbit and hare (exotic) activity was associated with lower vegetation height, lower foliage density, and a higher cover of bare ground. Greater total herbi- vore activity was associated with greater weed cover. Neither deer (exotic) nor kangaroo and wallaby (native) activity was related to response variables. We demonstrate that exotic herbivores dominate mammalian herbivory in these grasslands, which evolved without analogous hooved species. Given the restricted distribution and high endemism of these ecosystems, and associa- tions between exotic herbivores and characteristics of degraded grasslands, we recommend landscape-scale exotic herbivore management, focusing on maintaining ground cover and vegetation structure. 2Southern Ranges Branch, National Parks and Wildlife Service NSW, Jindabyne, Australia 21 Revised: 20 November 2021 Accepted: 22 November 2021 Received: 25 July 2021 Revised: 20 November 2021 Accepted: 22 November 2021 DOI: 10.1111/csp2.601 DOI: 10.1111/csp2.601 Is activity of exotic and native herbivores associated with a more simple vegetation structure, differences in vegetation composition, and higher soil compaction? Both exotic (Eldridge, Ding, & Travers, 2020; Mutze et al., 2016) and native (Donaldson et al., 2018; Vandegehuchte et al., 2017) herbivores can modify veg- etation structure by reducing biomass, lowering vegeta- tion height, altering plant composition, and increasing soil compaction. The associations between horses, rab- bits, and hares and reduced vegetation biomass in grass- lands of the Alps are well documented (e.g., Eldridge et al., 2019; Leigh et al., 1987), whereas such associa- tions have not been investigated for the majority of her- bivores. As ecosystem responses to herbivory can depend on whether a herbivore is novel to an ecosystem (Antonelli et al., 2011; Bailey & Schweitzer, 2010; Spear & Chown, 2009), we hypothesize that exotic her- bivore activity will be associated with lower vegetation height, lower foliage density, greater weed and forb cover, and greater soil compaction, while native herbi- vores will have neutral associations with these charac- teristics. Native grasslands are among the most extensive and threatened ecosystems worldwide (Sala et al., 2013). The high-elevation grasslands of the Australian Alps Bioregion (hereafter the Alps) evolved with limited mammalian her- bivory and are highly vulnerable to exotic herbivore impacts (Pickering et al., 2010). In this region, large native herbivores are mostly absent from elevations higher than 1650 meters above sea level (a.s.l.) (Green, 2016). In recent years, there have been marked increases in exotic herbivore populations in the Alps, and there is growing evidence of their negative impacts on native grassland ecosystems (Cairns, 2019; Claridge, 2016a; Eldridge, Travers, et al., 2020; Leigh et al., 1987; Matthews & Spooner, 2014; Schulz et al., 2019). However, investigations to date have typically considered a single target herbivore taxon, e.g. horse (Driscoll et al., 2019), deer (Davis et al., 2016), pig (Hone, 2002), hare (Green & Pickering, 2013), or rabbit (Leigh et al., 1987). Multispecies research is therefore required to determine the differing effects of co-occurring native and exotic herbivores in grasslands of the Alps. This research provides critical information on the effects of co-occurring native and exotic herbivore spe- cies in high-elevation grasslands that can be used to guide conservation and management in these areas of social and conservation importance. Williams et al. 2 of 11 HARTLEY ET AL. Herbivore habitat use can be influenced by proximity to natural and constructed landscape elements, such as roads, ecotones, and waterbodies (Davies et al., 2020; Forsyth et al., 2009; Girard et al., 2013; Green et al., 2014; Matthews & Spooner, 2014). We therefore predict herbivore presence and total herbi- vore activity will be positively associated with proxim- ity to woodland as a form of shelter, roads and tracks that may facilitate movement, water sources and low positions in the landscape, and negatively associated with higher elevations due to snow cover. We predict that each herbivore family will have a different proba- bility of presence associated with each grassland type. Herbivore habitat use can be influenced by proximity to natural and constructed landscape elements, such as roads, ecotones, and waterbodies (Davies et al., 2020; Forsyth et al., 2009; Girard et al., 2013; Green et al., 2014; Matthews & Spooner, 2014). We therefore predict herbivore presence and total herbi- vore activity will be positively associated with proxim- ity to woodland as a form of shelter, roads and tracks that may facilitate movement, water sources and low positions in the landscape, and negatively associated with higher elevations due to snow cover. We predict that each herbivore family will have a different proba- bility of presence associated with each grassland type. exotic herbivores have primarily detrimental effects when introduced to new areas (Barbar & Lambertucci, 2018; Nuñez et al., 2010). However, the differing and combined effects of exotic and native herbivores on ecosystems requires further research to better inform biodiversity conservation and effective herbivore management (Davis et al., 2016; Foster et al., 2014; Nuñez et al., 2010). The impacts of herbivores on ecosystem characteris- tics, such as habitat structure, species assemblages, and soils, are numerous and nuanced, often influenced by cli- mate and grazing history (Eldridge, Ding, & Travers, 2020; Sitters et al., 2020; Spake et al., 2020). For example, livestock grazing is used to maintain high- elevation grasslands for conservation and agriculture in some regions with a long history of herbivory (e.g., Komac et al., 2014; Pakeman et al., 2019). However, in ecosystems that evolved in the absence of grazing, native species can lack defensive traits, and ecosystems may be vulnerable to the impacts of exotic herbivores (Antonelli et al., 2011; Bailey & Schweitzer, 2010). 2. 2. 1. Are landscape elements associated with: (a) herbivore presence and (b) total herbivore activity? (2014) identified exotic species as the most pressing threat to ecosystems in the Alps and rec- ommended control of invasive species populations. Man- agement of exotic herbivore populations can be guided by understanding herbivore habitat use and spatial variation in the effects of herbivory (Foster et al., 2014, 2021). How- ever, the associations between herbivore activity and land- scape elements have not been investigated in the Alps. 1 \| INTRODUCTION globally, with livestock now 14 times more abundant than wildlife per unit mass (Gordon & Prins, 2019). While native herbivores may play key roles in maintaining biodiversity and ecosystem functions (Gordon & Prins, 2019), there is strong evidence that Invasive species are a major contributor to ecosystem degradation and their impacts are accelerating (Pyšek et al., 2020). Exotic herbivore numbers are increasing wileyonlinelibrary.com/journal/csp2 1 of 11 2.2 \| Experimental design FIGURE 1 Study area, showing site locations (x) within Kosciuszko National Park (shaded gray). Dashed lines show jurisdiction boundaries. Inset shows location of study area within Australia We established 67 sites, each composed of one 50 m tran- sect, to quantify herbivore presence and activity. We used a stratified random sampling design to select sites by overlaying a 15 km2 grid onto a map of Kosciuszko National Park. We excluded grid cells containing less than 10% grassland (identified from satellite imagery) or less than 10% area above 1200 m a.s.l. We excluded areas beyond 300 m from a vehicle track, burnt in the preced- ing year, or subject to access restrictions or disturbance from infrastructure development. We randomly selected locations using a random number generator to provide distance measures from the northwest corner of each grid cell. If randomly located sites were not in grasslands, on a northern or minor slope, or separated by a minimum of 800 m, we moved the site to the nearest suitable location. FIGURE 1 Study area, showing site locations (x) within Kosciuszko National Park (shaded gray). Dashed lines show jurisdiction boundaries. Inset shows location of study area within Australia Kosciuszko National Park, supporting several endemic and threatened flora species (McDougall & Walsh, 2007). The landscape includes broad, flat valleys, and steep slopes with diverse climates (Doherty, Wright, & McDougall, 2015; McDougall & Walsh, 2007). Annual precipitation ranges from 770 mm in subalpine areas to 3100 mm in alpine areas (Green & Osborne, 2012). Areas above 1200 m a.s.l. receive regular winter snow, and above 1400 m a.s.l., the snow often persists for several months (McDougall et al., 2015). All sites, except two, were dominated by grass (>60% overlapping cover; predominantly Poa spp.). We catego- rized the grassland type at each site according to whether (i) a midstorey was present (“shrub grassland”), (ii) a midstorey was absent and there was greater than 15% nongrass graminoid cover (“graminoid grassland”), or (iii) a midstorey was absent and there was less than 15% nongrass graminoid cover (“tussock grassland”). Our sites were between 1220 and 1951 m elevation, and spanned gradients of landscape attributes (landscape position, dis- tance from the nearest waterbody, distance from the nearest road or track, distance from the nearest wood- land, latitude, and longitude) and herbivore activities (see Supporting Information S1, Figure S1.1–2). 3 of 11 HARTLEY ET AL. 2.2 \| Experimental design ) Only two native mammalian herbivores regularly graze in grasslands above 1500 m a.s.l.; Mastacomys fuscus (Thomas 1882 broad-toothed rat) and Vombatus ursinus (Shaw 1800 common wombat) (Green & Osborne, 2012). However, the reduction in snow depth and duration in recent decades due to climate change has facilitated macropod and wombat range expansions (Green, 2016; Matthews & Spooner, 2014). Domestic live- stock grazing by sheep and cattle occurred in Kosciuszko National Park from the early 1800s until 1967 (Good & Johnston, 2019). Horse, rabbit, and hare populations had established by the late 1800s (Dyring, 1990; Leigh et al., 1987). However, by the late 1900s, rabbits were res- ident throughout the year and associated with significant changes in vegetation in subalpine grasslands, while horses and hares remained infrequent occupants (Green & Osborne, 2012; Leigh et al., 1987). Deer were first observed in Kosciuszko National Park in the 1990s but not recorded in high-elevation grasslands until 2008 (T. Stubbs personal communication, January 2021). The New South Wales government has legislative responsibil- ities to manage invasive species in Kosciuszko National Park (Biosecurity Act, 2015; Kosciuszko Wild Horse Heritage Bill, 2018) and in 2020, undertook 23 Asset Pro- tection and Containment programs to manage exotic herbivores (NSW National Parks and Wildlife Service unpublished data). 3 of 11 3 o 2.1 \| Study area We conducted this study in the high-elevation grasslands of Kosciuszko National Park (Figure 1), a 690,660 ha United Nations Educational Scientific and Cultural Organi- zation (UNESCO) Biosphere Reserve on Dtjilmamidtang, Ngarigo, and Walgalu Country in south-eastern Australia (Doherty, Wright, & McDougall, 2015; UNESCO, 2019). Treeless vegetation communities total almost 90,000 ha of To address these key research gaps, we examined the following questions: 4 of 11 HARTLEY ET AL. (kPa) at 20 points along the transect using a handheld penetrometer according to manufacturer specifications (ST 315, Prospectors, Australia). As soil moisture can affect soil compaction, we classified the soil as wet or dry at the time of measurement according to whether soil stuck to our fingers upon touch. combined). We used a BGLM with a negative binomial distribution and a log link function. We selected the best- fitting model from all possible models (64 candidate models; Table S3.6e). We quantified associations between vegetation and soil variables and observed herbivore sign at each site. We modeled each response variable (mean groundstorey vegetation height, foliage density, forb cover, weed cover, proportion of bare ground, and soil compaction) with total herbivore activity on the log scale as the predictor variable. We then modeled horse (Equidae), deer (Cervidae), rabbit and hare (Leporidae), and kangaroo and wallaby (Macropodidae) activities on the log scale as additive predictor variables. We selected the best-fitting model from all possible models (17 candidate models; Table S4.7). For each analysis, we fit BGLM using a Gamma distribution for continuous response variables (vegetation height, foliage density, soil compaction) and a Beta distribution for proportion response variables (forbs, weeds, bare ground). We considered a “herbivore” to be any mammal with a diet consisting primarily of plant material. We used a relative index of evidence of herbivory, quantified by the frequency of herbivore sign at each site. We calculated herbivore activity by recording the presence or absence of herbivore sign, such as scats and diggings, within each half of a 2 m radius plot at 5 m intervals along the tran- sect to give a frequency measure (out of 20) per sign type (Forsyth et al., 2007; Hone, 2002). We moved elevated vegetation but did not disturb the litter layer or dense ground cover to uncover sign. We classified sign types to family according to Triggs (1996) and Claridge (2016b), as species-level differences between sign cannot always be reliably distinguished. We classified worn paths, con- sumed vegetation, and pulled vegetation as “generic". We did not include pellets with evidence of decay, nor did we convert indices to animal densities (Forsyth et al., 2007). We conducted all surveys within a 4-week period, in February and March 2020, to minimize seasonal and weather variability between sites. At this time, snow cover was absent and mammal populations were likely to be occupying their maximal elevational ranges within the bounds of their current distribution (Foster et al., 2021). Vegetation and soil variables may be associated with landscape elements; for example, plant height and weed species richness generally decline with increasing eleva- tion (Green, 2016; McDougall & Walsh, 2007). We checked for Pearson and Spearman correlations between vegetation and soil variables and landscape elements using a two-sided t-test with the “cor.test” function in the R package “stats” (R Core Team, 2020). While herbivory has the potential to mask underlying associations, we found that weed cover and soil compaction were signifi- cantly (negatively) correlated with elevation (Figure S1.2 and Table S1.1). 2.3 \| Data collection methods We derived landscape elements from publicly available spatial datasets. These included elevation (m a.s.l.), land- scape position (Topographic Wetness Index), distance from the nearest waterbody (m), distance from the nearest road or track (m), and distance from the nearest woodland (m) (see Supporting Information S2 for source information). We manually corrected a small number of distance measures where elements were present on- ground but not captured by the spatial classification. Along each 50 m transect, we recorded the over- lapping percent cover of nongrass graminoids, shrubs, forbs (native and exotic combined), grasses, weeds (exotic flora, all forms) and bare ground within circular plots of 2 m radius at 10 m intervals. We calculated the mean groundstorey vegetation height (cm) and foliage density (number of foliage intercepts in 10 cm bands of a struc- ture pole up to 50 cm), excluding flowering parts, for each site. We measured soil compaction in kilopascals FIGURE 2 Relative contribution of each herbivore family FIGURE 2 Relative contribution of each herbivore family FIGURE 2 Relative contribution of each herbivore family (continuous segment borders) and generic (dashed segment borders) sign to total herbivore activity across all sites. The shading shows the proportion of native (white) and exotic (gray) herbivore activity contributing to each sign type. Herbivores with the potential to contribute to each generic sign are listed in parentheses (Claridge, 2016a; Leigh et al., 1987; Triggs, 1996; Green & Osborne, 2012; Green et al., 2013, 2014; Eldridge et al., 2019; Schulz et al., 2019). The proportion of native and exotic herbivores within generic sign segments is inferred from the proportion of each herbivores' contribution to nongeneric sign. Roll pits and pugging (horses, deer, pigs; 0.08%), unidentified sign (0.17%), and broad- toothed rat sign (0.08%) are included but not visible in the chart due to their small contributions 3.1 \| Distribution of herbivore families We detected herbivore sign (exotic and native) at 98.5% of sites; with exotic species at 94.0% of sites, and native species at 32.8% of sites. Exotic and native herbivores co-occurred at 28.4% of sites. Exotic herbivores were rep- resented by more families than native herbivores and contributed the majority of the herbivore activity (Figure 2). For native herbivores, we detected Macropodidae (Macropus giganteus Shaw 1790 eastern gray kangaroo, Macropus rufogriseus Desmarest 1817 red- necked wallaby), Vombatidae (common wombat), and Muridae (broad-toothed rat) sign. For exotic herbivores, 2.4 \| Statistical analyses We conducted all BGLM in R (R Core Team, 2020), using the package “brms” (Bürkner, 2017) and an unin- formative prior distribution using a student-t distribution (df = 7, mean = 0, SD = 2.5). We used Markov chain Monte Carlo (MCMC) methods and the Gelman–Rubin convergence diagnostic (Ȓ) based on four MCMC chains (Gelman & Rubin, 1992). For all models, Ȓ was equal to 1.00. We used four Markov chains and ran them for 2000 iterations discarding the first 1000 as a burn-in with a thinning factor of 1. Where we performed model selec- tion, we used the Widely Applicable Information Crite- rion (WAIC) (Vehtari et al., 2017a, 2017b) to select the best-fitting model from the candidate models. Where the difference in the WAIC of the best-fitting models was <1, we selected the most parsimonious model. We checked for multicollinearity in all models with multiple predictor variables using the ‘car’ package to calculate the Vari- ance Inflation Factor (Fox & Weisberg, 2019). No values for the Generalized Variance Inflation Factor were ≥2. We used the “DHARMa” package to check the residuals for model fit and evidence of nonlinearity (Hartig & When included as predictor variables in models, we log- transformed distance and sign frequency to minimize the effect of extreme values. Prior to log transformation, we added a small value to the zero values to avoid taking the logarithm of zero. We excluded two sites from soil com- paction analyses and one site from height analyses due to missing data. We standardized data for continuous pre- dictor variables (mean = 0, SD = 1). We modeled the probability of presence of each herbivore family using a Bayesian Generalized Linear Model (BGLM) (Gelman et al., 2013) with a Bernoulli dis- tribution and a logistic link function. We included the following predictor variables: elevation, log-distance to woodland, log-distance to a waterbody, log-distance to a road or track, landscape position, and grassland type. We selected the best-fitting model (see below for details) from all possible models (64 candidate models; Table S3.6a–d). Using the same predictor variables, we examined associations between landscape elements and total herbi- vore activity (all herbivore families and generic sign HARTLEY ET AL. 5 of 11 Lohse, 2020). We found no clear indications of non- linearity and therefore did not investigate alternative functions. We conducted post hoc calculations of esti- mated marginal means and pairwise comparisons using the R package “emmeans” (Lenth et al., 2018). we detected the sign of Equidae (Equus caballus Linnaeus 1758 domesticated horse), Cervidae (Cervus unicolor Kerr 1792 sambar deer and Dama dama Linnaeus 1758 fallow deer), Suidae (Sus scrofa Linnaeus 1758 pig), and Leporidae (Lepus europaeus Pallas 1778 European hare and Oryctolagus cuniculus Linnaeus 1758 European rabbit). We detected rabbit and hare sign at the highest number of sites (85%), followed by horse (42%), deer (30%) and kangaroo and wallaby (28%) sign. The elevation range was greatest for the rabbits and hares (1221–1951 m a.s.l.), followed by deer (1221–1750 m), horses (1253–1721 m), and kangaroos and wallabies (1256–1537 m). Pigs, wombats, and broad- toothed rats were present at less than 10% of sites and we therefore excluded them from family-level statistical ana- lyses. We detected horses, deer, rabbits and hares, and kan- garoos and wallabies with each of the other families at three or more sites (Table S1.1). we detected the sign of Equidae (Equus caballus Linnaeus 1758 domesticated horse), Cervidae (Cervus unicolor Kerr 1792 sambar deer and Dama dama Linnaeus 1758 fallow deer), Suidae (Sus scrofa Linnaeus 1758 pig), and Leporidae (Lepus europaeus Pallas 1778 European hare and Oryctolagus cuniculus Linnaeus 1758 European rabbit). We detected rabbit and hare sign at the highest number of sites (85%), followed by horse (42%), deer (30%) and kangaroo and wallaby (28%) sign. The elevation range was greatest for the rabbits and hares (1221–1951 m a.s.l.), followed by deer (1221–1750 m), horses (1253–1721 m), and kangaroos and wallabies (1256–1537 m). Pigs, wombats, and broad- toothed rats were present at less than 10% of sites and we therefore excluded them from family-level statistical ana- lyses. We detected horses, deer, rabbits and hares, and kan- garoos and wallabies with each of the other families at three or more sites (Table S1.1). 3.2 \| Landscape elements associated with each herbivore family We present the associations between landscape elements and each herbivore family derived from the best-fitting models in Figure 3a–h (see Tables S1.3a and S1.4a for details). The probability of horse presence increased as elevation decreased and as distance to water increased. The probability of deer presence was lower in graminoid grasslands than in shrub or tussock grasslands. The prob- ability of deer presence also increased as elevation increased, and as distance from a waterbody decreased. The probability of rabbit and hare presence increased as distance from a waterbody increased. The probability of kangaroo and wallaby presence increased as elevation decreased, and was lower in shrub and tussock grass- lands. The probability of presence of any family was not significantly related to the log-distance to roads and tracks, log-distance to woodland or landscape position. However, we note that the range of distances to roads and tracks sampled in our study was limited to achieve rigorous sampling across a large spatial scale, and there- fore associations between roads and herbivore presence warrant further investigation across a greater range of distances. 3.3 \| Landscape elements associated with total herbivore activity 6 of 11 0.2 0.4 0.6 0.8 1400 1600 1800 Elevation (m a.s.l.) Probability of deer presence (c) Probability of deer presence 0.2 0.4 0.6 0.8 1400 1600 1800 Elevation (m a.s.l.) Probability of kangaroo/wallaby presence (g) Probability of kangaroo/wallaby presence Probability of deer presence 0.2 0.4 0.6 0.8 gr sh tu G l d t Probability of deer presence (d) ce (b) (c) Probability of rabbit/hare presence Probability of kangaroo/wallaby prese (f) Probability of kangaroo/wallaby pres y ( ) 20 40 60 1400 1600 1800 Elevation (m a.s.l.) Total herbivore activity (index) (i) 20 40 60 gr sh tu Grassland type Total herbivore activity (index) (j) Total herbivore activity (index) Total herbivore activity (in FIGURE 3 Associations between the probability of presence of each herbivore family (y-axes, a–h) and total herbivore activity (y-axes, i–j), and landscape elements (x-axes) derived from the best-fitting models. Grassland types are graminoid (gr), shrub (sh), and tussock (tu). Gray shading and error bars represent the upper and lower 95% credible intervals decreased. Total herbivore activity was greater in tussock grasslands than in graminoid grasslands. Total herbivore activity in shrub grasslands did not differ significantly from tussock or graminoid grasslands. and hare activity increased. Soil compaction increased as horse activity increased. Soil compaction was not associ- ated with soil moisture at the time of sampling. None of the response variables was significantly related to deer or kangaroo and wallaby activity. 3.3 \| Landscape elements associated with total herbivore activity We present the associations between landscape elements and total herbivore activity derived from the best-fitting model in Figure 3i,j) (see Tables S1.3b and S1.4b for details). Total herbivore activity increased as elevation 0.2 0.4 0.6 0.8 1400 1600 1800 Elevation (m a.s.l.) Probability of horse presence (a) 0.2 0.4 0.6 0.8 100 200 300 Nearest waterbody (m) Probability of horse presence (b) 0.2 0.4 0.6 0.8 1400 1600 1800 Elevation (m a.s.l.) Probability of deer presence (c) 0.2 0.4 0.6 0.8 gr sh tu Grassland type Probability of deer presence (d) 0.2 0.4 0.6 0.8 100 200 300 Nearest waterbody (m) Probability of deer presence (e) 0.2 0.4 0.6 0.8 100 200 300 Nearest waterbody (m) Probability of rabbit/hare presence (f) 0.2 0.4 0.6 0.8 1400 1600 1800 Elevation (m a.s.l.) Probability of kangaroo/wallaby presence (g) 0.2 0.4 0.6 0.8 gr sh tu Grassland type Probability of kangaroo/wallaby presence (h) 20 40 60 1400 1600 1800 Elevation (m a.s.l.) Total herbivore activity (index) (i) 20 40 60 gr sh tu Grassland type Total herbivore activity (index) (j) FIGURE 3 Associations between the probability of presence of each herbivore family (y-axes, a–h) and total herbivore activity (y-axes, i–j), and landscape elements (x-axes) derived from the best-fitting models. Grassland types are graminoid (gr), shrub (sh), and tussock (tu). Gray shading and error bars represent the upper and lower 95% credible intervals 6 of 11 HARTLEY ET AL. 0.2 0.4 0.6 0.8 1400 1600 1800 Elevation (m a.s.l.) Probability of horse presence (a) 0.2 0.4 0.6 0.8 100 200 300 Nearest waterbody (m) Probability of horse presence (b) 0. 0. 0. 0. Probability of deer presence 0.2 0.4 0.6 0.8 100 200 300 Nearest waterbody (m) Probability of deer presence (e) 0.2 0.4 0.6 0.8 100 200 300 Nearest waterbody (m) Probability of rabbit/hare presence (f) 0. 0. 0. 0. Probability of kangaroo/wallaby presence 6 of 11 HARTLEY ET AL. 4 \| DISCUSSION 7 of 11 x) 30 40 50 5 10 15 20 Horse activity (index) Foliage density <50cm (c) 30 40 50 5 10 15 20 Rabbit/hare activity (index) Foliage density <50cm (d) 7 of 11 HARTLEY ET AL. 7 of 11 10 15 20 25 5 10 15 20 Horse activity (index) Vegetation height (cm) (a) 10 15 20 25 5 10 15 20 Rabbit/hare activity (index) Vegetation height (cm) (b) 30 40 50 5 10 Horse act Foliage density <50cm (c) Horse activity (index) Horse activity (index) 0.00 0.01 0.02 10 20 Rabbit/hare activity (index) Proportion of bare ground (g) y ( ) 0.1 0.2 0.3 5 10 15 20 Horse activity (index) Proportion of forb cover (e) y ( ) 0.0 0.1 0.2 20 40 60 80 Total herbivore activity (index) Proportion of weed cover (f) 1 2 5 10 15 Horse activity (index) p ( ) (h) Proportion of bare ground (h) Proportion of weed cover (g) (e) Soil compaction (kPa) Total herbivore activity (index) FIGURE 4 Associations between the vegetation and soil response variables (y-axes) and the relative indices of herbivore activity (x-axes) derived from the best-fitting models. Gray shading represents the upper and lower 95% credible intervals Schweitzer, 2010; Milchunas & Lauenroth, 1993; Spear & Chown, 2009). Consistent with studies in high-elevation grasslands around the world (Davies & Boyd, 2019; Lu et al., 2017; Schütz et al., 2003), our results demonstrate an association between exotic herbivore taxa and lower grass- land height, lower foliage density, greater forb and weed covers, and greater soil compaction; characteristics that may result in substantial negative impacts on biodiversity and grassland ecosystems. Schweitzer, 2010; Milchunas & Lauenroth, 1993; Spear & Chown, 2009). Consistent with studies in high-elevation grasslands around the world (Davies & Boyd, 2019; Lu et al., 2017; Schütz et al., 2003), our results demonstrate an association between exotic herbivore taxa and lower grass- land height, lower foliage density, greater forb and weed covers, and greater soil compaction; characteristics that may result in substantial negative impacts on biodiversity and grassland ecosystems. wallaby establishment on sites at higher elevations in the future (Boone, 2019; but see Green, 2016). Deer were more likely to occur in shrub and tussock grasslands and closer to waterbodies. 4.1 \| Are landscape elements associated with herbivore presence? Each herbivore family had an association with one or more landscape element. As elevation increased, the probability of deer presence increased, while the proba- bility of horse and kangaroo and wallaby presence decreased. Kangaroos and wallabies are restricted to ele- vations with little or no snow cover due to their diet and locomotion (Green, 2016). The presence of exotic herbi- vores at higher elevations than kangaroos and wallabies (1540 m a.s.l.) suggests exotic herbivores have a greater ability to either migrate to higher elevations during snow-free periods, or move through snow and consume exposed woody browse (Green et al., 2014). The continu- ing reduction of snow cover (Green & Pickering, 2013) and increase in temperatures with climate change may facilitate grassland grazing in winter (Giroux et al., 2016) and enable horse, deer, kangaroo, and 4 \| DISCUSSION Previous investiga- tions of the impacts of deer (e.g., Brown et al., 2016; Claridge, 2016a) and horses (e.g., Foster & Scheele, 2019; Robertson et al., 2019) in our study region have often focused on waterbodies. However, our study showed there was a greater probability of horse presence away from, rather than close to, waterbodies. These findings do not detract from evidence of damage to waterbodies associated with horses, rather they highlight the potential for less visi- ble impacts away from the soft soils adjacent to waterbodies and the need to assess impact across the land- scape. The probability of rabbit and hare presence also increased further from waterbodies. The establishment of rabbits and hares throughout the study area is consistent with observations that these species have been successful invaders of ecosystems worldwide, largely due to their rapid dispersal and reproduction, diverse diet and ability to adapt to harsh conditions (Barbar & Lambertucci, 2018; Courchamp et al., 2003; Foster et al., 2021). 4 \| DISCUSSION 4 We present the associations between vegetation and soil variables and log-herbivore activity derived from the best-fitting models in Figure 4 (see Table S1.5 for details). Vegetation height and foliage density both decreased with increasing horse and rabbit and hare activity. Forb cover increased as horse activity increased. Weed cover increased as total herbivore activity increased. There was little bare ground recorded across the sites, with a mean proportion of <0.01 cover (min = 0, max = 0.18). How- ever, the proportion of bare ground increased as rabbit Using a landscape-scale, multispecies study, we show that exotic taxa are the dominant mammalian herbivores in Australian high-elevation grasslands. Native herbivore sign was detected at approximately one-third of sites, primarily below 1540 m a.s.l. In contrast, exotic herbivores were almost ubiquitous across the study grasslands, which have evolved largely in the absence of grazing (Green & Osborne, 2012). A lack of shared coevolutionary history indicates that such grasslands are likely to be highly vul- nerable to grazing impacts (Antonelli et al., 2011; Bailey & 10 15 20 25 5 10 15 20 Horse activity (index) Vegetation height (cm) (a) 10 15 20 25 5 10 15 20 Rabbit/hare activity (index) Vegetation height (cm) (b) 30 40 50 5 10 15 20 Horse activity (index) Foliage density <50cm (c) 30 40 50 5 10 15 20 Rabbit/hare activity (index) Foliage density <50cm (d) 0.1 0.2 0.3 5 10 15 20 Horse activity (index) Proportion of forb cover (e) 0.0 0.1 0.2 20 40 60 80 Total herbivore activity (index) Proportion of weed cover (f) 0.00 0.01 0.02 10 20 Rabbit/hare activity (index) Proportion of bare ground (g) 1 2 5 10 15 Horse activity (index) Soil compaction (kPa) (h) FIGURE 4 Associations between the vegetation and soil response variables (y-axes) and the relative indices of herbivore activity (x-axes) derived from the best-fitting models. Gray shading represents the upper and lower 95% credible intervals HARTLEY ET AL. 7 of 11 10 15 20 25 5 10 15 20 Horse activity (index) Vegetation height (cm) (a) 10 15 20 25 5 10 15 20 Rabbit/hare activity (index) Vegetation height (cm) (b) 30 40 50 5 10 15 20 Horse activity (index) Foliage density <50cm (c) 30 40 50 5 10 15 20 Rabbit/hare activity (index) Foliage density <50cm (d) HARTLEY ET AL. 4.3 \| Recommendations for management and policy The wide distribution of exotic herbivores in the Alps highlights the need for landscape-scale management, accounting for emigration and dispersal. Priority should be given to removal of horses and deer from elevations above 1500 m a.s.l. where ecosystems evolved with lim- ited mammalian grazing (Costin et al., 1982). Their impacts also should be greatly reduced in all other areas. Although we found no significant associations between deer and grassland characteristics, this may be a result of their relatively recent invasion of the region. There is evi- dence that even small populations of exotic horses (Driscoll et al., 2019) and deer (Tanentzap et al., 2009) can heavily impact ecosystems, and eradication is there- fore likely to be the most appropriate management goal (Prior et al., 2018). As leporids are almost ubiquitous, population suppression should be prioritized in areas of high conservation significance, such as where competi- tion with threatened species occurs, and above the treeline where ecosystems are most threatened by climate change and species endemism is high (Doherty, Wright, & McDougall, 2015). Leporids can support exotic predator persistence in habitats that are otherwise marginal for the predators, or increase exotic predator abundance (Courchamp et al., 2003; Foster et al., 2021). Exotic pred- ator suppression may need to accompany leporid control to limit “prey-switching” in habitats with conservation- significant fauna (Doherty, Dickman, et al., 2015). As the major contributors to total herbivore activity, we recom- mend leporid and horse population control be prioritized in areas where weed cover or weed invasion are of man- agement concern. 4.2.1 \| Is activity of exotic and native herbivores associated with a more simple vegetation structure, differences in vegetation composition and higher soil compaction? 4.2.1 \| Is activity of exotic and native herbivores associated with a more simple vegetation structure, differences in vegetation composition and higher soil compaction? Our study is consistent with research from other regions showing exotic horse activity is associated with lower vegetation height and density, and higher forb cover and soil compaction in grasslands (Beever et al., 2018; de Villalobos & Schwerdt, 2020). Soil compaction can alter plant composition and hinder plant growth by reducing gas exchange and water infiltration (Czortek et al., 2018; Pineiro et al., 2010). HARTLEY ET AL. vegetation composition (Pyšek et al., 2020) simplify struc- ture, and reduce habitat suitability for native fauna (Sato et al., 2014), in high-elevation grasslands. considerably greater than of native herbivore sign. The higher total herbivore activity at lower elevations is unsurprising, as snow cover is ephemeral and herbivore populations can persist throughout the winter (Green et al., 2014). The greater total herbivore activity associ- ated with tussock grasslands is of significance to land managers as tussock grasslands are habitat for threatened species, such as Max Mueller's burr-daisy (Calotis pub- escens Muell. Ex Benth., Walsh & McDougall 2002) (McDougall & Walsh, 2007) and the broad-toothed rat, which depends on large, dense native grasses (Eldridge et al., 2019) for shelter. 4.2 \| Are landscape elements associated with total herbivore activity? Total herbivore activity was associated with elevation and grassland type, and the contribution of exotic herbi- vore sign to the measure of total herbivore activity was 8 of 11 REFERENCES Alexander, J. M., Lembrechts, J. J., Cavieres, L. A., Daehler, C., Haider, S., Kueffer, C., … Seipel, T. (2016). Plant invasions into mountains and alpine ecosystems: Current status and future challenges. Alpine Botany, 126(2), 89–103. https://doi.org/10. 1007/s00035-016-0172-8 Davies, C., Wright, W., Hogan, F. E., & Davies, H. (2020). Detect- ability and activity patterns of sambar deer (Rusa unicolor) in Baw Baw National Park, Victoria. Australian Mammalogy, 42(3), 312–320. https://doi.org/10.1071/AM19029 Antonelli, A., Humphreys, A. M., Lee, W. G., & Linder, H. P. (2011). Absence of mammals and the evolution of New Zealand grasses. Proceedings of the Royal Society B: Biological Sciences, 278(1706), 695–701. https://doi.org/10.1098/rspb.2010.1145 Davies, K. W., & Boyd, C. S. (2019). Ecological effects of free- roaming horses in north American rangelands. Bioscience, 69(7), 558–565. https://doi.org/10.1093/biosci/biz060 Davis, N. E., Bennett, A., Forsyth, D. M., Bowman, D. M. J. S., Lefroy, E. C., Wood, S. W., … Johnson, C. N. (2016). A system- atic review of the impacts and management of introduced deer (family Cervidae) in Australia. Wildlife Research, 43(6), 515–532. https://doi.org/10.1071/WR16148 Bailey, J. K., & Schweitzer, J. A. (2010). The role of plant resistance and tolerance to herbivory in mediating the effects of intro- duced herbivores. Biological Invasions, 12, 337–351. https://doi. org/10.1007/s10530-009-9630-1 Barbar, F., & Lambertucci, S. A. (2018). The roles of leporid species that have been translocated: A review of their ecosystem effects as native and exotic species. Mammal Review, 48(4), 245–260. https://doi.org/10.1111/mam.12126 Doherty, M. D., Wright, G. T., & McDougall, K. L. (2015). The flora of Kosciuszko National Park, New South Wales: Summary and overview. Cunninghamia, 9620, 13–68. https://doi.org/10.7751/ cunninghamia.2015.15.002 Beever, E. A., Huntsinger, L., & Petersen, S. L. (2018). Conservation challenges emerging from free-roaming horse management: A vexing social-ecological mismatch. Biological Conservation, 226, 321–328. https://doi.org/10.1016/j.biocon.2018.07.015 Doherty, T. S., Dickman, C. R., Nimmo, D. G., & Ritchie, E. G. (2015). Multiple threats, or multiplying the threats? Interac- tions between invasive predators and other ecological distur- bances. Biological Conservation, 190, 60–68. https://doi.org/10. 1016/j.biocon.2015.05.013 Biosecurity Act 2015. Government of New South Wales. https:// www.legislation.nsw.gov.au/view/html/inforce/current/act- 2015-024 (Accessed: March 18, 2021). Donaldson, J. E., Archibald, S., Govender, N., Pollard, D., Luhdo, Z., & Parr, C. L. (2018). Ecological engineering through fire-herbivory feedbacks drives the formation of savanna graz- ing lawns. Journal of Applied Ecology., 55(1), 225–235. https:// doi.org/10.1111/1365-2664.12956 Boone, R. B. (2019). Weather and climate impacts on browsing and grazing ungulates. In I. J. Gordon & H. H. T. ORCID Czortek, P., Delimat, A., Dyderski, M. K., Zięba, A., Jagodzinski, A. M., & Jaroszewicz, B. (2018). Climate change, tourism and historical grazing influence the distribution of Carex lachenalii Schkuhr—A rare arctic- alpine species in the Tatra Mts. Science of the Total Environ- ment, 618, 1628–1637. https://doi.org/10.1016/j.scitotenv. 2017.10.001 Renée Hartley https://orcid.org/0000-0002-2438-2053 Renée Hartley https://orcid.org/0000-0002-2438-2053 David B. Lindenmayer https://orcid.org/0000-0002- 4766-4088 David B. Lindenmayer https://orcid.org/0000-0002- 4766-4088 Chloe Sato https://orcid.org/0000-0001-7707-5068 Ben C. Scheele https://orcid.org/0000-0001-7284-629X Daskin, J. H., & Pringle, R. M. (2016). Does primary productivity modulate the indirect effects of large herbivores? A global meta-analysis. The Journal of Animal Ecology, 85(4), 857–868. https://doi.org/10.1111/1365-2656.12522 ACKNOWLEDGMENTS Cairns, S. (2019). Feral horses in the Australian Alps: The analysis of aerial surveys conducted in April–May, 2014 and April–May 2019. Armidale. The authors acknowledge the Djilamatang, Ngambri, Ngarigo, Ngunnawal, Ngunawal, Walgalu, and Wiradjuri people as the Traditional Custodians of the lands on which this study was conducted. The authors thanks Rob Gibbs, Nick Clemann, Geoffrey Robertson, and Dr Gillian Fuller for invaluable advice, local knowledge, and field assistance. Claridge, A. W. (2016a). Ecological and agricultural impacts of introduced deer in the Australian Alps. A final report to the Australian Alps Liaison Committee. Queanbeyan NSW. https:// theaustralianalps.files.wordpress.com/2018/03/0011-ecological- agricultural-impacts.pdf. Claridge, A. W. (2016b). Introduced deer field identification guide for the Australian Alps. Queanbeyan, New South Wales. https:// theaustralianalps.files.wordpress.com/2018/03/0012-introduce d-deer-field-identification-guide-2016.pdf. DATA AVAILABILITY STATEMENT Courchamp, F., Chapuis, J. L., & Pascal, M. (2003). Mammal invaders on islands: Impact, control and control impact. Biolog- ical Reviews of the Cambridge Philosophical Society, 78, 347–383. https://doi.org/10.1017/S1464793102006061 Data will be published on Dryad upon acceptance for publication; https://datadryad.org/stash. Data will be published on Dryad upon acceptance for publication; https://datadryad.org/stash. 4.3 \| Recommendations for management and policy Grazing by horses in the Alps differs most notably from domestic livestock grazing because there is no rest period to allow plants to complete their lifecycle, resulting in the potential extirpation of native species (Davies & Boyd, 2019). The resulting changes in grassland structure and composition can have cascading community-level impacts and negatively impact on threatened species (Daskin & Pringle, 2016; Eldridge et al., 2019; Vandegehuchte et al., 2017). Similarly, vege- tation height and density decreased and bare ground increased with greater rabbit and hare activity. Rabbits and hares can reduce biodiversity by removing palatable native flora before such impacts on plant structure are detected (Mutze et al., 2016). Further, exposed soils can lead to a shift from grasslands to shrublands (McDougall & Walsh, 2007). Therefore, uncontrolled rab- bit and hare populations could result in long-term loss of biodiversity and grassland ecosystems. We recommend strategic, long-term, and landscape- scale exotic herbivore management, integrated with weed and exotic predator management and accounting for cli- mate change. Specific aims should be to: (i) suppress or eradicate exotic herbivores from ecosystems that did not evolve with analogous herbivory (Norris, 2018), (ii) prevent expansion and establishment of new exotic herbivore populations, (iii) reduce total herbivore activity in all areas by focusing efforts on major exotic contribu- tors, and (iv) protect areas of high conservation value to avoid loss of native species. Long-term monitoring of her- bivore management activities is needed to assess whether management goals for ecosystems are being met (Prior et al., 2018). We found weed cover was best explained by total her- bivore activity. While mountain plant communities tend to be resistant to weed invasion largely due to climatic fil- tering and fewer disturbances, this result suggests the cumulative impacts of multiple herbivore taxa provides the necessary disturbance to facilitate weed establish- ment (Alexander et al., 2016). Under a warming climate, this facilitation may extend to higher elevations. High weed cover can reduce native plant species richness, alter HARTLEY ET AL. 9 of 11 CONFLICT OF INTEREST The authors do not have conflicts of interest to declare. The authors do not have conflicts of interest to declare. Costin, A., Gray, M., Totterdell, C. J., & Wimbush, D. J. (1982). Kosciuszko Alpine flora. CSIRO and William Collins Pty Ltd. DATA AVAILABILITY STATEMENT Data will be published on Dryad upon acceptance for publication; https://datadryad.org/stash. 10 of 11 Hone, J. (2002). Feral pigs in Namadgi National Park, Australia: Dynamics, impacts and management. Biological Conservation, 105(2), 231–242. https://doi.org/10.1016/S0006-3207(01)00185-9 Foster, C. N., Barton, P. S., & Lindenmayer, D. B. (2014). Effects of large native herbivores on other animals. Journal of Applied Ecol- ogy, 51(4), 929–938. https://doi.org/10.1111/1365-2664.12268 Komac, B., Domènech, M., & Fanlo, R. (2014). Effects of grazing on plant species diversity and pasture quality in subalpine grass- lands in the eastern Pyrenees (Andorra): Implications for con- servation. Journal for Nature Conservation, 22(3), 247–255. https://doi.org/10.1016/j.jnc.2014.01.005 Foster, C. N., & Scheele, B. C. (2019). Feral-horse impacts on cor- roboree frog habitat in the Australian Alps. Wildlife Research, 46(2), 184. https://doi.org/10.1071/WR18093 Foster, N. J., Maloney, R. F., Seddon, P. J., Recio, M. R., Khan, M. S. I., & van Heezik, Y. (2021). Altitudinal distribution of the entire invasive small mammal guild in the eastern dry- land zone of New Zealand's southern Alps. Biological Invasions, 23, 1837–1857. https://doi.org/10.1007/s10530-021-02474-y Kosciuszko Wild Horse Heritage Bill 2018. Government of New South Wales. Retrieved May 14, 2021 from https://www. parliament.nsw.gov.au/bills/Pages/bill-details.aspx?pk=3518. Leigh, J. H., Wimbush, D. J., Wood, D. H., Holgate, M. D., Slee, A. V., Stanger, M. G., & Forrester, R. I. (1987). Effects of rabbit grazing and fire on a sub-alpine environment. I. Herbaceous and shrubby vegetation. Australian Journal of Botany, 35(4), 433–464. https://doi.org/10.1071/BT9870433 Fox, J., & Weisberg, S. (2019). An R companion to applied regression (3rd ed.). Sage. Gelman, A., Carlin, J. B., Stern, H. S., Dunson, D. B., Vehtari, A., & Rubin, D. B. (2013). Bayesian data analysis (3rd ed.). CRC Press. Lenth, R., Singmann, H., Love, J., Buerkner, P. & Herve, M. (2018). Emmeans: Estimated marginal means, aka least-squares means. R package. p. 3. Gelman, A., & Rubin, D. B. (1992). Inference from iterative simula- tion using multiple sequences. Statistical Science, 7, 457–511. Lu, X., Kelsey, K. C., Yan, Y., Sun, J., Wang, X., Cheng, G., & Neff, J. C. (2017). Effects of grazing on ecosystem structure and function of alpine grasslands in Qinghai-Tibetan plateau: A synthesis. Ecosphere, 8(1), e01656. https://doi.org/10.1002/ecs2. 1656 Girard, T. L., Girard, T. L., Bork, E. W., Neilsen, S. E., & Alexander, M. J. (2013). Landscape-scale factors affecting feral horse habitat use during summer within the Rocky Mountain foothills. Environmental Management, 51, 435–447. https://doi. org/10.1007/s00267-012-9987-2 Giroux, M., Dussault, C., Tremblay, J.-P., & Côté, S. D. (2016). Win- ter severity modulates the benefits of using a habitat temporally uncoupled from browsing. 10 of 11 10 of 11 HARTLEY ET AL. (Eds.), The ecology of browsing and grazing II (pp. 405–445). Springer Nature. evidence-based solutions. Ecological Management & Restora- tion, 20(1), 63–72. https://doi.org/10.1111/emr.12357 evidence-based solutions. Ecological Management & Restora- tion, 20(1), 63–72. https://doi.org/10.1111/emr.12357 Dyring, J. (1990). The impact of feral horses (Equus caballus) on sub- alpine and montane environments (p. 160). University of Canberra. Green, K. (2016). Adaptations of large marsupials to survival in winter snow cover: Locomotion and foraging. Canadian Jour- nal of Zoology, 94(11), 753–759. https://doi.org/10.1139/cjz- 2016-0097 Eldridge, D. J., Ding, J., & Travers, S. K. (2020). Feral horse activity reduces environmental quality in ecosystems globally. Biologi- cal Conservation, 241, 108367. https://doi.org/10.1016/j.biocon. 2019.108367 Green, K., Davis, N. E., & Robinson, W. A. (2014). Does diet con- strain the occupation of high elevations by macropods? A com- parison between Macropus rufogriseus and Wallabia bicolor. Australian Mammalogy, 36(2), 219–228. https://doi.org/10. 1071/AM14007 Eldridge, D. J., Travers, S. K., Val, J., Wang, J., Liu, H., Singh, B. K., & Delgado-Baquerizo, M. (2020). Grazing regu- lates the spatial heterogeneity of soil microbial communities within ecological networks. Ecosystems, 23, 932–942. https:// doi.org/10.1007/s10021-019-00448-9 Green, K., Davis, N. E., Robinson, W. A., McAuliffe, J., & Good, R. B. (2013). Diet selection by European hares (Lepus europaeus) in the alpine zone of the Snowy Mountains, Australia. European Journal of Wildlife Research, 59(5), 693–703. https://doi.org/10.1007/s10344-013-0723-x Eldridge, D. J., Travers, S. K., Val, J., Zaja, A., & Veblen, K. E. (2019). Horse activity is associated with degraded subalpine grassland structure and reduced habitat for a threatened rodent. Rangeland Ecology and Management, 72(3), 467–473. https://doi.org/10.1016/j.rama.2018.12.008 Green, K., & Osborne, W. (2012). Field guide to wildlife of the Australian snow-country (2nd ed.). New Holland Publishers. Forsyth, D. M., Barker, R. J., Morriss, G., & Scroggie, M. P. (2007). Modeling the relationship between fecal pellet indices and deer density. Journal of Wildlife Management, 71(3), 964–970. https://doi.org/10.2193/2005-695 Green, K., & Pickering, C. (2013). Limited effect of hare grazing and short-term climatic variations on the most common alpine vegetation community in the Snowy Mountains, Australia. Plant Ecology and Diversity, 6(3–4), 511–522. https://doi.org/10. 1080/17550874.2013.787466 Forsyth, D. M., McLeod, S. R., Scroggie, M. P., & White, M. D. (2009). Modelling the abundance of wildlife using field surveys and GIS: Non-native sambar deer (Cervus unicolor) in the Yarra ranges, South-Eastern Australia. Wildlife Research, 36(3), 231–241. https://doi.org/10.1071/WR08075 Hartig, F. & Lohse, L. (2020). DHARMa: Residual diagnostics for hierarchical (multi-level/mixed) regression models. https:// cran.r-project.org/package=DHARMa. REFERENCES Prins (Eds.), The ecology of browsing and grazing II (pp. 197–213). Springer Nature. Brown, D., Brown, D., Thomas, E., Herbert, K., & Primrose, K. (2016). Evaluating the effects of feral deer management on endangered alpine peatlands: The Alpine National Park deer control trial. Plant Protection Quarterly, 31(2), 63. Driscoll, D. A., Worboys, G. L., Allan, H., Banks, S. C., Beeton, N. J., Cherubin, R. C., Doherty, T. S., Finlayson, C. M., Green, K., Hartley, R., Hope, G., Johnson, C. N., Lintermans, M., Mackey, B., Paull, D. J., Pittock, J., Porfirio, L. L., Ritchie, E. G., Sato, C. F., … Williams, R. M. (2019). Impacts of feral horses in the Australian Alps and Bürkner, P. C. (2017). Brms: An R package for Bayesian multilevel models using Stan. Journal of Statistical Software, 80(1), 1–28. https://doi.org/10.18637/jss.v080.i01 11 of 11 HARTLEY ET AL. Milchunas, D., & Lauenroth, W. (1993). Quantitative effects of graz- ing on vegetation and soils over a global range of environments. Ecological Monographs, 63(4), 327–366. Park. Forest Ecology and Management, 181(1–2), 177–188. https:// doi.org/10.1016/S0378-1127(03)00131-2 Park. Forest Ecology and Management, 181(1–2), 177–188. https:// doi.org/10.1016/S0378-1127(03)00131-2 Sitters, J., Wubs, E., Bakker, E., Crowther, T., Adler, P., Bagchi, S., … Veen, G. (2020). Nutrient availability controls the impact of mammalian herbivores on soil carbon and nitrogen pools in grasslands. Global Change Biology, 26(4), 2060–2071. https:// doi.org/10.1111/gcb.15023 Mutze, G., Cooke, B., & Jennings, S. (2016). Density-dependent grazing impacts of introduced European rabbits and sympatric kangaroos on Australian native pastures. Biological Invasions, 18(8), 2365–2376. https://doi.org/10.1007/s10530-016-1168-4 Spake, R., Bellamy, C., Gill, R., Watts, K., Wilson, T., Ditchburn, B., & Eigenbrod, F. (2020). Forest damage by deer depends on cross-scale interactions between climate, deer den- sity and landscape structure. Journal of Applied Ecology, 57(7), 1376–1390. https://doi.org/10.1111/1365-2664.13622 Norris, K. A. (2018). A review of contemporary U.S. wild horse and burro management policies relative to desired management outcomes. Human-Wildlife Interactions, 12(1), 18–30. Nuñez, M. A., Bailey, J. K., & Schweitzer, J. A. (2010). Population, community and ecosystem effects of exotic herbivores: A grow- ing global concern. Biological Invasions, 12(2), 297–301. https:// doi.org/10.1007/s10530-009-9626-x Spear, D., & Chown, S. L. (2009). Non-indigenous ungulates as a threat to biodiversity. Journal of Zoology, 279(1), 1–17. https:// doi.org/10.1111/j.1469-7998.2009.00604.x Pakeman, R. J., Fielding, D. A., Everts, L., & Littlewood, N. A. (2019). Long-term impacts of changed grazing regimes on the vegetation of heterogeneous upland grasslands. Journal of Applied Ecology, 56(7), 1794–1805. https://doi.org/10.1111/ 1365-2664.13420 Tanentzap, A. J., Burrows, L. E., Lee, W. G., Nugent, G., Maxwell, J. M., & Coomes, D. A. (2009). Landscape-level vege- tation recovery from herbivory: Progress after four decades of invasive red deer control. Journal of Applied Ecology, 46, 1064– 1072. https://doi.org/10.2307/25623087 Pickering, C. M., Hill, W., Newsome, D., & Leung, Y.-F. (2010). Comparing hiking, mountain biking and horse riding impacts on vegetation and soils in Australia and the United States of America. Journal of Environmental Management, 91(3), 551–562. https://doi.org/10.1016/j.jenvman.2009.09.025 Triggs, B. (1996). Tracks, scats and other traces: A field guide to Australian mammals. Oxford University Press. UNESCO (2019). Kosciuszko Biosphere Reserve, Australia, United Nations Educational Scientific and Cultural Organization. Retrieved November 23, 2020 from https://en.unesco.org/ biosphere/aspac/kosciuszko. Pineiro, G., Paruelo, J. M., Oesterheld, M., & Jobbagy, E. G. (2010). Pathways of grazing effects on soil organic carbon and nitrogen. Rangeland Ecology and Management, 63(1), 109–119. https:// doi.org/10.2111/08-255.1 Vandegehuchte, M. L., Schütz, M., de Schaetzen, F., & Risch, A. C. (2017). Mammal-induced trophic cascades in invertebrate food webs are modulated by grazing intensity in subalpine grass- land. Journal of Animal Ecology, 86(6), 1434–1446. https://doi. org/10.1111/1365-2656.12744 Prior, K. M., Adams, D. C., Klepzig, K. D., & Hulcr, J. (2018). When does invasive species removal lead to ecological recovery? Implications for management success. Biological Invasions, 20(2), 267–283. https://doi.org/10.1007/s10530-017-1542-x Vehtari, A., Gelman, A. & Gabry, J. (2017a). Pareto smoothed importance sampling. 1–41. Pyšek, P., Hulme, P. E., Simberloff, D., Bacher, S., Blackburn, T. M., Carlton, J. T., Dawson, W., Essl, F., Foxcroft, L. C., Genovesi, P., Jeschke, J. M., Kühn, I., Liebhold, A. M., Mandrak, N. E., Meyerson, L. A., Pauchard, A., Pergl, J., Roy, H. E., Seebens, H., … Richardson, D. M. (2020). Scientists' warning on invasive alien species. Biological Reviews, 95(6), 1511–1534. https://doi.org/10.1111/brv.12627 Vehtari, A., Gelman, A., & Gabry, J. (2017b). Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC. Statistics and Computing, 27(5), 1413–1432. https://doi. org/10.1007/s11222-016-9696-4 de Villalobos, A. E., & Schwerdt, L. (2020). Seasonality of feral horse grazing and invasion of Pinus halepensis in grasslands of the austral Pampean Mountains (Argentina): Management con- siderations. Biological Invasions, 22(10), 2941–2955. https://doi. org/10.1007/s10530-020-02300-x R Core Team. (2020). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https:// www.r-project.org/ Williams, R., Papst, W., McDougall, K., & Al, E. (2014). Alpine eco- systems. In D. Lindenmayer, E. Burns, N. Thurgate, & A. Lowe (Eds.), Biodiversity and environmental change: Monitoring, chal- lenges and direction (pp. 167–212). CSIRO Publishing. Robertson, G., Wright, J., Brown, D., Yuen, K., & Tongway, D. (2019). An assessment of feral horse impacts on treeless drain- age lines in the Australian Alps. Ecological Management & Res- toration, 20(1), 21–30. https://doi.org/10.1111/emr.12359 Sala, O. E., Vivanco, L., & Flombaum, P. (2013). Grassland ecosys- tems. In Encyclopedia of biodiversity (2nd ed., pp. 1–7). Elsevier. 10 of 11 Ecosphere, 7(8), e01432. https://doi. org/10.1002/ecs2.1432 Matthews, A., & Spooner, P. G. (2014). Using a resource selection approach to predict the suitability of alpine habitats for a com- mon herbivore. Austral Ecology, 39(2), 170–180. https://doi.org/ 10.1111/aec.12053 McDougall, K. L., & Walsh, N. G. (2007). Treeless vegetation of the Australian Alps. Cunninghamia, 10(1), 1–57. Good, R., & Johnston, S. (2019). Rehabilitation and revegetation of the Kosciuszko summit area, following the removal of grazing – An historic review. Ecological Management and Restoration, 20(1), 13–20. https://doi.org/10.1111/emr.12363 McDougall, K. L., Walsh, N. G., & Wright, G. T. (2015). Recovery of treeless subalpine vegetation in Kosciuszko National Park after the landscape-scale fire of 2003. Australian Journal of Botany, 63, 597–607. https://doi.org/10.1071/BT14319 Gordon, I. J., & Prins, H. H. T. (2019). Browsers and grazers drive the dynamics of ecosystems. In I. J. Gordon & H. H. T. Prins SUPPORTING INFORMATION Sato, C. F., Schroder, M., Green, K., Michael, D. R., Osborne, W. S., & Lindenmayer, D. B. (2014). Managing ski resorts to improve biodiversity conservation: Australian reptiles as a case study. Ecological Management & Restoration, 15(2), 147–154. https://doi.org/10.1111/emr.12111 Additional supporting information may be found in the online version of the article at the publisher's website. How to cite this article: Hartley, R., Blanchard, W., Schroder, M., Lindenmayer, D. B., Sato, C., & Scheele, B. C. (2022). Exotic herbivores dominate Australian high-elevation grasslands. Conservation Science and Practice, 4(2), e601. https://doi.org/10. 1111/csp2.601 Schulz, M., Schroder, M., & Green, K. (2019). The occurrence of the broad-toothed rat Mastacomys fuscus in relation to feral horse impacts. Ecological Management & Restoration, 20(1), 31–36. https://doi.org/10.1111/emr.12360 Schütz, M., Risch, A. C., Leuzinger, E., Krüsi, B. O., & Achermann, G. (2003). Impact of herbivory by red deer (Cervus elaphus L.) on pat- terns and processes in subalpine grasslands in the Swiss National
https://openalex.org/W2769318807	https://lib.ossn.ru/jour/article/download/284/285	Russian	null	Role of heart rate in mechanisms of compensation and decompensation in CHF patients with sinus rhythm and atrial fibrillation and methods for safe and efficient control of heart rhythm. Part 2. Atrial fibrillation	Kardiologiâ	2,017	cc-by	14,334	DOI: 10.18087/cardio.2401 Мареев В. Ю.1, Мареев Ю. В.2, 3 1 – ФГБОУ ВО «МГУ им. М. В. Ломоносова», 119991, Москва, ГСП-1, Ленинские горы, д. 1 2 – ФГБУ «Национальный медицинский исследовательский центр профилактической медицины» МЗ РФ, 101990, Москва, Петроверигский пер., д. 10, стр. 3 3 – «Робертсоновский центр биостатистики». Университет Глазго, G12 8QQ, Великобритания, Глазго, Университетское Авеню, Здание Бойд Орр Роль ЧСС в механизмах компенсации и декомпенсации у больных с ХСН при синусовом ритме и фибрилляции предсердий и методы безопасного и эффективного контроля сердечного ритма. Часть 2. Фибрилляция предсердий Ключевые слова: хроническая сердечная недостаточность, частота сердечных сокращений, синусовый ритм, фибрилляция предсердий, контроль сердечного ритма Ссылка для цитирования: Мареев В. Ю., Мареев Ю. В. Роль ЧСС в механизмах компенсации и декомпенсации у больных с ХСН при синусовом ритме и фибрилляции предсердий и методы безопасного и эффективного контроля сердечного ритма. Часть 2. Фибрилляция предсердий. Кардиология 2017;57(S2):351–366 Резюме Обзор литературы посвящен роли ЧСС у пациентов с сочетанием сердечной недостаточности и фибрилляции предсердий и методам контроля ЧСС у этих пациентов. Mareev V. Yu.1, Mareev Yu. V.2, 3 1 – M. V. Lomonosov Moscow State University, Leninskie Gory 1, GSP-1, Moscow 119991 2 – National Medical Research Center for Preventive Medicine, Petroverigskiy lane 10, Bldg. 3, Moscow 101990 3 – Robertson Centre for Biostatistics – University of Glasgow Boyd Orr Building University Avenue, Glasgow, Great Britain. G12 8QQ Role of heart rate in mechanisms of compensation and decompensation in CHF patients with sinus rhythm and atrial fibrillation and methods for safe and efficient control of heart rhythm. Part 2. Atrial fibrillation Keywords: chronic heart failure, heart rate, sinus rhythm, atrial fibrillation, control of heart rhythm For citation: Mareev V. Yu., Mareev Yu. V. Role of heart rate in mechanisms of compensation and decompensation in CHF patients with sinus rhythm and atrial fibrillation and methods for safe and efficient control of heart rhythm. Part 2. Atrial fibrillation. Kardiologiia. 2017;57(S2):351–366 Summary Thi i f th l f h t t i ti t ith bi ti f h t f il d t i l fib ill ti d th d f Мареев В. Ю.1, Мареев Ю. В.2, 3 р р р р 3 – «Робертсоновский центр биостатистики». Университет Глазго, G12 8QQ, Великобритания, Глазго, Университетское Авеню, Здание This review focuses on the role of heart rate in patients with a combination of heart failure and atrial fibrillation and on methods for heart rate control in such patients. This review focuses on the role of heart rate in patients with a combination of heart failure and atrial fibrillation and on methods for heart rate control in such patients. Основными механизмами негативного влияния ФП на прогноз больных ХСН являются: нарушение гемо‑ динамики ЛЖ (вследствие частого и нерегулярного ритма), исчезновение предсердного гемодинамическо‑ го вклада и, наконец, возможное ухудшение миокарди‑ альной перфузии [11]. Также из‑за нарушения гемо‑ динамики левого предсердия возрастает риск тромбо‑ за его полости и тромбоэмболических осложнений. Важнейшая проблема эффективности и безопасности современной антикоагулянтной терапии в улучшении прогноза больных с ФП [12], особенно при ХСН, тре‑ бует специального рассмотрения и выходит за рамки данной статьи [13–15]. Role of heart rate in mechanisms of compensation and decompensation in CHF patients with sinus rhythm and atrial fibrillation and methods for safe and efficient control of heart rhythm. Part 2. Atrial fibrillation Keywords: chronic heart failure, heart rate, sinus rhythm, atrial fibrillation, control of heart rhythm For citation: Mareev V. Yu., Mareev Yu. V. Role of heart rate in mechanisms of compensation and decompensation in CHF patients with sinus rhythm and atrial fibrillation and methods for safe and efficient control of heart rhythm. Part 2. Atrial fibrillation. Kardiologiia. 2017;57(S2):351–366 For citation: Mareev V. Yu., Mareev Yu. V. Role of heart rate in mechanisms of compensation and decompensation in CHF patients with sinus rhythm and atrial fibrillation and methods for safe and efficient control of heart rhythm. Part 2. Atrial fibrillation. Kardiologiia. 2017;57(S2):351–366 Ключевые слова: хроническая сердечная недостаточность, частота сердечных сокращений, синусовый ритм, фибрилляция предсердий, контроль сердечного ритма Ссылка для цитирования: Мареев В. Ю., Мареев Ю. В. Роль ЧСС в механизмах компенсации и декомпенсации у больных с ХСН при синусовом ритме и фибрилляции предсердий и методы безопасного и эффективного контроля сердечного ритма. Часть 2. Фибрилляция предсердий. Кардиология 2017;57(S2):351–366 ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § ОБЗОРЫ § ОБЗОРЫ § DOI: 10.18087/cardio.2401 Резюме Обзор литературы посвящен роли ЧСС у пациентов с сочетанием сердечной недостаточности и фибрилляции предсердий и методам контроля ЧСС у этих пациентов. Mareev V. Yu.1, Mareev Yu. V.2, 3 1 – M. V. Lomonosov Moscow State University, Leninskie Gory 1, GSP-1, Moscow 119991 . V. Lomonosov Moscow State University, Leninskie Gory 1, GSP-1, Moscow 119991 ational Medical Research Center for Preventive Medicine, Petroverigskiy lane 10, Bldg. 3, Moscow 101990 M. V. Lomonosov Moscow State University, Leninskie Gory 1, GSP 1, Moscow 119991 2 – National Medical Research Center for Preventive Medicine, Petroverigskiy lane 10, Bldg. 3, Moscow 1 2 – National Medical Research Center for Preventive Medicine, Petroverigskiy lane 10, Bldg. 3, , g y , g , obertson Centre for Biostatistics – University of Glasgow Boyd Orr Building University Avenue, Glasgow, G Role of heart rate in mechanisms of compensation and decompensation in CHF patients with sinus rhythm and atrial fibrillation and methods for safe and efficient control of heart rhythm. Part 2. Atrial fibrillation Роль ЧСС в механизмах компенсации и декомпенсации у больных с ХСН при синусовом ритме и фибрилляции предсердий и методы безопасного и эффективного контроля сердечного ритма. Часть 2. Фибрилляция предсердий Ключевые слова: хроническая сердечная недостаточность, частота сердечных сокращений, синусовый ритм, фибрилляция предсердий, контроль сердечного ритма Введение В предыдущей статье мы начали разговор о роли ЧСС у пациентов с ХСН [1]. В прошлый раз акцент был сделан на синусовом ритме, а сейчас речь пойдет о паци‑ ентах с фибрилляцией предсердий (ФП). В п у У пациентов с ХСН ФП является самой распростра‑ ненной аритмией. В среднем у 30–40 % пациентов с ХСН основным ритмом сердца является ФП, а среди пациен‑ тов с III–IV ФК число пациентов с ФП, по данным ряда работ, может достигать 50 % [2–7]. Еще большое число пациентов может иметь пароксизмальную форму ФП [8]. При этом пациенты с ФП имеют выраженные особенно‑ сти и требует специфических подходов к лечению [9, 10]. 351 ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § А вот вопрос о целесообразности и вариантах контро‑ ля ЧСС у больных с ХСН и постоянной формой ФП мы проанализируем подробно, т. к. гемодинамические нару‑ шения и потенциальное ухудшение миокардиальной пер‑ фузии напрямую связаны с ЧСС и регулярностью ритма у пациентов с ХСН и постоянной формой ФП. кардиологии (2005–2008 гг.), получавших оптимальную современную терапию ХСН, как и в обсервационном исследовании в Великобритании [22, 23]. Анализ данных Фрамингемского исследования проде‑ монстрировал, что предшествующая ХСН отрицательно влияет на прогноз больных с ФП, в то время как нали‑ чие ФП не ассоциируется с увеличенным риском смерти пациентов с ХСН [24]. Более того, ряд анализов, прове‑ денных у пациентов, ожидающих трансплантации сердца, также не выявил увеличения риска смерти у пациентов с ФП в сравнении с синусовым ритмом [25]. Мета-анализ исследований, сравнивших прогноз пациентов с ФП и синусовым ритмом, имеющих ХСН, показал, что ФП оказывает негативное влияние на прогноз пациентов с легкой и умеренной декомпенсацией, но не с тяжелой стадией (ФК ХСН) [26]. При этом в одной из работ, включенных в мета-анализ, проводилось определение давления заклинивания легочной артерии (ДЗЛА), кото‑ рое показало, что ФП оказывает негативный эффект на прогноз при уровне ДЗЛА менее 16 мм рт. ст. (уме‑ ренная декомпенсация) и не оказывало такого влияния при более высоком ДЗЛА (тяжелая декомпенсация) [27]. Одним из механизмов того, что ФП не оказывает нега‑ тивного влияния на прогноз у пациентов с тяжелой ХСН, авторы предположили избыточную активность симпато‑ адреналовой системы (САС), сопровождающуюся уве‑ личением ЧСС как при синусовом ритме, так и при ФП, при которой тахикардия становится более «ритмичной» (с уменьшением дефицита пульса). ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § терапии (AFFIRM [28], RACE [29, 30]). Причем специаль‑ ное проспективное исследование проводилось и у пациен‑ тов с систолической ХСН и ФП, в этом случае попытки восстановления синусового ритма оказались неудачными и не влияли на заболеваемость и смертность [31]. Однако дискуссии об эффективности восстановления синусового ритма у пациентов с ХСН не стихают. Часть исследователей отмечают, что традиционно применяемые для удержания синусового ритма антиаритмические пре‑ параты (у пациентов с ХСН в первую очередь – амиода‑ рон) сами могут оказывать негативный эффект на состоя‑ ние пациентов с выраженной ХСН [32]. Поэтому в настоя‑ щее время проводится оценка эффективности катетерной аблации левого предсердия с целью удержания синусового ритма у пациентов с ФП и сопутствующей ХСН. В неболь‑ шом исследование AATAC-AF (203 пациента) было пока‑ зано бóльшее удержание синусового ритма (70 % против 34 %), что сопровождалось лучшей 2‑летней выживаемо‑ стью и меньшим числом госпитализаций в группе кате‑ терной аблации в сравнении с приемом амиодарона [33]. В настоящее время проводится оценка безопасности и эффективности катетерной аблации у больных с ХСН и ФП в более крупном (n=420) исследовании CASTLE- HF [34]. Сходные результаты были получены в пилотном проекте CABANA (60 больных с сердечно-сосудистой патологией, включающей ХСН) – 65 % больных без ФП в группе катетерной аблации против 41 % в группе анти‑ аритмического лечения [35]. В настоящее время закончен набор более 2 200 боль‑ ных в основное исследование CABANA (ClinicalTrials.gov Identifier: NCT00911508). Но даже удачное завершение указанных исследований не снимает с повестки дня мето‑ ды медикаментозного контроля ритма сердца у многих пациентов с ХСН, у которых проведение катетерной абла‑ ции затруднено, удержание синусового ритма невозмож‑ но (лучшие результаты 65–70 %) и требуется эффективное лечение самой ХСН в условиях сохраняющейся ФП. у прогноз пациентов с ХСН? Первый вопрос, который может выглядеть странным: ухудшает ли наличие ФП прогноз больных с систоличе‑ ской ХСН? Тот факт, что встречаемость ФП увеличивает‑ ся с прогрессированием тяжести заболевания (ФК ХСН) может означать как то, что ФП ухудшает течение ХСН, так и то, что возникновение ФП один из маркеров про‑ грессирования декомпенсации. В целом ряде исследований (SOLVD [16], CHARM [17], VALIANT [18]) наличие ФП достоверно ухудша‑ ло прогноз ХСН, в то время как в других (V-HeFT I и II [19], PRIME II [20], COMET [21]) различий в прогнозе пациентов с ХСН в зависимости от характера основно‑ го ритма выявлено не было. Данные обсервационных работ также говорят в пользу того, что сам факт наличия ФП не ухудшает прогноз пациентов с ХСН. Так, по дан‑ ным ретроспективного анализа базы данных Отдела Заболеваний Миокарда и Сердечной Недостаточности (ОЗМиСН) НИИ кардиологии (1976–2005 гг.), не было выявлено различий в прогнозе пациентов с ХСН III– IV ФК с синусовым ритмом и ФП, причем достоверные различия в прогнозе больных с синусовым ритмом и ФП отсутствовали как при низкой, так и при относитель‑ но сохраненной ФВ ЛЖ (рис. 1, оригинальные данные, публикуемые впервые). Такой же результат был получен при анализе пациентов, лечившихся в ОЗМиСН НИИ Кроме того, сравнительные исследования не доказали улучшения прогноза больных с постоянной формой ФП при попытках восстановления и последующего удержания синусового ритма в сравнении с контролем ритма сердца, естественно, на фоне рекомендованной антикоагулянтной 352 ISSN 0022-9040. Кардиология. 2017;57(S2) Выживаемость ФВ ЛЖ≤34 % (n=400) 1,0 0,9 0,7 0,5 Дни 2500 0 500 1000 1500 2000 0,8 0,6 0,4 0,3 Фибрилляция предсердий (n=221) Синусовый ритм (n=179) p=0,898 Б Выживаемость ФВ ЛЖ>34 % (n=437) p=0,162 1,0 0,9 0,7 0,5 0,3 Дни 2500 0 500 1000 1500 2000 0,8 0,6 0,4 0,2 А Фибрилляция предсердий (n=128) Синусовый ритм (n=309) Рисунок 1. Влияние фибрилляции предсердий на прогноз пациентов с клинически выраженной ХСН III–IV ФК Выживаемость ФВ ЛЖ≤34 % (n=400) 1,0 0,9 0,7 0,5 Дни 2500 0 500 1000 1500 2000 0,8 0,6 0,4 0,3 Фибрилляция предсердий (n=221) Синусовый ритм (n=179) p=0,898 Б Выживаемость ФВ ЛЖ>34 % (n=437) p=0,162 1,0 0,9 0,7 0,5 0,3 Дни 2500 0 500 1000 1500 2000 0,8 0,6 0,4 0,2 А Фибрилляция предсердий (n=128) Синусовый ритм (n=309) Рисунок 1. Влияние фибрилляции предсердий на прогноз пациентов с клинически выраженной ХСН III–IV ФК ISSN 0022-9040. Кардиология. 2017;57(S2) 352 ОБЗОРЫ § Долгое время считалось, что у пациентов с ФП нуж‑ но стремиться к ЧСС менее 80 уд. / мин [37], но с появ‑ лением данных о том, что «порог безопасности» может быть выше [38], было проведено исследование RACE II (The Rate Control Efficacy in Permanent Atrial Fibrillation: a Comparison between Lenient versus Strict Rate Control II) [39]. В этом исследовании сравнили тактику «строгого» контроля ЧСС (ЧСС покоя ≤80 уд. / мин и ЧСС во время 6-мин теста ходьбы ≤110 уд. / мин) и «нестрогого» (ЧСС покоя <110 уд. / мин) у 614 пациентов с ФП (не обязатель‑ но при наличии ХСН). Реальная достигнутая в процессе лечения средняя ЧСС в группе «строгого» контроля при длительном наблюдении составляла 75–76 уд. / мин в сравнении с 84–85 уд. / мин (разница около 10 уд. / мин) в группе «нестрогого» контроля. Главным выводом исследования было отсутствие достоверных различий в частоте сердечно-сосудистой заболеваемости и смерт‑ ности в группах различного контроля ЧСС у пациентов с ФП, в том числе и при сопутствующей ХСН. ность» или степень регулярности ритма на протяжении 24 часов наблюдения и его вариабельность в течение все‑ го периода наблюдения. Кроме того, у пациентов в чет‑ вертой квартили имело место снижение средней ЧСС за 24 часа в течение первых 4 лет наблюдения (со 102 до 83 уд. / мин), в отличие от пациентов с исходно невысо‑ кой ЧСС, которая достоверно не менялась на протя‑ жении наблюдения (1 квартиль). Эти диспропорции ISSN 0022-9040. Кардиология. 2017;57(S2) жении наблюдения (1 квартиль). Эти диспропорции Пропорция выживших по медиане ЧСС (92 уд./мин) 1,0 0,9 0,6 0,4 Время (дни) 2500 0 500 1000 1500 2000 0,8 0,7 0,5 0,3 OP=0,932; p=0,590 А ЧСС>92 уд./мин (n=254) ЧСС≤92 уд./мин (n=240) Пропорция выживших по 25% квартилю ЧСС (80 уд./мин) 1,0 0,9 0,6 0,4 Время (дни) 2500 0 500 1000 1500 2000 0,8 0,7 0,5 0,3 OP=0,970; p=0,837 Б ЧСС>80 уд./мин (n=351) ЧСС≤80 уд./мин (n=143) Пропорция выживших по 75% квартилю ЧСС (105 уд./мин) 1,0 0,9 0,6 0,4 Время (дни) 2500 0 500 1000 1500 2000 0,8 0,7 0,5 0,3 OP=0,771; p=0,082 ЧСС >105 уд./мин (n=147) ЧСС ≤105 уд./мин (n=347) В Рисунок 2. ОБЗОРЫ § Выживаемость пациентов с систолической ХСН и ФП в зависимости от ЧСС Пропорция выживших по медиане ЧСС (92 уд./мин) 1,0 0,9 0,6 0,4 Время (дни) 2500 0 500 1000 1500 2000 0,8 0,7 0,5 0,3 OP=0,932; p=0,590 А ЧСС>92 уд./мин (n=254) ЧСС≤92 уд./мин (n=240) по медиане ЧСС (92 уд./мин) по медиане ЧСС (92 уд./мин) При сравнении риска сердечно-сосудистых осложне‑ ний у пациентов, у которых «строгий» контроль достиг цели (ЧСС <80 уд. / мин), с риском у пациентов «нестро‑ гого» контроля ЧСС и у пациентов «строгого» контро‑ ля, у которых не были достигнуты целевые показатели ЧСС (>80 уд. / мин), не обнаружено разницы в исходах между группами [40]. А дополнительный анализ резуль‑ татов RACE II не выявил и различий в процессах ремоде‑ лирования предсердий и желудочков у пациентов «стро‑ гого» и «нестрогого» контроля ЧСС [41]. Пропорция выживших по 25% квартилю ЧСС (80 уд./мин) 1,0 0,9 0,6 0,4 Время (дни) 2500 0 500 1000 1500 2000 0,8 0,7 0,5 0,3 OP=0,970; p=0,837 Б ЧСС>80 уд./мин (n=351) ЧСС≤80 уд./мин (n=143) Пропорция выживших р р Учитывая, что офисный контроль ЧСС, особенно при ФП, представляет известные трудности и подвержен влиянию многих дополнительных факторов, наиболее демонстративным может быть анализ при учете средне‑ суточного показателя ЧСС по данным Холтеровского монитора ЭКГ (ХМЭКГ). Такие результаты были полу‑ чены в относительно небольшом по объему (252 пациен‑ тов с ФП без клиники ХСН) исследовании, проведенном в Японии [42]. В этой работе был проведен анализ связи ЧСС по данным 24‑часового ХМ ЭКГ и исходов (сердеч‑ но-сосудистая смерть и / или госпитализация по поводу ХСН) у пациентов с постоянной формой ФП. Работа показала обратную зависимость: чем меньше ЧСС у лиц с ФП, тем большим был риск сердечно-сосудистых ослож‑ нений в течение 7 лет наблюдения. При средней ЧСС 62 уд. / мин (1 квартиль) 23,9 % умерли от ХСН или умер‑ ли внезапно (внезапная сердечная смерть – ВСС), или госпитализировались с обострением ХСН при сред‑ ней ЧСС 76 уд. / мин (2 квартиль) – 12,2 %, при средней ЧСС 85 уд. / мин (3 квартиль) – 19,4 % и при средней ЧСС 102 уд. / мин (4 квартиль) – только 7,2 %. у больных с ФП и ХСН Учитывая, что удержание синусового ритма, по дан‑ ным завершенных исследований, не лучше (по влиянию на заболеваемость и смертность) контроля ЧСС при ФП у больных с ХСН, особенно при увеличенных размерах левого предсердия, тактика контроля ЧСС часто исполь‑ зуется у таких пациентов [36]. Возникает главный вопрос: какая ЧСС оптимальна для пациентов с клинически выраженной ХСН и ФП с точки зрения достижения компенсации и возможного улучшения прогноза? Ответив на него, станет понятно, к какой ЧСС нужно стремиться при лечении больных с ХСН и ФП. ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § Следует заметить, что показатель среднесуточной ЧСС у лиц с ФП учитывает не только истинную ЧСС, но и «стабиль‑ Пропорция выживших по 75% квартилю ЧСС (105 уд./мин) 1,0 0,9 0,6 0,4 Время (дни) 2500 0 500 1000 1500 2000 0,8 0,7 0,5 0,3 OP=0,771; p=0,082 ЧСС >105 уд./мин (n=147) ЧСС ≤105 уд./мин (n=347) В Пропорция выживших Рисунок 2. Выживаемость пациентов с систолической ХСН и ФП в зависимости от ЧСС 354 ISSN 0022-9040. Кардиология. 2017;57(S2) ISSN 0022-9040. Кардиология. 2017;57(S2) ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § могут частично объяснить кажущиеся неожиданными результаты. Если же посмотреть на результаты исследо‑ вания в целом, то мы увидим ту же закономерность, что и в исследовании RACE II – больные с ФП и ЧСС более 80 уд. / мин имеют не худший (если и не лучший) прогноз, чем пациенты с ЧСС менее 70 уд. / мин. Данные вышеприведенных работ касаются пациен‑ тов, не имевших исходно ХСН. Но что мы знаем о паци‑ ентах с сочетанием ХСН и ФП? Какая ЧСС оптимальна у этой группы больных? Как мы отмечали в предыдущей статье [1], у пациентов с ХСН на фоне синусового рит‑ ма оптимальной ЧСС является диапазон 70–75 уд. / мин. Но данные ряда исследований показывают, что перенос данных пациентов с ХСН и синусовым ритмом на паци‑ ентов с ХСН и ФП может быть некорректным. Так, ана‑ лиз длительной выживаемости пациентов, госпитализи‑ рованных в ОЗМиСН НИИ кардиологии им. Мясникова за 30 лет (1977–2005 годы), показал, что большая ЧСС достоверно ухудшает прогноз больных с ХСН при сину‑ совом ритме, но не при ФП [1]. Необходимо отметить, что не было выявлено разницы в выживаемости больных с ХСН II–IV ФК и постоянной формой ФП при разделении пациентов ни по медиане ЧСС (92 уд. / мин), ни по низшей (80 уд. / мин), ни по выс‑ шей квартилям (105 уд. / мин) (рис. 2А, 2Б и 2В соответ‑ ственно), хотя в последнем случае наблюдалась тенден‑ ция к лучшему прогнозу у пациентов с бóльшей ЧСС. Сходные данные были получены и в крупном рандо‑ мизированном контролируемом исследовании (РКИ) CHARM, в которое включались как пациенты с низ‑ кой ФВ ЛЖ (СНнФВ), так и с промежуточной ФВ ЛЖ (СНпФВ) и сохраненной ФВ ЛЖ (СНсФВ) [6]. В толь‑ ко что опубликованном анализе пациентов, включенных в 11 основных РКИ, изучавших применение β-АБ у паци‑ ентов с СНнФВ (14 166 пациентов с синусовым ритмом и 3 065 пациентов с ФП) [43]. Во всех этих работах отме‑ чалась тенденция, хотя и не достигшая достоверности, к тому, что пациенты с более высокой ЧСС при ХСН, про‑ текающей с ФП, имели лучшую выживаемость, чем паци‑ енты с низкой ЧСС. Крупный Шведский регистр также показал, что у паци‑ ентов с ФП и СНнФВ нет связи между значениями ЧСС и риском смерти при ЧСС менее 100 уд. / мин, и только у пациентов, имеющих ЧСС >100 уд. / мин, повышен риск смерти в сравнении с пациентами с ЧСС ≤60 уд. / мин [44]. ОБЗОРЫ § Во-первых, в верхнюю квартиль пациентов входило большое число пациентов с ЧСС значимо выше 100 уд. / мин. Во-вторых, дополнительный анализ показал, что повышение риска смерти отмечалось у пациентов с ФВ ЛЖ >40 % (относительный риск (ОР) =1,08 при увеличении ЧСС на 10 уд. / мин; p<0,001). В то же время при ФВ ЛЖ ≤40 % ЧСС ≥75 уд. / мин не ассо‑ циировалась с повышением риска смерти и при этом была связана с более низким риском комбинированной точки смерть и госпитализация из‑за ХСН (ОР=0,95; 95 % ДИ: 0,91–0,99 при увеличении ЧСС на 10 уд. / мин; p=0,018). Нами было проведено небольшое исследование с уча‑ стием 89 больных с ХСН с постоянной формой ФП на фоне оптимальной терапии. ЧСС, позволяющая опти‑ мально разделить пациентов по исходам, была опреде‑ лена при помощи ROC-анализа. Исследование показа‑ ло, что пациенты с ЧСС <75 уд. / мин при поступлении (медиана – 67 уд. / мин [интерквантильный размах (ИР) 62; 72]) имели более высокий риск смерти, чем паци‑ енты с ЧСС ≥75 уд. / мин (медиана 90 уд. / мин [ИР 80; 105]). И после лечения эта закономерность сохрани‑ лась. Достоверно лучший прогноз отмечен у пациентов с ЧСС ≥70 уд. / мин (медиана 78 уд. / мин [ИР 74; 83]), то есть ЧСС снизилась в процессе лечения на 12 уд. / мин, в сравнении с теми, у кого достигнута ЧСС <70 уд. / мин (медиана – 67 уд. / мин [ИР 64;70]). Несмотря на разни‑ цу ЧСС, равную 11 уд. / мин, относительный риск смерти для пациентов с ХСН с постоянной формой ФП и ЧСС <70 уд. / мин составил 1,91 (р=0,0085). Следует отметить, что есть работы, показавшие, что высокая ЧСС при ФП не только не увеличивает риск осложнений и смерти у пациентов с ХСН, но и может свидетельствовать о более благоприятном прогнозе. Так, например, в анализе исследования PRIME II (Second Prospective Randomised Study of Ibopamine on Mortality and Efficacy) при проведении многофакторного ана‑ лиза было продемонстрировано, что пациенты с ФП и низкой ЧСС (медиана – 72 уд. / мин) имели достовер‑ но более высокий риск смерти (ОР=2,9), чем пациенты с высокой ЧСС (медиана – 90 уд. / мин) [47]. ОБЗОРЫ § Также следует отметить две работы, посвященные программе американских госпиталей по практическо‑ му соблюдению рекомендаций при ХСН (Get With The Guidelines-HeartFailure = GWTG-HF). В первом из этих исследований проводился анализ связи ЧСС при госпитализации и риска внутрибольничной смерти у пациентов с ХСН в зависимости от характера основного Крупный Шведский регистр также показал, что у паци‑ ентов с ФП и СНнФВ нет связи между значениями ЧСС и риском смерти при ЧСС менее 100 уд. / мин, и только у пациентов, имеющих ЧСС >100 уд. / мин, повышен риск смерти в сравнении с пациентами с ЧСС ≤60 уд. / мин [44]. Также следует отметить две работы, посвященные программе американских госпиталей по практическо‑ му соблюдению рекомендаций при ХСН (Get With The Guidelines-HeartFailure = GWTG-HF). В первом из этих исследований проводился анализ связи ЧСС при госпитализации и риска внутрибольничной смерти у пациентов с ХСН в зависимости от характера основного ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § Вероятность остаться в живых ЧСС при поступлении 1,0 0,9 0,6 0,4 Дни наблюдения 500 1000 1500 2000 2500 3000 0,8 0,7 0,5 0,2 А 0,3 0 p=0,0023 ЧСС>75 уд./мин Медиана=90 (80;105) ЧСС≤75 уд./мин Медиана=67 (62;72) ОБЗОРЫ § Вероятность остаться в живых ЧСС при выписке (оптимальная терапия) 1,0 0,9 0,6 0,4 Дни наблюдения 500 1000 1500 2000 2500 3000 0,8 0,7 0,5 0,2 Б 0,3 0 p=0,0085 ЧСС>70 уд./мин Медиана=78 (74;83) ЧСС≤70 уд./мин Медиана=67 (64;70) Вероятность остаться в живых ЧСС при поступлении 1,0 0,9 0,6 0,4 Дни наблюдения 500 1000 1500 2000 2500 3000 0,8 0,7 0,5 0,2 А 0,3 0 p=0,0023 ЧСС>75 уд./мин Медиана=90 (80;105) ЧСС≤75 уд./мин Медиана=67 (62;72) Рисунок 3. Взаимосвязь частоты сокращения желудочков у больных с ХСН и ФП и смертностью (когорта ОЗМ и СН S2002–2007 гг., n=89). 3А – ЧСС при поступлении. 3Б – ЧСС при выписке. Вероятность остаться в живых ЧСС при выписке (оптимальная терапия) 1,0 0,9 0,6 0,4 Дни наблюдения 500 1000 1500 2000 2500 3000 0,8 0,7 0,5 0,2 Б 0,3 0 p=0,0085 ЧСС>70 уд./мин Медиана=78 (74;83) ЧСС≤70 уд./мин Медиана=67 (64;70) Рисунок 3. Взаимосвязь частоты сокращения желудочков у больных с ХСН и ФП и смертностью (когорта ОЗМ и СН S2002–2007 гг., n=89). 3А – ЧСС при поступлении 3Б – ЧСС при выписке 3А – ЧСС при поступлении. 3Б – ЧСС при выписке. ЧСС. ОБЗОРЫ § Это свидетельствует о том, что высокая ЧСС при ФП у пациентов с ХСН не связана напрямую со сте‑ пенью гиперактивации САС и имеются другие факторы, определяющие характер ритма и его частоту у деком‑ пенсированных больных. ритма [45]. Минимальный риск смерти регистрировался при синусовом ритме с ЧСС 66–70 уд. / мин, а у пациен‑ тов с ФП при ЧСС 78–80 уд. / мин (т.е разница составляет 10–12 уд. / мин). уд ) Второй анализ был посвящен изучению связи между достигнутой ЧСС при выписке и риском СС осложнений (смерти и госпитализации из‑за ХСН) в течение 12 мес. [46]. Этот анализ, в отличие от предыдущих работ, пока‑ зал, что ЧСС более 75 уд. / мин при ФП связана с умерен‑ ным увеличением риска смерти пациентов с ХСН. Но сле‑ дует отметить 2 момента. Во-первых, в верхнюю квартиль пациентов входило большое число пациентов с ЧСС значимо выше 100 уд. / мин. Во-вторых, дополнительный анализ показал, что повышение риска смерти отмечалось у пациентов с ФВ ЛЖ >40 % (относительный риск (ОР) =1,08 при увеличении ЧСС на 10 уд. / мин; p<0,001). В то же время при ФВ ЛЖ ≤40 % ЧСС ≥75 уд. / мин не ассо‑ циировалась с повышением риска смерти и при этом была связана с более низким риском комбинированной точки смерть и госпитализация из‑за ХСН (ОР=0,95; 95 % ДИ: 0,91–0,99 при увеличении ЧСС на 10 уд. / мин; p=0,018). Следует отметить, что есть работы, показавшие, что высокая ЧСС при ФП не только не увеличивает риск осложнений и смерти у пациентов с ХСН, но и может свидетельствовать о более благоприятном прогнозе. Так, например, в анализе исследования PRIME II (Second Prospective Randomised Study of Ibopamine on Mortality and Efficacy) при проведении многофакторного ана‑ лиза было продемонстрировано, что пациенты с ФП и низкой ЧСС (медиана – 72 уд. / мин) имели достовер‑ но более высокий риск смерти (ОР=2,9), чем пациенты с высокой ЧСС (медиана – 90 уд. / мин) [47]. Важным является и тот факт, что в исследовании PRIME II про‑ водился анализ концентрации норадреналина, которая значимо не отличалась в группах с высокой и низкой Второй анализ был посвящен изучению связи между достигнутой ЧСС при выписке и риском СС осложнений (смерти и госпитализации из‑за ХСН) в течение 12 мес. [46]. Этот анализ, в отличие от предыдущих работ, пока‑ зал, что ЧСС более 75 уд. / мин при ФП связана с умерен‑ ным увеличением риска смерти пациентов с ХСН. Но сле‑ дует отметить 2 момента. ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § Важным является и тот факт, что в исследовании PRIME II про‑ водился анализ концентрации норадреналина, которая значимо не отличалась в группах с высокой и низкой Важно, что группы пациентов не отличались по воз‑ расту, соотношению мужчин и женщин, ФК по NYHA, уровню АД, этиологии XCН (за исключением рубцовых изменений на ЭКГ, которые встречались чаще у пациен‑ тов с меньшей ЧСС; р=0,045), длительности QRS, числу пациентов с блокадой левой ножки пучка Гиса (БЛНПГ) и медикаментозной терапии, включающей ингибито‑ ры АПФ, β-АБ и антагонисты минералокортикоидных рецепторов. Сложность анализа исследований, изучавших роль ЧСС у пациентов с ФП и ХСН, заключается в том, что большая часть – это ретроспективные анализы уже 356 ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § 89 пациентов Ретроспективный ана- лиз выживаемости б-х, госпитализированных с диагнозом ХСН Среднее время наблю- дения – 7 лет При поступлении Низкая ЧСС: ЧСС <75 уд. / мин Медиана – 67 уд. / мин [62; 72] Высокая ЧСС: ЧСС ≥75 уд. / мин Медиана – 90 уд. / мин [80; 105] При выписке Низкая ЧСС: ЧСС <70 уд. / мин Медиана – 67 уд. / мин [64; 70] Высокая ЧСС: ЧСС ≥70 уд. / мин Медиана – 78 уд. / мин [74; 83] Выживаемость в группе пациентов с высокой ЧСС была выше как при анализе пациентов при поступлении, так и при выписке Пациенты с любой ФВ ЛЖ Средняя ФВ ЛЖ 33 % [23; 46] II–III ФК по NYHA CHARM [6] 1148 пациентов Субанализ РКИ Время наблюдения – 2,8–3,4 года Пациенты были разделены на три терцили (Т): 1 Т=Медиана – 64; [60–68] 2 Т=Медиана – 76 [72–80] 3 Т=Медиана – 90 [86–100] Смерть от всех причин T1 ОР=1,00 T2 ОР=1,13; ДИ: 0,89–1,43 T3 ОР=0,89; ДИ: 0,69–1,15 Сердечно-сосудистая смерть или госпитализация из‑за ХСН T1 ОР=1,00 T2 ОР=0,96; ДИ: 0,78–1,19 T3 ОР=0,79; ДИ: 0,63–0,99 Любая ФВ ЛЖ Средняя ФВ ЛЖ 37,5 % II–IV ФК по NYHA Достоверность пропала по- сле учета различий в характе- ристиках больных Hull LifeLab [23] 488 пациентов Анализ пациентов, включенных в програм- му по изучению ХСН Время наблюдения – 3,6 года (1,7–6,9 лет). ЧСС при поступлении (деление на квартили) Q1 <69 уд. / мин; Q2 69–81 уд. / мин Q3 82–98 уд. / мин; Q4 >98 уд. / мин ЧСС во время визита через год Q1 <62 уд. / мин; Q2 62–72 уд. / мин Q3 73–82 уд. / мин; Q4 >82 уд. / мин Связь ЧСС при поступлении и прогноза Q1 референс Q2 ОР=1,01; ДИ: 0,69–1,47; p=0,97 Q3 ОР=0,94; ДИ: 0,64–1,38; p=0,75 Q4 ОР=0,68; ДИ: 0,45–1,03; p=0,07 Связь ЧСС во время контрольных визитов и прогноза Q1 референс Q2 ОР=0,99; ДИ: 0,56–1,72; p=0,96 Q3 ОР=0,71; ДИ: 0,39–1,27; p=0,24 Q4 ОР=1,07; ДИ: 0,60–1,90; p=0,82 Пациенты с ФВ ЛЖ <50 % I–IV ФК по NYHA По однофакторному анали- зу – снижение риска смерти на 7 % при повышении ЧСС на 10 уд. / мин (p=0,02). Достоверность пропала по- сле учета различий в характе- ристиках пациентов Исследование, число больных с ФП, тип исследования, длительность Результаты (относительный риск – ОР) и 95 % доверительный интервал (ДИ) ОБЗОРЫ § ОБЗОРЫ § ОБЗОРЫ § проведенных работ и / или наблюдательных исследова‑ ний. Единственный анализ, являющийся post-hoc анали‑ зом РКИ, где сравнивалась тактика менее строго и более строгого контролей ЧСС – это дополнительный анализ исходов 287 пациентов с ХСН, участвовавших в исследо‑ вании RACE II. По данным этого анализа, риск наступле‑ ния первичной конечной точки составил 18,2 % в группе строгого контроля ЧСС и 15 % в группе нестрогого кон‑ троля ЧСС (p=0,53) [48]. Интересно, что исходная раз‑ ница ЧСС в группе нестрогого контроля – 94±11 уд. / мин и строгого контроля – 76±12 уд. / мин (Δ=18 уд. / мин; р<0,001) уменьшилась к концу наблюдения (через 3 года). Достигнутая ЧСС составила 85±14 уд. / мин в группе нестрого контроля и 76±15 уд. / мин там, где контроль ЧСС был приоритетом (Δ=9 уд. / мин; р<0,001). Эти данные показывают, что на фоне лечения ЧСС снизилась в большей степени в группе нестрого контроля, что, воз‑ можно, связано с тем, что лечащие врачи традиционно пытались снижать ЧСС больным с тахикардией. В ито‑ ге разница в достигнутой ЧСС у больных с ХСН и ФП в процессе лечения снизилась вдвое. В таблице 1 мы суммировали основные работы, прове‑ денные в этой области, включая проспективные исследо‑ вания, ретроспективные анализы регистров и баз данных исследований и госпиталей, одноцентровые многолетние наблюдения, что позволит проанализировать имеющийся массив данных более подробно. Таким образом, анализ многочисленных РКИ и баз данных показывает примерно одинаковую закономер‑ ность – оптимальное число желудочковых сокраще‑ Таблица 1. Результаты исследований, изучавших оптимальную ЧСС у пациентов с ХСН и ФП (анализы влияния ЧСС при ФП на прогноз больных ХСН) (анализы влияния ЧСС при ФП на прогноз больных ХСН) Исследование, число больных с ФП, тип исследования, длительность Анализированные группы Результаты (относительный риск – ОР) и 95 % доверительный интервал (ДИ) Комментарии PRIME II [47] 77 пациентов Субанализ РКИ Наблюдение ≈ 3,3±0,9 лет Низкая ЧСС Медиана – 72 уд. / мин (разброс 57–80 уд. / мин) Высокая ЧСС Медиана – 90 уд. / мин (разброс 81–163 уд. / мин) По однофакторному анализу не было разницы в риске смерти между группой низкой и высокой ЧСС (ОР 1,3 ДИ: 0,7–2,4; p=0,58), а по многофакторному – большая смертность в группе Низкой ЧСС (ОР 2,9 ДИ: 1,4–5,8; p=0,002) Пациенты с СНнФВ Средняя ФВ ЛЖ 23+8 % III–IV ФК по NYHA Мареев Ю. В., 2014 г. Анализированные группы Комментарии Низкая ЧСС Медиана – 72 уд. / мин (разброс 57–80 уд. / мин) При поступлении Низкая ЧСС: ЧСС <75 уд. / мин Медиана – 67 уд. / мин [62; 72] Высокая ЧСС: ЧСС ≥75 уд. / мин Медиана – 90 уд. / мин [80; 105] При выписке Низкая ЧСС: ЧСС <70 уд. / мин Медиана – 67 уд. / мин [64; 70] Высокая ЧСС: ЧСС ≥70 уд. / мин Медиана – 78 уд. / мин [74; 83] Выживаемость в группе пациентов с высокой ЧСС была выше как при анализе пациентов при поступлении, так и при выписке Выживаемость в группе пациентов с высокой ЧСС была выше как при анализе пациентов при поступлении, так и при выписке Смерть от всех причин T1 ОР=1,00 T2 ОР=1,13; ДИ: 0,89–1,43 T3 ОР=0,89; ДИ: 0,69–1,15 Сердечно-сосудистая смерть или госпитализация из‑за ХСН T1 ОР=1,00 T2 ОР=0,96; ДИ: 0,78–1,19 T3 ОР=0,79; ДИ: 0,63–0,99 Смерть от всех причин T1 ОР=1,00 T2 ОР=1,13; ДИ: 0,89–1,43 T3 ОР=0,89; ДИ: 0,69–1,15 Сердечно-сосудистая смерть или госпитализация из‑за ХСН T1 ОР=1,00 T2 ОР=0,96; ДИ: 0,78–1,19 T3 ОР=0,79; ДИ: 0,63–0,99 Пациенты были разделены на три терцили (Т): 1 Т=Медиана – 64; [60–68] 2 Т=Медиана – 76 [72–80] 3 Т=Медиана – 90 [86–100] CHARM [6] 1148 пациентов Субанализ РКИ Время наблюдения – 2,8–3,4 года ЧСС при поступлении (деление на квартили) Q1 <69 уд. / мин; Q2 69–81 уд. / мин Q3 82–98 уд. / мин; Q4 >98 уд. / мин ЧСС во время визита через год Q1 <62 уд. / мин; Q2 62–72 уд. / мин Q3 73–82 уд. / мин; Q4 >82 уд. / мин ЧСС при поступлении (деление на квартили) Q1 <69 уд. / мин; Q2 69–81 уд. / мин Q3 82–98 уд. / мин; Q4 >98 уд. / мин ЧСС во время визита через год Q1 <62 уд. / мин; Q2 62–72 уд. / мин Q3 73–82 уд. / мин; Q4 >82 уд. / мин Связь ЧСС при поступлении и прогноза Q1 референс Q2 ОР=1,01; ДИ: 0,69–1,47; p=0,97 Q3 ОР=0,94; ДИ: 0,64–1,38; p=0,75 Q4 ОР=0,68; ДИ: 0,45–1,03; p=0,07 Связь ЧСС во время контрольных визитов и прогноза Q1 референс Q2 ОР=0,99; ДИ: 0,56–1,72; p=0,96 Q3 ОР=0,71; ДИ: 0,39–1,27; p=0,24 Q4 ОР=1,07; ДИ: 0,60–1,90; p=0,82 Hull LifeLab [23] 488 пациентов Анализ пациентов, включенных в програм- му по изучению ХСН Время наблюдения – 3,6 года (1,7–6,9 лет). По однофакторному анали- зу – снижение риска смерти на 7 % при повышении ЧСС на 10 уд. Главной целью работы было сравнение β-АБ и плацебо. По данным последнего ана- лиза, прием β-АБ не влиял на риск смерти пациентов с ФП ОБЗОРЫ § ОБЗОРЫ Анализированные группы / мин (p=0,02). Достоверность пропала по- сле учета различий в характе- ристиках пациентов 357 ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § Таблица 1. Результаты исследований, из (анализы влияния ЧСС при ФП на прог Исследование, число больных с ФП, тип исследования, длительность Анали г Петрухина А. А., 2008 г. 494 пациента Ретроспективный анализ длительной вы- живаемости пациентов Среднее время наблю- дения – 6 лет По медиане Низкая ЧСС – Высокая ЧСС По 25-му проце Низкая ЧСС – Высокая ЧСС По 75‑му проце Низкая ЧСС – Высокая ЧСС Шведский регистр 2015 год [44] 7 392 пациента Среднее время наблю- дения – 2,4 года Пациенты были в зависимости о A – ≤60 уд. / мин B – 61–70 уд. / м C – 71–80 уд. / м D – 81–90 уд. / м E – 91–100 уд. / м F – >100 уд. / ми У пациентов с ФВ ЛЖ >40 % ЧСС ≥75 уд в мин коррелировала с по- вышенным риском общей смерти ОР 1,080 ДИ 1,035–1,126, при росте ЧСС на 10 уд. / мин; p<0,001) У пациентов с ФВ ЛЖ ≤40 % ЧСС ≥75 уд была связана с более низким риском смерти или госпита- лизация из‑за ХСН (ОР=0,95; 95 % ДИ: 0,910–0,991 при увеличении ЧСС на 10 уд. / мин; p=0,018) Таблица 1. Результаты исследований, изучавших оптимальную ЧСС у пациентов с ХСН и ФП (анализы влияния ЧСС при ФП на прогноз больных ХСН) (продолжение) Таблица 1. Результаты исследований, изучавших оптимальную ЧСС у пациентов с ХСН и ФП (анализы влияния ЧСС при ФП на прогноз больных ХСН) (продолжение) Исследование, число больных с ФП, тип исследования, длительность Петрухина А. А., 2008 г. 494 пациента Ретроспективный анализ длительной вы- живаемости пациентов Среднее время наблю- дения – 6 лет Результаты (относительный риск – ОР) и 95 % доверительный интервал (ДИ) Риск смерти от всех причин высокая ЧСС vs низкая ЧСС По медиане ОР=0,93; p=0,59 По 25‑му процентилю ОР=0,97; p=0,837 По 75‑му процентилю ОР=0,77; p=0,082 Результаты (относительный риск – ОР) и 95 % доверительный интервал (ДИ) Результаты (относительный риск – ОР) и 95 % доверительный интервал (ДИ) Риск смерти от всех причин высокая ЧСС vs низкая ЧСС По медиане ОР=0,93; p=0,59 По 25‑му процентилю ОР=0,97; p=0,837 По 75‑му процентилю ОР=0,77; p=0,082 5 % ) Комментарии Пациенты с любой ФВ ЛЖ Медина ФВ ЛЖ – 40 [ИР 30; 50] II–IV ФК по NYHA ⅔ больных, наблюда- лись в 80‑е и 90‑е годы. Соответственно, в группе было много пороков сердца за не ь- си- му ци- е- ин СНнФВ I–IV ФК по NYHA ной тро- о» =0,53 той оля н- Пациенты с любой ФВ ЛЖ Средняя ФВ ЛЖ 48±14 I–III ФК по NYHA Группа строгого контроля не отличалась от группы не- строгого по качеству жизни по- ОР СС Пациенты с любой ФВ ЛЖ Медиана ФВ ЛЖ – 48 [ИР 30; 55] Анализированные группы / мин Пациенты имели высокую ЧСС Медиана 81 [ИР 70; 98] уд. / мин Анализированные группы Комментарии Пациенты с любой ФВ ЛЖ Медина ФВ ЛЖ – 40 [ИР 30; 50] II–IV ФК по NYHA ⅔ больных, наблюда- лись в 80‑е и 90‑е годы. Соответственно, в группе было много пороков сердца СНнФВ I–IV ФК по NYHA Анализированные группы Комментарии Пациенты с любой ФВ ЛЖ Медина ФВ ЛЖ – 40 [ИР 30; 50] По медиане Низкая ЧСС – ЧСС ≤92 уд. / мин Высокая ЧСС – ЧСС >92 уд. / мин По 25-му процентилю Низкая ЧСС – ЧСС ≤80 уд. / мин Высокая ЧСС – ЧСС >80 уд. / мин По 75‑му процентилю Низкая ЧСС – ЧСС ≤105 уд. / мин Высокая ЧСС – ЧСС >105 уд. / мин Риск смерти от всех причин высокая ЧСС vs низкая ЧСС По медиане ОР=0,93; p=0,59 По 25‑му процентилю ОР=0,97; p=0,837 По 75‑му процентилю ОР=0,77; p=0,082 Пациенты были разделены на 6 групп в зависимости от ЧСС: По данным однофакторного анализа не было разницы в риске смерти у боль- ных, разделенных на группы в зависи- мости от ЧСС. По многофакторному анализу – бóльшая смертность у паци- ентов с ЧСС >100 уд. / мин в сравне- нии с пациентами с ЧСС ≤60 уд. / мин (ОР=1,3; ДИ 1,11–1,52; p=0,001) Пациенты были разделены на 6 групп в зависимости от ЧСС: A – ≤60 уд. / мин B – 61–70 уд. / мин C – 71–80 уд. / мин D – 81–90 уд. / мин E – 91–100 уд. / мин F – >100 уд. / мин Пациенты были разделены на 6 групп в зависимости от ЧСС: A – ≤60 уд. / мин B – 61–70 уд. / мин C – 71–80 уд. / мин D – 81–90 уд. / мин E – 91–100 уд. / мин F – >100 уд. / мин A – ≤60 уд. / мин B – 61–70 уд. / мин C – 71–80 уд. / мин D – 81–90 уд. / мин E – 91–100 уд. / мин F – >100 уд. / мин ОР наступления первичной конечной точки* в группе «нестрогого» контро- ля в сравнении с группой «строгого» контроля – 0,83; ДИ: 0,46–1,49; р=0,53 ОР наступления сердечно-сосудистой смерти в группе нестрогого контроля в сравнении с группой строгого кон- троля – 1,10; ДИ: 0,35–3,40 Субанализ RACE II [48] 287 пациентов Среднее время наблю- дения – 3 года Пациенты с любой ФВ ЛЖ ЧСС <75 уд. / мин ЧСС ≥75 уд. ИР – интерквантильный размах (25‑й и 75‑й процентили). * – Первичная точка в исследовании RACE II: сердечно-сосудистая смерть и госпитализация из‑за ХСН, инсульт, системная эмболия, большое кро‑ вотечение, аритмическое событие, включающее синкопе и ЖТ, остановка сердца, жизнеопасные реакции на ритмурежающие препараты, имплан‑ тация ЭКС или ИКД. ОБЗОРЫ § ОБЗОРЫ § мальной частотой сокращений желудочков при ХСН, протекающей на фоне ФП, можно считать диапазон 85–87 уд. / мин. В этом случае становится понятным, почему больные с ХСН и ФП в нашей когорте (так же, как в исследованиях PRIME II и RACE II) имели не худ‑ шую (лучшую?) выживаемость при ЧСС 85–90 уд. / мин в сравнении с ЧСС 67–72 уд. / мин. Видимо, пациентам с СНнФВ и ФП не требуется столь же строгий контроль ЧСС, как пациентам с синусовым ритмом. ISSN 0022-9040. Кардиология. 2017;57(S2) GWTG-HF [46] 20 197 пациентов Медиана ФВ ЛЖ – 48 [ИР 30; 55] Не было связи между ЧСС ФП и ри- ском смерти от всех причин ИР – интерквантильный размах (25‑й и 75‑й процентили). * – Первичная точка в исследовании RACE II: сердечно-сосудистая смерть и госпитализация из‑за ХСН, инсульт, системная эмболия, большое кро‑ вотечение, аритмическое событие, включающее синкопе и ЖТ, остановка сердца, жизнеопасные реакции на ритмурежающие препараты, имплан‑ тация ЭКС или ИКД. ний при ФП у больных с клинически выраженной ХСН на 10–12 ударов больше, чем при синусовом ритме. Как уже говорилось выше, увеличение ЧСС при ХСН является компенсаторным механизмом поддержания ний при ФП у больных с клинически выраженной ХСН на 10–12 ударов больше, чем при синусовом ритме. Как уже говорилось выше, увеличение ЧСС при ХСН является компенсаторным механизмом поддержания сердечного выброса. Однако чрезмерное повышение ЧСС (более 75 уд. / мин при синусовом ритме) начина‑ ет оказывать негативное влияние, в том числе на кро‑ воснабжение миокарда [1, 49]. Следовательно, опти‑ 358 ISSN 0022-9040. Кардиология. 2017;57(S2) ISSN 0022-9040. Кардиология. 2017;57(S2) ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § Только нерегулярность стимуляции при ее одинаковой частоте (неравномерность наполнения и опорожнения желудочков сердца) сопровождается снижением сердечно‑ го выброса на 12 % [59]. Наличие ФП в сравнении с сину‑ совым ритмом при одинаковой ЧСС приводит к снижению показателей сердечного выброса на 18–25 % [60]. При упор‑ ной ФП с частотой сокращений желудочков 102 уд. / мин после аблации атрио-вентрикулярного соединения и сти‑ муляции с такой же ЧСС (устранение только иррегуляр‑ ности ритма) отмечено достоверное повышение сердеч‑ ного индекса на 18 % и снижение ДЗЛА с 17 до 14 мм рт. ст. (р<0,002) [61]. Это сопровождается улучшением клини‑ ческого состояния и при сохранении минимальной вариа‑ бельности интервалов R–R позволяет выровнять показате‑ ли сердечного выброса от сокращения к сокращению [62]. Даже при использовании наиболее современной бивентри‑ кулярной стимуляции (сердечная ресинхронизирующая терапия) достижение оптимальных показателей гемодина‑ мики связано как с оптимизацией ЧСС, так и с устранени‑ ем иррегулярности сокращений желудочков при ФП [63]. Поэтому, по данным ряда работ, при проведении сердечной ресинхронизирующей терапии больным с ХСН и ФП необ‑ ходимо добиваться жесткого контроля ЧСС (с помощью β-АБ и их комбинации с дигоксином), что позволяет увели‑ чить частоту навязанных (ритмичных) импульсов и полу‑ чить результаты, сравнимые с данными пациентов с сохра‑ ненным синусовым ритмом [64, 65]. диастолического наполнения желудочков и ухудшением коронарного кровоснабжения. 3. Третий вопрос о способах оптимального контроля ЧСС при ХСН с ФП требует специального анализа. диастолического наполнения желудочков и ухудшением коронарного кровоснабжения. 3. Третий вопрос о способах оптимального контроля ЧСС при ХСН с ФП требует специального анализа. β-адреноблокаторы Отрицательный хронотропный и инотропный эффек‑ ты этого класса препаратов связаны со снижением симпа‑ тической стимуляции β-адренорецепторов и умеренным снижением атрио-вентрикулярной проводимости. р р у р р В настоящий момент основная доказательная база по лечению ХСН и влиянию на прогноз этих больных существуют только для пациентов с СНнФВ. При этом до недавнего времени считалось, что пациентам с СНнФВ независимо от ритма сердца показано лечение β-АБ с целью улучшения прогноза и снижения риска госпита‑ лизации. Как мы подробно обсудили это в предыдущей статье, при синусовом ритме имеется четкая корреляция между снижением ЧСС и улучшением прогноза боль‑ ных с СНнФВ [1, 67]. Но у больных с ФП картина выгля‑ дит не столь благополучно. В ретроспективных анализах и регистрах, где нельзя исключать влияние врачебного фак‑ тора на назначение или неназначение β-АБ, их прием ассо‑ циировался с лучшей выживаемостью [44, 68]. Но 2 мета- анализа РКИ, опубликованных в 2013 и 2014 гг., показали, что применение β-АБ у пациентов с СНнФВ и постоян‑ ной формой ФП не оказывает влияния на риск смерти этой группы пациентов, несмотря на уменьшение числа сокращений желудочков [57, 58]. Учитывая неожиданные результаты, «переворачивающие представление» о лече‑ нии значительной части больных с ХСН и ФП, была созда‑ на специальная группа по изучению эффективности β-АБ у больных ХСН в зависимости от ЧСС и характера основ‑ ного ритма (Beta-Blockers in Heart Failure Collaborative Group). Проведенный этой группой индивидуальный гармонизированный анализ результатов лечения больных в 11 РКИ подтвердил различия во влиянии β-АБ на про‑ гноз [58]. Среди 14 166 больных с синусовым ритмом β-АБ Практически это выглядит следующим образом: при ХСН и ФП нужно контролировать не только ЧСС, но и величину дефицита пульса, достигающую 11–16 уд. / мин и являющуюся маркером нерегулярных и неэффективных сокращений сердца [37, 66]. Именно эта разница в числе неэффективных сокращений и может быть причиной того, что оптимальные величины ЧСС при ХСН с синусовым ритмом и ФП отличаются в пределах указанной величины дефицита пульса (неэффективных сердечных циклов). Каков практический смысл этих данных? Что должен помнить врач, лечащий пациента с систолической ХСН и ФП? Должен ли он стремиться к контролю ЧСС? Если да, то до какой степени и с помощью каких лекарств? На первые два вопроса легко дать ответ: 1. Контроль ЧСС при ХСН, протекающей с ФП, необхо‑ дим. Ведь мы говорим о ЧСС покоя, а чем ниже этот показатель, тем больше «резерв» безопасного выпол‑ нения физических нагрузок. 2. Цель – снижение ЧСС по крайней мере до 85–90 уд. Почему отличаются оптимальные показатели ЧСС при ХСН с ФП и с синусовым ритмом? Как объяснить различия в оптимальной ЧСС при синусовом ритме и ФП и почему при ФП связь между величиной частоты сокращения желудочков и прогнозом больных ХСН намного менее тесная? Первая проблема в том, что между ХСН и ФП имеется тесная двусторонняя связь. Во Фрамингемском исследо‑ вании было показано, что среди 921 больного с ФП 26 % исходно имели ХСН и еще у 16 % она развилась в процес‑ се 7‑летнего наблюдения. В свою очередь, среди 931 боль‑ ного с имеющейся ХСН 24 % имели в качестве основно‑ го ритма ФП и у 17 % она появилась в течение 7‑летнего наблюдения [24]. Большинство экспертов считают ФП и клиническую ХСН двумя «лицами» одной болезни, имеющей общие ФР: генетическую предрасположен‑ ность, возраст, артериальную гипертензию, ИБС, СД, ожирение, курение, злоупотребление алкоголем [50–52]. Так как частота этих ФР в XXI веке неуклонно увеличи‑ вается, мы имеем дело с эпидемиологическим ростом числа больных как с ХСН, так и с ФП. При этом разви‑ вается электрическое и структурное ремоделирование левого предсердия, поддерживающее ФП, а меняющиеся от цикла к циклу интервалы R-R характеризуются нерав‑ номерным диастолическим наполнением ЛЖ, ухудшени‑ ем гемодинамики, развитием и прогрессированием ХСН [4]. Кроме этого, нельзя сбрасывать со счетов и потерю «предсердного вклада» в диастолическое наполнение желудочков сердца. Причем чем больше ЧСС при ФП, тем меньше вклад сокращения предсердий в наполне‑ ние ЛЖ [53]. Все это требует одновременного контро‑ ля ритма и лечения самой ХСН, если мы хотим повлиять на заболеваемость и смертность [54–56]. При попытке восстановления ритма он удерживается лишь у 70 % паци‑ ентов в связи со структурными изменениями миокарда желудочков и предсердий [28, 29, 31]. При попытках контроля только ЧСС, например, с при‑ менением β-АБ, добиться улучшения прогноза у больных с ХСН и ФП не удается [43, 57, 58]. Это происходит потому, что нельзя недооценивать крайне важный фактор – при ФП течение ХСН утяжеляет не только частота, но и нерегуляр‑ ность наполнения и сокращения желудочков сердца. ISSN 0022-9040. Кардиология. 2017;57(S2) Методы оптимального контроля ЧСС при ХСН с ФП В рекомендациях по лечению ФП (независимо от нали‑ чия или отсутствия ХСН) для уменьшения числа сокра‑ щений желудочков рекомендуется использование β-АБ с классом рекомендаций IВ [36]. Если эта терапия недо‑ статочно эффективна, необходимо рассмотреть приме‑ нение дигоксина, причем при ФВ ЛЖ <40 % и симптомах ХСН дигоксин также имеет класс рекомендаций IB [36]. Несмотря на минимальное количество проведенных иссле‑ дований, предполагается и комбинирование этих двух пре‑ паратов. Мы попытаемся представить имеющиеся данные об использовании этих 3 опций у больных с ФП и ХСН. ОБЗОРЫ § Кстати, «чистый» брадикардитиче‑ ский эффект ивабрадина без влияния на нейрогормоны не позволял достоверно снижать риск смерти больных с СНнФВ и синусовым ритмом, хотя приводил к досто‑ верному улучшению комбинированной конечной точки, включающей риск смерти и декомпенсации, требующей госпитализации больных, что подтверждается результа‑ тами применения ивабрадина, изолированно снижающе‑ го ЧСС [79]. Но при ФП ситуация не столь однозначна. Как было продемонстрировано в исследовании PRIME II, концентрация норадреналина значимо не отличались у пациентов с высокой и низкой ЧСС на фоне ФП [47]. Это свидетельствует о том, что высокая ЧСС при ФП у пациентов с ХСН не связана напрямую со степенью гиперактивации САС и имеются другие факторы, опре‑ деляющие характер ритма у этих пациентов. Особенно важную роль играет нерегулярность сердечного рит‑ ма, проявляющаяся наличием дефицита пульса. В этом случае β-АБ, снижая ЧСС в такой же степени, как и при синусовом ритме, но не устраняя нерегулярности ритма, приводят к увеличению интервалов R–R [80] с возможным развитием пауз, которые потенциально могут провоцировать остановку сердца [81, 82]. Можно предположить, что положительное влияние от блокады САС у больных с СНнФВ и ФП при лечении β-АБ могло нивелироваться отрицательным эффектом от снижения ЧСС [9]. Кроме того, во всех исследованиях с β-АБ боль‑ шинство пациентов находилось на терапии дигоксином, который обеспечивает блокаду САС и снижает уровень норадреналина в крови, в частности, при применении в малых дозах (0,25 мг / сут. и концентрации 1,17 нг / мл) [83, 84]. Лишь дальнейшие исследования смогут про‑ яснить диссонанс между урежением ЧСС и отсутстви‑ ем положительного влияния β-АБ на прогноз больных с ХСН и ФП. Пока же в рекомендациях по контролю 1. Исходная ЧСС. В оптимальном (лучшем из возмож‑ ных мета-анализов) она составляла 80 уд. / мин (ИР 72–88 уд. / мин) при синусовом ритме и 81 уд. / мин при ФП (ИР 72–92 уд. / мин), т. е. была абсолютно одинаковой! 2. Достигнутая ЧСС в процессе лечения. При синусо‑ вом ритме – оптимальные 69 уд. / мин и при ФП те же 69 уд. / мин, в то время как оптимальной выглядит часто‑ та сокращений желудочков 85–87 уд. / мин. Иными сло‑ вами, исходно до лечения β-АБ ЧСС у больных с ХСН и ФП находилась в нормальном диапазоне, т.е было бы странным ожидать снижения смертности при уменьше‑ нии ЧСС ниже оптимальных значений. 3. Может ли быть пульсурежающий эффект единствен‑ ным фактором, определяющим влияние β-АБ (или дру‑ гой терапии) на прогноз больных ХСН? В этом существуют большие сомнения. ОБЗОРЫ § ОБЗОРЫ § ОБЗОРЫ § ведет к дисфункции кардиомиоцитов, риску развития дисфункции миокарда и нарушений сердечного ритма [71–73]. Именно за счет блокады активности САС β-АБ блокируют ремоделирование сердца и снижают риск прогрессирования ХСН и желудочковых нарушений ритма сердца [74–76]. Более того, в только что завер‑ шившихся экспериментальных исследованиях было продемонстрировано, что β-АБ в значительной степени устраняют патологическую картину экспрессии генов, наблюдаемую при СН [77]. Имеются предположения, что разные β-АБ имеют различную степень воздей‑ ствия на парасимпатическое звено автономной нервной системы и, соответственно, дополнительную способ‑ ность блокировать ремоделирование сердца [78]. достоверно на 27 % снижали смертность (ОР=0,73 против плацебо; 95 % ДИ: 0,67–0,79; p<0,001), причем независи‑ мо от исходной ЧСС (р=0,35). Лечение β-АБ 3 034 боль‑ ных с ФП никак не влияло на прогноз (ОР смерти = 0,96; 95 % ДИ: 0,81–1,12; p=0,58), также независимо от исход‑ ной частоты сокращения желудочков. Более того, наблю‑ далась тенденция к более высокому риску смерти у паци‑ ентов с ЧСС ниже 60 уд. / мин. Правда, нужно отметить, что достигнутая ЧСС измеряется дискретно, как правило, на одном из ранних визитов, и связывать ее с показателями смертности (линейным показателем) не совсем корректно. р ( ) рр Самое интересное, что среднее снижение ЧСС состав‑ ляло 11 уд. / мин при синусовом ритме и 12 уд. / мин при ФП, то есть было совершенно одинаковым! Возникает философский вопрос: чего мы хотим? Если контроля ЧСС – мы его достигаем при лечении β-АБ больных с ХСН как при синусовом ритме, так и при ФП. Как мы отмечали выше, это важно, так как определяет лучшее клиническое состояние и резерв для выполнения физических нагрузок. Кстати, в уже упоминавшихся Европейских рекоменда‑ циях по лечению ФП указаны параметры контроля ЧСС: «менее 110 уд. / мин или более строго при сохраняющихся симптомах ХСН» [36]. Однако уменьшения смертности при применении β-АБ у больных с ХСН и ФП не дости‑ гается. Авторы мета-анализа и комментаторы в передо‑ вой статье продолжают задаваться вопросом: почему при синусовом ритме снижение ЧСС связано с уменьше‑ нием смертности, а при ФП – нет? [43, 69]. На сегодня этому есть несколько объяснений: При синусовом ритме снижение ЧСС, отражающее симпатолитическую активность β-АБ, является важным дополнительным механизмом положительного влия‑ ния на прогноз. β-адреноблокаторы / мин (в отличие от менее 75 уд. / мин при ХСН с синусовым ритмом и менее 60 уд. / мин для больных с синусовым ритмом без ХСН). По всей видимости, при этой ЧСС достигается максимальный вклад частоты в поддержа‑ ние минутного объема сердца и отсутствует крити‑ ческое укорочение диастолы, чреватое уменьшением 360 360 ISSN 0022-9040. Кардиология. 2017;57(S2) ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § Так, в исследовании COMET при прямом сравнении 2 β-АБ, примерно оди‑ наково снижающих ЧСС, было продемонстрировано достоверное преимущество карведилола над метопро‑ лолом по влиянию на прогноз больных с СНнФВ [70]. Известно, что дисбаланс симпатической и парасимпа‑ тической нервных систем с резким преобладанием 1‑й ISSN 0022-9040. Кардиология. 2017;57(S2) 361 ОБЗОРЫ § с уменьшением дифицита пульса, то есть увеличивает коли‑ чество эффективных сокращений сердца [37, 66, 92, 93]. ритма у пациентов с ФП с СНнФВ указано «использо‑ вать минимальную дозу β-АБ, позволяющую контроли‑ ровать число сердечных сокращений, при необходимо‑ сти добавить дигоксин и избегать брадикардии» [36]. В исследованиях ОЗМиСН НИИ кардиологии изуча‑ лось влияние дигоксина на ЧСС у больных с ХСН с сину‑ совым ритмом и ФП по данным 72‑часового монитори‑ рования ЭКГ [94]. В отличие от β-АБ дигоксин достовер‑ но больше снижал среднюю ЧСС при ФП (–18 уд. / мин), чем при синусовом ритме (–11 уд. / мин), что связано с уменьшением дефицита пульса и более «регулярным» ритмом при ФП. Максимальное снижение ЧСС про‑ исходило при более высокой концентрации дигокси‑ на (>1,4 нг / мл) на 9 уд. / мин больше, чем при низкой и в ночные часы (с 3 до 6 часов) на 13 уд. / мин, что свиде‑ тельствует о мощном парасимпатическом действии [94]. Однако превышение концентрации дигоксина >1,1 нг / мл у больных СНнФВ и синусовым ритмом, сопровождалось увеличением риска смерти [95]. Поэтому сегодня реко‑ мендуется аккуратное дозирование дигоксина. При этом еще до появления отрицательного дромотропного дей‑ ствия гликозиды способны уменьшать основной период предсердных волн F и в результате их уменьшения способ‑ ствовать увеличению скрытого проведения через атрио- вентрикулярный узел и упорядочиванию ритма [96]. Дигоксин В 1785 году в свой знаменитой книге W. Whithering отмечал максимальный клинический и диуретический эффекты дигиталиса «у пациентов с одышкой, отеками, частым и нерегулярным ритмом» и токсический эффект «у больных с редким и регулярным ритмом»* После выделения кристаллического дигоксина [85] и создания метода его контроля в крови [86] препарат стал основным средством лечения СНнФВ, протекающей с ФП, хотя проспективных РКИ, изучивших применение дигок‑ сина у пациентов с ХСН и ФП не проводилось [87, 88]. В Европейских и Американских рекомендациях отсутству‑ ют раздельные рекомендации по применению дигноксина у пациентов с ХСН при наличии синусового ритма и ФП. В последнее время появились ретроспективные обсерва‑ ционные исследования, показавшие, что пациенты с ХСН, получавшие дигоксин, имели выше риск смерти, чем паци‑ енты, не получавшие его [89, 90]. Но в 2015 году вышел системный мета-анализ по применению дигоксина у паци‑ ентов с ХСН [91]. Мета-анализ показал, что по данным РКИ, применение дигоксина не увеличивает риск смерти, но снижает риск госпитализации из‑за обострения ХСН. Что касается обсервационных исследований, то в них паци‑ енты, получавшие дигоксин, были старше (на 2,4 года), имели более низкое значение ФВ ЛЖ (33 против 42 %) и среди них было больше число пациентов с СД, чем сре‑ ди пациентов в группе контроля. Применение статисти‑ ческой обработки данных с учетом разницы в исходных харатктеристиках между пациентами, принимавшими и не принимавшими дигоксин, уменьшало влияние препа‑ рата на риск смерти. Мета-регрессия показала связь между погрешностями в исследованиях (bias score) и повышени‑ ем риска смерти на фоне дигоксина. В связи с этим авторы поставили под сомнение отрицательное влияние дигок‑ сина на прогноз, указав, что необходимо проведение про‑ спективных рандомизированных исследований. Учитывая, что «безопасная» концентрация дигоксина в плазме крови не всегда позволяет достигать выраженного пульсурежающего эффекта [97, 98], целесообразна комби‑ нация с β-АБ, которые больше влияют на атриовентрику‑ лярную проводимость, не упорядочивая ритм при ФП [96]. * – W. Withering, An account of the foxglove, and Some of its Medical Uses: with practical remarks on dropsy, and other diseases. London: Birmingham: printed by M. Swinney; for G. G. J. and J. Robinson, Paternoster-Row, 1785 С П И С О К Л И Т Е Р А Т У Р Ы 1. Мареев В. Ю., Мареев Ю. В. Роль ЧСС в механизмах компен- сации и декомпенсации у больных ХСН при синусовом ритме и фибрилляции предсердий и методы безопасного и эффектив- ного контроля сердечного ритма. Часть 1. Синусовый ритм. Журнал Сердечная Недостаточность. 2017;18 (3):213–24. DOI:10.18087 / rhfj.2017.3.2352 [Mareev V. Yu., Mareev Yu. V. Rol` ChSS v mexanizmax kompensaczii i dekompensaczii u bol`ny`x XSN pri sinusovom ritme i fibrillyaczii predserdij i metody` bezo- pasnogo i e`ffektivnogo kontrolya serdechnogo ritma. Chast` 1. Sinusovy`j ritm. Zhurnal Serdechnaya Nedostatochnost`. 2017;18 (3):213–24. DOI:10.18087 / rhfj.2017.3.2352]. 1. Мареев В. Ю., Мареев Ю. В. Роль ЧСС в механизмах компен- сации и декомпенсации у больных ХСН при синусовом ритме и фибрилляции предсердий и методы безопасного и эффектив- ного контроля сердечного ритма. Часть 1. Синусовый ритм. Журнал Сердечная Недостаточность. 2017;18 (3):213–24. DOI:10.18087 / rhfj.2017.3.2352 [Mareev V. Yu., Mareev Yu. V. Rol` ChSS v mexanizmax kompensaczii i dekompensaczii u bol`ny`x XSN pri sinusovom ritme i fibrillyaczii predserdij i metody` bezo- pasnogo i e`ffektivnogo kontrolya serdechnogo ritma. Chast` 1. Sinusovy`j ritm. Zhurnal Serdechnaya Nedostatochnost`. 2017;18 (3):213–24. DOI:10.18087 / rhfj.2017.3.2352]. ной терапии, влияющей на прогноз у больных с фибрилля- цией предсердий в сочетании с артериальной гипертонией, ишемической болезнью сердца, хронической сердечной недо- статочностью по данным регистра РЕКВАЗА-КЛИНИКА. Кардиоваскулярная терапия и профилактика. 2017;16 (2):33– 8. DOI:10.15829 / 1728‑8800‑2017‑2‑33‑38 [Stepina E. V., Luk`yanov M. M., Bojczov S. A., Bichurina M. A., Belova E. N., Kudryashov E. V i dr. Naznachenie medikamentoznoj terapii, vli- yayushhej na prognoz u bol`ny`x s fibrillyacziej predserdij v soche tanii s arterial`noj gipertoniej, ishemicheskoj bolezn`yu serdcza, xronicheskoj serdechnoj nedostatochnost`yu po danny`m registra REKVAZA-KLINIKA. Kardiovaskulyarnaya terapiya i profilaktika. 2017;16 (2):33–8. DOI:10.15829 / 1728‑8800‑2017‑2‑33‑38]. ной терапии, влияющей на прогноз у больных с фибрилля- цией предсердий в сочетании с артериальной гипертонией, ишемической болезнью сердца, хронической сердечной недо- статочностью по данным регистра РЕКВАЗА-КЛИНИКА. Кардиоваскулярная терапия и профилактика. 2017;16 (2):33– 8. DOI:10.15829 / 1728‑8800‑2017‑2‑33‑38 [Stepina E. V., Luk`yanov M. M., Bojczov S. A., Bichurina M. A., Belova E. N., Kudryashov E. V i dr. Naznachenie medikamentoznoj terapii, vli- yayushhej na prognoz u bol`ny`x s fibrillyacziej predserdij v soche tanii s arterial`noj gipertoniej, ishemicheskoj bolezn`yu serdcza, xronicheskoj serdechnoj nedostatochnost`yu po danny`m registra REKVAZA-KLINIKA. Kardiovaskulyarnaya terapiya i profilaktika. 2017;16 (2):33–8. DOI:10.15829 / 1728‑8800‑2017‑2‑33‑38]. ( ) fj ] 2. Поляков Д. С., Фомин И. В., Валикулова Ф. Ю., Вайсберг А. Р., Краием Н., Бадин Ю. В. и др. С П И С О К Л И Т Е Р А Т У Р Ы Эпидемиологическая программа ЭПОХА–ХСН: декомпенсация хронической сердечной недо- статочности в реальной клинической практике (ЭПОХА–Д– ХСН). Журнал Сердечная Недостаточность. 2016;17 (6):299– 305. DOI:10.18087 / rhfj.2016.5.2239 [Polyakov D. S., Fomin I. V., Valikulova F. Yu., Vajsberg A. R., Kraiem N., Badin Yu. V. i dr. E`pidemiologicheskaya programma E`POXA–XSN: dekom- pensacziya xronicheskoj serdechnoj nedostatochnosti v real`noj klinicheskoj praktike (E`POXA–D–XSN). Zhurnal Serdechnaya Nedostatochnost`. 2016;17 (6):299–305. DOI:10.18087 / rhfj.20 16.5.2239]. 11. Shahid F, Lip GYH. Atrial Fibrillation and Heart Failure: How Should We Manage Our Patients? Arrhythm Electrophysiol Rev. 2016;5 (3):162–3. DOI:10.15420 / AER.2016.5.3. ED3. ( ) 12. Мелехов А. В., Гендлин Г. Е., Дадашова Э. Ф., Никитин И. Г., Алексеева Е. М., Захарова И. И. и др. Практика ведения больных с фибрилляцией предсердий: реальная ситуация на примере Московского регистра в период 2009–2015 гг. Сердце: Журнал Для Практикующих Врачей. 2017;16 (1):73–80 [Melexov A. V., Gendlin G. E., Dadashova E`. F., Nikitin I. G., Alekseeva E. M., Zaxarova I. I. i dr. Praktika vedeniya bol`ny`x s fibrillyacziej pred- serdij: real`naya situacziya na primere Moskovskogo registra v peri- od 2009–2015 gg. Serdcze: Zhurnal Dlya Praktikuyushhix Vrachej. 2017;16 (1):73–80]. 3. Guha K, McDonagh T. Heart failure epidemiology: European per- spective. Curr Cardiol Rev. 2013;9 (2):123–7.i 4. Anter E, Jessup M, Callans DJ. Atrial fibrillation and heart fail- ure: treatment considerations for a dual epidemic. Circulation. 2009;119 (18):2516–25. DOI:10.1161 / CIRCULATIONAHA.1 08.821306. 13. Козиолова Н. А., Никонова Ю. Н., Шилова Я. Э., Агафонов А. В., Полянская Е. А. Характеристика хронической сердечной недо- статочности на фоне перманентной формы фибрилляции предсердий. Журнал Сердечная Недостаточность. 2013;14 (1):14–21. DOI:10.18087 / rhfj.2013.1.1806 [Koziolova N. A., Nikonova Yu. N., Shilova Ya. E`., Agafonov A. V., Polyanskaya E. A. Xarakteristika xronicheskoj serdechnoj nedostatochnosti na fone permanentnoj formy` fibrillyaczii predserdij. Zhurnal Serdechnaya Nedostatochnost`. 2013;14 (1):14–21. DOI:10.18087 / rhfj.2013. 1.1806]. 5. Khand AU, Cleland JGF, Deedwania PC. Prevention of and medi- cal therapy for atrial arrhythmias in heart failure. Heart Fail Rev. 2002;7 (3):267–83. 6. Castagno D, Skali H, Takeuchi M, Swedberg K, Yusuf S, Granger CB et al. Association of heart rate and outcomes in a broad spectrum of patients with chronic heart failure: results from the CHARM (Candesartan in Heart Failure: Assessment of Reduction in Mortality and morbidity) program. J Am Coll Cardiol. 2012;59 (20):1785–95. DOI:10.1016 / j.jacc.2011.12.044. 14. Арутюнов Г. П. Место и значение наблюдательных исследова- ний НОАК у пациентов с неклапанной фибрилляцией пред- сердий с позиции медицины, основанной на доказательствах. Сердце: Журнал для практикующих врачей. ОБЗОРЫ § Конечно, имеющихся на сегодня данных крайне мало для того, чторбы сделать окончательные выводы. Требуются крупные проспективные исследования, но, по нашим дан‑ ным, такие протоколы в настоящее время не проводятся. Поэтому хочется еще раз отметить тот факт, что не всегда данных доказательной медицины достаточно для формули‑ рования выводов, которые могли бы использоваться в кли‑ нической практике. И несмотря на то, что уменьшение ЧСС не всегда сопровождается улучшением прогноза больных, особенно при СНнФВ и ФП, с клинической точки зрения целесообразно уменьшение ЧСС покоя до 85–90 уд. / мин у этой категории пациентов. Представленный нами обзор подтверждает, на наш взгляд, оправданность аккуратно‑ го комбинированного применения небольших доз β-АБ и дигоксина для контроля числа желудочковых сокращений у больных с ХСН и ФП. По всей видимости, максимальный эффект от этой терапии могут получить пациенты с кли‑ нически выраженной ХСН [100] и / или ЧСС покоя выше 90–100 уд. / мин, и эта тактика прописана в современных рекомендациях по лечению ФП [36]. Комбинация дигоксина и β-АБ Для безопасного и эффективного контроля числа желу‑ дочковых сокращений у больных с ХСН и ФП эта комби‑ нация выглядит абсолютно оправданной. Однако, на удив‑ ление, имеется всего 1 небольшое по объему (47 больных) проспективное сравнительное исследование, сравнившее 3 тактики контроля ЧСС у больных с ХСН и ФП [99]. Результаты этого исследования не выявили преимущества β-АБ карведиола (47,6 мг / сут.) над дигоксином (концен‑ трация 1,2 нг / мл) в контроле числа сокращений желудоч‑ ков, более того, дигоксин достоверно более значимо влиял на ночную ЧСС. При этом комбинированная терапия про‑ демонстрировала однозначно бóльшую эффективность в сравнении с монотерапией карведилолом или дигокси‑ ном. Важно, что при концентрации дигоксина 1,4 нг / мл его комбинация с карведилолом хотя и увеличивала среднюю продолжительность пауз (2,9±0,8 против 2,4±0,6 секунд, на монотерапии дигоксином; p<0,05), но не провоциро‑ вала появления опасных пауз продолжительностью более 3 секунд ни в дневное, ни в ночное время [99]. Два основных эффекта дигоксина при лечении ХСН, особенно протекающей с ФП – это положительное ино‑ тропное действие и отрицательный хронотропный эффект, связанный с умеренной блокадой катехоламинов и одно‑ временной активацией парасимпатической нервной систе‑ мы. Замедление атрио-вентрикулярной проводимости (отрицательный дромотропный эффект гликозидов) позво‑ ляет не только снижать ЧСС, но и «упорядочивать ритм» ISSN 0022-9040. Кардиология. 2017;57(S2) 362 ОБЗОРЫ § Impact of implantable cardioverter-defibrillator, ami- odarone, and placebo on the mode of death in stable patients with heart failure: analysis from the sudden cardiac death in heart failure trial. Circulation. 2009;120 (22):2170–6. DOI:10.1161 / CIRCUL ATIONAHA.109.853689. j j 18. Nilsson KR, Al-Khatib SM, Zhou Y, Pieper K, White HD, Maggioni AP et al. Atrial fibrillation management strategies and ear- ly mortality after myocardial infarction: results from the Valsartan in Acute Myocardial Infarction (VALIANT) Trial. Heart. 2010;96 (11):838–42. DOI:10.1136 / hrt.2009.180182. 33. Di Biase L, Mohanty P, Mohanty S, Santangeli P, Trivedi C, Lakkireddy D et al. Ablation Versus Amiodarone for Treatment of Persistent Atrial Fibrillation in Patients With Congestive Heart Failure and an Implanted Device: Results From the AATAC Multicenter Randomized Trial. Circulation. 2016;133 (17):1637– 44. DOI:10.1161 / CIRCULATIONAHA.115.019406.t 19. Carson PE, Johnson GR, Dunkman WB, Fletcher RD, Farrell L, Cohn JN. The influence of atrial fibrillation on prognosis in mild to moderate heart failure. The V-HeFT Studies. The V-HeFT VA Cooperative Studies Group. Circulation. 1993;87 (6 Suppl): VI102–110. 34. Marrouche NF, Brachmann J, CASTLE-AF Steering Committee. Catheter ablation versus standard conventional treatment in patients with left ventricular dysfunction and atrial fibrillation (CASTLE-AF) – study design. Pacing Clin Electrophysiol. 2009;32 (8):987–94. DOI:10.1111 / j.1540–8159.2009.02428.x.t 20. Crijns HJ, Tjeerdsma G, de Kam PJ, Boomsma F, van Gelder IC, van den Berg MP et al. Prognostic value of the presence and development of atrial fibrillation in patients with advanced chronic heart failure. Eur Heart J. 2000;21 (15):1238–45. DOI:10.1053 / euhj.1999.2107. j 35. Cleland JGF, Coletta AP, Buga L, Ahmed D, Clark AL. Clinical tri- als update from the American College of Cardiology meeting 2010: DOSE, ASPIRE, CONNECT, STICH, STOP-AF, CABANA, RACE II, EVEREST II, ACCORD, and NAVIGATOR. Eur J Heart Fail. 2010;12 (6):623–9. DOI:10.1093 / eurjhf / hfq083. j 21. Swedberg K, Olsson LG, Charlesworth A, Cleland J, Hanrath P, Komajda M et al. Prognostic relevance of atrial fibrillation in patients with chronic heart failure on long-term treatment with beta-blockers: results from COMET. Eur Heart J. 2005;26 (13):1303–8. DOI:10.1093 / eurheartj / ehi166. 36. Kirchhof P, Benussi S, Kotecha D, Ahlsson A, Atar D, Casadei B et al. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Eur J Cardiothorac Surg. 2016;50 (5):e1–88. DOI:10.1093 / ejcts / ezw313. ( ) j 22. Мареев Ю. В., Герасимова В. В., Горюнова Т. В., Петрухина А. А., Даниелян М. ОБЗОРЫ § в Кардиологии. 2017;13 (2):146–54. DOI:10.20996 / 1819 ‑6446‑2017‑13‑2‑146‑154 [Stepina E. V., Luk`yanov M. M., Bichurina M. A., Belova E. N., Kudryashov E. V., Yuz`kov Yu. V i dr. Terapiya oral`ny`mi antikoagulyantami u bol`ny`x s fibrillyacziej predserdij v sochetanii s arterial`noj gipertoniej, ishemicheskoj bolezn`yu serdcza, xronicheskoj serdechnoj nedostatochnost`yu na gospital`nom i ambulatornom e`tapax lecheniya po danny`m registra REKVAZA-KLINIKA. Raczional`naya Farmakoterapiya v Kardiologii. 2017;13 (2):146–54. DOI:10.20996 / 1819‑6446‑2 017‑13‑2‑146‑154]. 27. Middlekauff HR, Stevenson WG, Stevenson LW. Prognostic sig- nificance of atrial fibrillation in advanced heart failure. A study of 390 patients. Circulation. 1991;84 (1):40–8. 28. Wyse DG, Waldo AL, DiMarco JP, Domanski MJ, Rosenberg Y, Schron EB et al. A comparison of rate control and rhythm con- trol in patients with atrial fibrillation. N Engl J Med. 2002;347 (23):1825–33. DOI:10.1056 / NEJMoa021328. 29. Van Gelder IC, Hagens VE, Bosker HA, Kingma JH, Kamp O, Kingma T et al. A comparison of rate control and rhythm control in patients with recurrent persistent atrial fibrillation. N Engl J Med. 2002;347 (23):1834–40. DOI:10.1056 / NEJMoa021375. 16. Dries DL, Exner DV, Gersh BJ, Domanski MJ, Waclawiw MA, Stevenson LW. Atrial fibrillation is associated with an increased risk for mortality and heart failure progression in patients with asymp- tomatic and symptomatic left ventricular systolic dysfunction: a ret- rospective analysis of the SOLVD trials. Studies of Left Ventricular Dysfunction. J Am Coll Cardiol. 1998;32 (3):695–703. J 30. Chatterjee S, Sardar P, Lichstein E, Mukherjee D, Aikat S. Pharmacologic rate versus rhythm-control strategies in atrial fibril- lation: an updated comprehensive review and meta-analysis. Pacing Clin Electrophysiol. 2013;36 (1):122–33. DOI:10.1111 / j.1540– 8159.2012.03513.x. 31. Roy D, Talajic M, Nattel S, Wyse DG, Dorian P, Lee KL et al. Rhythm control versus rate control for atrial fibrillation and heart failure. N Engl J Med. 2008;358 (25):2667–77. DOI:10.1056 / NEJMoa0708789. 17. Olsson LG, Swedberg K, Ducharme A, Granger CB, Michelson EL, McMurray JJV et al. Atrial fibrillation and risk of clinical events in chronic heart failure with and without left ventricular systolic dysfunction: results from the Candesartan in Heart failure-Assess- ment of Reduction in Mortality and morbidity (CHARM) pro- gram. J Am Coll Cardiol. 2006;47 (10):1997–2004. DOI:10.1016 / j.jacc.2006.01.060. 32. Packer DL, Prutkin JM, Hellkamp AS, Mitchell LB, Bernstein RC, Wood F et al. ОБЗОРЫ § О., Капанадзе Л. Г. и др. Факторы, определя- ющие прогноз при хронической сердечной недостаточно- сти: роль ширины и морфологии комплекса QRS. Журнал Сердечная Недостаточность. 2012;13 (5):255–66 [Mareev Yu. V., Gerasimova V. V., Goryunova T. V., Petruxina A. A., Danielyan M. O., Kapanadze L. G. i dr. Faktory`, opredelyayushhie prognoz pri xronicheskoj serdechnoj nedostatochnosti: rol` shiriny` i mor- fologii kompleksa QRS. Zhurnal Serdechnaya Nedostatochnost`. 2012;13 (5):255–66]. j 37. Mukharliamov NM, Tsibekmakher TD. [Clinical evalua- tion of the effectiveness of treatment with cardiac glycosides]. Kardiologiia. 1975;15 (7):141–53. 38. Van Gelder IC, Wyse DG, Chandler ML, Cooper HA, Olshansky B, Hagens VE et al. Does intensity of rate-control influence out- come in atrial fibrillation? An analysis of pooled data from the RACE and AFFIRM studies. Europace. 2006;8 (11):935–42. DOI:10.1093 / europace / eul106. 23. Cullington D, Goode KM, Zhang J, Cleland JGF, Clark AL. Is heart rate important for patients with heart failure in atrial fibril- lation? JACC Heart Fail. 2014;2 (3):213–20. DOI:10.1016 / j.jchf.2014.01.005. 39. Van Gelder IC, Groenveld HF, Crijns HJGM, Tuininga YS, Tijssen JGP, Alings AM et al. Lenient versus strict rate control in patients with atrial fibrillation. N Engl J Med. 2010;362 (15):1363– 73. DOI:10.1056 / NEJMoa1001337. j j 24. Wang TJ, Larson MG, Levy D, Vasan RS, Leip EP, Wolf PA et al. Temporal relations of atrial fibrillation and congestive heart fail- ure and their joint influence on mortality: the Framingham Heart Study. Circulation. 2003;107 (23):2920–5. DOI:10.1161 / 01. CIR.0000072767.89944.6E. 40. Groenveld HF, Tijssen JGP, Crijns HJGM, Van den Berg MP, Hillege HL, Alings M et al. Rate control efficacy in permanent atrial fibrillation: successful and failed strict rate control against a back- ground of lenient rate control: data from RACE II (Rate Control Efficacy in Permanent Atrial Fibrillation). J Am Coll Cardiol. 2013;61 (7):741–8. DOI:10.1016 / j.jacc.2012.11.038. 25. Mahoney P, Kimmel S, DeNofrio D, Wahl P, Loh E. Prognostic sig- nificance of atrial fibrillation in patients at a tertiary medical center referred for heart transplantation because of severe heart failure. Am J Cardiol. 1999;83 (11):1544–7. j j 41. Smit MD, Crijns HJGM, Tijssen JGP, Hillege HL, Alings M, Tuininga YS et al. Effect of lenient versus strict rate control on cardi- ac remodeling in patients with atrial fibrillation data of the RACE II (RAte Control Efficacy in permanent atrial fibrillation II) study. J Am Coll Cardiol. 2011;58 (9):942–9. DOI:10.1016 / j.jacc.2011.04.030. J 26. С П И С О К Л И Т Е Р А Т У Р Ы 2016;15 (6 (92)): 441–7. DOI:10.18087 / RHJ.2016.6.2297 [Arutyunov G. P. Mesto i znachenie nablyudatel`ny`x issledovanij NOAK u paczien- tov s neklapannoj fibrillyacziej predserdij s poziczii medicziny`, osnovannoj na dokazatel`stvax. Serdcze: Zhurnal dlya praktikuy- ushhix vrachej. 2016;15 (6 (92)): 441–7. DOI:10.18087 / RHJ.20 16.6.2297]. 7. Lip GYH, Laroche C, Popescu MI, Rasmussen LH, Vitali- Serdoz L, Dan G-A et al. Heart failure in patients with atrial fibril- lation in Europe: a report from the EURObservational Research Programme Pilot survey on Atrial Fibrillation. Eur J Heart Fail. 2015;17 (6):570–82. DOI:10.1002 / ejhf.254. j 8. Cleland JGF, Swedberg K, Follath F, Komajda M, Cohen-Solal A, Aguilar JC et al. The EuroHeart Failure survey programme – a sur- vey on the quality of care among patients with heart failure in Europe. Part 1: patient characteristics and diagnosis. Eur Heart J. 2003;24 (5):442–63. 15. Степина Е. В., Лукьянов М. М., Бичурина М. А., Белова Е. Н., Кудряшов Е. В., Юзьков Ю. В и др. Терапия оральными анти- коагулянтами у больных с фибрилляцией предсердий в соче- тании с артериальной гипертонией, ишемической болезнью сердца, хронической сердечной недостаточностью на госпи- тальном и амбулаторном этапах лечения по данным реги- стра РЕКВАЗА-КЛИНИКА. Рациональная Фармакотерапия 9. Mareev Y, Cleland JGF. Should β-blockers be used in patients with heart failure and atrial fibrillation? Clin Ther. 2015;37 (10):2215– 24. DOI:10.1016 / j.clinthera.2015.08.017. 10. Степина Е. В., Лукьянов М. М., Бойцов С. А., Бичурина М. А., Белова Е. Н., Кудряшов Е. В и др. Назначение медикаментоз- ISSN 0022-9040. Кардиология. 2017;57(S2) 363 ОБЗОРЫ § Naito M, David D, Michelson EL, Schaffenburg M, Dreifus LS. The hemodynamic consequences of cardiac arrhythmias: evalua- tion of the relative roles of abnormal atrioventricular sequencing, irregularity of ventricular rhythm and atrial fibrillation in a canine model. Am Heart J. 1983;106 (2):284–91.f 61. Clark DM, Plumb VJ, Epstein AE, Kay GN. Hemodynamic effects of an irregular sequence of ventricular cycle lengths during atrial fibrillation. J Am Coll Cardiol. 1997;30 (4):1039–45. 47. Rienstra M, Van Gelder IC, Van den Berg MP, Boomsma F, Hillege HL, Van Veldhuisen DJ. A comparison of low versus high heart rate in patients with atrial fibrillation and advanced chronic heart failure: effects on clinical profile, neurohormones and survival. Int J Cardiol. 2006;109 (1):95–100. DOI:10.1016 / j.ijcard.2005.05.054. i J 62. Muntinga HJ, Gosselink AT, Blanksma PK, De Kam PJ, Van Der Wall EE, Crijns HJ. Left ventricular beat to beat performance in atrial fibrillation: dependence on contractility, preload, and after- load. Heart. 1999;82 (5):575–80. 48. Mulder BA, Van Veldhuisen DJ, Crijns HJGM, Tijssen JGP, Hillege HL, Alings M et al. Lenient vs. strict rate control in patients with atrial fibrillation and heart failure: a post-hoc analysis of the RACE II study. Eur J Heart Fail. 2013;15 (11):1311–8. DOI:10.1093 / eurjhf / hft093. ( ) 63. Melenovsky V, Hay I, Fetics BJ, Borlaug BA, Kramer A, Pastore JM et al. Functional impact of rate irregularity in patients with heart failure and atrial fibrillation receiving cardiac resynchronization therapy. Eur Heart J. 2005;26 (7):705–11. DOI:10.1093 / eur- heartj / ehi066. jt 49. Boersma E, Pieper KS, Steyerberg EW, Wilcox RG, Chang WC, Lee KL et al. Predictors of outcome in patients with acute coronary syndromes without persistent ST-segment elevation. Results from an international trial of 9461 patients. The PURSUIT Investigators. Circulation. 2000;101 (22):2557–67. 64. Tolosana JM, Arnau AM, Madrid AH, Macias A, Lozano IF, Osca J et al. Cardiac resynchronization therapy in patients with perma- nent atrial fibrillation. Is it mandatory to ablate the atrioventricu- lar junction to obtain a good response? Eur J Heart Fail. 2012;14 (6):635–41. DOI:10.1093 / eurjhf / hfs024. 50. Lubitz SA, Benjamin EJ, Ellinor PT. Atrial fibrillation in congestive heart failure. Heart Fail Clin. 2010;6 (2):187–200. DOI:10.1016 / j.hfc.2009.11.001. 65. Мареев Ю. В., Шитов В. Н., Киктев В. Г., Сапельников О. В., Латыпов Р. С., Гришин И. Р. и др. Сердечная ресинхронизиру- ющая терапия при постоянной форме мерцательной аритмии. Кардиология. 2013;53 (5):43–9 [Mareev Yu. ОБЗОРЫ § 55. Neuberger H-R, Mewis C, van Veldhuisen DJ, Schotten U, van Gelder IC, Allessie MA et al. Management of atrial fibrillation in patients with heart failure. Eur Heart J. 2007;28 (21):2568–77. DOI:10.1093 / eurheartj / ehm341.i 42. Hirano K, Yamashita T, Suzuki S, Hayama E, Matsuoka J, Otsuka T et al. Relationship between 24‑h Holter recordings and clinical outcomes in patients with permanent atrial fibrillation. J Cardiol. 2012;60 (1):42–6. DOI:10.1016 / j.jjcc.2012.02.002. p DOI:10.1093 / eurheartj / ehm341. j jj 43. Kotecha D, Flather MD, Altman DG, Holmes J, Rosano G, Wikstrand J et al. Heart Rate and Rhythm and the Benefit of Beta- Blockers in Patients With Heart Failure. J Am Coll Cardiol. 2017;69 (24):2885–96. DOI:10.1016 / j.jacc.2017.04.001. 56. Heist EK, Mansour M, Ruskin JN. Rate control in atrial fibrillation: targets, methods, resynchronization considerations. Circulation. 2011;124 (24):2746–55. DOI:10.1161 / CIRCULATIONAHA. 111.019919. 57. Rienstra M, Damman K, Mulder BA, Van Gelder IC, McMurray JJV, Van Veldhuisen DJ. Beta-Blockers and Outcome in Heart Failure and Atrial Fibrillation. JACC: Heart Failure. 2013;1 (1):21–8. DOI:10.1016 / j.jchf.2012.09.002. 44. Li S-J, Sartipy U, Lund LH, Dahlström U, Adiels M, Petzold M et al. Prognostic Significance of Resting Heart Rate and Use of β-Blockers in Atrial Fibrillation and Sinus Rhythm in Patients With Heart Failure and Reduced Ejection FractionCLINICAL PERSPECTIVE: Findings From the Swedish Heart Failure Registry. Circulation: Heart Failure. 2015;8 (5):871–9. DOI:10.1 161 / CIRCHEARTFAILURE.115.002285. 58. Kotecha D, Holmes J, Krum H, Altman DG, Manzano L, Cleland JGF et al. Efficacy of β blockers in patients with heart fail- ure plus atrial fibrillation: an individual-patient data meta-analysis. The Lancet. 2014;384 (9961):2235–43. DOI:10.1016 / S0140– 6736 (14) 61373–8. 45. Bui AL, Grau-Sepulveda MV, Hernandez AF, Peterson ED, Yancy CW, Bhatt DL et al. Admission heart rate and in-hospital outcomes in patients hospitalized for heart failure in sinus rhythm and in atrial fibrillation. Am Heart J. 2013;165 (4):567–574.e6. DOI:10.1016 / j.ahj.2013.01.007. 59. Daoud EG, Weiss R, Bahu M, Knight BP, Bogun F, Goyal R et al. Effect of an irregular ventricular rhythm on cardiac output. Am J Cardiol. 1996;78 (12):1433–6. j j 46. Laskey WK, Alomari I, Cox M, Schulte PJ, Zhao X, Hernandez AF et al. Heart Rate at Hospital Discharge in Patients With Heart Failure Is Associated With Mortality and Rehospitalization. Journal of the American Heart Association. 2015;4 (4):e001626 – e001626. DOI:10.1161 / JAHA.114.001626. J 60. ОБЗОРЫ § van den Berg MP, van Gelder IC, van Veldhuisen DJ. Impact of atri- al fibrillation on mortality in patients with chronic heart failure. Eur J Heart Fail. 2002;4 (5):571–5. 364 ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § 85. Smith S. LXXII. – Digoxin, a new digitalis glucoside. J Chem Soc. 1930;508–10. 71. Kiseleva Z. M., Orlova A. I. [Functional state of the sympathetic- adrenal system in patients with infectious-allergic myocarditis and idiopathic myocardial lesions under the effect of treatment with car- diac glycosides and beta-blockaders]. Ter Arkh. 1978;50 (8):38–43. 86. Smith TW, Haber E. Current techniques for serum or plasma digi- talis assay and their potential clinical application. Am J Med Sci. 1970;259 (5):301–8. 72. Kiseleva Z. M., Orlova A. I. [Study of catecholamines and DOPA in obsidan treatment of patients with circulatory insufficiency]. Ter Arkh. 1979;51 (1):49–53.h 87. Мареев В. Ю., Агеев Ф. Т., Арутюнов Г. П., Коротеев А. В., Мареев Ю. В., Овчинников А. Г и др. Национальные реко- мендации ОССН, РКО и РНМОТ по диагностике и лече- нию ХСН (четвертый пересмотр). Утверждены на Конгрессе ОССН 7 декабря 2012 года, на Правлении ОССН 31 мар- та 2013 и Конгрессе РКО 25 сентября 2013 года. Журнал Сердечная Недостаточность. 2013;14 (7):379–472. DOI:10.18087 / rhfj.2013.7.1860 [Mareev V. Yu., Ageev F. T., Arutyunov G. P., Koroteev A. V., Mareev Yu. V., Ovchinnikov A. G i dr. Naczional`ny`e rekomendaczii OSSN, RKO i RNMOT po diagnostike i lecheniyu XSN (chetverty`j peresmotr). Utverzhdeny` na Kongresse OSSN 7 dekabrya 2012 goda, na Pravlenii OSSN 31 marta 2013 i Kongresse RKO 25 sentyabrya 2013 goda. Zhurnal Serdechnaya Nedostatochnost`. 2013;14 (7):379–472. DOI:10.18087 / rhfj.2013.7.1860]. 73. Floras JS, Ponikowski P. The sympathetic / parasympathetic imba lance in heart failure with reduced ejection fraction. Eur Heart J. 2015;36 (30):1974–1982b. DOI:10.1093 / eurheartj / ehv087. j 74. Bristow MR, Kantrowitz NE, Ginsburg R, Fowler MB. Beta- adrenergic function in heart muscle disease and heart failure. J Mol Cell Cardiol. 1985;17 Suppl 2:41–52. 75. Heilbrunn SM, Shah P, Bristow MR, Valantine HA, Ginsburg R, Fowler MB. Increased beta-receptor density and improved hemo- dynamic response to catecholamine stimulation during long-term metoprolol therapy in heart failure from dilated cardiomyopathy. Circulation. 1989;79 (3):483–90. 76. Скворцов А. А., Челмакина С. М., Пожарская Н. И., Мареев В. Ю. Модулирование активности системы нейрогуморальной регу- ляции при хронической сердечной недостаточности Блокада ренин-ангиотензин-альдостероновой системы. Россий ский медицинский журнал. 2000; (2):87 [Skvorczov A. A., Chelmakina S. M., Pozharskaya N. I., Mareev V. Yu. Modulirovanie aktivnosti sistemy` nejrogumoral`noj regulyaczii pri xroniches- koj serdechnoj nedostatochnosti Blokada renin-angiotenzin- al`dosteronovoj sistemy`. Rossijskij mediczinskij zhurnal. 2000; (2):87]. j 88. Мареев В. Ю., Фомин И. В., Агеев Ф. Т., Арутюнов Г. П., Беграмбекова Ю. ОБЗОРЫ § Л., Васюк Ю. А. и др. Хроническая сердечная недостаточность (ХСН). Журнал Сердечная Недостаточность. 2017;18 (1):3–40. DOI:10.18087 / rhfj.2017.1.2346 [Mareev V. Yu., Fomin I. V., Ageev F. T., Arutyunov G. P., Begrambekova Yu. L., Vasyuk Yu. A. i dr. Xronicheskaya serdechnaya nedostatochnost` (XSN). Zhurnal Serdechnaya Nedostatochnost`. 2017;18 (1):3– 40. DOI:10.18087 / rhfj.2017.1.2346]. 77. Tobin SW, Hashemi S, Dadson K, Turdi S, Ebrahimian K, Zhao J et al. Heart Failure and MEF2 Transcriptome Dynamics in Response to β-Blockers. Sci Rep. 2017;7 (1):4476. DOI:10.10 38 / s41598‑017‑04762‑x. fj 89. Adedinsewo D, Xu J, Agasthi P, Oderinde A, Adekeye O, Sachdeva R et al. Effect of Digoxin Use Among Medicaid Enrollees With Atrial Fibrillation. Circ Arrhythm Electrophysiol. 2017;10 (5):e004573. DOI:10.1161 / CIRCEP.116.004573. 78. Hjalmarson A. Cardioprotection with beta-adrenoceptor blockers. Does lipophilicity matter? Basic Res Cardiol. 2000;95 Suppl 1: I41–45. 90. Vamos M, Erath JW, Hohnloser SH. Digoxin-associated mortality: a systematic review and meta-analysis of the literature. Eur Heart J. 2015;36 (28):1831–8. DOI:10.1093 / eurheartj / ehv143.ffi t 79. Swedberg K, Komajda M, Böhm M, Borer JS, Ford I, Dubost- Brama A et al. Ivabradine and outcomes in chronic heart failure (SHIFT): a randomised placebo-controlled study. The Lancet. 2010;376 (9744):875–85. DOI:10.1016 / S0140–6736 (10) 61198–1. 91. Ziff OJ, Lane DA, Samra M, Griffith M, Kirchhof P, Lip GYH et al. Safety and efficacy of digoxin: systematic review and meta-analy- sis of observational and controlled trial data. BMJ. 2015; h4451. DOI:10.1136 / bmj.h4451. 80. Corino VDA, Ulimoen SR, Enger S, Mainardi LT, Tveit A, Platonov PG. Rate-control drugs affect variability and irregula- rity measures of RR intervals in patients with permanent atrial fibrillation. J Cardiovasc Electrophysiol. 2015;26 (2):137–41. DOI:10.1111 / jce.12580. 92. Zener JC, Anggard EE, Harrison DC, Kalman SM. Persistence of digoxin effect in atrial fibrillation. JAMA. 1973;224 (2):239–41.i gfi 93. Rawles JM, Rowland E. Is the pulse in atrial fibrillation irregularly irregular? Br Heart J. 1986;56 (1):4–11. 81. Bloch Thomsen PE, Jons C, Raatikainen MJP, Moerch Joergensen R, Hartikainen J, Virtanen V et al. Long-term recording of car- diac arrhythmias with an implantable cardiac monitor in patients with reduced ejection fraction after acute myocardial infarction: the Cardiac Arrhythmias and Risk Stratification After Acute Myocardial Infarction (CARISMA) study. Circulation. 2010;122 (13):1258–64. DOI:10.1161 / CIRCULATIONAHA.109.902148.h 94. Vangeli RS, Naumov VG, Blank ML, Sergakova LM, Grigoriants RA. Anti-arrhythmia and arrhythmogenic actions of digoxin in ven- tricular rhythm disorders in patients with circulatory insufficiency. Ter Arkh. 1987;59 (9):81–6. ОБЗОРЫ § V., Shitov V. N., Kiktev V. G., Sapel`nikov O. V., Laty`pov R. S., Grishin I. R. i dr. Serdechnaya resinxroniziruyushhaya terapiya pri postoyannoj forme merczatel`noj aritmii. Kardiologiya. 2013;53 (5):43–9].f 51. de Vos CB, Pisters R, Nieuwlaat R, Prins MH, Tieleman RG, Coelen R-JS et al. Progression from paroxysmal to persistent atrial fibrilla- tion clinical correlates and prognosis. J Am Coll Cardiol. 2010;55 (8):725–31. DOI:10.1016 / j.jacc.2009.11.040. ( ) j j 52. Степина Е. В., Лукьянов М. М., Бичурина М. А., Бойцов С. А., Клочкова Н. Н., Белова Е. Н. и др. Структура сердечно-сосу- дистой патологии и медикаментозного лечения у больных фибрилляцией предсердий в сочетании с артериальной гипер- тонией, ишемической болезнью сердца, хронической сердечной недостаточностью: данные госпитального регистра Рекваза- Клиника. Профилактическая Медицина. 2016;19 (2–2):73–4 [Stepina E. V., Luk`yanov M. M., Bichurina M. A., Bojczov S. A., Klochkova N. N., Belova E. N. i dr. Struktura serdechno-sosudistoj patologii i medikamentoznogo lecheniya u bol`ny`x fibrillyacziej predserdij v sochetanii s arterial`noj gipertoniej, ishemicheskoj bolezn`yu serdcza, xronicheskoj serdechnoj nedostatochnost`yu: danny`e gospital`nogo registra Rekvaza-Klinika. Profilakticheskaya Mediczina. 2016;19 (2–2):73–4]. j g 66. Soloff LA, Zatuchni J. Clinical experiences with digoxin. Am Heart J. 1959;57 (5):674–83. 67. McAlister FA, Wiebe N, Ezekowitz JA, Leung AA, Armstrong PW. Meta-analysis: beta-blocker dose, heart rate reduction, and death in patients with heart failure. Ann Intern Med. 2009;150 (11):784–94. 68. Cadrin-Tourigny J, Shohoudi A, Roy D, Talajic M, Tadros R, Mondésert B et al. Decreased Mortality With Beta-Blockers in Patients With Heart Failure and Coexisting Atrial Fibrillation: An AF-CHF Substudy. JACC Heart Fail. 2017;5 (2):99–106. DOI:10.1016/j.jchf.2016.10.015. 69. Filippatos G, Farmakis D. How to Use Beta-Blockers in Heart Failure With Reduced Ejection Fraction and Atrial Fibrillation. J Am Coll Cardiol. 2017;69 (24):2897–2900. doi:10.1016/ j.jacc.2017.04.049. 53. Kruse I, Arnman K, Conradson TB, Rydén L. A comparison of the acute and long-term hemodynamic effects of ventricular inhibited and atrial synchronous ventricular inhibited pacing. Circulation. 1982;65 (5). 70. Poole-Wilson PA, Swedberg K, Cleland JGF, Di Lenarda A, Hanrath P, Komajda M et al. Comparison of carvedilol and meto- prolol on clinical outcomes in patients with chronic heart failure in the Carvedilol Or Metoprolol European Trial (COMET): randomised controlled trial. Lancet. 2003;362 (9377):7–13. DOI:10.1016 / S0140–6736 (03) 13800–7. 54. Khand AU, Rankin AC, Kaye GC, Cleland JG. Systematic review of the management of atrial fibrillation in patients with heart failure. Eur Heart J. 2000;21 (8):614–32. DOI:10.1053 / euhj.1999.1767. 365 ISSN 0022-9040. Кардиология. 2017;57(S2) ОБЗОРЫ § 95. Rathore SS, Curtis JP, Wang Y, Bristow MR, Krumholz HM. Association of serum digoxin concentration and outcomes in patients with heart failure. JAMA. 2003;289 (7):871–8. 96. Nedostup A. V., Blagova O. V., Bogdanova E. A., Platonova A. A. Medication reducing therapy of atrial fibrillation: a new approach to an old problem. Ter Arkh. 2002;74 (8):35–42.fi 82. Ruwald A-CH, Bloch Thomsen PE, Gang U, Jørgensen RM, Huikuri HV, Jons C. New-onset atrial fibrillation predicts malig- nant arrhythmias in post-myocardial infarction patients -a Cardiac Arrhythmias and RIsk Stratification after acute Myocardial infarc- tion (CARISMA) substudy. Am Heart J. 2013;166 (5):855–863.e3. DOI:10.1016 / j.ahj.2013.08.017. p 97. Goldman S, Probst P, Selzer A, Cohn K. Inefficacy of ‘therapeutic’ serum levels of digoxin in controlling the ventricular rate in atrial fibrillation. Am J Cardiol. 1975;35 (5):651–5. i 98. Naumov V. G., Rossel’s A. N. [Clinical value of determination of digoxin concentration in blood during treatment of cardiac insuf- ficiency]. Kardiologiia. 1977;17 (9):34–9. 83. Нарусов О. В., Мареев В. Ю., Скворцов А. А. Клинические, гемодинамические и нейрогуморальные эффекты дигокси- на у больных с ХСН. ЖСН. 2000;1 (1):26–30 [Narusov O. V., Mareev V. Yu., Skvorczov A. A. Klinicheskie, gemodinamicheskie i nejrogumoral`ny`e e`ffekty` digoksina u bol`ny`x s XSN. ZhSN. 2000;1 (1):26–30]. i 99. Khand AU, Rankin AC, Martin W, Taylor J, Gemmell I, Cleland JGF. Carvedilol alone or in combination with digoxin for the manage- ment of atrial fibrillation in patients with heart failure? J Am Coll Cardiol. 2003;42 (11):1944–51. 84. Нарусов О. В., Мареев В. Ю., Скворцов А. А. Клинические, гемодинамические и нейрогуморальные эффекты дигоксина у больных с ХСН. Журнал Сердечная Недостаточность. 2000;1 (1):26–30 [Narusov OV, Mareev VYu, Skvorczov AA. Klinicheskie, gemodinamicheskie i nejrogumoral`ny`e e`ffekty` digoksina u bol`ny`x s XSN. Zhurnal Serdechnaya Nedostatochnost`. 2000;1 (1):26–30]. 100. Gheorghiade M, Patel K, Filippatos G, Anker SD, van Veldhuisen DJ, Cleland JGF et al. Effect of oral digoxin in high- risk heart failure patients: a pre-specified subgroup analysis of the DIG trial. European Journal of Heart Failure. 2013;15 (5):551–9. DOI:10.1093 / eurjhf / hft010. Материал поступил в редакцию 16/08/2017 366 ISSN 0022-9040. Кардиология. 2017;57(S2) ISSN 0022-9040. Кардиология. 2017;57(S2)
https://openalex.org/W4390144327	https://revistagastroperu.com/index.php/rgp/article/download/1602/1211	es		Colecistectomía como factor de riesgo para reflujo duodenogástrico	Revista de gastroenterología del Perú	2,023	cc-by	5,009	© 2023 Sociedad de Gastroenterología del Perú ART ÍC ULO ORIG INAL Colecistectomía como factor de riesgo para reflujo duodenogástrico Cholecystectomy as a risk factor for duodenogastric reflux Edgard Chávez-Mendoza1,2 , Victor Parra-Pérez1,3 1 2 3 Universidad Nacional Mayor de San Marcos. Lima, Perú. Servicio de Gastroenterología, Hospital San José del Callao. Lima, Perú. Servicio de Gastroenterología, Hospital Hipólito Unanue. Lima, Perú. Recibido: 05/09/2023 - Aprobado: 10/11/2023 RESUMEN El reflujo biliar o reflujo duodenogástrico (RDG) es el flujo retrógrado del contenido duodenal (principalmente bilis) hacia el estómago, capaz de producir daño químico a la mucosa y desencadenar mutaciones hacia el desarrollo de metaplasia intestinal, displasia e incluso, cáncer gástrico. Objetivo: El objetivo de este estudio fue estimar la prevalencia del reflujo biliar primario en pacientes colecistectomizados e identificar si la colecistectomía es un factor de riesgo para el RDG. Materiales y métodos: Se realizó un estudio observacional transversal analítico, en el que fueron incluidos todos los pacientes que se realizaron endoscopía digestiva alta entre febrero y junio del 2023, en un centro endoscópico privado en la ciudad de Lima, Perú. De acuerdo al reporte endoscópico, los pacientes fueron divididos en dos grupos: con RDG y sin RDG. Se analizaron estadísticamente las características demográficas, el antecedente de colecistectomía y los hallazgos endoscópicos. Fueron incluidos 408 pacientes. Resultados: La edad media de la población fue 48,18 ± 16,82años; el 61,52% fueron mujeres. La prevalencia de RDG fue de 25,74% en la población y de 52,11% en pacientes colecistectomizados. La prevalencia de RDG en pacientes colecistectomizados fue 2,58 veces en comparación a los pacientes sin colecistectomía (p< 0,001). La edad ≥50 años también se comportó como factor de riesgo para RDG (p=0,025). No hubo diferencias significativas respecto a diabetes, infección por Helicobacter pylori ni consumo de tabaco. Conclusión: En conclusión, el antecedente de colecistectomía y la edad demostraron ser factores de riesgo para el desarrollo de RDG primario. Palabras clave: Reflujo duodenogástrico; Factores de riesgo; Gastritis; Colecistectomía; Metaplasia (fuente: DeCS-Bireme). ABSTRACT Bile reflux or duodenogastric reflux (DGR), refers to the retrograde flow of duodenal contents (mainly bile) into the stomach; capable of producing chemical damage to the mucosa, and triggering mutations towards the development of intestinal metaplasia, dysplasia and even gastric cancer. Objective: This study aimed to estimate the prevalence of primary bile reflux in cholecystectomized patients and to identify whether cholecystectomy is a risk factor for development of DGR. Materials and methods: An analytical cross-sectional and observational study was conducted, in which all patients who underwent upper digestive endoscopy from February to June 2023 in a private endoscopic center in Lima, Peru, were included. According to the endoscopic report, patients were divided into two groups as those with DGR and those without DGR. Demographic characteristics, history of cholecystectomy, and endoscopic findings were statistically analyzed. 408 patients were included. Results: The mean age of the population was 48.18 ± 16.82 years; 61.52% were female. The prevalence of DGR was 25.74% in the population, while in cholecystectomized patients it was 52.11%. The prevalence of DRG in patients with a history of cholecystectomy was 2.58 times compared to patients without cholecystectomy (p<0.001). Age ≥50 years also behaved as a risk factor for RDG (p=0.025). No significant difference in diabetes, Helicobacter pylori infection or smoking were found. Conclusion: In conclusion, a history of cholecystectomy as well as age were found to be risk factors for development of primary DGR. Keywords: Duodenogastric reflux; Risk factors; Gastritis; Cholecystectomy; Metaplasia (source: MeSH NLM). INTRODUCCIÓN El reflujo biliar o reflujo duodenogástrico (RDG) se define como el flujo retrógrado del contenido duodenal, incluida la bilis y jugo pancreático, hacia el estómago; fenómeno que puede conducir a la ruptura de la barrera mucosa y causar daño químico directo al epitelio gástrico (1-3). Fue descrito por primera vez por Beaumont, en 1883, en un paciente con fístula gastrocutánea (4); pero recién empieza a cobrar importancia con el desarrollo creciente de las cirugías gástricas que alteran el píloro (5,6). Derwr y col describieron los cambios patológicos asociados al reflujo biliar en pacientes sometidos a gastrectomía e introdujeron el término de “gastritis por reflujo biliar” (1,7). El RDG fisiológico puede presentarse en las primeras horas de la mañana, en el período posprandial y después del ayuno prolongado. Cuando se hace más intenso, Citar como: Chávez-Mendoza E, Parra-Pérez V. Colecistectomía como factor de riesgo para reflujo duodenogástrico. Rev Gastroenterol Peru. 2023;43(4):334-40. doi: 10.47892/rgp.2023.434.1602. 334 Rev Gastroenterol Peru. 2023;43(4):334-40 https://doi.org/10.47892/rgp.2023.434.1602 Colecistectomía como factor de riesgo para reflujo duodenogástrico Chávez-Mendoza E, et al. puede producir síntomas clínicos, cambios endoscópicos e histológicos, convirtiéndose en RDG patológico, causante de la “gastritis por reflujo biliar” o ”gastritis alcalina” (8,9). Existen barreras anatómicas (píloro y ángulo bulboduodenal) y funcionales (motilidad antroduodenal y pilórica) que restringen el paso de la bilis hacia el estómago, clasificando al RDG según su mecanismo en dos tipos: RDG primario y secundario. El reflujo biliar secundario ocurre después de una cirugía gástrica o biliar, mientras que el reflujo primario se presenta en ausencia de la misma (3,9). colecistectomía es un factor de riesgo para el desarrollo de RDG (1,13,25). La gastritis por reflujo biliar representa aproximadamente el 22,6% de las gastritis crónicas (10). La prevalencia general de RDG es de 17 a 24%, pudiendo llegar hasta 60-78% en pacientes con factores de riesgo (11-13). Diversos estudios han investigado potenciales factores de riesgo asociados a reflujo biliar primario, entre ellos: edad, diabetes mellitus, obesidad, consumo de tabaco, antecedente de colecistectomía e incluso factores psicológicos; sin embargo, el RDG primario continúa siendo un fenómeno poco conocido (10). La evaluación diagnóstica del RDG es muy importante, pues existe evidencia creciente de que los ácidos biliares pueden desencadenar mutaciones en la replicación celular y progresar hacia el desarrollo de metaplasia intestinal gástrica (MIG), displasia, e incluso cáncer gástrico (CG) (3,8,14,15). Los síntomas más comunes del RDG son: náuseas, vómitos biliosos, dolor abdominal y sensación de plenitud, pero no son específicos (9,16). Se han descrito varios métodos para el diagnóstico de RDG, sin embargo, hasta la actualidad, no existe un estándar de oro, ya que ninguno de ellos es totalmente confiable, teniendo cada uno sus propias ventajas y desventajas (8,17). En nuestro país, la endoscopía digestiva alta es el método más accesible para el diagnóstico de RDG, y ha sido utilizada en varios estudios internacionales (3,11,15,18,19). Los hallazgos endoscópicos incluyen: eritema, pliegues gástricos engrosados y presencia de bilis en el estómago. El reflujo biliar endoscópico es uno de los principales criterios diagnósticos de gastritis por reflujo biliar (10); sin embargo, los cambios histológicos no son específicos (9). La endoscopía tiene la ventaja de diagnosticar otras patologías gástricas y, además, por medio de la toma de biopsia, permite la detección de lesiones preneoplásicas como la MIG y de la infección por Helicobacter pylori (HP) (16,20). La colecistectomía es el tratamiento de elección para la colelitiasis sintomática, y a la vez, una de las cirugías más frecuentemente realizadas a nivel mundial (21). La mayoría de los pacientes experimenta una resolución de sus síntomas después de la operación; sin embargo, el 15 a 20% pueden presentar molestias como náuseas, vómitos biliosos y síntomas tipo dispepsia, fenómeno conocido como “síndrome poscolecistectomía”, lo cual se atribuye en gran parte al RDG excesivo (22,23). Un estudio en Egipto reportó una prevalencia de gastritis por reflujo biliar de 61,8% en pacientes poscolecistectomizados, comparado con 16,7% en pacientes no operados (13,24). Estudios en México, Turquía e Irak también demostraron que la https://doi.org/10.47892/rgp.2023.434.1602 A pesar de que fue descrito desde hace muchos años, y de los estudios que evidencian su asociación con el antecedente de colecistectomía; hasta la fecha, la etiología y los factores de riesgo del reflujo biliar primario no han sido bien comprendidos, y no existe un tratamiento 100% efectivo (10,26). La incidencia del RDG ha aumentado gradualmente en las últimas décadas en pacientes sin antecedente de cirugía gástrica (9,15); sin embargo, en nuestro país no existen investigaciones al respecto, a pesar de la evidencia creciente de los efectos nocivos de los ácidos biliares, y su potencial riesgo de desarrollar CG (15,19,27). El objetivo del presente estudio es determinar la prevalencia del reflujo biliar primario endoscópico, e identificar si la colecistectomía es un factor de riesgo para su desarrollo. MATERIALES Y MÉTODOS Se realizó un estudio observacional, transversal analítico, que incluyó a todos los pacientes ambulatorios con síntomas digestivos e indicación de endoscopía digestiva alta, que acudieron al Centro endoscópico privado Gastromedic, en la ciudad de Lima, Perú, durante el período comprendido entre febrero y junio del 2023. El estudio fue aprobado por el Comité de Ética de la Universidad Nacional Mayor San Marcos. Es necesario recalcar que se cumplieron todos los principios éticos delineados en la Declaración de Helsinki. Los criterios de inclusión fueron: Pacientes de ambos sexos que se realizaron endoscopía alta con biopsia, entre los meses de febrero y junio del 2023, y que firmaron el consentimiento informado para participar en el estudio. Los criterios de exclusión fueron: pacientes con antecedentes personales de cirugía gastrointestinal o de vías biliares, enfermedades mentales o enfermedades crónicas descompensadas, embarazo, diagnóstico histológico de adenocarcinoma gástrico y aquellos que no firmaron el consentimiento informado. Se diseñó una encuesta a cada participante previo a la realización de la endoscopía, donde se registró la siguiente información: edad, sexo, antecedentes personales de diabetes, consumo de alcohol, consumo de tabaco y antecedente de colecistectomía. El principio de selección de los factores de riesgo potenciales mencionados anteriormente, dependió principalmente de la literatura previa y de nuestra experiencia en la práctica clínica (1,10,19,24). Los estudios endoscópicos estuvieron a cargo de dos gastroenterólogos con amplia experiencia en endoscopía. Fueron realizados bajo sedación consciente (MidazolamPropofol) y con un mínimo de 6 horas de ayuno. Se utilizaron equipos de videoendoscopía de luz blanca marca Olympus modelo CV-160®. El diagnóstico de RDG se realizó en base a las características del contenido biliar en la cavidad gástrica durante la endoscopía y se clasificó tomando como referencia dos estudios recientemente Rev Gastroenterol Peru. 2023;43(4):334-40 335 Chávez-Mendoza E, et al. Colecistectomía como factor de riesgo para reflujo duodenogástrico realizados (15,19) en: ausente (lago mucoso claro), grado 1 (amarillo claro), grado 2 (verde amarillento) y grado 3 (amarillo oscuro y turbio). A todos los pacientes se les realizó toma de biopsia, siguiendo el protocolo de Sidney actualizado (2 de antro: curvatura menor y mayor, 2 de cuerpo: curvatura menor y mayor y 1 de incisura angularis) (14) . El diagnóstico de infección por HP, fue realizado en base a la biopsia. En caso de sospecha de neoplasia maligna se tomaron 8 biopsias y fueron excluidos del estudio. Los datos del cuestionario, así como el reporte endoscópico, con el respectivo informe del grado de reflujo biliar y el informe anatomopatológico, fueron registrados en la ficha de recolección de datos. Una vez terminada la recolección de datos, se clasificó a los pacientes en dos grupos: con RDG y sin RDG, procediéndose con el análisis estadístico. Análisis estadístico En el análisis univariado, las variables numéricas fueron expresadas en media +- desviación estándar, o mediana dependiendo la distribución de los datos que fueron evaluados mediante la prueba de Shapiro–Wilk e histograma. Por otro lado, las variables categóricas fueron expresadas en número y porcentajes o frecuencias. En el análisis bivariado, se evaluó la asociación entre la variable dependiente reflujo duodenogástrico y la variable independiente colecistectomía, mediante la prueba de chi cuadrado (X2) o exacta de Fisher, a un nivel de confianza de 95%, previa evaluación de los supuestos estadísticos. Los resultados fueron considerados como estadísticamente significativos si p<0,05. En el análisis bivariado, la prevalencia de RDG en pacientes operados de vesícula fue de 52,11%, mientras que en los no operados fue de 20,18%. Fueron factores asociados a la presencia de reflujo biliar la edad mayor o igual a 50 años (p=0,025) y el antecedente de colecistectomía (p<0,001). Las demás variables no tuvieron diferencia estadísticamente significativa. Tabla 2, Figura 1. Para determinar la fuerza de asociación se realizó mediante un modelo lineal generalizable (GLM) donde se determinó las razones de prevalencia (RP) con sus respectivos intervalos de confianza al 95%. La prevalencia de tener reflujo biliar en varones fue 1,02 veces mayor en comparación a las mujeres y no fue estadísticamente significativo (p=0,890). Tabla 1. Características de la población. Variables n (%) Sexo Femenino 251 (61,52) Masculino 157 (38,48) Edad (años) 48,18 ± 16,82 Edad categorizada <50 años 221 (54,17) ≥ 50 años 187 (45,83) Consumo de tabaco No 352 (86,27) Si, antes 31 (7,60) Sí, ahora 25 (6,13) El análisis estadístico se realizó con el paquete estadístico STATA v 14. Consumo de alcohol No 306 (75) Aspectos éticos Si, antes 64 (15,69) Sí, ahora 25 (6,13) Se obtuvo la aprobación del Comité de Ética de la Universidad Nacional Mayor San Marcos y el permiso de la institución para la realización del estudio. Los pacientes que ingresaron al estudio firmaron consentimiento informado. Diabetes No 371 (90,93) Sí 37 (9,07) Helicobacter pylori RESULTADOS Fueron evaluados inicialmente un total de 448 participantes que se realizaron endoscopía digestiva alta entre febrero y junio del 2023, de los cuales fueron excluidos 40 pacientes por no cumplir con los criterios de inclusión y/o tener al menos un criterio de exclusión, por lo que el grupo de estudio quedó conformado por 408 pacientes. La edad media de la población fue 48,18 ± 16,82 años. Respecto al género, fueron 251 (61,52%) mujeres. De los 408 pacientes estudiados se encontró RDG en 105 (25,74%) y ausencia del mismo en 303 (74,26%). Tuvieron antecedente de colecistectomía 71 (17,40%). El resto de características de la población estudiada se describe en la Tabla 1. 336 Rev Gastroenterol Peru. 2023;43(4):334-40 No 158 (38,97) 1+ 136 (33,33) 2+ 69 (16,91) 3+ 44 (10,78) Colecistectomía No 337 (82,60) Sí 71(17,40) Reflujo biliar No 303 (74,26) Grado I 82 (20,10) Grado II 23 (5,64) https://doi.org/10.47892/rgp.2023.434.1602 Chávez-Mendoza E, et al. Colecistectomía como factor de riesgo para reflujo duodenogástrico Tabla 2. Análisis bivariado de los factores de riesgo relacionado a reflujo biliar. Reflujo biliar Variables No (n=303) Sí (n=105) Sexo Valor de p 0,890 Femenino 187 (74,50) 64 (25,50) Masculino 116 (73,89) 41(26,11) Edad (años)β 45 (34 - 59) 53 (36 - 67) Edad categorizada 0,080 0,025* <50 años 174 (78,73) 47 (21,27) ≥ 50 años 129 (68,98) 58 (31,02) No 264 (75) 88 (25) Sí, antes 23 (74,19) 8 (25,81) Sí ahora 16 (64) 9 (36) Consumo de tabaco 0,478 Consumo de alcohol 0,318 No 223 (72,88) 83 (27,12) Sí antes 48 (75) 16 (25) Sí ahora 32 (84,21) 6 (15,79) Diabetes 0,560 No 277 (74,66) 94 (25,34) Sí 26 (70,27) 11 (29,73) No 118 (74,21) 41 (25,79) Sí 185 (74,30) 64 (25,70) No 269 (79,82) 68 (20,18) Sí 34 (47,89) 37 (52,11) 0,985 Helicobacter pylori Colecistectomía <0,001* β: Los datos fueron expresados en mediana (p25-p75); Prueba T de Student. * Se considero estadísticamente significativo, un valor p<0,05 90 80 La prevalencia de tener reflujo biliar en pacientes ≥ 50 años fue 1,45 veces mayor en comparación a las pacientes < 50 años y fue estadísticamente significativo (p=0,025). 79,82 70 Porcentaje (%) 60 52,11 47,89 50 40 30 20,18 20 10 0 No Si Colecistectomía Reflujo biliar No Reflujo biliar Sí Figura 1. Proporción de pacientes colecistectomizados con y sin reflujo duodenogástrico. https://doi.org/10.47892/rgp.2023.434.1602 La prevalencia de tener reflujo biliar en pacientes con colecistectomía fue 2,58 veces en comparación a los pacientes sin colecistectomía y fue estadísticamente significativo (p<0,001). Ver Tabla 3. La prevalencia de tener reflujo biliar en pacientes con Helicobacter pylori fue 1,07 veces en comparación a los pacientes sin Helicobacter pylori ajustado por la variable grupo etario, diabetes y colecistectomía y no fue estadísticamente significativo (p=0,644). La prevalencia de tener reflujo biliar en pacientes con colecistectomía fue 2,47 veces en comparación a los pacientes sin colecistectomía ajustado por la variable, grupo etario, diabetes y Helicobacter pylori y fue estadísticamente significativo (p<0,001). Ver Tabla 4. Rev Gastroenterol Peru. 2023;43(4):334-40 337 Colecistectomía como factor de riesgo para reflujo duodenogástrico Chávez-Mendoza E, et al. Tabla 3. Razones de prevalencias crudas para características de reflujo biliar. Tabla 4. Razones de prevalencias ajustadas para características de reflujo biliar. Variables Análisis crudo RP (95% IC) Valor p Sexo Femenino Referencia Masculino 1,02 (0,73 -1,43) Referencia ≥ 50 años 1,45 (1,04 - 2,03) 0,890 Referencia Sí 1,17 (0,69 - 1,98) 0,026* Referencia Sí 0,99 (0,71 - 1,39) 0,551 Referencia Sí 2,58 (1,89 - 3,51) 0,985 <0,001* * Se considero estadísticamente significativo, un valor p<0.05 RP: Riesgo de prevalencia; IC: Intervalo de confianza DISCUSIÓN Es conocida la relación entre el RDG secundario y el antecedente de cirugía gástrica o de vías biliares; sin embargo, el RDG primario continúa siendo hasta la actualidad un fenómeno poco comprendido (9). El presente estudio, realizado en una institución privada, demostró asociación entre la colecistectomía y la edad mayor o igual a 50 años, como factores de riesgo para RDG primario endoscópico. No se encontró una correlación significativa entre sexo, infección por HP, diabetes, consumo de alcohol, consumo de tabaco, y RDG primario. De acuerdo a los reportes internacionales, la prevalencia general de RDG oscila entre 17 y 24%. Un estudio que incluyó 804 casos en niños, en los que se realizó endoscopía alta por dolor abdominal, encontró reflujo biliar en el 23,9% de pacientes (11), mientras que en nuestro estudio, que incluyó 408 pacientes, la prevalencia global fue de 25,74%, ligeramente mayor al promedio. Se sabe que en ciertos grupos de riesgo, la prevalencia de RDG puede aumentar considerablemente, como en el caso de pacientes diabéticos y operados de vesícula (24). Nuestro estudio demostró que la edad mayor o igual a 50 años es un factor de riesgo para RDG, similar al estudio de Chen et al., quien encontró que la edad mayor de 45 años es un factor de riesgo para RDG (10). Estos datos contrastan con el estudio de Alnajjar (25), que reportó una mayor prevalencia de RDG en mujeres jóvenes, al igual que lo descrito por la literatura (28) . En nuestro estudio no hubo asociación entre sexo y RDG, similar a lo reportado por Othman et al. (29). Sin 338 Rev Gastroenterol Peru. 2023;43(4):334-40 <50 años Referencia ≥50 años 1,28 (0,92 - 1,78) 0,133 No Referencia Sí 1 (0,61 - 1,64) 0,994 No Referencia Sí 1,07 (0,78 - 1,47) 0,644 Colecistectomía Colecistectomía No Valor p Helicobater pylori Helicobater pylori No RP (95% IC) Diabetes Diabetes No Análisis ajustado Grupo etario Edad categorizada <50 años Variables No Referencia Sí 2,47 (1,80 - 3,39) <0,001* * Se considero estadísticamente significativo, un valor p<0,05 RP: Riesgo de prevalencia; IC: Intervalo de confianza embargo, otro estudios sí mostraron diferencias, como en el caso de Chen et al, quien demostró que el sexo masculino es un factor de riesgo para RDG primario (10), contrastando a Dan Li et al. (19) , quien reporta una tasa de RDG mucho mayor en mujeres que en hombres, acorde a lo descrito por la literatura (28), lo cual es comprensible, ya que el sexo femenino es un factor de riesgo para la colelitiasis, y ésta es a su vez es considerada como factor de riesgo para RDG (10,28). Diversos estudios han demostrado que el RDG puede aumentar hasta un 60 a 78% en pacientes operados de vesícula. Othman et al. encontró una prevalencia de RDG de 61,8% en pacientes colecistectomizados (29), una tasa mayor en comparación a nuestro estudio, que fue de 52,11%; por otro lado, Mercan et al. (16) reportó una prevalencia de RDG de 78% después de la colecistectomía, frente a 48% antes de la misma. Capi-Rizo et al. encontró que el 40,6% de pacientes con RDG presentaban el antecedente de colecistectomía, frente al 9,5% en los no operados (1). La mayor prevalencia encontrada en los estudios internacionales probablemente esté en relación al tipo de población, ya que fueron realizados en hospitales, donde generalmente se atiende a pacientes con mayor comorbilidad y posiblemente presenten varios factores asociados a RDG. La pérdida de la función de reservorio de la vesícula en pacientes operados, produce una alteración en el patrón cíclico de excreción de la bilis; mientras que la pérdida de respuesta humoral causa una alteración en la motilidad antroduodenal y pilórica, aumentando el RDG (13). El reflujo biliar tiene efectos nocivos en la mucosa gástrica, que comienzan 2 a 6 meses después de la colecistectomía. Los ácidos biliares, productos del metabolismo del colesterol, son sintetizados en el hígado (ácidos biliares primarios) y luego transformados por las bacterias intestinales (ácidos https://doi.org/10.47892/rgp.2023.434.1602 Colecistectomía como factor de riesgo para reflujo duodenogástrico Chávez-Mendoza E, et al. biliares secundarios). Generalmente, los ácidos biliares hidrófobos son citotóxicos, y ejercen sus efectos fisiológicos al unirse a receptores de membrana nuclear, induciendo estrés oxidativo y daño al ácido desoxirribonucleico (ADN), lisis de las membranas celulares, y daño tisular, pudiendo progresar hasta una gastritis atrófica, MIG y CG (27) . en cuanto a sintomatología y comorbilidad. A pesar de las limitaciones descritas, creemos que este estudio debe ser un punto de partida para futuras investigaciones en nuestro país. Es probable que estudios multicéntricos a largo plazo nos puedan brindar más claridad sobre la causalidad del RDG primario. Othman reportó en un estudio retrospectivo, que la diabetes y la obesidad son factores de riesgo para el reflujo biliar en pacientes colecistectomizados (24,29). En nuestro estudio, la diabetes no demostró esta asociación, probablemente debido al corto período de estudio y al tiempo de evolución de la enfermedad, pues se sabe que la gastroparesia diabética ,que es la causante del RDG, se presenta en pacientes con diabetes de larga data (28) . Sería importante valorar resultados en un mayor período de tiempo y considerar el tiempo de evolución de la diabetes. En conclusión, los resultados de este estudio demostraron que la colecistectomía y la edad mayor o igual a 50 años, pueden ser factores de riesgo asociados a RDG primario. Teniendo en cuenta que el RDG es un potencial inductor para el desarrollo de lesiones precancerosas y cáncer gástrico (15), consideramos de suma importancia investigar, además de la colecistectomía, otros probables factores asociados, como la obesidad, la diabetes y la colelitiasis. Los resultados de los estudios que investigan la relación entre HP y el reflujo biliar son contradictorios. Asfeldt et al. (30) reportó que la colonización por HP disminuye como resultado de ácidos biliares elevados , contrastando a Li et al. (19), quien demostró que el RDG aumenta la presencia del HP. En nuestro estudio no se demostró asociación entre HP y RDG. Sería muy importante que se realicen mayores investigaciones al respecto, considerando que ambos son los principales inductores de MIG y CG. Es importante reconocer algunas limitaciones de nuestro estudio. En primer lugar, el diagnóstico de RDG fue realizado sólo basándose en el hallazgo de bilis en la endoscopía. Un estudio que utilizó un segundo método para confirmar el RDG, reportó que el grado 1 de reflujo biliar endoscópico no es representativo de RDG, mientras que el grado 3 se asoció a un excesivo RDG (16), lo cual debe ser tomado en cuenta para estudios posteriores. La gammagrafía hepática, la impedanciometría de pH de 24 horas, y el monitoreo de bilirrubina de 24 horas (Bilitec 2000), son otros métodos conocidos, pero ninguno es lo suficientemente válido para el diagnóstico de RDG, ni están a nuestro alcance (8,17), lo que constituye una limitación importante. El presente estudio también estuvo limitado por las características de los estudios transversales, en los cuales, tanto la variable desenlace como exposición son medidas simultáneamente, sin poderse establecer una adecuada relación de temporalidad ni causalidad. Es posible que algunos pacientes hayan presentado reflujo biliar desde antes de la colecistectomía y recién se haya detectado al momento de la endoscopía. Por ello, sería importante considerar en estudios posteriores un diagnóstico de reflujo biliar, previo y posterior a la cirugía de vesícula. Si bien la población fue de 408 pacientes, el período de recolección de datos fue sólo de 5 meses. Por último, los estudios fueron realizados en un centro privado con pacientes ambulatorios, pudiendo ser una subrepresentación de la realidad, ya que el tipo de población del hospital es distinta a la de un centro privado, https://doi.org/10.47892/rgp.2023.434.1602 Conflicto de intereses: Los autores declaran no tener conflicto de intereses de ningún tipo. Financiamiento: El trabajo se solventó con recursos propios. Los autores declaran no haber recibido financiación económica de ningún tipo. Contribución de los autores: EC y VP contribuyeron igualmente en lo siguiente: adquisición de datos, análisis de datos y redacción del manuscrito. EC fue responsable de la concepción y diseño del estudio, interpretación de datos y revisión crítica del manuscrito. Ambos autores aprobaron la versión final del manuscrito. REFERENCIAS BIBLIOGRÁFICAS 1. 2. 3. 4. 5. 6. 7. 8. 9. Rizo CRC, García F, Fernández SZ, Sánchez DV. Gastritis alcalina y reflujo duodenogástrico en pacientes colecistectomizados. Endoscopia. 2018;30(2):55-9. Kuran S, Parlak E, Aydog G, Kacar S, Sasmaz N, Ozden A, et al. Bile reflux index after therapeutic biliary procedures. BMC Gastroenterol. 2008;8:4. doi: 10.1186/1471-230X-8-4. Szőke A, Mocan S, Negovan A. Helicobacter pylori infection over bile reflux: No influence on the severity of endoscopic or premalignant gastric lesion development. Exp Ther Med. 2021;22(1):766. doi: 10.3892/etm.2021.10198. Beaumont W. Nutrition Classics. Experiments and observations on the gastric juice and the physiology of digestion. By William Beaumont. Plattsburgh. Printed by F. P. Allen. 1833. Nutr Rev. 1977;35(6):144-5. doi: 10.1111/j.1753-4887.1977. tb06570.x. Madura JA. Primary bile reflux gastritis: diagnosis and surgical treatment. Am J Surg. 2003;186(3):269-73. doi: 10.1016/ s0002-9610(03)00213-7. Rojas Díaz RO, Pérez Sosa JA, Contreras Ruiz Velasco R. Reflujo duodenogástrico. Definición y diagnóstico. Cir Gen. 2015;37(3-4):91-8. Hinojosa Ruiz A, Valdés Lías R, Corral Medina A, García Guerrero VA, Zárate Guzmán AM, Majalca Martínez C, et al. Correlación de gastropatía reactiva biliar con antecedentes de colecistectomía y grupo control. Endoscopía. 2008;20(4):277-84. Livzan MA, Mozgovoi SI, Gaus OV, Bordin DS, Kononov AV. Diagnostic Principles for Chronic Gastritis Associated with Duodenogastric Reflux. Diagnostics (Basel). 2023;13(2):186. doi: 10.3390/diagnostics13020186. McCabe ME, Dilly CK. New Causes for the Old Problem of Bile Reflux Gastritis. Clinical Gastroenterology and Hepatology. 2018;16(9):1389-92. doi: 10.1016/j.cgh.2018.02.034. Rev Gastroenterol Peru. 2023;43(4):334-40 339 Chávez-Mendoza E, et al. Colecistectomía como factor de riesgo para reflujo duodenogástrico 10. Chen L, Zhu G, She L, Ding Y, Yang C, Zhu F. Analysis of Risk Factors and Establishment of a Prediction Model for Endoscopic Primary Bile Reflux: A Single-Center Retrospective Study. Front Med (Lausanne). 2021;8:758771. doi: 10.3389/ fmed.2021.758771. 11. Taşcı EK, Karakoyun M, Sezak M, Doğanavsargil B, Çetin F, Aydoğdu S. Does bile reflux reduce Helicobacter pylori gastritis? Turk J Pediatr. 2022;64(1):122-6. doi: 10.24953/turkjped.2020.2839. 12. Basnayake C, Geeraerts A, Pauwels A, Koek G, Vaezi M, Vanuytsel T, et al. Systematic review: duodenogastroesophageal (biliary) reflux prevalence, symptoms, oesophageal lesions and treatment. Aliment Pharmacol Ther. 2021;54(6):755-778. doi: 10.1111/apt.16533. 13. Mercan E, Duman U, Tihan D, Dilektasli E, Senol K. Cholecystectomy and duodenogastric reflux: interacting effects over the gastric mucosa. Springerplus. 2016;5(1):1970. doi: 10.1186/ s40064-016-3641-z. 14. Dixon MF, Genta RM, Yardley JH, Correa P. Classification and grading of gastritis. The updated Sydney System. International Workshop on the Histopathology of Gastritis, Houston 1994. Am J Surg Pathol. 1996;20(10):1161-81. doi: 10.1097/00000478-199610000-00001. 15. Zhang LY, Zhang J, Li D, Liu Y, Zhang DL, Liu CF, et al. Bile reflux is an independent risk factor for precancerous gastric lesions and gastric cancer: An observational cross-sectional study. J Dig Dis. 2021;22(5):282-90. doi: 10.1111/1751-2980.12986. 16. Chang WK, Lin CK, Chuan DC, Chao YC. Duodenogastric Reflux: Proposed New Endoscopic Classification in Symptomatic Patients. J Med Sci [Internet]. 2016 [citado 2 de julio de 2023];36(1):1-5. Disponible en: https://journals.lww.com/ joms/Fulltext/2016/36010/Duodenogastric_Reflux__Proposed_New_Endoscopic.1.aspx 17. Eldredge TA, Myers JC, Kiroff GK, Shenfine J. Detecting Bile Reflux-the Enigma of Bariatric Surgery. Obes Surg. 2018;28(2):559-66. doi: 10.1007/s11695-017-3026-6. 18. Agin M, Kayar Y. The Effect of Primary Duodenogastric Bile Reflux on the Presence and Density of Helicobacter pylori and on Gastritis in Childhood. Medicina (Kaunas). 2019;55(12):775. doi: 10.3390/medicina55120775. 19. Li D, Zhang J, Yao WZ, Zhang DL, Feng CC, He Q, et al. The relationship between gastric cancer, its precancerous lesions and bile reflux: A retrospective study. J Dig Dis. 2020;21(4):222229. doi: 10.1111/1751-2980.12858. 20. Pimentel-Nunes P, Libânio D, Marcos-Pinto R, Areia M, Leja M, Esposito G, et al. Management of epithelial precancerous conditions and lesions in the stomach (MAPS II): European Society of Gastrointestinal Endoscopy (ESGE), European Helicobacter and Microbiota Study Group (EHMSG), European Society of Pathology (ESP), and Sociedade Portuguesa de Endoscopia Digestiva (SPED) guideline update 2019. Endoscopy. 2019 Apr;51(4):365-388. doi: 10.1055/a-0859-1883. 340 Rev Gastroenterol Peru. 2023;43(4):334-40 21. Sami-Bostan M, Ugurlu C, Yildirim M, Koca B, Sami-Bostan M, Ugurlu C, et al. The relationship between bile reflux and common bile duct diameter after cholecystectomy: a clinical case-control study. Cir Cir. 2023;91(1):9-14. doi: 10.24875/ ciru.22000490. 22. Wilson P, Jamieson JR, Hinder RA, Anselmino M, Perdikis G, Ueda RK, et al. Pathologic duodenogastric reflux associated with persistence of symptoms after cholecystectomy. Surgery. 1995;117(4):421-8. doi: 10.1016/s0039-6060(05)80062-5. 23. Manifold DK, Anggiansah A, Owen WJ. Effect of cholecystectomy on gastroesophageal and duodenogastric reflux. Am J Gastroenterol. 2000;95(10):2746-50. doi: 10.1111/j.15720241.2000.02298.x. 24. Othman AA, Dwedar AA, ElSadek HM, AbdElAziz HR, Abdelrahman AA. Post-cholecystectomy bile reflux gastritis: Prevalence, risk factors, and clinical characteristics. Chronic Illn. 2023;19(3):529-538. doi: 10.1177/17423953221097440. 25. Alnajjar AS, Al‑Bayat S. Evaluation of the Gastrointestinal Clinical, Endoscopic, and Histological Findings in Patients with Bile Reflux Diseases: A Cross‑Sectional Study. Mustansiriya Med J. 2019;18(1):10-5. 26. Murillo-Godínez G. Reflujo duodenogástrico (biliar, alcalino). Med Int Mex. 2023;39(1):127-40. 27. Qu X, Shi Y. Bile reflux and bile acids in the progression of gastric intestinal metaplasia. Chin Med J (Engl). 2022;135(14):166472. doi: 10.24245/mim.v39i1.5120. 28. Shi X, Chen Z, Yang Y, Yan S. Bile Reflux Gastritis: Insights into Pathogenesis, Relevant Factors, Carcinomatous Risk, Diagnosis, and Management. Gastroenterol Res Pract. 2022;2022:2642551. doi: 10.1155/2022/2642551. 29. Othman AAA, Dewedar AAZ, ElSadek HM, AbdelAziz HR, AdelRahman AAF. Do obesity and diabetes increase the frequency and risk of bile reflux gastritis post-cholecystectomy? Pol Przegl Chir. 2022;94(5):1-8. doi: 10.5604/01.3001.0015.6980. 30. Asfeldt AM, Steigen SE, Løchen ML, Straume B, Johnsen R, Bernersen B, et al. The natural course of Helicobacter pylori infection on endoscopic findings in a population during 17 years of follow-up: the Sørreisa gastrointestinal disorder study. Eur J Epidemiol. 2009;24(10):649-58. doi: 10.1007/s10654009-9371-6. Correspondencia: Edgard Mariano Chávez Mendoza. Dirección: Av. Alberto Alexander No 2427, Dpto 102, Lince, Lima, Perú. Código postal: 15073. Teléfono: (+51) 994544962. E-mail: edgard888@hotmail.com https://doi.org/10.47892/rgp.2023.434.1602
https://openalex.org/W3183492042	http://www.jiem.org/index.php/jiem/article/download/3446/981	English	null	A machine learning approach to enable bulk orders of critical spare-parts in the shipping industry	Journal of industrial engineering and management	2,021	cc-by	9,558	Abstract: Purpose: The main purpose of this paper is to propose a methodological approach and a decision support tool, based on prescriptive analytics, to enable bulk ordering of spare parts for shipping companies operating fleets of vessels. The developed tool utilises Machine Learning (ML) and operations research algorithms, to forecast and optimize bulk spare parts orders needed to cover planned maintenance requirements on an annual basis and optimize the company’s purchasing decisions. Design/methodology/approach: The proposed approach consists of three discrete methodological steps, each one supported by a decision support tool based on clustering and Machine Learning (ML) algorithms. In the first step, clustering is applied in order to identify high interest items. Next, a forecasting tool is developed for estimating the expected needs of the fleet and to test whether the needed quantity is influenced by the source of purchase. Finally, the selected items are cost-effectively allocated to a group of vendors. The performance of the tool is assessed by running a simulation of a bulk order process on a mixed fleet totaling 75 vessels. Findings: The overall findings and approach are quite promising Indicatively, shifting demand planning focus to critical spares, via clustering, can reduce administrative workload. Furthermore, the proposed forecasting approach results in a Mean Absolute Percentage Error of 10% for specific components, with a potential for further reduction, as data availability increases. Finally, the cost optimizer can prescribe spare part acquisition scenarios that yield a 9% overall cost reduction over the span of two years. Originality/value: By adopting the proposed approach, shipping companies have the potential to produce meaningful results ranging from soft benefits, such as the rationalization of the workload of the purchasing department and its third party collaborators to hard, quantitative benefits, such as reducing the cost of the bulk ordering process, directly affecting a company’s bottom line. Keywords: shipping industry, spare parts management, bulk ordering, machine learning, forecasting Anglou, F.Z., Ponis, S., & Spanos, A. (2021). A machine learning approach to enable bulk orders of critical spare-parts in the shipping industry. Journal of Industrial Engineering and Management, 14(3), 604-621. https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management JIEM, 2021 – 14(3): 604-621 – Online ISSN: 2013-0953 – Print ISSN: 2013-8423 https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management JIEM, 2021 – 14(3): 604-621 – Online ISSN: 2013-0953 – Print ISSN: 2013-8423 https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management JIEM, 2021 – 14(3): 604-621 – Online ISSN: 2013-0953 – Print ISSN: 2013-8423 https://doi.org/10.3926/jiem.3446 A Machine Learning Approach to Enable Bulk Orders of Critical Spare-Parts in the Shipping Industry Fiorentia Zoi Anglou , Stavros Ponis , Athanasios Spanos National Technical University Athens, School of Mechanical Engineering (Greece) fz.anglou@gmail.com, staponis@gmail.com, atspanos@gmail.com fz.anglou@gmail.com, staponis@gmail.com, atspanos@gmail.com Received: December 2020 Accepted: June 2021 Received: December 2020 Accepted: June 2021 1. Introduction The maintenance of the machinery onboard a vessel is a critical task, since any engine failure results in delays and down-times in the voyage of a vessel, which translates into additional cost and penalties (Kian, Bektaş & Ouelhadj, 2019). Therefore, ship management companies establish full proof and robust planned maintenance frameworks and systems, whilst giving high prioritization to the planned maintenance of their vessels and undertaking cross departmental projects to ensure the timely delivery of high-quality spare parts, with as low a cost to the ownership as possible. Depending on the age of the vessel and the type of machinery, those needs may vary but the overhaul needs in parts and services when accumulated for the whole fleet may amount to a considerable expense for the company. At the same time, costs of maintenance and administration increase in a diminishing proportion as the size of a ship increases (Lun, Lai & Cheng, 2010), thus deeming the fleet mix an important parameter, further complexing the design, plan and execution of the maintenance strategy. In the research presented in this paper, the authors focus on the planned maintenance needs of a fleet of ships and more specifically on sixteen (16) components commonly found when addressing the technical needs of the vessels, e.g. connecting rods, pumps etc. The process of simultaneously assessing spare part needs, and consequent purchasing, for a fleet of ships is commonly known as a bulk ordering process. Our research simulated the process of bulk ordering and the subsequent planned maintenance needs of a ship management company operating approximately 75 vessels. The bulk ordering process is far from a trivial task, since the number of distinct items ordered each year is considerably high- amounting to several thousand different spare parts- and the delivery locations are not constant and are subject to the vessels’ movements. Furthermore, the suppliers that can provide the necessary parts in the necessary volumes for overhauling processes are few and are mostly concentrated in two geographic regions, i.e. Europe and Asia. In that context, the main challenges that shipping companies encounter are uncertainty, volume and administrative workload. Uncertainty is inherent in the process since the demand for ship spares possesses an erratic nature as it can arise at any time (Jiang, Kong & Liu, 2011). The need for ordering in high volumes is also essential, in order for the ordering process to achieve economies of scale. To cite this article: -604- 2. Literature Review Despite its importance, spare parts management literature has paid little attention to its integration in supply chain management to optimize ordering policy and reduce costs of spares in the maritime industry (Vukić, Stazić, Pijaca & Peronja, 2021). Few are the authors who identify and highlight the significance of the subject, one of them being Nenni and Schiraldi (2013), who state that spare parts management in the maritime industry is indeed a very important issue due to the complexity and uniqueness of the ship operational environment, where reliability and safety are particularly essential. Still, according to our review and to the best of our knowledge, only three journal papers are published in the last five years that are directly relevant to the research presented in this paper. First, the paper by Eruguz, Tan and van Houtum (2018) attempts to minimize the expected total discounted cost of spare part deliveries, part replacements, and inventory holding over an infinite planning horizon. To do so, the authors formulate the problem as a Markov decision process and use numerical experiments to show that the cost savings obtained by the integrated optimization of spare part inventory and part replacement decisions are significant. Finally, they attempt to validate their approach, by using real-life data from a collaborative company, i.e. Fugro Marine Services. In the research presented in this paper, we also attempt to provide an integrated approach to forecast and optimize the spare parts quantities in bulk needed to cover planned maintenance requirements on an annual basis. Still, this is done quite differently, by utilizing ML algorithms for clustering and forecasting, while introducing cost optimization only in the third step of the approach. By evaluating these two approaches in combination, one can actually note that there is an increased level of complementarity. The model proposed by Eruguz et al. (2018), deals with Condition Based Maintenance, hence the Markov chain, while this paper deals with tactical Planned Maintenance. The CBM model in Eruguz et al. (2018) seems to have difficulties in scaling, while the one proposed in this paper utilises a more condition-agnostic approach, thus is able to scale more efficiently on fleet level and for numerous components. In our opinion, these two approaches could work in tandem, offering a very strong predictive maintenance approach, albeit using telemetry data on P-F curves, instead of Markov chains. Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 such as ship age, which in a way determine demand for spare parts. To further reduce the administrative workload and generate cost optimal scenario in steps, blending of analytics with traditional operations research, i.e. prescriptive analytics, will be examined in an attempt to drive the optimum sourcing decisions on the basis of minimum Total Cost of Ownership (TCO). As a result, a decision- support tool is formulated that can lead to significant decrease of the administrative workload and the total time needed to complete the project and provides viable scenarios for spare parts acquisition strategies that overall yield a total cost saving of 9%. 1. Introduction In that sense, bulk orders ideally would refer to more than thirty (30) and sometimes reaching up to ninety (90) vessels with a span of several thousand unique items every year, making it very time-consuming to negotiate with the implicated suppliers and conclude the selection process. Finally, administrative workload is undoubtedly high due to the large number of interconnected parties and stakeholders participating in the process, which makes it very unwieldy and slow-moving. It is essential that the administrative costs do not increase unreasonably as a proportion of the value of the purchased items (Huiskonen, 2001). Empirical research results indicate that even in cases with very mature planned maintenance systems installed, significant additional FTE (Full Time Equivalent) effort is needed for the smooth completion of the process. The above challenges have triggered the case company, presented in this study, to scour for ways to optimally address them and unlock further value of the bulk order process. It has to be noted that current practice in the shipping industry is to leverage experience and simple analytics to determine the optimum quantities, timing and allocation of suppliers among the procured items. The absence of a sophisticated and well-structured process drives the total timing of this project to almost nine months (this project is undertaken every year) with the involvement of multiple departments. The case company looked to ML due to its current strong standing and high maturity profile in deploying advanced analytics to increase effectiveness and boost efficiency in supply chain areas such as general consumables forecasting, crew scheduling and strategic network design. This gave rise to the topic of this research paper which will aim to address the aforementioned key challenges by taming a very sizeable and overly complex dataset, providing ways to extract useful information and insights from historical data, facilitating the ability to forecast the needs of the fleet, reducing administrative workload and support the decision-making process by generating indicative solutions. In the following sections, the applicability of ML in dealing with similar business issues will be examined and the design of an integrated tool that aims to tackle challenges throughout the bulk ordering process, will be attempted. More specifically, clustering and forecasting of the quantities needed by the vessels will be undertaken, to provide a laser focused and current view of the critical needs of vessels. This will be done by integrating exogenous factors, -605- 3. The Clustering Component The ordering of spare parts for planned maintenance purposes is a time-consuming project, in part due to the high number of items that comprise a bulk order. In the interest of industry-wide standardization, each item in the spare parts industry can be referred with a unique number called a maker reference. In each bulk order, there could be over 4,000 distinct product codes making it an arduous process to compare the items or even to systemically log the prices of each supplier. Therefore, it seems important to be able to narrow down the high-interest product codes for each bulk order to facilitate and expedite the process. From an analytics perspective this is a task best tackled by the clustering approach, which is one of the most common unsupervised ML techniques (Hinton & Sejnowski 1999). This way the analysis can be focused only on product codes that have been identified as high-interest and therefore the volume of administrative workload for the departments will be smaller. Identifying the input variables in the unsupervised learning algorithm is of great importance as the relationship between them will determine the product codes upon which forecasting will be attempted. In this study, we identify the following input variables: a) Price: which indicates the acquisition price of the item (also accounting for discounts- if any), b) Quantity: which indicates the number of times the item was bought in the past for planned maintenance purposes, c) Total Volume: as provided by the product of price times quantity. This variable highlights the importance of items that have a medium price but are ordered in considerable quantities, thus making the total volume quite high, d) Number of Unique Vessels: which indicates the number of different vessels that the item is installed on. This variable increases the importance of an item, even if it doesn’t have a considerable volume, price or quantity, if it is installed on many vessels and therefore has an increased influence in the uniformity and possible problems across several vessels and e) Average Age: which indicates the average age of the vessels this item is installed on. The main goal is to determine the product codes that have an abnormally high price, quantity, combination of both or/and are installed on several vessels. Therefore, a clustering process is required in order to identify the ‘outliers’ of the dataset, thus labelling the items that have the characteristics described above. Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 with the one presented in this paper, still they are highlighted in the remainder of this section, for completeness purposes. Azizah and Subiono (2018) focus exclusively on the spare parts of the ship engine and propose a Petri-Net approach representing its component’s spare part ordering. Then they elaborate a max-plus algebra model to obtain the date when the spare part should be ordered. Efficiency measures or cost benefits from applying the method are not mentioned in the paper. Hmida, Regan and Lee (2013), propose a multicriteria inventory policy using inventory classification method integrated with a preventive maintenance program. The authors report savings of thirty-four (34) service days as a result of the decrease in the number of downtime days due to pumps failure and 10% reduction in inventory, by the application of their approach. Last, but not least, we have to note the emergence of Additive Manufacturing (AM) in the scientific field of spare parts management for the shipping industry, which promises to provide disruptive solutions, enhanced flexibility and significant economies for the whole process (Kostidi & Nikitakos, 2008). 2. Literature Review Second, we have to note the very recent efforts by Jimenez, Bouhmala and Gausdal (2020). In their paper, the authors develop a predictive maintenance solution based on a computational artificial intelligence model using real- time monitoring data in the shipping industry. In doing so, they analyze a set of historical sensor data, using the statistical programming language R. Their results highlight the potential of using big data analytics for developing a predictive vessel maintenance model. Still, the authors state that there is a number of further issues that have to be addressed prior to designing the algorithms and a solution based on artificial intelligence. Third, the work of Kian et al., (2019), is considered marginally related to the work presented in this paper in the sense that it also provides an integrated solution for the challenge of spare parts management for maintenance scheduling in the maritime industry. But this is where the relevancy with the research presented in our paper stops, since the authors focus on a specific problem of Condition-Based Monitoring predictive maintenance dealing with a vessel operating on a given route that is defined by a sequence of port visits. When a warning on part failure is received, the problem decides when and to which port each part should be ordered, where the latter is also the location at which the maintenance operation would be performed. The authors use a mathematical programming model of the problem and a shortest path dynamic programming formulation for a single part to solve the problem. Furthermore, the validation they use is based on simulation tests of different scenarios and not in actual case data. Finally, our research produced a small number of references, which attempt to address the issue of spare parts management in the shipping industry from different perspectives. Their approaches are not directly comparable -606- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 3. The Clustering Component This process is often called anomaly detection (Zimek & Filzmoser 2018). In this paper, the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm (Ester, Kriegel, Sander & Xu, 1996) is used, as it automatically creates a cluster containing the outliers. DBSCAN is a data clustering algorithm that given a set of points in some space, classifies in the same category, points that are closely packed together. At the same time, it marks as outlier points, those that lie in low-density regions. The items are clustered using DBSCAN algorithm with minPoints=10 and local radius for expanding cluster set to eps = 1.8. The results are visualized in Figure 1. Cluster 0, hereafter the “Outlier Cluster”, contains the outliers of the analysis. The mean quantity of the Outlier Cluster is considerably higher than the one of Cluster 1, which contains the clear majority of the data, here on after the “Average Cluster”. The same can be said for the price of the Outlier Cluster as compared to the price of the Average Cluster. Evidently, the total volume, which is computed as the product of the aforementioned characteristics (price and quantity), is also considerably higher. Finally, the number of unique vessels, that as mentioned before describes the number of distinct vessels that the specific item is installed on, is also considerably higher in the outlier cluster. However, the average age of the vessels is virtually the same for the two clusters. -607- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 The cluster of k-means was further used as an independent variable in the forecasting analysis of the next section. The product codes previously identified as cost drivers, will be used as a basis for the bulk order price collection and winner nomination. Simulation results indicate that approximately 4% of the total items can represent roughly 50% of the total cost. By leveraging this finding, the purchasing department of any ship management company can focus only on the pre-identified items to collect prices, assess the quotations, negotiate with and select the winners and therefore decrease its administrative workload. Concurrently, the product codes identified as cost drivers will provide the basis for the subsequent steps of this analysis, i.e. the predictive and prescriptive components. The cluster of k-means was further used as an independent variable in the forecasting analysis of the next section. The product codes previously identified as cost drivers, will be used as a basis for the bulk order price collection and winner nomination. Simulation results indicate that approximately 4% of the total items can represent roughly 50% of the total cost. By leveraging this finding, the purchasing department of any ship management company can focus only on the pre-identified items to collect prices, assess the quotations, negotiate with and select the winners and therefore decrease its administrative workload. Concurrently, the product codes identified as cost drivers will provide the basis for the subsequent steps of this analysis, i.e. the predictive and prescriptive components. 4. The Predictive Forecasting Component The proposed forecasting model aims to calculate the nominal quantity that a vessel needs in the coming year and maker reference. The model variables are: a) Average Age: The age of the vessel is one of the most important vessel characteristics and, as described in the introduction, the maintenance of the vessel and thus the quantities of the items that will be ordered are highly correlated; b) Average Price: The price of an item is one of the most important demand characteristics and in the sections below its relationship with the final quantity will be examined; c) DWT: This variable is an indicator to the size of the vessel and to its needs; d) Type: A categorical variable which, in combination with DWT, denominates the class of the vessel; e) Origin: This variable indicates the country of construction of the vessel. It is a categorical variable of three levels: South Korea, Japan and China and aims to unveil correlations between the shipyard and the quality of the vessel and f) k-Means cluster: denoting one of the three clusters the item belongs to as per above. The available dataset is divided into 16 core components. In the simulation dataset, there is an evident scarcity of data for this application therefore one of the main challenges that needs to be addressed is the overfitting of the models. The training set and the test set with a random 80-20 data partition, as for most of the components, lacks the number sufficient number of data points to create a validation set. This is a commonplace finding in the shipping industry where abundance of data availability and quality remains elusive. In addition, time series analysis, which seems as the most direct approach to such a problem is not applicable in this case due to factors commonly plaguing such algorithms e.g. multicollinearity, heteroscedasticity and autocorrelation (Hanke and Wichern, 2009) as well as due to the high degree of influence of the demand by exogenous factors. Considering the aforementioned limitations, the below models were chosen; • Random Forest (RF), due to its ability to avoid overfitting (Hastie, Tibshirani & Friedman, 2008) and its superior efficiency (Ho, 1995). • Random Forest (RF), due to its ability to avoid overfitting (Hastie, Tibshirani & Friedman, 2008) and its superior efficiency (Ho, 1995). Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Cluster 2 contains a small fraction of the total items that have a large quantity and are installed on several young vessels. The product codes that will be included in the analysis are those that are contained in the outlier cluster. To further analyse the data k-Means clustering on the previously identified high-interest items is performed. Figure 1. Product codes clustered in three clusters using DBSCAN Figure 1. Product codes clustered in three clusters using DBSCAN Figure 2. K-means clustering in the outlier cluster Figure 2. K-means clustering in the outlier cluster Cluster Description Data Points Percentage of Total Cost of Outlier Cluster 1 High volume – young vessels 220 73.27% 2 High volume – old vessels 128 21.68% 3 Low volume 141 5.04% Table 1. Summarized results for 2nd clustering -608- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 4. The Predictive Forecasting Component • Generalized Linear Model (GLM), due to its simplicity and its ability to handle error distributions other than normal ones (Nelder & Wedderburn, 1972). • Principal Component Regression (PCR); due to its ability to be applied when the number of variables is high in relation to the number of available data. (Jackson, 1991). For each component the mean absolute percentage error is computed as per the below formula: The results are presented in Table 2. For a number of components, forecasting did not take place as the entries were not enough to properly train and test the algorithms. As can be seen from Table 2 the forecasting error is, in some cases, considerable while in some cases, where it is below 40%, it qualifies as satisfactory. The satisfactory error level was determined after interviews with practitioners of the shipping industry, specifically in the purchasing department. To better visualize the performance of the algorithms, a snapshot for the forecasting results for the component ‘Compressors’ is depicted in Figure 3. -609- -609- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Component RF GLM PCR Number of Points Compressors 34% 119% 112% 272 Rods 77% 111% 272% 31 Cylinder Heads 140% 220% 221% 906 Cylinder Liners 111% 311% 301% 157 Fuel Injection Valves 98% 109% 128% 287 Fuel Oil System 22% 45% 34% 26 Pistons 64% 256% 259% 298 Shaft 64% 70% 85% 81 Turbo Chargers 48% 70% 48% 55 Table 2. Mean absolute errors for indicative components Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Figure 3. Forecasting results for the component ‘Compressors’ Figure 3. Forecasting results for the component ‘Compressors’ For the component ‘Fuel Injection Valves’, the error of the random forest algorithm is 98%. As seen in Figure 4, there are two data points that have order quantity over a hundred items, which is significantly higher than the average quantity observed. Therefore, a data cleansing method is used to determine such data points and eliminate them from the training and evaluating sets of the algorithms. In the specific dataset, it is common to come across data points that can be considered as outliers. In a business sense, this can be explained by a superintended engineer over/ under estimating actual demand or a vessel having abnormally high/low needs for a specific year due to a sequence of unplanned maintenance events. 4. The Predictive Forecasting Component Consequently, DBSCAN is used to determine the outliers and exclude those data points from the analysis, with minPts = 6 and eps = 0.5. Parameter minPts was set to six by multiplying the dimensions by two. The dimensions for which DBSCAN is applied are only the numeric ones (age of the vessel, DWT, price of the product code). Parameter eps was dictated by interpreting the k-NN graph (Ester et al., 1996). Next, the three models described before are applied, trained and tested on the ‘cleaned’ dataset. The training and testing of the algorithms in a sanitized dataset, produces better results, by reducing the mean average percentage error by around 36%. Still, there are some components that the outlier handling could not decrease the error to satisfactory levels, i.e. ‘Cylinder Heads’, ‘Cylinder Liners’. For those components, an analysis of the optimum eps value is presented in Figure 5. According to Ester et al. (1996), when the eps parameter decreases the number of data points included in the analysis is increased (the outliers are decreasing). At the same time the forecasting error decreases as well. As can be seen from the graph, the optimum error (without simultaneous elimination of a considerable amount of data points) is at eps=0.8. However, the MAPE, after the optimization process, is still considerably high making it very difficult to rely on the applied forecasting methods for these components with this dataset. -610- -610- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Figure 4. Forecasting results for the component ‘Fuel Injection Valves’ Figure 5. Change of MAPE (left) and number of data points (right) as eps parameter increases for component ‘Cylinder Heads’ Figure 4. Forecasting results for the component ‘Fuel Injection Valves’ Figure 4. Forecasting results for the component ‘Fuel Injection Valves’ Figure 4. Forecasting results for the component ‘Fuel Injection Valves’ Figure 5. Change of MAPE (left) and number of data points (right) as eps parameter increases for component ‘Cylinder Heads’ Figure 5. Change of MAPE (left) and number of data points (right) as eps parameter increases for component ‘Cylinder Heads’ The least mean absolute error for each component for all methods (RF, PCR, GLM) are presented in Table 3, showcasing that the best performing method for the vast majority of the components is the random forest algorithm. Table 3. Least mean absolute error for indicative components for all methods Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Mean absolute errors for extra needs for indicative components Component Minimum MAPE Method Rods 0% Random Forest No Outliers Drive Section 0% Random Forest Fuel Injection Valves 88% Random Forest No Outliers Pumps 14% Random Forest No Outliers Purifiers 6% Principal Component Regression No Outliers Turbo Chargers 0% Random Forest Table 4. Mean absolute errors for extra needs for indicative components 4. The Predictive Forecasting Component This result was expected as the random forest algorithm best handles the exogenous factors that influence the outcome in a stochastic manner that makes it impervious to over/ under fitting (Hastie et al., 2008). Component Minimum MAPE Method Compressors 31% Random Forest No Outliers Rods 0% Random Forest No Outliers Cylinder Heads 125% Random Forest No Outliers Cylinder Liners 111% Random Forest Fuel Injection Valves 44% Random Forest No Outliers Purifiers 46% Random Forest No Outliers Fuel Oil System 10% Random Forest No Outliers Pistons 53% Random Forest Shaft 29% Random Forest No Outliers Turbo Chargers 0% Random Forest No Outliers Table 3. Least mean absolute error for indicative components for all methods -611- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Finally, the exceptional case of items that are reordered during the same period is studied. This reordering is either due to miscalculation of the vessel’s needs, unplanned maintenance events or failure of previously bought equipment. To determine those additional quantities a forecasting process is used. The forecasts are generated by following the same sequence of actions as for the nominal quantities with one additional dimension, i.e. Bulk Market. This variable is a categorical variable with two levels: maker or parallel, i.e. purchased from original equipment manufacturer or from a source providing imitation spares, respectively. The variable will be used to explore possible correlation between the source of purchase and the additional quantities. The results are presented in Table 4. The forecasting of the extra quantities is performed only on a small number of components (12 out of 16) as the rest do not exhibit variations in the data making the completion of the training very challenging. Especially after outlier handling the number of forecastable components drops even further, from 12 to 8. This particular finding could generate an area for further research in the future. In conclusion, forecasting the nominal needs of the vessels exhibits satisfactory results (average MAPE 53%) and could, in the future when the training samples increase, become more accurate. For some specific components that show increased accuracy, e.g. fuel oil system (MAPE = 10%) the tool can be used to expedite the process, while decreasing the workload both for the vessel and for the shore- based engineers. However, the forecasting of the extra needs does not yield such results. The average MAPE is increased compared to the forecasting of the nominal quantities, while the number of components upon which forecasting is applied decreases. In the next section, the forecasting results are used to create the cost optimizer that leads to optimum allocation of items to vendors so as to minimize total cost of the bulk orders. Component Minimum MAPE Method Rods 0% Random Forest No Outliers Drive Section 0% Random Forest Fuel Injection Valves 88% Random Forest No Outliers Pumps 14% Random Forest No Outliers Purifiers 6% Principal Component Regression No Outliers Turbo Chargers 0% Random Forest Table 4. where: SS is the safety stock and WACC is the weighted average cost of capital with WACC • Stock out cost: this cost component represents the costs that are incurred when an item that should have been on board the vessel is not. For its computation the authors present a novel approach as an inverse function of the Safety Stock SL variable • Stock out cost: this cost component represents the costs that are incurred when an item that should have been on board the vessel is not. For its computation the authors present a novel approach as an inverse function of the Safety Stock SL variable 5. Cost Optimization Component The cost optimization prescriptive model ties in the entire bulk order analytics framework and shifts it toward the decision support domain by serving as a guideline on the optimal cost basis of spare parts procurement. The model facilitates the choices whether each spare part should be ordered more times than the nominal need of the vessel and whether it should be bought from maker or from the parallel market. The components of the cost function are the following: • Acquisition cost: it represents the cost of purchase for each item. It depends on the total quantities and on the acquisition price of each item. What needs to be noted here is that for the two main categories of suppliers, makers and parallels, the acquisition price changes considerably. • Acquisition cost: it represents the cost of purchase for each item. It depends on the total quantities and on the acquisition price of each item. What needs to be noted here is that for the two main categories of suppliers, makers and parallels, the acquisition price changes considerably. Where safety stock depends on the desired service level (SL). The safety stock will also be added to the acquisition and transportation cost as it is assumed that both the target inventory and the safety stock are bought together, considering that price fluctuations in the spare parts are not high. The safety stock follows the formula below (Ballou, 2003). -612- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Where LT is the lead time, Z¬SL is the inverse distribution function of a standard normal distribution with cumulative probability of the underlying service level and demand refers to the historical demand of the relevant item. For both the lead time and average demand there are more than 30 observations therefore by the central limit theorem it can be said that these variables satisfy the underlying assumptions (i.e. normal distribution) of the above formula. • Transportation cost: this cost component represents the cost of the transportation of each item on board the vessel. This cost depends on several parameters such as the location of the supplier, the trading route of the vessel, any specific requirements for clearance etc. For the purposes of this analysis it is assumed that the transportation cost depends mainly on the lead time which determines the transportation mode to main logistics hubs, e.g. the Netherlands. Empirical research indicates that there are two main regions from which ship spare parts can be sourced: Europe and Far East (Japan, Korea and China). Without loss of generality, we assume that around 30% of the spare parts are sourced in Europe wherein we will assume transportation costs to be zero, given the proximity of the vessels and the high frequency that they call European ports. Therefore, according to the above the final formula for the transportation cost is the below: • Inventory Cost: this cost component represents the costs that are incurred because of the inventory held on the vessel. The inventory cost follows the simple formula below • Inventory Cost: this cost component represents the costs that are incurred because of the inventory held on the vessel. The inventory cost follows the simple formula below where: SS is the safety stock and WACC is the weighted average cost of capital with WACC ∈ [3%, 8% 5.1. Acquisition Cost The main components of the acquisition cost are the total quantities and the purchase price. To determine the level of the nominal needs and any additional quantities for each product code the best performing model (the one exhibiting the smallest MAPE) was used (see Table 3). Figure 6, depicts the final quantities (nominal + extra) for ‘Fuel Injection Valves’, and the differences created from the sourcing parameter (maker/parallel). Next, the safety stock for each item is calculated. The average demand and standard deviation of are computed regardless of the market, using past data. The lead time assumes on-hand stock availability from makers and a range between 5 and 35 days from the parallel market. -613- -613- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Figure 6. Extra needs depending on the market for component ‘Fuel Injection Valves’ Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Figure 6. Extra needs depending on the market for component ‘Fuel Injection Valves’ Lastly, to make the calculation stochastic in nature rather than deterministic, and account for the forecast errors of the previous models for each component the forecast bias is computed, and it is determined whether there is an over-forecasting or an under-forecasting bias. If there is an over-forecasting bias, then the safety stock computed is multiplied by the accuracy of the forecast of nominal quantities. The forecast bias is computed using the following formula: This forecast bias, is sometimes called the normalized forecast metric. As can be seen, the metric [-1, +1] where 0 indicates the absence of forecast bias (Singh, 2017). Negative values show a tendency to under-forecast and positive values to over-forecast. In a business sense, the safety stock is needed to cover needs arising from either demand and/ or lead time variability, e.g. unplanned maintenance events or manufacturing issues at supplier. However, if the demand has been forecasted with a method that indicates positive bias the final quantity that will be purchased will be unnecessarily high. This reasoning explains the final formula of the safety stock. An example of a positive forecast bias for the component ‘Cylinder Liners’ is shown in Figure 7. As seen in Figure 7, the forecast for some components tends to overestimate the quantities that will ultimately be needed to cover the planned maintenance needs of the vessel. Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Random service levels were used to compute the total cost of the items and the service level having the minimum total cost was identified as the optimum service level and was used in the final step of the prescriptive model. The random service levels were chosen in the range of 95% to 99.9%. As the items ordered in the bulk process are critical for the smooth running of machinery, this range was chosen to address business and technical needs often found in the shipping industry. In Figure 8, one can see a high fluctuation of the total cost (around 20%) as the service level changes, highlighting the need to determine the optimum service level. Minimum cost is achieved for different service levels for makers (maker optimum service level = 95.18%) and for non-makers (parallel optimum service level = 96.31%), which is explained by the changes in the underlying quantities which can be observed in Figure 9. To calculate the acquisition cost of each item the prices of the items depending on the market were determined using historical data. Figure 8. Cost of an indicative item of component ‘Connecting Rods’ as a function of the service level Figure 9. Total quantity of an indicative item of component ‘Connecting Rods’ as a function of the service level Figure 8. Cost of an indicative item of component ‘Connecting Rods’ as a function of the service level Figure 8. Cost of an indicative item of component ‘Connecting Rods’ as a function of the service level Figure 8. Cost of an indicative item of component ‘Connecting Rods’ as a function of the service level Figure 9. Total quantity of an indicative item of component ‘Connecting Rods’ as a function of the service level Figure 9. Total quantity of an indicative item of component ‘Connecting Rods’ as a function of the service level 5.1. Acquisition Cost Therefore, the over forecasted quantities can be used as safety stock. This will avoid over-stocking the vessels with unnecessarily high quantities of items that have been forecasted with methods that exhibit high positive forecast bias. To determine the optimum service level of each product code exhaustive enumeration was used, as the problem size is limited and there are no problem-specific heuristics to reduce the set of candidate solutions to a manageable size. As seen in Figure 7, the forecast for some components tends to overestimate the quantities that will ultimately be needed to cover the planned maintenance needs of the vessel. Therefore, the over forecasted quantities can be used as safety stock. This will avoid over-stocking the vessels with unnecessarily high quantities of items that have been forecasted with methods that exhibit high positive forecast bias. To determine the optimum service level of each product code exhaustive enumeration was used, as the problem size is limited and there are no problem-specific heuristics to reduce the set of candidate solutions to a manageable size. Figure 7. Example of positive forecast bias in the component ‘Cylinder Liners’ Figure 7. Example of positive forecast bias in the component ‘Cylinder Liners’ -614- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 where is the percentage of the times that on an urgent basis the parts are bought from the maker. For the purposes of this simulation we will assume a=0.5 and where SL is the service level for each item. where SL is the service level for each item. where the urgency factor can be defined as - 5.2. Stock-out Cost This cost component represents the cost of urgently re-supplying the vessel with the item, if the existing stock runs out. Stock-out cost is normally higher than the previous costs since there is no time to receive quotations from several vendors or to make extensive price negotiations. When a requisition is made on an urgent basis, lead time is of critical importance and the transportation cost of the shipment can be higher due to an inconvenient delivery port and/or because there might be no additional orders to achieve economies of scales in either airfreight, road freight or launch boat cost. The components of the stock out cost are analysed below: -615- -615- 6. Results To evaluate the performance of the proposed model, here on after ‘Model 1’, we are including a comparison with the common practice in the maritime industry here on after ‘Model 2’. The results are shown in Table 5. From empirical research, we have concluded that ship management companies do not take stock out cost into consideration when determining the quantities needed for the fleet and the underlying stockout probability on average is 25%. As can be derived from Table 6 there is a well expected difference in total costs after the application of the two alternatives, i.e. Model 1 and Model 2. Figure 10, visualizes the difference between the total costs in the four cost components after the implementation of Models 1 and 2. It is evident that the cost components that drive this difference are the acquisition and the stock out costs. The increased acquisition cost can be attributed to two main factors: quantity increase due to safety stock estimation and alternate choice between more expensive spare parts purchased from original manufacturers over imitation spare parts. As previously said the stock out cost of Model 2 was calculated assuming that the actual quantities did not account for safety stocks. Figure 11, visualizes the effect of the probability of stock out in actual savings from the implementation of Model 1. The implementation of the model is expected to generate savings from the first year of implementation when the underlying probability of stock out for Model 2 is more than 30%. Parameters/ Variables Model 1 Model 2 Service Level Variable [95% - 99.9%] Constant [75%] Extra Quantities Included Not Included Stock Out Cost Included Not included Inventory Cost Included Not Included Transportation Cost Included Included Table 5. Comparison of the Models Table 5. Comparison of the Models -616- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem Component Delta between Model 1 and Model 2 Compressors 37% Camshaft -16% Rods -1% Cylinder Heads 24% Cylinder Liners 4% Drive Section -28% Fuel Injection Valves -35% Pumps -48% Purifiers 13% Fuel Oil System -31% Lubricating Oil Purifiers 6% Diesel Generator 5% Mechanical System 170% Pistons 35% Shaft 160% Turbo Chargers 9% Total 4% Table 6. 6. Results Comparison of total costs incurred after application of Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem Component Delta between Model 1 and Model 2 Compressors 37% Camshaft -16% Rods -1% Cylinder Heads 24% Cylinder Liners 4% Drive Section -28% Fuel Injection Valves -35% Pumps -48% Purifiers 13% Fuel Oil System -31% Lubricating Oil Purifiers 6% Diesel Generator 5% Mechanical System 170% Pistons 35% Shaft 160% Turbo Chargers 9% Total 4% Table 6. Comparison of total costs incurred after application of M d l 1 d M d l 2 Component Delta between Model 1 and Model 2 Compressors 37% Camshaft -16% Rods -1% Cylinder Heads 24% Cylinder Liners 4% Drive Section -28% Fuel Injection Valves -35% Pumps -48% Purifiers 13% Fuel Oil System -31% Lubricating Oil Purifiers 6% Diesel Generator 5% Mechanical System 170% Pistons 35% Shaft 160% Turbo Chargers 9% Total 4% Table 6. Comparison of total costs incurred after application of Model 1 and Model 2 Figure 10. Percentage change across Model 1 and Model 2 Figure 10. Percentage change across Model 1 and Model 2 Figure 10. Percentage change across Model 1 and Model 2 What needs to be noted here is that the increased quantities purchased in the first year can be considered as an investment which, keeping all other factors stable, would be paid back in full in the 2nd year of implementation of Model 1 regardless of the probability of stock out. In the first year of implementation, the purchase of higher quantities to serve as inventory is proposed increasing the acquisition cost by 27% compared to Model 2. In total, the acquisition cost is increased 13% accounting for the nominal purchases of the 2 years and the creation of a safety stock on board the vessels. On the other hand, by implementing Model 1 the stock out cost dramatically decreases (22% decrease) resulting in total cost savings of around 9% as shown in Figure 12. -617- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Figure 11. Sensitivity analysis for stock out probability Figure 12. Final total costs incurred for 2 years Figure 11. Sensitivity analysis for stock out probability Figure 11. Sensitivity analysis for stock out probability Figure 12. Final total costs incurred for 2 years Figure 12. 6. Results Final total costs incurred for 2 years In conclusion, the analysis indicates that the purchase of safety stock will lead to decreased costs from the second year of implementation. The safety stock demands an initial investment of almost one quarter of the total acquisition cost that, according to the model, will decrease the stock out probability, thus decreasing the total costs of the process. In conclusion, the analysis indicates that the purchase of safety stock will lead to decreased costs from the second year of implementation. The safety stock demands an initial investment of almost one quarter of the total acquisition cost that, according to the model, will decrease the stock out probability, thus decreasing the total costs of the process. Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Based on the results described above, the authors conclude that the adoption and application of their proposed methodology from shipping companies managing and operating a large fleet of vessels has the potential to produce meaningful results ranging from soft benefits, such as the rationalization of the workload of the purchasing department and its third party collaborators to hard, quantitative benefits, such as reducing the cost of the bulk ordering process, which directly affects a company’s bottom line. Inherent limitations do exist in the proposed approach, such as the concept of endogeneity. In purchasing, traditionally prices are ‘manipulated’ by the choice of the supplier. The proposed approach indirectly assumes that quantity is, among others, a function of price, which often is not accurate, since the price suffers from shocks relative to the choice of the supplier. The authors, in coordination with the case company, decided not to treat the endogeneity issue. First, because the actual business impact of such an intervention was deemed by purchasing executives as not practical and efficient, since year-to-year price surges were spotted in just a few large suppliers and second because development was at risk in falling into an unnecessary and not cost-effective vicious cycle, performing an iteration process in order to correct the error between the quantity that was forecasted with a certain price and the actual quantity that might have a different price minimum. In other words, a change in the supplier of an item, would create a new price, and the combination of price-item as an input in the forecasting tool, could generate a different quantity, potentially leading back to a different choice of supplier/ service level etc. Further research on the subject includes the re-evaluation of the entire algorithmic framework using datasets with larger breadth in terms of independent variables and depth in terms of observations. More specifically, with a sizeable enough critical mass of data, deep learning algorithms could be performed and assessed to see if they would further improve accuracy and performance indicators, such as the Mean Absolute Percentage Error produced by the statistical forecasting tool. Moreover, the companies adopting the tools proposed in this paper should re-evaluate the P-F curves of the critical items identified by the clustering algorithms, especially if their origin is from the parallel market. Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 in the third methodological step, a cost-related decision support tool is developed to cost-effectively allocate the selected items to a group of vendors. in the third methodological step, a cost-related decision support tool is developed to cost-effectively allocate the selected items to a group of vendors. The application of our proposed approach produced several interesting findings that can fuel fruitful discussions at the decision-making level of the companies in the shipping industry. First, one has to note the value of clustering, when applied to the item codes dataset. In the benchmarking dataset chosen in this paper, focusing on the 4% of high- interest items which contribute to 50% of the total process cost, results in a significant reduction of the administrative workload for both the internal departments involved in such a process as well as the suppliers. The use of the proposed statistical forecasting tool for the nominal needs of the vessels, initially produces satisfactory results (average MAPE 53%), which can potentially further improve with the training of the samples. Additionally, for specific components that show increased accuracy, e.g. fuel oil system (MAPE = 10%), the tool can be used to expedite the process while decreasing the workload both for the vessel and for the shore- based engineers. However, a further analysis including more independent variables, related to the types of machinery and their nominal running hours etc. could be performed in the future as they were out of the scope of this paper. In addition, the proposed prescriptive model suggests that increased quantities will potentially lead to total cost savings of 9% versus the baseline case depicted in Model 2. This comes as a result of the high contribution of the stock out cost to the total cost function. Indeed, the cost to deliver an item on board at expedite/urgent conditions is very high. Even if the previous exercise showed no influence of the source of purchase on the extra quantities, the prescriptive model suggests an increased allocation of items to makers. This, however, mainly stems from the fact the makers have lower delivery lead times, thus driving the safety stock down. In summary, the application of the proposed prescriptive model suggests an initial investment on increased safety stock that would be paid back in full (all other factors constant) after the 2nd year of implementation. 7. Conclusions The main objective of this paper is to propose a methodological approach supporting the spare parts’ bulk ordering process of companies managing and operating a fleet of vessels in the shipping industry. Studying the literature on spare parts management in the shipping industry and its intersection with ML- enhanced forecasting techniques and tools, confirmed the authors’ initial assumption, that the shipping industry lags behind in both understanding and further more utilizing these techniques and tools in the every-day business practice. Literature proves that spare parts forecasting with the use of ML is still an area at its infancy for the companies of the shipping industry and that is exactly where the research presented in this paper sets its focus. That is to provide proof to shipping industry professionals that ML can be a useful and efficient tool that they can understand, master and apply to their everyday practice. The proposed approach is based on the development of a comprehensive decision support tool able to facilitate the process of the bulk orders and optimize the purchasing decisions. The approach consists of three discrete methodological steps, each one supported by a decision support tool based on clustering and ML algorithms. In the first step, the initial dataset is rationalized using clustering techniques to reduce the base of analysis by identifying the high interest items. The rationalization refers directly to the reduction of workload for the departments involved in the process and the creation of a targeted and added value subset for further analyses. The next step involves the development of a forecasting tool for estimating the expected needs of the fleet regarding the previously identified items and to test whether the needed quantity is influenced by the source of purchase. Finally, -618- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Retrieving and consolidating this information in a structured format from the engineering crew onboard, in between overhauls with the bulk order spare parts, would subsequently pave the way for a more holistic predictive/ condition-based maintenance model. The latter should gauge this improved visibility into the spare part reliability, when predicting demand or prescribing outcomes for optimum total cost of ownership. -619- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Declaration of Conflicting Interests The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding This research has been co-financed by the European Union and Greek national funds through the Operational Program Competitiveness, Entrepreneurship, and Innovation, under the call RESEARCH—CREATE—INNOVATE (project code T1EDK-01168). References Azizah, F., & Subiono, S. (2018). Max Plus Algebra and Petri Net Application on Scheduling of Ship Engine Component’s Spare Part Ordering. International Journal of Computing Science and Applied Mathematics, 4(1), 1-9. https://doi.org/10.12962/j24775401.v4i1.3555 Ballou, R.H. (2003). Business Logistics/Supply Chain Management. New York: Prentice Hall. Eruguz, A.S., Tan, T., & van Houtum, G.J. (2018). Integrated maintenance and spare part optimization for moving assets. IISE Transactions, 50(3), 230-245. https://doi.org/10.1080/24725854.2017.1312037 Ester, M., Kriegel, H.P., Sander, J., & Xu, X. (1996). A density-based algorithm for discovering clusters a density- based algorithm for discovering clusters in large spatial databases with noise. Proceedings of the Second International Conference on Knowledge Discovery and Data Mining (226-231). Portland, U.S.A. anke, J., & Wichern, D. (2009). Business Forecasting. New Jersey: Pearson. Hanke, J., & Wichern, D. (2009). Business Forecasting. New Jersey: Pearson. Hastie, T, Tibshirani, R., & Friedman, J. (2008). The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Stanford: Springer. Hinton, G., & Sejnowski, T. (1999). Unsupervised learning: Foundation Computation. Massachusetts: MIT Press. https://doi.org/10.7551/mitpress/7011.001.0001 Hmida, J.B., Regan, G., & Lee, J. (2013). Inventory management and maintenance in offshore vessel industry. Journal of Industrial Engineering. https://doi.org/10.1155/2013/851092 Ho, T.K. (1995). Random Decision Forests. Proceedings of the 3rd International Conference on Document Analysis and Recognition (1, 278-282). IEEE. Huiskonen, J. (2001). Maintenance spare parts logistics: Special characteristics and strategic choices. International Journal of Production Economics, 71(1-3), 125-133. https://doi.org/10.1016/S0925-5273(00)00112-2 Jackson, E. (1991). A User's Guide to Principal Components. New York: John Wiley & Sons. https://doi.org/10.1002/0471725331 Jiang, C., Kong, M., & Liu, M. (2011). Research on the storage scheme of spare parts for long voyage vessel. Proceedings of the 9th International Conference on Reliability, Maintainability and Safety (1173-1178). IEEE. https://doi.org/10.1109/ICRMS.2011.5979475 Jimenez, V.J., Bouhmala, N., & Gausdal, A.H. (2020). Developing a predictive maintenance model for vessel machinery. Journal of Ocean Engineering and Science, 5(4), 358-386. https://doi.org/10.1016/j.joes.2020.03.003 Kian, R., Bektaş, T., & Ouelhadj, D. (2019). Optimal spare parts management for vessel maintenance scheduling. Annals of Operations Research, 272(1-2), 323-353. https://doi.org/10.1007/s10479-018-2907-y Kostidi, E., & Nikitakos, N. (2018). Is it time for the maritime industry to embrace 3d printed spare parts?. TransNav: International Journal on Marine Navigation and Safety of Sea Transportation, 12(3), 557-564. https://doi.org/10.12716/1001.12.03.16 Lun, Y.H.V., Lai, K.-H., & Cheng, T.C.E. (2010). Shipping and Logistics Management. London, UK: Springer-Verlag. https://doi.org/10.1007/978-1-84882-997-8 -620- Journal of Industrial Engineering and Management – https://doi.org/10.3926/jiem.3446 Nelder, J.A., & Wedderburn, R.W. (1972). Generalized linear models. Journal of the Royal Statistical Society: Series A (General), 135(3), 370-384. https://doi.org/10.2307/2344614 Nenni, M., & Schiraldi, M.M. (2013). Optimizing spare parts inventory in shipping Industry. International Journal of Engineering and Technology, 5(3), 3152-3157. Singh, S. (2017). A Critical Look at Measuring and Calculating Forecast Bias. The Institute of Business Forecasting & Planning [online]. https://demand-planning.com/2017/01/13/a-critical-look-at-measuring-and-calculating-forecast- bias/ (Accessed: October 2020). Vukić, L., Stazić, L., Pijaca, M., & Peronja, I. (2021). Modelling the optimal delivery of spare parts to vessels: Comparison of three different scenarios. Cogent Engineering, 8(1), 1920565. Vukić, L., Stazić, L., Pijaca, M., & Peronja, I. (2021). Modelling the optimal delivery of spare parts to vessels: Comparison of three different scenarios. Cogent Engineering, 8(1), 1920565. https://doi.org/10.1080/23311916.2021.1920565 Zimek, A., & Filzmoser, P. (2018). There and back again: Outlier detection between statistical reasoning and data mining algorithms. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 8(6), 1-26. https://doi.org/10.1002/widm.1280 Journal of Industrial Engineering and Management, 2021 (www.jiem.org) Journal of Industrial Engineering and Management, 2021 (www.jiem.org) Journal of Industrial Engineering and Management, 2021 (www.jiem.org) Article’s contents are provided on an Attribution-Non Commercial 4.0 Creative commons International License. Readers are allowed to copy, distribute and communicate article’s contents, provided the author’s and Journal of Industrial Engineering and Management’s names are included. It must not be used for commercial purposes. To see the complete license contents, please visit https://creativecommons.org/licenses/by-nc/4.0/. -621-
https://openalex.org/W2970599780	https://www.iiste.org/Journals/index.php/JHMN/article/download/49092/50725	English	null	Assessment of Knowledge of Obstetric Danger Signs During Pregnancy Among Pregnant Mothers in Wolaita Sodo Hospitals and Sodo Town Health Centers in 2017	Journal of Health, Medicine and Nursing	2,019	cc-by	6,563	www.iiste.org www.iiste.org Journal of Health, Medicine and Nursing Journal of Health, Medicine and Nursing ISSN 2422-8419 An International Peer-reviewed Journal DOI: 10.7176/JHMN Vol.65, 2019 DOI: 10.7176/JHMN Abstract Background: Pregnancy complications are the major health problems among women in developing countries. Approximately 529,000 women die from pregnancy related causes annually and almost all (99%) of these maternal deaths occur in developing countries. Objective: To assess knowledge of pregnant mothers towards obstetric danger signs during pregnancy in Wolaita Sodo town health centers and hospital, April- May, 2017.Methods: Institutional based cross-sectional quantitative study was conducted in April-May,2017, on a sample of 365 mothers who are at least 4 months pregnant for prim gravid and multpara pregnant mothers selected from,11 urban Kebeles of Wolaita sodo town and simple random sampling technique was employed to select the study participants.Results: A total of 365 pregnant mothers were enrolled in the study giving a response rate of 100%. The study participants from selected urban kebeles constituted 100%. (46%) of the study participants were found to have poor knowledge. Respondents who got obstetric danger signs information from health institution had 2 times good knowledge of obstetrics danger signs than those respondents who didn’t get information from the same source. Participants who had formal education have 2.5 times knowledgeable than a participants who had informal education obstetric danger signs.Conclusion: Knowledge of study participants was assessed that 54% of participants have knowledge about obstetric danger signs during pregnancy and only 34% have knowledge about obstetrics danger signs during labor and child birth. Publication date: August 31st 2019 Publication date: August 31st 2019 Assessment of Knowledge of Obstetric Danger Signs During Pregnancy Among Pregnant Mothers in Wolaita Sodo Hospitals and Sodo Town Health Centers in 2017 Tilahun Saol Tura* Afflation: Department of Neonatology and Pediatric Nursing, College of Health Sciences and Medicine,Wolaita Sodo University, Wolaita Sodo, Ethiopia Masresha Assefa Demissie Afflation: Department of Neonatology and Pediatric Nursing, College of Health Sciences and Medicine,Wolaita Sodo University, Wolaita Sodo, Ethiopia Masresha Assefa Demissie Afflation: Department of Neonatology and Pediatric Nursing, College of Health Sciences and Medicine,Wolaita Sodo University, Wolaita Sodo, Ethiopia Statement of the problem Every minute, a woman dies due to causes related to pregnancy, childbirth and postnatal period [13].Maternal deaths are avoidable, if women with complications are able to identify and seek appropriate emergency obstetric care which makes a difference between life and death [14]. Maternal deaths have both direct and indirect causes. Around 80% of maternal deaths worldwide are brought about by direct obstetric complications the five major global causes of maternal death are:- severe bleeding (mostly bleeding postpartum), infections (also mostly soon after delivery), unsafe induced abortion, hypertensive disorders in pregnancy (eclampsia) and obstructed labour. Globally, about 80% of maternal deaths are due to these causes [6]. Hemorrhage alone accounts for one third of all maternal deaths in Africa, yet many of these deaths are preventable. Severe bleeding after birth can kill a healthy woman within two hours if she is unattended. Obstetric fistula resulting from obstructed labor is a long term complication suffered by as many as two million women (7). Indirect causes such as malaria, diabetes, hepatitis, anemia and other cardiovascular disorders which are aggravated by pregnancy can also lead to maternal death [1, 3].Awareness of the danger signs of obstetric Hemorrhage alone accounts for one third of all maternal deaths in Africa, yet many of these deaths are preventable. Severe bleeding after birth can kill a healthy woman within two hours if she is unattended. Obstetric fistula resulting from obstructed labor is a long term complication suffered by as many as two million women (7). Indirect causes such as malaria, diabetes, hepatitis, anemia and other cardiovascular disorders which are aggravated by pregnancy can also lead to maternal death [1, 3].Awareness of the danger signs of obstetric complications is the essential first step in accepting appropriate and timely referral to obstetric and newborn care. Raising awareness of women on danger signs of pregnancy, childbirth and the postpartum period improve mothers’ attitude to seek medical care and is crucial for safe motherhood. (11) When mothers do not recognize the danger signs in pregnancy, adverse effects can occur to the mother, the unborn baby, or the pregnancy itself. Adverse effects include: Illness or death of the mother, for instance, severe bleeding can lead to anemia or death of the mother, infection to the unborn baby through prematurely ruptured membranes, when amniotic fluid leaks from the vagina. Background Danger signs during the postpartum period include severe bleeding following childbirth, loss of consciousness after 1 Journal of Health, Medicine and Nursing www.iiste.org DOI: 10.7176/JHMN ISSN 2422-8419 An International Peer-reviewed Journal Vol.65, 2019 childbirth, and fever. Hemorrhage remains the leading cause of maternal mortality, accounting for approximately one third of deaths [7].Many of the complications that result in maternal deaths contributing to prenatal deaths are unpredictable, and their onset can be both sudden and severe. The complications leading to maternal death can occur without warning at any time during pregnancy and childbirth [8].Low awareness of danger signs and symptoms during pregnancy, labor, delivery and postpartum contribute to delays in seeking and receiving skilled care. Awareness of the danger signs of obstetric complications is the essential first step in accepting appropriate and timely referral to obstetric and newborn care. Knowledge of obstetric danger signs and birth preparedness are strategies aimed at enhancing the utilization of skilled care during low risk births and emergency obstetric care in complicated cases in low income countries. Increased knowledge and awareness is essential for reducing delays in seeking health care and in reaching a health facility. Communities and individuals should be empowered not only to recognize pregnancy related risks, but they must also have the means to react quickly and effectively once such problems arise [9-11] p The national reproductive strategy of Ethiopia has given emphasis to maternal and newborn health so as to reduce the high maternal and neonatal mortality. The strategy focuses on the need to empower women, men, families and communities to recognize pregnancy related risks, and to take responsibility for developing and implementing appropriate response to them. One of the targets in the strategies is to ensure that 80% of all families recognize at least three danger signs associated with pregnancy related complications by 2010 in areas where health extension program is fully implemented [12]. Background Pregnancy complications are the major health problems among women in developing countries. Approximately 529,000 women die from pregnancy related causes annually and almost all (99%) of these maternal deaths occur in developing countries. The global maternal mortality is unacceptably high [1]. Although there was significant progress in all developing regions, the average annual percentage decline in the global maternal mortality ratio (MMR) was 3.1%, short of the millennium Development Goals (MDGs) target of 5.5%. Every day, almost 800 women still die due to pregnancy or child birth, and for every woman who dies 20 or more experience serious complications. One of the United Nations‟ MDGs is to reduce MMR by 75% by 2015 [2]. Developing countries account for 99% (284,000) of the global maternal deaths, the majority of which are in sub-Saharan Africa (162,000) and Southern Asia (83,000). These two regions accounted for 85% of global burden, with sub-Saharan Africa alone accounting for 56%.The average maternal mortality ratio in developing countries in 2010 was 240 per 100,000 births versus 16 per 100,000 in developed countries reflecting inequities in access to health services, and highlighting the gap between rich and poor. Sub-Saharan Africa had the highest maternal mortality ratio at 500 maternal deaths per 100,000 live births. According to a systematic analysis of progress towards Millennium Development Goal 5 more than 50% of all maternal deaths in 2008 were in only six countries (India, Nigeria, Pakistan, Afghanistan, Ethiopia, and the Democratic Republic of the Congo[3].The situation in Ethiopia is similar to the situation in many developing countries. In Ethiopia the levels of maternal mortality and morbidity are among the highest in the world and the current estimate of MMR is 353per 100, 000 live births and it is reported that Maternal deaths accounted for 30% of all deaths to women age 15-49 [4].Reduction of mortality and morbidity of both mother and newborn have been identified as priority areas need urgent attention by the health sector. Maternal morbidity and mortality could be prevented and minimized significantly if women and their families recognize obstetric danger signs and promptly seek health care [5-6].The commonest danger signs during pregnancy include severe vaginal bleeding, swollen hands/face and blurred vision. Key danger signs during labor and childbirth include severe vaginal bleeding, prolonged labor, convulsions, and retained placenta. Significance of the study As there is no adequate information on obstetric danger signs Knowledge, the study result will be vital and can be used as an input for maternal health curriculum, strategy and package establishment. This study will provide basic data on the issue that may help policy makers and as baseline data for MOH to reduce the highest maternal mortality rate of Ethiopia. In addition to this, this study can have the following importance to different stake holders. The outcome of the study can be an input for concerned policy makers in decision making process regarding obstetric danger signs for pregnant and delivered mothers. And also it serves as an input for health education program undertaken by different organizations so as to keep the pregnant and delivered mothers being aware of the consequence of obstetric danger signs. OBJECTIVE To assess knowledge of obstetric danger signs during pregnancy among mothers In Wolaita S Health centers and wolaita sodo teaching and referral hospital, SNNPR, Ethiopia April- May, 2 Statement of the problem In spite of great potential of knowledge, attitude and practice of obstetric danger signs in reducing the maternal and newborn deaths its status are not well known in most of Sub-Saharan Africa including Ethiopia [15] [ ] The study therefore aims to fill this gap by assessing the current status of knowledge of danger signs among mothers in the study area. Statement of the problem If not attended to, this can lead to fetal or neonatal morbidity and mortality, termination of a pregnancy before term in vaginal bleeding (9). Maternal hypertension or fever, can lead to increased numbers of neonatal deaths or prematurely born babies who may eventually die due to inadequate facilities to care for them [2].A mothers’ death in childbirth denies her children their natural, primary care giver and significantly increases the risk that her infant will die or fail to survive to age 5. A mothers’ death also has an extremely detrimental effect on her children’s access to education and health care. Many children who survive without mothers also risk being emotionally lost [1-3]. Most maternal deaths are avoidable, as the health care solutions to prevent or manage complications are well known. All women need access to antenatal care in pregnancy, skilled care during childbirth, and care and support in the weeks after childbirth. It is particularly important that all births are attended by skilled health professionals, as timely management and treatment can make the difference between life and death [1,2]. According to the Ethiopian Federal Ministry of Health, only10% of the deliveries is attended by health professionals. In one nation where the maternal mortality ratio is 353 per 100,000 live and IMR 51.1/1000 and NMR 27.7/1000 live births which are the highest in the world. In Ethiopia, there is little information about the knowledge of obstetric danger signs during pregnancy since the introduction of Health Extension Workers (HEWs), despite the national Reproductive strategy aim to raise the awareness to 80% in the area in which HEW are deployed [12].Studies conducted in Aleta Wondo district, indicated that the knowledge level of pregnant women about obstetric danger signs (during pregnancy, childbirth and postpartum period) was low and affected by residential area. Therefore, the identified deficiencies in awareness should be addressed through maternal and child health services by designing an appropriate strategies including provision of targeted information, education and 2 Journal of Health, Medicine and Nursing g g ISSN 2422-8419 An International Peer-reviewed Journal DOI: 10.7176/JHMN Vol.65, 2019 DOI: 10.7176/JHMN communication. In spite of great potential of knowledge, attitude and practice of obstetric danger signs in reducing the maternal and newborn deaths its status are not well known in most of Sub-Saharan Africa including Ethiopia [15]. communication. Study design An institution based cross- sectional study design was conducted. Source population p p Women who were attending health centers and hospital at Sodo town within the study period in 2017. Study population y p p All pregnant mothers who attend ANC follow up at Sodo town administration health centers and hospital which are included under study. S l i d t i ti All pregnant mothers who attend ANC follow up at Sodo town administration health centers and hospital which are included under study. Study area and period y p The study was conducted in SNNPR, in Wolaita zone, Sodo town. WolaitaSodo town is 380km and 156km from Addis Ababa and Hawassa respectively. Sodo town is 82.1square km and consists of 11 urban kebeles. The weather condition of the town is “dega” with mean annual temperature of 15.1 to 20 degree Celsius. The town is found within the latitude of 6.48 up to 6.53 north and longitude of 37.44 up to 37.46 east. The mean annual rain fall of the town is about 1200 to 1400 milt. The total population of the town is 121,538. Among which 64,149 (52.8%)are males and 57,389 (47.2%) are females. WolaitaSodo town has 28,725 households. Among the women, 33,132 are in reproductive age group (15 to 49 years). There are various governmental, non-governmental and private health institutions in the town. There are two hospitals, three health centers, six health posts and thirty private clinics. This study was conducted in 3 health centers and 1 governmental hospital in SNNPR,SODDO TOWN , from April- MAY , 2017 . Study design Sample size determination The following assumptions were made to determine the sample size: The formula to calculate the sample size ; n= (zα/2)²p (1 p)/ D² The formula to calculate the sample size ; n= (zα/2)²p (1-p)/ D², METHODS AND MATERIALS Study area and period Where n=number of the study subjects Where n=number of the study subjects y j Z= the standardized normal distribution curve value for the 95% confidence interval (1.96) Z= the standardized normal distribution curve value fo P= the level of Knowledge of obstetric danger signs during pregnancy in Tigray region the study areas which is 61.1% have knowledge towards obstetrics danger signs during pregnancy. P= the level of Knowledge of obstetric danger signs during pregnancy in Tigray region the study area 61 1% h k l d t d b t t i d i d i 61.1% have knowledge towards obstetrics danger signs during pregnancy. g g g g p g y d=the desired precision of the estimate (the margin of error between the sample and population, 5%) = (1.96)²x 0.611 (1-0.611) / (0.05)² d=the desired precision of the estimate (the margin of error between the sample and population, 5%) = (1.96)²x 0.611 (1-0.611) / (0.05)² 365+18( 5% N d t t ) 383 i th fi l l i =365+18( 5% None respondent rate)=383 is the final sample size. 3 3 Journal of Health, Medicine and Nursing ISSN 2422-8419 An International Peer-reviewed Journal DOI: 10.7176/JHMN Vol.65, 2019 Sodo town Health centers and WSUTRH (560 total eligible population per month) Soddo HC 146 Geneme HC 120 Wadu HC 103 WSUTRH 187 100 83 71 129 Total sample size 383 Figure 1 Schematic presentations of Sample allocationInclusion criteria and Exclusion criteria Mothers who are mentally and physically capable of were interviewed and those who were volunteer to participate in the study were included and those mothers who were critically sick, pregnant health worker, mothers who are in labor and delivery service were excluded from study. Figure 1 Schematic presentations of Sample allocationInclusion criteria and Exclusion criteria Figure 1 Schematic presentations of Sample allocationInclusion criteria and Exclusion criteria Mothers who are mentally and physically capable of were interviewed and those who were volunteer to participate in the study were included and those mothers who were critically sick, pregnant health worker, mothers who are in labor and delivery service were excluded from study. g p p Mothers who are mentally and physically capable of were interviewed and those who were volunteer to participate in the study were included and those mothers who were critically sick, pregnant health worker, mothers who are in labor and delivery service were excluded from study. Data collection and Data quality control Data collection and Data quality control q y Data collection tool: data collection was carried out by structured questionnaire which consisting of two parts: Socio demographic and knowledge towards obstetric danger sign The response was cared or written with pencil in provided space. Data processing and analysis The collected data was checked for its completeness by investigators. The data was tallied using a tally sheet and analyzed manually using scientific calculator after its completeness checked. Then analyzed data was presented using table, charts and texts. g The nature of the study and associated risk and benefit was explained and then written consent was obtained for the participants. Where n=number of the study subjects Ethical clearance and consideration Ethical clearance letter was obtained from Research Ethics Committee (RIC) of Wolaita Sodo University health science and Medical College. Official letter of cooperation was obtained from WSU before conducting the interviews and observation was given to the participants and participants were assured of voluntary participation, confidentiality, anonymity and freedom to withdraw from the study at any time. Procedure for data collection Data collectors were interviewed the respondents and filled the questionnaires. The collectors were explained unclear questions if necessary during filling of the questionnaire. Finally, the filled questionnaires were collected by principal data collectors; it was tallied manually on tally sheet. The data was analyzed by using scientific calculator. Their statistically significance was assessed with the dependent variables. Finally using international and national data the result was compared, discussed and conclusion and recommendation was forwarded. The questionnaire was first Prepared in English and then translated into Amharic language for better understanding by respondents. The Amharic version of the questionnaire was then translated back to English language to check for its consistency. Prior to the beginning of data collection each question was reached as to understand what variables were measured in each question and pretest was done by 10% of sample size. Questionnaires were reviewed for completeness at end of administering questions by group members. Students were cross checked the questionnaires for completeness every day. for completeness every day. Data processing and analysis Obstetric danger signs Knowledge, during pregnancy, child birth and postnatal period in Wolaita Sodo town. Obstetric danger signs Knowledge, during pregnancy, child birth and postnatal period in Wolaita Sodo town Socio-demographic characteristics of the study subjects Socio-demographic characteristics of the study subjects Socio demographic characteristics of the study subjects A total of 365 pregnant mothers were enrolled in the study giving a response rate of 95.3%. The study participants from selected urban kebeles were participated by volunteer respondents. The mean ages of the respondents were 28 years. 136(37.2%) of the participant were Orthodox, 190 (52%) were protestant,29(8%)were musilims,10(2.8%)were others by religion. Majority, 356 (97%) of the women were married and most 316(86.5%) 4 4 Journal of Health, Medicine and Nursing www.iiste.org of the respondents were wolaita by ethnicity. 188(51.5%)were housewives .128(35%) had completed primary schoo,90(24.6%)had completed secondary school. 101 (27.7%) of the respondents had income between 1001 to 1500 birr during the study period. (See Table 1) g y p ( ) Table1: Distribution of socio-demographic and economic variables of respondents, of selected kebeles of wolaita sodo Town, 2017. Distribution of socio-demographic and economic variables of respondents, of selected keb sodo Town, 2017. Variable Frequency Percentages Age in years 15-19 50 13.7 20-24 122 33.4 25-29 120 32.9 30-39 68 18.6 >40 5 1.4 Total 365 100 Marital Status Single 4 1.1 Married 352 96.4 Widowed 5 1.4 Divorced 4 1.1 Total 365 100 Religion Orthodox 136 37.3 Protestant 190 52 Muslim 29 8 Others 10 2.7 Total 365 100 Ethnicity Wolaita 316 86.5 Gamo 22 6 Amhara 9 2.5 Kambata 7 2 Others 11 3 Total 365 100 Occupation Housewife 188 51.5 Governmental employ 89 24.4 Private employ 35 9.6 Own business 53 14.5 Total 365 100 Educational status Not able to read and write 30 8.2 able to read and write 11 3 Primary School 128 35.5 Secondary School 85 23.3 Diploma 79 21.6 Degree 32 8.8 Total 365 100 Monthly incomes <500 69 19 501-1000 79 21.6 1001-1500 83 22.7 1501-2000 101 27.7 >2000 33 9 Total 365 100 Time taken to reach the nearest health center? 10-20 minutes 162 44.4 20-30 minutes 203 55.5 Total 365 100 5 Journal of Health, Medicine and Nursing iste.org DOI: 10.7176/JHMN Obstetric characteristics of the respondents p From total number of respondents 291(79.7%) had history of 2-4 pregnancies and 35(5.5%) mothers were pregnant for more than four times. Regarding first pregnancy age mothers got their first pregnancy at 20-29years. From the total number of mothers 291(79.7%) had 1-4 number of children. Regarding history of previous of pregnancy, 263(90.4%) had ANC follow up, among those who had ANC follow up 253(96%) had 4 visits and majority of mothers 347(95%) properly receiving ANC follow up for current pregnancy. From those attended ANC 237(81.4%) gave birth at health institutions were 136(69%) of the study participants had got danger signs information from health personnel followed by media ,friends and relatives with respective frequencies of 11.7%, 10.7% and 8.6% respectively. p y FIGURE 2 Total number of pregnancies among pregnant mothers who had ANC follow up at Sodo town administration health centers and hospitals, April _May, 2017. 0 100 200 300 400 ONE PREGNANCY 2-4 PREGNANCIES MORE THAN FOUR PREGNANCIES FIGURE 2 Total number of pregnancies among pregnant mothers who had ANC follow up at Sodo town administration health centers and hospitals, April _May, 2017. Obstetric danger signs Knowledge, during pregnancy, child birth and postnatal period in Wolaita Sodo c danger signs Knowledge, during pregnancy, child birth and postnatal period in Wolaita Sodo town. Out of the 365 respondents, 197(54%) reported that they had got information about obstetric danger sign during pregnancy. From those who had the information 78(39.6%) identified severe vaginal bleeding at any time during pregnancy as danger sign while decreased or absence of fetal movement was indicated by 40(20.3%). Swelling of the body 25(12.7%), feeling of tired 17(8.6%), unusual abdominal pain 14(7.1%), persistent headache and blurred vision 12(6%), persistent back pain 11(5.6 %) were also indicated by the study subjects as danger sign. Foul smelling vaginal discharge was mentioned by 11(5.6%) of the respondents and 14(7.1%) respondents mention other danger signs not mention here. (Table 2). g g ( ) Figure 3: Knowledge of pregnant mothers who had ANC follow up at Sodo town administration health centers and hospitals, April-May, 2017. 1st Qtr 72% 2nd Qtr 28% Figure 3: Knowledge of pregnant mothers who had ANC follow up at Sodo town administration health centers and hospitals, April-May, 2017. 6 www.iiste.org DOI: 10.7176/JHMN Table 2: knowledge of obstetric danger signs during pregnancy among pregnant mothers in wolaita Sodo town health centers and wolaita Sodo university teaching and referral hospital selected Kebeles of Wolaita Sodo town April-May, 2017. Obstetric danger signs Knowledge, during pregnancy, child birth and postnatal period in Wolaita Sodo Variable Category Frequency Percent Heard of “obstetric danger signs during pregnancy Yes 197 54 No 168 46 Total 365 100 Mentioned severe bleeding as danger sign during pregnancy Yes 78 39.6 No 119 60.4 Total 197 100 Mentioned swelling of the body as danger sign during pregnancy Yes 25 12.7 No 172 87.3 Total 197 100 Mentioned unusual abdominal pain as danger sign during pregnancy Yes 14 7.1 NO 183 92.9 Total 197 100 Mentioned severe headache or blurred vision as danger sign during pregnancy Yes 12 6 No 185 94 Total 197 100 Mentioned feeling very tired as danger sign during pregnancy Yes 17 8.6 NO 180 91.4 Total 197 100 Mentioned absent or decreased fetal movement as danger sign during pregnancy Yes 40 20.3 NO 157 79.7 Total 197 100 Mentioned persistent back pain as danger sign during Pregnancy Yes 11 5.6 No 186 94.4 Total 197 100 Mentioned foul smelling vaginal discharge as danger sign during pregnancy Yes 11 5.6 NO 186 94.4 Total 197 100 Mentioned high fever as danger sign during pregnancy Yes 12 6 NO 185 94 Total 196 100 The source of information about obstetric danger signs during pregnancy Health personnel 136 69 Relatives 17 8.6 Friends 21 10.7 Media 23 11.7 Total 197 100 Can any woman develop obstetric danger signs during pregnancy? Yes 14 7 No 134 68 Don’t know 49 25 Total 197 100 Table 2: knowledge of obstetric danger signs during pregnancy among pregnant mothers in wolaita Sodo town health centers and wolaita Sodo university teaching and referral hospital selected Kebeles of Wolaita Sodo town April-May 2017 7 7 Journal of Health, Medicine and Nursing www.iiste.org ISSN 2422-8419 An International Peer-reviewed Journal DOI: 10.7176/JHMN Vol.65, 2019 Variable Category Frequency Percent Is Information given about obstetric danger sign/s at the clinic? Yes 150 76.1 No 47 23.9 Total 197 100 Experienced any obstetric danger sign during pregnancy Yes 29 14.7 No 168 85.3 Total 197 100 Received antenatal care during last pregnancy (For multi gravid and delivered mothers only. Obstetric danger signs Knowledge, during pregnancy, child birth and postnatal period in Wolaita Sodo Yes 263 90.4 NO 28 9.6 Total 291 100 Properly receive antenatal care during current pregnancy Yes 347 95 No 18 5 Total 365 100 Number of antenatal visits during current pregnancy First 54 14.8 Second 86 23.5 Third 103 28.2 Fourth 108 29.5 >Fourth 14 4 Total 365 100 Place of delivery during last pregnancy (For multi gravid and delivered mothers only). Health center 237 81.4 Home 41 14 TBA 13 4.6 Total 291 100 Knowledge on danger signs during child birth and postnatal period Knowledge on danger signs during child birth Out of the 365 respondents, 124(34%) reported that they had got information about obstetric danger during child birth and postnatal period. From those who had the information 52(42%) identified severe vaginal bleeding at any time during child birth and postnatal period as danger sign, while prolonged labor 31(25%) ,mali position 24(19.3%), retained placenta 17(13.7%) , were indicated by respondents. Table 3: knowledge of obstetric danger signs during child birth among pregnant mothers in Wolaita Sodo town health centers and Wolaita sodo university teaching and referral hospital Apri-May, 2017. Variable Category Frequency Percent(%) Have information about danger sign during labor and child-birth? Yes 124 34 NO 241 66 Total 365 100 Mentioned Sever vaginal bleeding as danger signs Yes 52 42 NO 72 58 Total 124 100 Mentioned Prolonged labor as danger signs Yes 31 25 NO 93 75 Total 124 100 Mentioned Mali position as danger signs Yes 24 19.3 NO 100 80.7 Total 124 100 Mentioned Retained placenta as danger signs Yes 17 13.7 NO 107 86.3 Total 124 100 Knowledge on danger signs during child birth and postnatal period Knowledge on danger signs during child birth and postnatal period 8 www.iiste.org DOI: 10.7176/JHMN Table 4: knowledge of obstetric danger signs during postnatal period among pregnant mothers in wolaita sodo town health centers and wolaita sodo university teaching and referral hospital selected kebeles of Wolaita sodo town Apri-May, 2017. Table 4: knowledge of obstetric danger signs during postnatal period among pregnant mothers in wolaita sodo town health centers and wolaita sodo university teaching and referral hospital selected kebeles of Wolaita sodo town Apri-May, 2017. DISCUSSION According to this study 54% had heard obstetric danger signs and from those mothers about 150(76.1%) got information from clinic and it is less than the study done in Mekele that 79.6% respondents had information about danger signs of pregnancy (17). This difference might be due to socio-economic and geographical location. A di hi d 69% f h d i i h d d i i f i f h l h l According to this study 54% had heard obstetric danger signs and from those mothers about 150(76.1%) got information from clinic and it is less than the study done in Mekele that 79.6% respondents had information about danger signs of pregnancy (17). This difference might be due to socio-economic and geographical location. According to this study 69% of the study participants had got danger signs information from health personnel followed by media, friends and relatives with respective frequencies of 11.7%, 10.7% and 8.6% respectively was higher than the study done in Northern Nigeria, about a quarter of respondents in Birnin Gwari cluster and 13% in Kunchi cluster indicated that danger sign songs in the media were the sources of information about safe motherhood. Over 10% of women in Birnin Gwari cluster and almost 30% in Kunchi cluster indicated that friends and neighbors were their sources of information of safe motherhood (37). The differences might be due to cultural and geographic location variations and the area where the study conducted. Based on this study from those who had the information of obstetric danger signs 39.6% identified severe vaginal bleeding at any time during pregnancy which is lower than the findings in Aleta Wondo district(45.9%) (15), higher than the findings in Burkina Faso (39.4%) and Guatemala (31.0%) (38, 39).This difference might be due to socio-cultural difference and difference of sample size used. According to this study 46% of the study respondents were unable to mention an obstetric danger sign which is more than the study done in Aleta Wondo district 39.1% and Tsegedie district 35.1% and (15,16).The differences might be due to socio economic and health education provided. This study revealed that the respondents‟ occupation, educational level, monthly income, first pregnancy age, having ANC follow up, time to nearest health facility, source of information and place of residence were significantly associated with knowledge of obstetric danger signs during pregnancy. Obstetric danger signs Knowledge, during pregnancy, child birth and postnatal period in Wolaita Sodo Heard information about danger sign during postnatal period Yes 104 28.5% NO 261 71.5% Total 365 100% Mentioned Fever as danger signs Yes 21 20.2% NO 87 79.8% Total 104 100% Mentioned vaginal Bleeding as danger sins Yes 39 37.5% NO 71 62.5% Total 104 100% Mentioned Offensive vaginal discharge as danger signs Yes 26 25% NO 83 75% Total 104 100% Mentioned Loss of consciousness as danger signs Yes 18 17.3% No 91 83.7% Total 104 100% REFERENCES 1. UNICEF. A Global Overview of Maternal Mortality.2010 1. UNICEF. A Global Overview of Maternal Mortality.2010 1. UNICEF. A Global Overview of Maternal Mortality.2010 2. UNFPA. Safe Motherhood Stepping up Efforts to Save Mothers‟ Lives 2011. 2. UNFPA. Safe Motherhood Stepping up Efforts to Save Mothers‟ Lives 2011. 3. WHO. Trends in Maternal Mortality: 1990 to 2010, WHO, UNFPA, UNICEF and World Bank. Geneva 2012. 4. Central Statistical Agency and ORC Marco. Ethiopian Demographic and Health Survey, 2011. Addis Ababa, Ethiopia and Calverton, Maryland, USA; 2012. y 4. Central Statistical Agency and ORC Marco. Ethiopian Demographic and Health Survey, 2011. Addi Ethiopia and Calverton, Maryland, USA; 2012. 5. Thaddeus S, Maine D. Too far to walk: maternal mortality in context. Social Science Medicine 1994; 38: 1091- 1110 6. Jackson D. Loveday M, Doherty T, Mbombo N, Wington A, Matizirifa L,et al. Community Based Situation Analysis: Maternal and Neonatal Follow-up Care. Durban: Health Systems Trust.2006. p. 9-13. 7. Dutta DC. Text Book of Obstetrics. 6th ed. Calcutta: New Central Book Agency (P) Ltd; 2004.p. 121 g y ( ) p 8. Rashad WA, Essa RM.Women‟s Awareness of Danger Signs of Obstetrics Complications in Egypt. Journal of American Science 2010; 6(10): 1299-1306. 9. Kabakyenga JK, Ostergren PO, TuryakiraE,Pettersson K. Knowledge of obstetric danger signs and birth preparedness practices among women in rural Uganda. BMC pregnancy and Child birth 2011,11: 73 10. Mihret H, Mesganaw F: Birth preparedness and complication readiness among women in Adig northern Ethiopia. EthiopJHealth Dev 2008, 22(1):14–20. p p ( ) 11. JHPIEGO. Monitoring birth preparedness and complication readiness: tools and indicators for Maternal and Newborn health Baltimore.2004. 12. Federal Democratic Republic of Ethiopia, Ministry of Health. National Reproductive Strategy, 2006-2015. Addis Ababa, Ethiopia: FMOH; 2006.p. 16-21. 13. Abouzahr C. Safe Motherhood:a brief history of the global movement. World Health Organization, Geneva, Switzerland 1947-2002. 14. WHO. Why do so many women still die in pregnancy or childbirth? 2005. 15. Mesay Hailu et al, knowledge about obstetric danger signs among pregnant women in Aleta Wondo district, Sidama zone, Southern Ethiopia. Ethiopian Journal of Health Development 2007; 20(1); 25-32. p43 16. Hailu D, Berhe H (2014) Knowledge about Obstetric Danger Signs and Associated Factors among Mothers in Tsegedie District, Tigray Region, Ethiopia 17. Gebrehiwot et al. Awareness of Danger Signs of Pregnancy and its Associated Factors among Pregnant Women who Visit ANC in Mekelle Public Hospitals, 2014. 18. USAID. DISCUSSION It was found that residence had significant association with knowledge of obstetric danger signs during pregnancy that urban residents had 2.2 times good knowledge than rural respondents [2.2) . This is greater than the study done in Aleta Wondo(15). Based on this study it was found that the study respondents’ occupation was significantly associated with obstetric danger signs knowledge of the respondents: those respondents who were government employee had 2.1 times good knowledge than housewives. It is less than the study done in Egypt (8) found that occupation appeared to influence women’s awareness of danger signs of obstetric complications. However, this finding is contrasted with study done in rural Tanzania (19). In this study, the respondents‟ educational level seems to play a role in having knowledge of obstetric danger signs that respondents who have had formal education had 2.1 times good knowledge than who have informal education and respondents who have had high level education had 3 times good knowledge than those respondents with informal education. This study is similar with the study done in Egypt (8) and rural Tanzania (19) which found that women with higher levels of education were more aware of danger signs of obstetric complications than women with lower or no formal education. This study found no statistically significant difference in knowledge of danger signs of obstetric danger signs with respondent’s number of ANC visits, similar to the study done in Tanzania (19), Pembe et al (2009:6) found that women who made four or more ANC visits were more aware of danger signs of obstetric complications than those who made less than four ANC visits, independent of gestational age at booking. This difference might be due health education given during ANC follow up. 9 9 Journal of Health, Medicine and Nursing ww.iiste.org DOI: 10.7176/JHMN ISSN 2422-8419 An International Peer-reviewed Journal DOI: 10.7176/JHMN Vol.65, 2019 ISSN 2422-8419 An International Peer-reviewed Journal DOI: 10.7176/JHMN Vol.65, 2019 Declaration: We declare that this article is our original work. It was not copied from any other research and all references used for the preparation of this article were cited properly. Regarding fund of research; it was fully covered by researchers. CONCLUSION Based on this study finding majority of mothers had got information about obstetric danger sign different areas but few mothers have knowledge about the complications of obstetric danger signs and its severity. are significantly associated with obstetric danger signs knowledge. The researcher recommend that concerned body should act to increase the knowledge of the community regarding obstetric danger sign. Researcher recommend also other scholars to conduct in this area because this research only describes t y but it does not show the factor associate with it Acknowledgement First of we would like to thank Almighty God. Also we would like acknowledge our friends and senior staff who helped us during of this article preparations. REFERENCES Family Planning and Reproductive Health Indicators Database. Existence of a safe motherhood strategic plan to promote access and quality of safe motherhood services. Measure Evaluation Population and Reproductive health 2010. 19. Pembe et al: Rural Tanzanian women‟s awareness of danger signs of obstetric complications.BMC Pregnancy Child birth March 2009, 9; 12. 20. Nepal. Improving Maternal, Newborn and Child Health in the South East Asia Region. 2005. p. 63. 21. Perreira KM,Bailey PE, De Bocaletti E, Hurtado E, Recinos de Villagran S, Matute J, Increasing awareness of danger signs in pregnancy through community and clinic based education in Guatemala. Maternal Child Health Journal 2002; 6(1): 19-28. ( ) 22. Urassa DP, Pembe AB and Mganga F. Birth preparedness and complication readiness among women in Mpwapwa district, Tanzania. Tanzania Journal of Health Research January 2012; 14(1): 3-6 10 Journal of Health, Medicine and Nursing www.iiste.org DOI: 10.7176/JHMN 23. Ali AA, Rayis DA , Abaker AO, Adam I. Awareness of danger signs and nutritional education among pregnant women in Kassala, Eastern Sudan. Sudanese Journal of Public Health 2010; 5(4): 179-181. A, Rayis DA , Abaker AO, Adam I. Awareness of danger signs and nutritional education among pregnant men in Kassala, Eastern Sudan. Sudanese Journal of Public Health 2010; 5(4): 179-181. 24. Alam AY,Qureshi AA, Adil MM, Ali H. Comparative study of Knowledge, Attitude and Practices among Antenatal Care Facilities utilizing and non- utilizing women in Islamabad. Journal of Pakistan Medical Association 2004; 55(2): 2-4. 25. Anya SE, Hydara A, Jaiteh LES. Antenatal care in The Gambia: Missed opportunity for information, education and communication. BMC Pregnancy and Childbirth 2007, 8:9. 26. Agboatwalla M, Akram DS. Impact of Health education on mothers‟ knowledge of preventive health practices, Karachi Pakistan. PubMed Journal of Pakistan Medical Association 1997; 27(4): 199- 202. 27. Cham M, Sundby J, Vangen S. Maternal mortality in rural Gambia, a qualitative study on access to emergency obstetric care BMC Reproductive Health 2005; 2:3.44 28. Sarode VM. Does illiteracy influence pregnancy complications? Among women in the slums o Mumbai. International Journal of Sociology and Anthropology 2010; 2(5): 82-94. gy p gy ( ) 29. WHO. Maternal, Neonatal, and child health. Strategy Global Health program 2009. 29. WHO. Maternal, Neonatal, and child health. Strategy Global Health program 2009. 30. Nikiema B, Beninguisse G, Haggerty JL.Providing information on pregnancy complications during antenatal visits: unmet educational needs in sub- Saharan Africa. REFERENCES Health policy and planning 2009; 34: 370-374. 31. Chongo C. A study to determine knowledge attitude and practice of women in child bearing age towards danger signs in pregnancy in Lusaka, urban Lusaka 1885. 32. Marco A. Knowledge and practice of maternal health care in Indonesia 2003. 33. I. J. Hasan, N. Nisar. Women‟s Perceptions regarding Obstetric Complications and Care in a Poor Fishing Community in Karachi, 1999. 34. Making pregnancy and childbirth safer. (1997) Policy Brief. Washington, DC: USA World Population reference bureau. Available: www.measurecommunication.org. 35. World health organization. (2005) make every mother and child count. Switzerland: Worl Organization. Available: http://www.who.int/Whr/ 2005/ hr 2005_en.pdf. 36. Debre Birhan Woreda Health Office (2013 report). 37. Okereke et al. Reproductive Health 2013, 10:57. Knowledge of safe motherhood among women in rural communities in northern Nigeria: implications for maternal mortality reduction. 38. Baya B, Sangli G, Maiga A. Measuring the effects of behavior change interventions in Burkina F population-based survey results. Maryland, USA: JHPIEGO; 2004 39. Becker F, Y.glesias C. Measuring the effects of behavior change and service deliver interventions in Guatemala with population-based survey. Maryland, USA:JHPIEGO; 2004. 11 11
https://openalex.org/W4362429666	https://aacr.figshare.com/articles/journal_contribution/Supplementary_Figure_5_from_Novel_Tumor_Subgroups_of_Urothelial_Carcinoma_of_the_Bladder_Defined_by_Integrated_Genomic_Analysis/22446227/1/files/39897275.pdf	English	null	Supplementary Figure 5 from Novel Tumor Subgroups of Urothelial Carcinoma of the Bladder Defined by Integrated Genomic Analysis	null	2,023	cc-by	49	Supplementary Figure 5. Unsupervised hierarchical cluster analysis of aCGH data from all tumor samples and genome-wide frequency plots of copy number alterations in individual ≥T2 clusters. Unsupervised hierarchical cluster analysis of aCGH data from all tumor samples and genome-wide frequency plots of copy number alterations in individual ≥T2 clusters.
https://openalex.org/W2775021600	https://bmcgenet.biomedcentral.com/track/pdf/10.1186/s12863-017-0569-4	English	null	Identification and integrated analysis of differentially expressed lncRNAs and circRNAs reveal the potential ceRNA networks during PDLSC osteogenic differentiation	BMC genomic data	2,017	cc-by	8,546	Gu et al. BMC Genetics (2017) 18:100 DOI 10.1186/s12863-017-0569-4 Gu et al. BMC Genetics (2017) 18:100 DOI 10.1186/s12863-017-0569-4 Identification and integrated analysis of differentially expressed lncRNAs and circRNAs reveal the potential ceRNA networks during PDLSC osteogenic differentiation Xiuge Gu1,2, Mengying Li1,2, Ye Jin1,2, Dongxu Liu1,2* and Fulan Wei1,2* Xiuge Gu1,2, Mengying Li1,2, Ye Jin1,2, Dongxu Liu1,2* and Fulan Wei1,2* © The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. * Correspondence: liudongxu@sdu.edu.cn; weifl@sdu.edu.cn 1Department of Orthodontics, Shandong Provincial Key Laboratory of Oral Tissue Regeneration, School of Stomatology, Shandong University, Wenhua Xi Road No. 44-1, Jinan, Shandong 250012, People’s Republic of China Full list of author information is available at the end of the article Background differentiation, there is not much data published on their potential networks and functions. To fully understand the impact of ceRNA crosstalk on PDLSC osteogenic differentiation, it will be crucial to inte- grate the lncRNA/circRNA-miRNA-mRNA competi- tive regulatory networks. In this study, we developed RNA sequencing (RNA-seq) with Illumina HiSeq2000 to comprehensively identify differentially expressed lncRNAs and circRNAs in normal and osteogenic inductive PDLSCs. Subsequently, the representative lncRNAs and circRNAs were further confirmed using quantitative real-time polymerase chain reaction (qRT-PCR). Finally, we predicted the ceRNA net- works of lncRNAs, circRNAs, miRNAs and mRNAs based on miRanda and investigated their potential regulatory roles via gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) ana- lysis. Our findings might provide new evidence for exploring the molecular mechanism of PDLSC osteo- genic differentiation. Periodontitis is a slowly progressive disease character- ized by the loss of periodontal tissue, which is the main cause of tooth loss [1]. With the development of stem-cell delivery therapeutics, periodontal ligament stem cells (PDLSCs) have been shown to produce typical periodontal ligament-like tissue to regenerate tissue damaged by periodontitis [2, 3]. Our previous studies also have shown the great potential of PDLSCs to regenerate periodontal tissue and form a bioengineered tooth root in miniature pigs [4–6]. The regeneration potency of PDLSCs con- tributes to their self-renewal and multi-differentiation capacity, especially osteogenic differentiation [7]. Exploring the molecular mechanisms of PDLSC osteogenic differen- tiation might provide new genetic strategies for periodontal regenerative medicine. Recently, through microarray analysis, microRNAs (miR- NAs) were identified and predicted to be involved in PDLSC osteogenic differentiation [8]. In addition, miR- NA146a, miRNA17 and miRNA22 have been demonstrated to regulate PDLSC osteogenic differentiation by modulating the expression of target genes at the post-transcriptional level [9–11]. However, recent studies have revealed that a new player, competing endogenous RNA (ceRNA), is essential for the circuitry of miRNAs and target genes [12]. By competing for common miRNA response elements (MREs), ceRNAs can break the balance between miRNAs and target genes to regulate the physiological and patho- physiological process [13]. These ceRNAs include various types of RNAs, such as long non-coding RNAs (lncRNAs), circular RNAs (circRNAs), messenger RNAs (mRNAs) and pseudogenes. Abstract Background: Researchers have been exploring the molecular mechanisms underlying the control of periodontal ligament stem cell (PDLSC) osteogenic differentiation. Recently, long noncoding RNAs (lncRNAs) and circular RNAs (circRNAs) were shown to function as competitive endogenous RNAs (ceRNAs) to regulate the effect of microRNAs (miRNAs) on their target genes during cell differentiation. However, comprehensive identification and integrated analysis of lncRNAs and circRNAs acting as ceRNAs during PDLSC osteogenic differentiation have not been performed. Results: PDLSCs were derived from healthy human periodontal ligament and cultured separately with osteogenic induction and normal media for 7 days. Cultured PDLSCs were positive for STRO-1 and CD146 and negative for CD31 and CD45. Osteo-induced PDLSCs showed increased ALP (alkaline phosphatase) activity and up-regulated expression levels of the osteogenesis-related markers ALP, Runt-related transcription factor 2 and osteocalcin. Then, a total of 960 lncRNAs and 1456 circRNAs were found to be differentially expressed by RNA sequencing. The expression profiles of eight lncRNAs and eight circRNAs were measured with quantitative real-time polymerase chain reaction and were shown to agree with the RNA-seq results. Furthermore, the potential functions of lncRNAs and circRNAs as ceRNAs were predicted based on miRanda and were investigated using Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analysis. In total, 147 lncRNAs and 1382 circRNAs were predicted to combine with 148 common miRNAs and compete for miRNA binding sites with 744 messenger RNAs. These mRNAs were predicted to significantly participate in osteoblast differentiation, the MAPK pathway, the Wnt pathway and the signaling pathways regulating pluripotency of stem cells. Among them, lncRNAs coded as TCONS_00212979 and TCONS_00212984, as well as circRNA BANP and circRNA ITCH, might interact with miRNA34a and miRNA146a to regulate PDLSC osteogenic differentiation via the MAPK pathway. Conclusions: This study comprehensively identified lncRNAs/circRNAs and first integrated their potential ceRNA function during PDLSC osteogenic differentiation. These findings suggest that specific lncRNAs and circRNAs might function as ceRNAs to promote PDLSC osteogenic differentiation and periodontal regeneration. Keywords: Long non-coding RNAs, Circular RNA, Competitive endogenous RNA, Periodontal ligament stem cells, Osteogenic differentiation, RNA sequencing Gu et al. BMC Genetics (2017) 18:100 Page 2 of 13 Page 2 of 13 Methods BMC Genetics (2017) 18:100 Gu et al. BMC Genetics (2017) 18:100 Alkaline Phosphatase (ALP) staining and ALP activity assay Alkaline Phosphatase (ALP) staining and ALP activity assay incubated for 15 min with citrate buffer and phosphatase substrate (Sigma-Aldrich), and then quantified by spec- trophotometric absorbance at 405 nm. ALP staining and activity assay were performed using an Alkaline Phosphatase Kit (Sigma-Aldrich) as described previously [24]. Briefly, PDLSCs were fixed with 70% ethanol for 30 min and stained with a solution of so- dium nitrite, FRV alkaline and naphthol AS-BI alkaline for 15 min away from light. For the ALP activity assays, total protein was isolated from two groups of PDLSCs, Methods All protocols for the handling of human tissues were ap- proved by the Research Ethics Committee of Stomatology Hospital of Shandong University, China (G201401601). Informed consent was obtained from all donors. Cell culture and osteogenic induction In this study, PDLSCs were derived from the middle third of the root surface of normal human impacted third molars, which were collected from 18- to 20-year- old patients at the Department of Oral Maxillofacial Surgery, Stomatology Hospital of Shandong University, using the explant culture method. Then, they were cultured with normal media, consisting of α-modification of Eagle’s media (HyClone, South Logan, UT, USA), 10% fetal calf serum (Gibco BRL, Grand Island, NY, USA), 100 U/ml penicillin and 100 μg/ml streptomycin (Invitrogen, Carlsbad, CA, USA) at 37 °C in 5% carbon dioxide. The stemness of PDLSCs was characterized by the scanning of cell surface markers (STRO-1, CD146, CD31, and CD45) through flow cytometric analysis (Becton, Dickinson and Company, Franklin Lakes, NJ, USA). For osteogenic differentiation, PDLSCs were cultured with osteogenic inductive media supplemented with 10 nM dexamethasone, 10 mM β-glycerophosphate and 50 μg/ml vitamin C (Sigma-Aldrich, St. Louis, MO, USA). Through separately culture with osteogenic induction and normal media for 7 days, PDLSCs were divided into two groups: induced and non-induced groups. All cells in this study were used at passage number 3. LncRNA is a class of non-coding RNA (ncRNA) tran- scripts longer than 200 nucleotides [14]. Recent studies have reported that lncRNAs were involved in the osteogenic differentiation process [15]. For example, up-regulation of lncRNA HIF 1α-anti-sense 1 induced by TGFβ-mediated targeting of sirtuin 1 promotes the osteogenic differentiation of human bone marrow stromal cells [16]. Furthermore, lncRNA ANCR was proved to be critical in regulating the PDLSC osteogenic differentiation via the Wnt signaling pathway [17]. CircRNA is another new class of RNA composed of more than one exon with a covalently closed loop [18]. Com- pared to linear RNA, this circular structure is more stable and resistant to RNase R [19]. Emerging evidence has re- vealed that circRNAs participate in osteogenic differenti- ation [20, 21]. For instance, 154 differentially expressed circRNAs were found to associate with bone morpho- genetic protein 2-induced osteogenic differentiation of MC3T3-E1 cells [22]. Moreover, circRNAs were predicted to have potential roles in osteogenesis of PDLSCs [23]. Although several lncRNAs, circRNAs and miRNAs are suggested to be involved in PDLSC osteogenic Page 3 of 13 Gu et al. Quantitative real-time PCR (qRT-PCR) Quantitative real-time PCR (qRT-PCR) Total RNA was isolated from two groups of PDLSCs. For analysis of linear transcripts, 1 μg of RNA per sample was reverse transcribed into cDNA using a cDNA Reverse Transcription Kit (Takara, Tokyo, Japan). Fig. 1 Identification and osteogenic differentiation of PDLSCs. a PDLSCs were derived from periodontal ligament explants. b PDLSCs were cultured with non-induced media at passage number 3. c-f PDLSCs were positive for STRO-1 and CD146 and negative for CD31 and CD45. g, h ALP activity was enhanced in osteo-induced PDLSCs (Induced), as evidenced by ALP staining and ALP activity assay. i-k Compared with the non-induced group, the induced group showed up-regulated expression of the osteogenic genes ALP, Runx2 and OCN by qRT-PCR. All PCRs were performed in triplicate. The data are represented as means ± SD. , P < 0.05; , P < 0.01 Fig. 1 Identification and osteogenic differentiation of PDLSCs. a PDLSCs were derived from periodontal ligament explants. b PDLSCs were cultured with non-induced media at passage number 3. c-f PDLSCs were positive for STRO-1 and CD146 and negative for CD31 and CD45. g, h ALP activity was enhanced in osteo-induced PDLSCs (Induced), as evidenced by ALP staining and ALP activity assay. i-k Compared with the non-induced group, the induced group showed up-regulated expression of the osteogenic genes ALP, Runx2 and OCN by qRT-PCR. All PCRs were performed in triplicate. The data are represented as means ± SD. , P < 0.05; *, P < 0.01 Page 4 of 13 Gu et al. BMC Genetics (2017) 18:100 Convergent primers were designed to detect lncRNAs and mRNAs. Divergent primers were designed to detect the circular form of circRNAs. Relative transcript levels were measured with quantitative PCR using a Roche LightCycler®480 sequence detection system (Roche, Basel, Switzerland) following the manufacturer’s protocol. Each Fig. 2 The apparent variations of differentially expressed lncRNAs and circRNAs. (a, b) LncRNAs and circRNAs were broadly distributed across the 24 pairs of human chromosomes according to their locations. The inner blue ring corresponds to the non-induced group; the outer yellow ring corresponds to the induced group. c Among differentially expressed lncRNAs, 17 common lncRNAs and 180 specific lncRNAs in the non-induced group and 763 specific lncRNAs in the induced group were identified. d Among differentially expressed circRNAs, 95 common circRNAs and 642 specific circRNAs in the non-induced group and 719 specific circRNAs in the induced group were identified. Total RNA was extracted from two groups using Trizol (Invitrogen) according to the manufacturer’s protocols. Strand-specific cDNA libraries were constructed following a previously described protocol [25] and were sequenced using an Illumina HiSeq2000 sequencer (LC Biotech, Hangzhou, China) by performing a paired-end run with a 100 bp read length. The raw reads were processed by removing the adaptor reads and low-quality tags. All sub- sequent analyses were performed using clean reads. Quantitative real-time PCR (qRT-PCR) e Differentially expressed lncRNAs, consisting of 777 up-regulated lncRNAs and 183 down-regulated lncRNAs, are displayed in the heatmap. f Differentially expressed circRNAs, consisting of 766 up-regulated circRNAs and 690 down-regulated circRNAs, are displayed in the heatmap Fig. 2 The apparent variations of differentially expressed lncRNAs and circRNAs. (a, b) LncRNAs and circRNAs were broadly distributed across the 24 pairs of human chromosomes according to their locations. The inner blue ring corresponds to the non-induced group; the outer yellow ring corresponds to the induced group. c Among differentially expressed lncRNAs, 17 common lncRNAs and 180 specific lncRNAs in the non-induced group and 763 specific lncRNAs in the induced group were identified. d Among differentially expressed circRNAs, 95 common circRNAs and 642 specific circRNAs in the non-induced group and 719 specific circRNAs in the induced group were identified. e Differentially expressed lncRNAs, consisting of 777 up-regulated lncRNAs and 183 down-regulated lncRNAs, are displayed in the heatmap. f Differentially expressed circRNAs, consisting of 766 up-regulated circRNAs and 690 down-regulated circRNAs, are displayed in the heatmap Convergent primers were designed to detect lncRNAs and mRNAs. Divergent primers were designed to detect the circular form of circRNAs. Relative transcript levels were measured with quantitative PCR using a Roche LightCycler®480 sequence detection system (Roche, Basel, Switzerland) following the manufacturer’s protocol. Each Gu et al. BMC Genetics (2017) 18:100 Page 5 of 13 Table 1 Statistical data of RNA-Seq reads for two samples Sample Raw reads Clean reads Unique lncRNAs Unique circRNAs Unique mRNAs Non-induced 91.4 million 90.9 million 11,529 5913 66,134 Induced 113.8 million 112.4 million 9511 3162 52,227 Table 1 Statistical data of RNA-Seq reads for two samples Fig. 3 Differentially expressed lncRNAs/circRNAs validated by qRT-PCR. a Convergent primers were designed to detect eight lncRNAs with P < 0.05, FC ≥2 and FPKM in at least one of the samples ≥4. LncRNAs coded as TCONS_00019601, TCONS_00227764, TCONS_00254538, TCONS_00198784 and TCONS_00136898 were up-regulated in the induced group compared with the non-induced group, and TCONS_00085268, TCONS_00125934, and TCONS_00115113 were down-regulated. b Divergent primers were designed to detect the circular form of circRNAs with P < 0.05, FC ≥2 and RPM in at least one of the samples ≥4. CircRNAs named CDR1-AS, NCOA3, and SKIL were up-regulated in the induced group compared to the non-induced group, and the circRNAs IFF01, NTNG1, PLOD2, SMO, and SMURF2 were down-regulated. Quantitative real-time PCR (qRT-PCR) The results agreed with the normalized expression of validated lncRNAs and circRNAs shown in Tables 2 and 3. All PCRs were performed in triplicate. The data are represented 20 μl reaction volume contained 10 μl SYBR® Premix Ex Taq™(Takara), 0.4 μl 10 μM forward primer (0.4 μM final), 0.4 μl 10 μM reverse primer (0.4 μM final), 200 ng of template cDNA and DEPC-treated water. GAPDH was used as an internal control to quantify and normalize the results. The primer pairs are listed in Additional file 1. The specificity of the reaction was determined by detec- tion of the Tms of the amplification products immediately after the last reaction cycle. The 2-△△CT value was used for comparative quantitation. All qRT-PCRs were performed in triplicate. 20 μl reaction volume contained 10 μl SYBR® Premix Ex Taq™(Takara), 0.4 μl 10 μM forward primer (0.4 μM final), 0.4 μl 10 μM reverse primer (0.4 μM final), 200 ng of template cDNA and DEPC-treated water. GAPDH was used as an internal control to quantify and normalize the results. The primer pairs are listed in Additional file 1. The specificity of the reaction was determined by detec- tion of the Tms of the amplification products immediately after the last reaction cycle. The 2-△△CT value was used for comparative quantitation. All qRT-PCRs were performed in triplicate. Identification of differentially expressed lncRNAs, circRNAs and mRNAs The clean reads from two cDNA libraries were mapped to the human genome sequence in GENCODE Release 19 using TopHat version 2.0.9 [26]. The transcripts were then assembled and annotated using Cufflinks [27]. According to the annotation of the human genome sequence, the known lncRNAs and mRNAs were identified. The coding potential of the remaining unknown transcripts was calculated with a coding potential calculator based on Fig. 3 Differentially expressed lncRNAs/circRNAs validated by qRT-PCR. a Convergent primers were designed to detect eight lncRNAs with P < 0.05, FC ≥2 and FPKM in at least one of the samples ≥4. LncRNAs coded as TCONS_00019601, TCONS_00227764, TCONS_00254538, TCONS_00198784 and TCONS_00136898 were up-regulated in the induced group compared with the non-induced group, and TCONS_00085268, TCONS_00125934, and TCONS_00115113 were down-regulated. b Divergent primers were designed to detect the circular form of circRNAs with P < 0.05, FC ≥2 and RPM in at least one of the samples ≥4. CircRNAs named CDR1-AS, NCOA3, and SKIL were up-regulated in the induced group compared to the non-induced group, and the circRNAs IFF01, NTNG1, PLOD2, SMO, and SMURF2 were down-regulated. The results agreed with the normalized expression of validated lncRNAs and circRNAs shown in Tables 2 and 3. All PCRs were performed in triplicate. The data are represented as means ± SD. , P < 0.05; **, P < 0.01 Table 2 Expression profiles of lncRNAs validated by RNA-seq Test_id. Induced (FPKM) Non-induced (FPKM) Regulation TCONS_00019601 93.6452 0 up TCONS_00227764 830.447 47.6825 up TCONS_00085268 0 26.6246 down TCONS_00254538 18,683.1 3962.3 up TCONS_00198784 9.25995 0.0230801 up TCONS_00136898 8.67356 0.0105046 up TCONS_00125934 0 4.65289 down TCONS_00115113 0 4.05314 down Table 2 Expression profiles of lncRNAs validated by RNA-seq Test_id. Induced (FPKM) Non-induced (FPKM) Regulation Gu et al. BMC Genetics (2017) 18:100 Page 6 of 13 Table 3 Expression profiles of circRNAs validated by RNA-seq Name Induced (RPM) Non-induced (RPM) Regulation Table 3 Expression profiles of circRNAs validated by RNA-seq Name Induced (RPM) Non-induced (RPM) Regulation circRNA CDR1-AS 3038.695 1102.345 up circRNA IFF01 0 337.2772 down circRNA NCOA3 1421.615 0 up circRNA NTNG1 0 1286.478 down circRNA PLOD2 0 3742.131 down circRNA SKIL 1956.098 0 up circRNA SMO 0 783.6154 down circRNA SMURF2 0 282.7267 down capacity. The expression levels of lncRNAs and mRNAs were quantified as FPKM (fragments per kilobase of exon per million fragments mapped) using the Cuffdiff program [27]. Identification of differentially expressed lncRNAs, circRNAs and mRNAs The differential expression of lncRNAs and mRNAs was determined using DESeq, with P < 0.05 and fold change (FC) ≥2 [29]. Compared with linear RNAs, circRNAs exhibit distinct patterns of alternative back-splicing and alternative splicing. An upgraded computational pipeline (CIRCexplorer2) was used to systematically identify and annotate circRNAs [30]. The expression levels of circRNAs were quantified as RPM (mapped backsplicing junction reads per million mapped reads) using the EBSeq package [31]. The differential expression of circRNAs was determined with P < 0.05, FC ≥2, and circRNA junction reads ≥5 [31]. the quality, completeness, and sequence similarity of the open reading frame to the proteins in the protein databases [28]. The remaining unknown transcripts of more than 200 base pairs (bp) were considered novel lncRNAs with a coding potential score of less than −1, which suggested that a transcript had no protein-coding Functional analysis The ceRNA networks among lncRNAs, circRNAs, miRNAs and mRNAs were predicted based on miRanda with a Fig. 4 CeRNA networks of 430 lncRNAs and 779 miRNAs based on miRanda Fig. 4 CeRNA networks of 430 lncRNAs and 779 miRNAs based on miRanda Page 7 of 13 Gu et al. BMC Genetics (2017) 18:100 Fig. 5 CeRNA networks of selected circRNAs and miRNAs based on miRanda. Considering that graphics cannot display the enormous amount of network information between 1401 circRNAs and 855 miRNAs, we selected circRNAs with more miRNA binding sites and less binding free energy to make the network diagram using Cytoscape3.5.1 Fig. 5 CeRNA networks of selected circRNAs and miRNAs based on miRanda. Considering that graphics cannot display the enormous amount of network information between 1401 circRNAs and 855 miRNAs, we selected circRNAs with more miRNA binding sites and less binding free energy to make the network diagram using Cytoscape3.5.1 (Fig. 1a, b). Cultured PDLSCs were positive for STRO-1 and CD146 and negative for CD31 and CD45 (Fig. 1c-f). Increased ALP activity identified via ALP staining and ALP activity assay indicated osteogenic differentiation of osteo-induced PDLSCs (Fig. 1g, h). Subsequently, the up-regulated expression levels of the osteogenesis- related markers ALP, Runt-related transcription factor 2 (Runx2) and osteocalcin (OCN) provided further proof for the occurrence of PDLSC osteogenic differentiation (Fig. 2i-k). These findings agreed with previous reports on PDLSC differentiation into osteoblasts [7]. maximum binding free energy of −20 [32]. First, we pre- dicted and selected miRNAs binding with both differentially expressed lncRNAs and circRNAs. Then, target mRNAs of these selected miRNAs were predicted and compared to the differentially expressed mRNAs that were identified in the RNA-seq results. Subsequently, we selected the intersecting elements of target mRNAs and differen- tially expressed mRNAs to analyze their potential functions through GO functional annotation and KEGG pathway analysis. GO terms were enriched using Blast2GO [33] by referring to GO databases. Meanwhile, KEGG pathway analysis was performed by referring to KEGG pathway databases. Cytoscape3.5.1 was used to display the lncRNA/circRNA-miRNA-mRNA networks. Differential expression of lncRNAs, circRNAs and mRNAs upon osteogenic differentiation of PDLSCs p g Ribosome-depleted RNA-seq generated 91.4 million raw reads of the non-induced group and 113.8 million raw reads of the induced group (Table 1). After filtering the adaptor reads and low-quality tags, we separately obtained 90.9 million and 112.4 million clean reads. More than 98% of the raw reads per sample were clean reads. According to the annotation of the human genome sequence GENCODE Release 19, 11,529 lncRNAs (non-induced group) and 9511 lncRNAs (induced group) were identified from the two cDNA libraries (Table 2) (Additional file 2). We also identified 5913 circRNAs in the non-induced group and 3162 circRNAs in the induced group (Table 2) (Additional file 3). In addition, 66,134 mRNAs (non- induced group) and 52,227 mRNAs (induced group) Statistical analysis Quantitative qRT-PCR datasets are presented as the means ± standard deviation (SD) of at least three inde- pendent experiments. The statistical calculations were performed with SPSS statistics software version 17.0. Student’s t-test was performed for normally distributed data to determine statistical significance. A P-value less than 0.05 was considered statistically significant. Results Identification and Osteogenic differentiation of PDLSCs PDLSCs derived from periodontal ligament explants were cultured with normal media to passage number 3 Gu et al. BMC Genetics (2017) 18:100 Page 8 of 13 Page 8 of 13 were annotated (Table 2) (Additional file 4). Both lncRNAs and circRNAs were broadly distributed across the 24 pairs of human chromosomes (Fig. 2a, b). A total of 960 lncRNAs, 1161 circRNAs and 1887 mRNAs were found to be differentially expressed, with P-value <0.05 and FC ≥2 (Additional file 5, Additional file 6 and Additional file 7). Among the differentially expressed lncRNAs, 17 common lncRNAs and 180 specific lncRNAs in the non-induced group and 763 specific lncRNAs in the induced group were detected, with 777 up-regulated lncRNAs and 183 down-regulated lncRNAs (Fig. 2c). Meanwhile, we also identified 95 common circRNAs and 642 specific circRNAs in the non-induced group and 719 specific circRNAs in the induced group, with 766 up- regulated circRNAs and 690 down-regulated circRNAs (Fig. 2d). The apparent variations in transcripts between the two groups are visually displayed with heatmaps (Fig. 2e, f). FPKM/RPM in at least one of the samples ≥4 were selected for qRT-PCR validation. The lncRNAs were amplified with convergent primers and the circRNAs were ampli- fied with divergent primers (Fig. 3a, b). Compared to the non-induced group, the induced group showed increased expression of the lncRNAs coded as TCONS_00019601, TCONS_00227764, TCONS_00254538, TCONS_00198784 and TCONS_00136898 and decreased expression of TCONS_00085268, TCONS_00125934, and TCONS_00 115113 (Fig. 3a). The circRNAs named CDR1-AS, NCOA3, and SKIL were up-regulated in the induced group compared to the non-induced group, and the circRNAs IFF01, NTNG1, PLOD2, SMO, and SMURF2 were down-regulated (Fig. 3b). All the results were consistent with the normalized expression of RNA-seq shown in Tables 2 and 3. Differentially expressed lncRNAs/circRNAs validated by qRT-PCR Based on miRanda with the maximum binding free energy of −20, 430 lncRNAs were predicted to share at least two binding sites with 779 miRNAs (Fig. 4) (Additional file 8). We also predicted that 1401 circRNAs bind 855 miRNAs q To verify the results of the RNA-seq experiments, eight lncRNAs and eight circRNAs with P < 0.05, FC ≥2 and Fig. 6 CeRNA networks of 744 mRNAs and 148 common miRNAs based on miRanda Fig. 6 CeRNA networks of 744 mRNAs and 148 common miRNAs based on miRanda Gu et al. BMC Genetics (2017) 18:100 Page 9 of 13 Page 9 of 13 with at least two binding sites (Additional file 9). Consid- ering that graphics cannot display the enormous amount of network information between 1401 circRNAs and 855 miRNAs, we selected circRNAs with more miRNA bind- ing sites and less binding free energy to make the network diagram (Fig. 5). Through analysis of the common binding MREs of lncRNAs and circRNAs, 165 miRNAs were predicted to combine with both 158 lncRNAs and 1385 circRNAs (data not shown). expressed mRNAs that were found to combine with 148 common miRNAs along with 147 lncRNAs and 1382 circRNAs (Additional file 10). The networks of 744 mRNAs and 148 common miRNAs are shown in Fig. 6. The potential regulatory roles of the ceRNA networks were predicted by analyzing the functions of 744 mRNAs through GO and KEGG pathway analysis (Additional file 11 and Additional file 12). GO annotations (P < 0.05) involv- ing the top 60 mRNAs are displayed in Fig. 7a and include multiple biological processes, cellular components and molecular functions. Among these GO terms, we obtained GO: 0001649 (osteoblast differentiation), which was significantly enriched by 21 mRNAs (Additional file 11). To reveal their potential function, we predicted target mRNAs of these miRNAs based on miRanda and compared these target mRNAs to the differentially expressed mRNAs that were identified in the RNA-seq results (Additional file 7). There were 744 differentially Fig. 7 GO annotations and KEGG pathway analysis of 744 differentially expressed mRNAs. a GO annotations (P < 0.05) involving the top sixty numbers of mRNAs included multiple biological processes, cellular components and molecular functions. b These mRNAs were significantly enriched in 34 KEGG pathways, including the the MAPK pathway, the Wnt pathway and the signaling pathways regulating pluripotency of stem cells Fig. Differentially expressed lncRNAs/circRNAs validated by qRT-PCR 7 GO annotations and KEGG pathway analysis of 744 differentially expressed mRNAs. a GO annotations (P < 0.05) involving the top sixty numbers of mRNAs included multiple biological processes, cellular components and molecular functions. b These mRNAs were significantly enriched in 34 KEGG pathways, including the the MAPK pathway, the Wnt pathway and the signaling pathways regulating pluripotency of stem cells Fig. 7 GO annotations and KEGG pathway analysis of 744 differentially expressed mRNAs. a GO annotations (P < 0.05) involving the top sixty numbers of mRNAs included multiple biological processes, cellular components and molecular functions. b These mRNAs were significantly enriched in 34 KEGG pathways, including the the MAPK pathway, the Wnt pathway and the signaling pathways regulating pluripotency of stem cells Page 10 of 13 Page 10 of 13 Gu et al. BMC Genetics (2017) 18:100 Gu et al. BMC Genetics (2017) 18:100 The complex mRNA networks involved in GO: 0001649 (osteoblast differentiation) and related miRNAs, lncRNAs, and circRNAs are displayed in Fig. 8. Through KEGG analysis, mRNAs were predicted to participte in 34 path- ways (Fig. 7b). Among these KEGG pathways, the MAPK pathway, the Wnt pathway and the signaling pathways regulating pluripotency of stem cells were closely related to osteogenic differentiation. Discussion LncRNAs, circRNAs, miRNAs and mRNAs form large- scale ceRNA cross-talk networks through MREs, which has exciting implications for gene regulation at the post-transcriptional level during multiple physiological and pathophysiological processes [12, 34]. In recent years, studies have documented the functions and clinical impli- cations of ceRNAs in cancer, systemic lupus erythematosus and other diseases, which may present opportunities for new therapeutic approaches for diseases [35, 36]. Based on the above results, we selected several lncRNAs, circRNAs, miRNAs and mRNAs associated with the MAPK pathway to further display the ceRNA networks (Fig. 9). The lncRNAs coded as TCONS_00212979 and TCONS_00212984, circRNA BANP and circRNA ITCH were predicted to combine with miRNA34a and miR- NA146a. DUSP1, FAS and RAC1 were predicted to be tar- get genes of miRNA34a. In addition, PDGFRA, TGFBR2 and MYC were predicted to be targeted by miRNA146a. These six mRNAs were the pivotal genes of the MAPK pathway according to the KEGG analysis. This compli- cated ceRNA network suggested that TCONS_00212979, TCONS_00212984, circRNA BANP and circRNA ITCH might play regulatory roles in the MAPK pathway through miRNA34a, miRNA146a and their targets during PDLSC osteogenic differentiation. p pp Recently, researchers have systematically constructed ceRNA networks through RNA-seq and bioinformatics in mouse germline stem cells to reveal functions and mechanisms of lncRNAs and circRNAs in mouse germline stem cell self-renewal and differentiation [37]. Moreover, lncRNA POIR was demonstrated to form a regulatory network with miRNA182 and FoxO1 to up-regulate PDLSC osteogenic differentiation in periodontitis patients [38]. However, the ceRNA networks were revealed to be intertwined [39]. To fully understand the impact of ceRNA crosstalk on PDLSC osteogenic differentiation, it will be crucial to integrate the competitive lncRNA/ circRNA-miRNA–mRNA regulatory networks. In our study, 744 mRNAs were predicted to combine with 148 Fig. 8 CeRNA networks of lncRNAs/circRNAs-miRNAs-mRNAs significantly participated in GO: 0001649 (osteoblast differentiation) 8 CeRNA networks of lncRNAs/circRNAs-miRNAs-mRNAs significantly participated in GO: 0001649 (osteoblast differentiation) Page 11 of 13 Page 11 of 13 Gu et al. BMC Genetics (2017) 18:100 Fig. 9 CeRNA networks of lncRNAs/circRNAs-miRNAs-mRNAs significantly participated in the MAPK pathway. The lncRNAs coded as TCONS_00212979 and TCONS_00212984, as well as circRNA BANP and circRNA ITCH, were predicted to interact with miRNA34a and miRNA146a. DUSP1, FAS and RAC1 were predicted to be target genes of miRNA34a. PDGFRA, TGFBR2 and MYC were predicted to be targeted by miRNA146a. Discussion These six mRNAs are pivotal genes in the MAPK pathway according to KEGG analysis Fig. 9 CeRNA networks of lncRNAs/circRNAs-miRNAs-mRNAs significantly participated in the MAPK pathway. The lncRNAs coded as TCONS_00212979 and TCONS_00212984, as well as circRNA BANP and circRNA ITCH, were predicted to interact with miRNA34a and miRNA146a. DUSP1, FAS and RAC1 were predicted to be target genes of miRNA34a. PDGFRA, TGFBR2 and MYC were predicted to be targeted by miRNA146a. These six mRNAs are pivotal genes in the MAPK pathway according to KEGG analysis common miRNAs, along with 147 lncRNAs and 1382 circRNAs. Through GO analysis, 21 mRNAs were predicted to significantly participate in osteoblast differentiation (GO: 0001649) (Fig. 8). Among them, ALPL, also called ALP, was reported to be an osteogenesis-related marker and was up-regulated during PDLSC osteogenic differenti- ation [7]. The up-regulated expression level of ALP was also detected by qRT-PCR in our study (Fig. 1i). SMAD3 and SMAD5, members of the SMAD family, were also predicted to form ceRNA networks and participate in osteoblast differentiation. Both SMAD3 and SMAD5 were demonstrated to regulate PDLSC osteogenic differentiation by modulating TGF-β signals [40, 41]. Additionally, Notch1, another participant in osteoblast differentiation (GO: 0001649), is part of the Notch signaling pathway, which is important for maintaining osteogenic differentiation of PDLSCs [42, 43]. These crucial osteogenic genes formed ceRNA networks with lncRNAs and circRNAs by targeting common miRNAs, and these networks might provide evidence of new regula- tory mechanisms in PDLSC osteogenic differentiation. Discussion These six mRNAs are pivotal genes in the MAPK pathway according to KEGG analysis Through KEGG pathway analysis, mRNAs of ceRNA networks were predicted to be involved in the Wnt path- way, MAPK pathway and signaling pathways regulating the pluripotency of stem cells. Previous studies have demonstrated that Wnt signaling contributes to the differentiation of periodontal ligament fibroblasts into osteoblasts [44]. In addition, the MAPK pathway was found to be critical for skeleton development and bone homeostasis [45]. Moreover, it plays significant roles in osteogenic differentiation of PDLSCs [46, 47]. Furthermore, we constructed a ceRNA network of TCONS_00212979, TCONS_00212984, circRNA BANP, circRNA ITCH, miR- NA34a, miRNA146a, DUSP1, FAS, RAC1, PDGFRA, TGFBR2 and MYC. These mRNAs are important elements of the MAPK pathway based on KEGG analysis. Among them, DUSP1, FAS and RAC1 are targeted by miRNA34a, while PDGFRA, TGFBR2 and MYC are targeted by miR- NA146a. Studies have illustrated that both miRNA34a and miRNA146a are closely related to osteogenic differentiation of human mesenchymal stem cells [48, 49]. In addition, miRNA146a was revealed to promote differentiation of periodontal ligament cells [9]. TCONS_00212979, TCONS_00212984, circRNA BANP and circRNA ITCH were predicted to bind miRNA34a and miRNA146a. TCONS_00212979, known as CARMEN, has been reported to be a cardiac mesoderm enhancer-associated lncRNA that modulates cardiac differentiation through miRNA143 and miRNA145 [50]. TCONS_00212984 is a novel lncRNA with a genomic origin similar to that of TCONS_00212979 ac- cording to RNA-seq. circRNA BANP and circRNA ITCH have both been reported to contribute to carcinogenesis and might serve as cancer biomarkers [51, 52]. However, the regulatory roles of these two lncRNAs and circRNAs in osteogenic differentiation remain unclear. In summary, the ceRNA network suggested that TCONS_00212979, TCONS_00212984, circRNA BANP and circRNA ITCH might interact with miRNA34a and miRNA146a to regulate PDLSC osteogenic differentiation via the MAPK pathway. However, their regulatory mechanisms need to be further investigated. Our future study plan will be to validate their differential expression profiles, verify their ceRNA networks and specify their effects on PDLSC osteogenic differentiation using knockdown and overex- pression experiments. Fig. 9 CeRNA networks of lncRNAs/circRNAs-miRNAs-mRNAs significantly participated in the MAPK pathway. The lncRNAs coded as TCONS_00212979 and TCONS_00212984, as well as circRNA BANP and circRNA ITCH, were predicted to interact with miRNA34a and miRNA146a. DUSP1, FAS and RAC1 were predicted to be target genes of miRNA34a. PDGFRA, TGFBR2 and MYC were predicted to be targeted by miRNA146a. Competing interests Th h d l h Competing interests The authors declare that they have no competing interests. Additional file 4: mRNA_all expression_.xlsx The expression profiles of all mRNAs identified by RNA-seq (XLSX 8078 kb) Additional file 5: lncRNA_differential expression.xlsx Differentially expressed lncRNAs identified by RNA-seq (XLSX 87 kb) Ethics approval and consent to participate PDLSCs were derived from the extracted teeth, which were collected from patients at the Department of Oral Maxillofacial Surgery, Stomatology Hospital of Shandong University. All protocols for handling dental tissues were performed in accordance with relevant guidelines and regulations. All protocols for the handling of human tissues were approved by the Research Ethics Committee of Stomatology Hospital of Shandong University, China (G201401601). The participant informed consent was written. Funding g The performance of RNA-seq, analysis of the sequenced data and revision of the manuscript were supported by grant from the National Natural Science Foundation of China (81470709). All the cell biology experiments and mo- lecular biology experiments were supported by grant from the Construction Engineering Special Fund of “Taishan Scholars” (tsqn201611068). 9. Hung PS, Chen FC, Kuang SH, Kao SY, Lin SC, Chang KW. miR-146a induces differentiation of periodontal ligament cells. J Dent Res. 2010;89(3):252–7. 10. Liu Y, Liu W, Hu C, Xue Z, Wang G, Ding B, Luo H, Tang L, Kong X, Chen X, et al. MiR-17 modulates osteogenic differentiation through a coherent feed- forward loop in mesenchymal stem cells isolated from periodontal ligaments of patients with periodontitis. Stem Cells. 2011;29(11):1804–16. Abbreviations 3. Park JY, Jeon SH, Choung PH. Efficacy of periodontal stem cell transplantation in the treatment of advanced periodontitis. Cell Transplant. 2011;20(2):271–85. 3. Park JY, Jeon SH, Choung PH. Efficacy of periodontal stem cell transplantation in the treatment of advanced periodontitis. Cell Transplant. 2011;20(2):271–85. ALP: Alkaline phosphatase; bp: Base pairs; ceRNAs: Competitive endogenous RNAs; circRNAs: Circular RNAs; FC: Fold change; FPKM: Fragments per kilobase of exon per million fragments mapped; GO: Gene Ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; lncRNAs: Long non- coding RNAs; miRNAs: MicroRNAs; MREs: miRNA response elements; mRNAs: Messenger RNAs; ncRNA: Non-coding RNA; OCN: Osteocalcin; PDLSC: Periodontal ligament stem cell; qRT-PCR: Quantitative real-time poly- merase chain reaction; RNA-seq: RNA sequencing; RPM: Mapped back-splice junction reads per million mapped reads; Runx2: Runt-related transcription factor 2; SD: Standard deviation 4. Wei F, Qu C, Song T, Ding G, Fan Z, Liu D, Liu Y, Zhang C, Shi S, Wang S, Vitamin C. Treatment promotes mesenchymal stem cell sheet formation and tissue regeneration by elevating telomerase activity. J Cell Physiol. 2012; 227(9):3216–24. 4. Wei F, Qu C, Song T, Ding G, Fan Z, Liu D, Liu Y, Zhang C, Shi S, Wang S, Vitamin C. Treatment promotes mesenchymal stem cell sheet formation and tissue regeneration by elevating telomerase activity. J Cell Physiol. 2012; 227(9):3216–24. 5. Wei F, Song T, Ding G, Xu J, Liu Y, Liu D, Fan Z, Zhang C, Shi S, Wang S. Functional tooth restoration by allogeneic mesenchymal stem cell-based bio-root regeneration in swine. Stem Cells Dev. 2013;22(12):1752–62. 6. Ding G, Liu Y, Wang W, Wei F, Liu D, Fan Z, An Y, Zhang C, Wang S. Allogeneic periodontal ligament stem cell therapy for periodontitis in swine. Stem Cells. 2010;28(10):1829–38. References D Additional file 11: target mRNAs_GO enrichment.xlsx GO items enriched by target mRNAs through GO analysis (XLSX 95 kb) Additional file 11: target mRNAs_GO enrichment.xlsx GO items enriched by target mRNAs through GO analysis (XLSX 95 kb) References 1. Damgaard C, Holmstrup P, Van Dyke TE, Nielsen CH. The complement system and its role in the pathogenesis of periodontitis: current concepts. J Periodontal Res. 2015;50(3):283–93. References 1. Damgaard C, Holmstrup P, Van Dyke TE, Nielsen CH. The complement system and its role in the pathogenesis of periodontitis: current concepts. J Periodontal Res. 2015;50(3):283–93. Additional file 12: target mRNAs_KEGG pathway enrichment.xlsx Pathways enriched by target mRNAs through KEGG analysis (XLSX 20 kb) Additional file 12: target mRNAs_KEGG pathway enrichment.xlsx Pathways enriched by target mRNAs through KEGG analysis (XLSX 20 kb) Additional file 12: target mRNAs_KEGG pathway enrichment.xlsx Pathways enriched by target mRNAs through KEGG analysis (XLSX 20 2. Liu Y, Zheng Y, Ding G, Fang D, Zhang C, Bartold PM, Gronthos S, Shi S, Wang S. Periodontal ligament stem cell-mediated treatment for periodontitis in miniature swine. Stem Cells. 2008;26(4):1065–73. 2. Liu Y, Zheng Y, Ding G, Fang D, Zhang C, Bartold PM, Gronthos S, Shi S, Wang S. Periodontal ligament stem cell-mediated treatment for periodontitis in miniature swine. Stem Cells. 2008;26(4):1065–73. Consent for publication Not applicable. Consent for publication Not applicable. Additional file 3: circRNA_all expression_.xlsx The expression profiles of all circRNAs identified by RNA-seq (XLSX 1577 kb) Competing interests The authors declare that they have no competing interests. Authors’ contributions FW and DL conceived, designed and revised this study. XG analyzed the RNA-seq data and wrote the manuscript. ML detected the expression of genes using qRT-PCR. YJ performed the cell culture and osteogenic induction treatments. All authors read and approved the final manuscript. Acknowledgments k l d h We acknowledge the anonymous reviewers and academic editors for their constructive suggestions on the previous version of the manuscript. We acknowledge laboratory members for their contributions. 7. Seo BM, Miura M, Gronthos S, Bartold PM, Batouli S, Brahim J, Young M, Robey PG, Wang CY, Shi S. Investigation of multipotent postnatal stem cells from human periodontal ligament. Lancet. 2004;364(9429):149–55. 7. Seo BM, Miura M, Gronthos S, Bartold PM, Batouli S, Brahim J, Young M, Robey PG, Wang CY, Shi S. Investigation of multipotent postnatal stem cells from human periodontal ligament. Lancet. 2004;364(9429):149–55. 8. Hao Y, Ge Y, Li J, Hu Y, Wu B, Fang F. Identification of MicroRNAs by microarray analysis and prediction of target genes involved in the Osteogenic differentiation of human periodontal ligament stem cells. J Periodontol. 2017:1–11. 8. Hao Y, Ge Y, Li J, Hu Y, Wu B, Fang F. Identification of MicroRNAs by microarray analysis and prediction of target genes involved in the Osteogenic differentiation of human periodontal ligament stem cells. J Periodontol. 2017:1–11. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Additional file 6: circRNA_differential expression.xlsx Differentially expressed circRNAs identified by RNA-seq (XLSX 312 kb) Received: 19 August 2017 Accepted: 16 November 2017 Received: 19 August 2017 Accepted: 16 November 2017 Received: 19 August 2017 Accepted: 16 November 2017 Additional file 10: lncRNA-circRNA-miRNA-mRNA.xlsx LncRNAs and circRNAs predicted to share miRNA binding sites in common with target mRNAs (XLSX 8696 kb) Conclusion This study identified differentially expressed lncRNAs, circRNAs and mRNAs during osteogenic differentiation of PDLSCs. Competitive lncRNA/circRNA-miRNA–mRNA regulatory networks were comprehensively integrated and predicted to be involved in osteoblast differentiation by GO and KEGG pathway analysis. Moreover, the lncRNAs coded as TCONS_00212979 and TCONS_00212984, circRNA BANP and circRNA ITCH were predicted to Page 12 of 13 Page 12 of 13 Gu et al. BMC Genetics (2017) 18:100 Page 12 of 13 Gu et al. BMC Genetics (2017) 18:100 interact with miRNA34a and miRNA146a to regulate PDLSC osteogenic differentiation via the MAPK pathway. Our study suggested that specific lncRNAs and circRNAs might function as ceRNAs to promote PDLSC osteogenic differentiation and periodontal regeneration. 11. Yan GQ, Wang X, Yang F, Yang ML, Zhang GR, Wang GK. MicroRNA-22 promoted Osteogenic differentiation of human periodontal ligament stem cells by targeting HDAC6. J Cell Biochem. 2017;118(7):1653–8. Additional files Additional file 1: Gene Primers.doc Gene primers used in qRT-PCR (DOC 38 kb) Additional file 2: lncRNA_all expression_.xlsx The expression profiles of all lncRNAs identified by RNA-seq (XLSX 1653 kb) Additional file 3: circRNA_all expression_.xlsx The expression profiles of all circRNAs identified by RNA-seq (XLSX 1577 kb) Additional file 4: mRNA_all expression_.xlsx The expression profiles of all mRNAs identified by RNA-seq (XLSX 8078 kb) Additional file 5: lncRNA_differential expression.xlsx Differentially expressed lncRNAs identified by RNA-seq (XLSX 87 kb) Additional file 6: circRNA_differential expression.xlsx Differentially expressed circRNAs identified by RNA-seq (XLSX 312 kb) Additional file 7: mRNA_differential expression.xlsx Differentially expressed mRNAs identified by RNA-seq (XLSX 240 kb) Additional file 8: lncRNA_miRNA_miRanda_2BindingSites.xlsx LncRNAs predicted to share at least two miRNA binding sites (XLSX 115 kb) Additional file 9: circRNA_miRNA_miRanda_2BindingSites.xlsx CircRNAs predicted to share at least two miRNA binding sites (XLSX 23183 kb) Additional file 10: lncRNA-circRNA-miRNA-mRNA.xlsx LncRNAs and circRNAs predicted to share miRNA binding sites in common with target mRNAs (XLSX 8696 kb) Additional file 11: target mRNAs_GO enrichment.xlsx GO items enriched by target mRNAs through GO analysis (XLSX 95 kb) Additional file 12: target mRNAs_KEGG pathway enrichment.xlsx Pathways enriched by target mRNAs through KEGG analysis (XLSX 20 kb) Additional file 1: Gene Primers.doc Gene primers used in qRT-PCR (DOC 38 kb) Additional file 2: lncRNA_all expression_.xlsx The expression profiles of all lncRNAs identified by RNA-seq (XLSX 1653 kb) Author details 1 f Additional file 7: mRNA_differential expression.xlsx Differentially expressed mRNAs identified by RNA-seq (XLSX 240 kb) 1Department of Orthodontics, Shandong Provincial Key Laboratory of Oral Tissue Regeneration, School of Stomatology, Shandong University, Wenhua Xi Road No. 44-1, Jinan, Shandong 250012, People’s Republic of China. 2Shandong Provincial Key Laboratory of Oral Tissue Regeneration, School of Stomatology, Shandong University, Jinan, People’s Republic of China. Additional file 8: lncRNA_miRNA_miRanda_2BindingSites.xlsx LncRNAs predicted to share at least two miRNA binding sites (XLSX 115 kb) Additional file 9: circRNA_miRNA_miRanda_2BindingSites.xlsx CircRNAs predicted to share at least two miRNA binding sites (XLSX 23183 kb) Availability of data and materials All data analyzed during this study are included in this article and its supplementary information files. RNA-seq data sets are available in the Gene Expression Omnibus database under accession number GSM2856840 (for non-induced PDLSCs) and GSM2856841 (for induced PDLSCs). 11. Yan GQ, Wang X, Yang F, Yang ML, Zhang GR, Wang GK. MicroRNA-22 promoted Osteogenic differentiation of human periodontal ligament stem cells by targeting HDAC6. J Cell Biochem. 2017;118(7):1653–8. Page 13 of 13 Page 13 of 13 Gu et al. BMC Genetics (2017) 18:100 Gu et al. BMC Genetics (2017) 18:100 12. Salmena L, Poliseno L, Tay Y, Kats L, Pandolfi PP. A ceRNA hypothesis: the Rosetta stone of a hidden RNA language? Cell. 2011;146(3):353–8. 38. Wang L, Wu F, Song Y, Li X, Wu Q, Duan Y, Jin Z. Long noncoding RNA related to periodontitis interacts with miR-182 to upregulate osteogenic differentiation in periodontal mesenchymal stem cells of periodontitis patients. Cell Death Dis. 2016;7(8):e2327. 13. Sen R, Ghosal S, Das S, Balti S, Chakrabarti J. Competing endogenous RNA: 13. Sen R, Ghosal S, Das S, Balti S, Chakrabarti J. Competing endogenous RNA: the key to posttranscriptional regulation. TheScientificWorldJOURNAL. 2014; 2014:896206. the key to posttranscriptional regulation. TheScientificWorldJOURNAL. 2014; 2014:896206. 39. Tay Y, Rinn J, Pandolfi PP. The multilayered complexity of ceRNA crosstalk and competition. Nature. 2014;505(7483):344–52. 14. Kapranov P, Cheng J, Dike S, Nix DA, Duttagupta R, Willingham AT, Stadler PF, Hertel J, Hackermuller J, Hofacker IL, et al. RNA maps reveal new RNA classes and a possible function for pervasive transcription. Science. 2007; 316(5830):1484–8. 40. Huang Y, Zheng Y, Jia L, Li W. Long noncoding RNA H19 promotes Osteoblast differentiation via TGF-beta1/Smad3/HDAC signaling pathway by deriving miR-675. Stem Cells. 2015;33(12):3481–92. 41. Ishibashi O, Ikegame M, Takizawa F, Yoshizawa T, Moksed MA, Iizawa F, Mera H, Matsuda A, Kawashima H: Endoglin is involved in BMP-2-induced osteogenic differentiation of periodontal ligament cells through a pathway independent of Smad-1/5/8 phosphorylation. J Cell Physiol 2010, 222(2):465-473. 15. Fatica A, Bozzoni I. Long non-coding RNAs: new players in cell differentiation and development. Nat Rev Genet. 2014;15(1):7–21. 16. WS XY, Tang C, Chen W. Upregulation of long non-coding RNA HIF 1α-anti- sense 1 induced by transforming growth factor-β-mediated targeting of sirtuin 1 promotes osteoblastic differentiation of human bone marrow stromal cells. Mol Med Rep. 2015;12(5):7233–8. 42. Osathanon T, Ritprajak P, Nowwarote N, Manokawinchoke J, Giachelli C, Pavasant P. Availability of data and materials Surface-bound orientated Jagged-1 enhances osteogenic differentiation of human periodontal ligament-derived mesenchymal stem cells. J Biomed Mater Res A. 2013;101(2):358–67. 17. Jia Q, Jiang W, Ni L. Down-regulated non-coding RNA (lncRNA-ANCR) promotes osteogenic differentiation of periodontal ligament stem cells. Arch Oral Biol. 2015;60(2):234–41. 43. Li Y, Li SQ, Gao YM, Li J, Zhang B. Crucial role of notch signaling in osteogenic differentiation of periodontal ligament stem cells in osteoporotic rats. Cell Biol Int. 2014;38(6):729–36. 18. Wilusz JE, Sharp PA. Molecular biology. A circuitous route to noncoding RNA. Science. 2013;340(6131):440–1. 44. Heo JS, Lee SY, Lee JC. Wnt/beta-catenin signaling enhances osteoblastogenic differentiation from human periodontal ligament fibroblasts. Molecules and cells. 2010;30(5):449–54. 19. Qu S, Yang X, Li X, Wang J, Gao Y, Shang R, Sun W, Dou K, Li H. Circular RNA: a new star of noncoding RNAs. Cancer Lett. 2015;365(2):141–8. 20. Barrett SP, Salzman J. Circular RNAs: analysis, expression and potential functions. Development. 2016;143(11):1838–47. 45. Thouverey C, Caverzasio J. Focus on the p38 MAPK signaling pathway in bone development and maintenance. BoneKEy reports. 2015;4:711. 21. Lasda E, Parker R. Circular RNAs: diversity of form and function. RNA. 2014; 20(12):1829–42. 46. Tang M, Peng Z, Mai Z, Chen L, Mao Q, Chen Z, Chen Q, Liu L, Wang Y, Ai H. Fluid shear stress stimulates osteogenic differentiation of human periodontal ligament cells via the extracellular signal-regulated kinase 1/2 and p38 mitogen-activated protein kinase signaling pathways. J Periodontol. 2014;85(12):1806–13. 22. Qian DY, Yan GB, Bai B, Chen Y, Zhang SJ, Yao YC, Xia H. Differential circRNA expression profiles during the BMP2-induced osteogenic differentiation of MC3T3-E1 cells. Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie. 2017;90:492–9. 47. Li L, Han MX, Li S, Xu Y, Wang L. Hypoxia regulates the proliferation and osteogenic differentiation of human periodontal ligament cells under cyclic tensile stress via mitogen-activated protein kinase pathways. J Periodontol. 2014;85(3):498–508. 23. Zheng Y, Li X, Huang Y, Jia L, Li W. The circular RNA landscape of periodontal ligament stem cells during Osteogenesis. J Periodontol. 2017;88(9):906–14. 24. Wei FL, Wang JH, Ding G, Yang SY, Li Y, YJ H, Wang SL. Mechanical force- induced specific MicroRNA expression in human periodontal ligament stem cells. Cells Tissues Organs. 2014;199(5–6):353–63. 48. Chen L, Holmstrom K, Qiu W, Ditzel N, Shi K, Hokland L, Kassem M. MicroRNA-34a inhibits osteoblast differentiation and in vivo bone formation of human stromal stem cells. Stem Cells. 2014;32(4):902–12. g 25. Availability of data and materials Borodina T, Adjaye J, Sultan M. A strand-specific library preparation protocol for RNA sequencing. Methods Enzymol. 2011;500:79–98. 49. Huszar JM, Payne CJ. MIR146A inhibits JMJD3 expression and osteogenic differentiation in human mesenchymal stem cells. FEBS Lett. 2014;588(9):1850–6. 26. Kim D, Pertea G, Trapnell C, Pimentel H, Kelley R, Salzberg SL. TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions. Genome Biol. 2013;14(4):R36. 50. Ounzain S, Micheletti R, Arnan C, Plaisance I, Cecchi D, Schroen B, Reverter F, Alexanian M, Gonzales C, Ng SY, et al. CARMEN, a human super enhancer- associated long noncoding RNA controlling cardiac specification, differentiation and homeostasis. J Mol Cell Cardiol. 2015;89(Pt A):98–112. 27. Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ, Pachter L. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol. 2010;28(5):511–5. 51. Zhu M, Xu Y, Chen Y, Yan F, Circular BANP. An upregulated circular RNA that modulates cell proliferation in colorectal cancer. Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie. 2017;88:138–44. 28. Kong L, Zhang Y, Ye ZQ, Liu XQ, Zhao SQ, Wei L, Gao G. CPC: assess the protein-coding potential of transcripts using sequence features and support vector machine. Nucleic Acids Res. 2007;35(Web Server issue):W345–9. 52. Guo W, Zhang J, Zhang D, Cao S, Li G, Zhang S, Wang Z, Wen P, Yang H, Shi X, et al. Polymorphisms and expression pattern of circular RNA circ-ITCH contributes to the carcinogenesis of hepatocellular carcinoma. Oncotarget. 2017;8(29):48169–77. 29. Anders S, Huber W. Differential expression analysis for sequence count data. Genome Biol. 2010;11(10):R106. 30. Zhang XO, Dong R, Zhang Y, Zhang JL, Luo Z, Zhang J, Chen LL, Yang L. Diverse alternative back-splicing and alternative splicing landscape of circular RNAs. Genome Res. 2016;26(9):1277–87. 31. Leng N, Dawson JA, Thomson JA, Ruotti V, Rissman AI, Smits BM, Haag JD, Gould MN, Stewart RM, Kendziorski C. EBSeq: an empirical Bayes hierarchical model for inference in RNA-seq experiments. Bioinformatics. 2013;29(8):1035–43. model for inference in RNA-seq experiments. Bioinformatics. 2013;29(8):1035–43. 32. miRanda: http://www.microrna.org/microrna/home.do Assessed August 2010. 32. miRanda: http://www.microrna.org/microrna/home.do Assessed August 20 33. Conesa A, Gotz S, Garcia-Gomez JM, Terol J, Talon M, Robles M. Blast2GO: a universal tool for annotation, visualization and analysis in functional genomics research. Bioinformatics. 2005;21(18):3674–6. Submit your next manuscript to BioMed Central and we will help you at every step: Submit your next manuscript to BioMed Central and we will help you at every step: 34. Ala U, Karreth FA, Bosia C, Pagnani A, Taulli R, Leopold V, Tay Y, Provero P, Zecchina R, Pandolfi PP. Integrated transcriptional and competitive endogenous RNA networks are cross-regulated in permissive molecular environments. Proc Natl Acad Sci U S A. 2013;110(18):7154–9. 34. Ala U, Karreth FA, Bosia C, Pagnani A, Taulli R, Leopold V, Tay Y, Provero P, Zecchina R, Pandolfi PP. Integrated transcriptional and competitive endogenous RNA networks are cross-regulated in permissive molecular environments. Proc Natl Acad Sci U S A. 2013;110(18):7154–9. • We accept pre-submission inquiries • Our selector tool helps you to ﬁnd the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit and we will help you at every step: 35. Qi X, Zhang D-H, Wu N, Xiao J-H, Wang X, Ma W. ceRNA in cancer: possible functions and clinical implications. J Med Genet. 2015;52(10):710–8. 36. Li LJ, Zhao W, Tao SS, Leng RX, Fan YG, Pan HF, Ye DQ. Competitive endogenous RNA network: potential implication for systemic lupus erythematosus. Expert Opin Ther Targets. 2017;21(6):639–48. 37. Li X, Ao J, Wu J. Systematic identification and comparison of expressed profiles of lncRNAs and circRNAs with associated co-expression and ceRNA networks in mouse germline stem cells. Oncotarget. 2017;8(16):26573–90. 37. Li X, Ao J, Wu J. Systematic identification and comparison of expressed profiles of lncRNAs and circRNAs with associated co-expression and ceRNA networks in mouse germline stem cells. Oncotarget. 2017;8(16):26573–90.
https://openalex.org/W2765491445	http://mediatum.ub.tum.de/doc/1432474/document.pdf	English	null	Binding patterns of homo-peptides on bare magnetic nanoparticles: insights into environmental dependence	Scientific reports	2,017	cc-by	9,162	Binding patterns of homo-peptides on bare magnetic nanoparticles: insights into environmental dependence Received: 23 February 2017 Accepted: 4 October 2017 Published: xx xx xxxx Received: 23 February 2017 Accepted: 4 October 2017 Published: xx xx xxxx Silvia A. Blank-Shim 1, Sebastian P. Schwaminger 1, Monika Borkowska-Panek2, Priya Anand2, Peyman Yamin2, Paula Fraga-García 1, Karin Fink2, Wolfgang Wenzel2 & Sonja Berensmeier 1 Magnetic nanoparticles (MNP) are intensively investigated for applications in nanomedicine, catalysis and biotechnology, where their interaction with peptides and proteins plays an important role. However, the characterisation of the interaction of individual amino acids with MNP remains challenging. Here, we classify the affinity of 20 amino acid homo-hexamers to unmodified iron oxide nanoparticles using peptide arrays in a variety of conditions as a basis to identify and rationally design selectively binding peptides. The choice of buffer system is shown to strongly influence the availability of peptide binding sites on the MNP surface. We find that under certain buffer conditions peptides of different charges can bind the MNP and that the relative strength of the interactions can be modulated by changing the buffer. We further present a model for the competition between the buffer and the MNP’s electrostatically binding to the adsorption sites. Thereby, we demonstrate that the charge distribution on the surface can be used to correlate the binding of positively and negatively charged peptides to the MNP. This analysis enables us to engineer the binding of MNP on peptides and contribute to better understand the bio-nano interactions, a step towards the design of affinity tags for advanced biomaterials. Magnetic nanoparticles (MNP) are widely used for the purification of nucleic acids and other biological mole- cules1–4. MNP are also employed in the immobilisation of enzymes5,6, for biomedical applications such as drug delivery and hyperthermia in cancer treatment and as contrast agents for magnetic resonance imaging7,8. Their superparamagnetic behaviour allows for their manipulation by an external magnetic field to easily accumulate MNP in a desired area9. In addition, MNP have also spiked interest in the field of catalytic reaction engineering10. For most applications, MNP have to be functionalised to allow for a selective binding of the target molecule. This is presently achieved by attaching metal-ion chelating molecules, e.g. nitrilotriacetic acid, to the MNP surface, which then bind His-tagged biomolecules11,12. Drawbacks of this method are the leakage of toxic metal ions and instability of the surface functionalisation13,14. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Received: 23 February 2017 Accepted: 4 October 2017 Published: xx xx xxxx SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 Binding patterns of homo-peptides on bare magnetic nanoparticles: insights into environmental dependence Alternative surface modifications for protein adsorption on magnetic particles are glutathione, streptavidin, biotin or protein A, all of which lead to high costs running in the thousands of Euros per gram15. Use of bare superparamagnetic iron oxides thus offers decisive advantages for industrial applications mainly due to the easy, rapid and low-cost synthesis and the absence of degradable func- tional surface groups. We undertake here the first systematic study of peptide-MNP interactions of bare MNP to ultimately develop peptides that can be genetically engineered into proteins as tags and strongly bind to the nonfunctionalised MNP. The key to the design of high-affinity peptide tags lies in an in-depth understanding of surface-peptide recognition patterns16. However, a rational development of models guiding the design of peptides with specific binding propensity is complicated by the lack of available methods for a quantitative assessment of the affinity of individual amino acids, which are the peptide building blocks. One of the reasons that comprehen- sive interaction data for MNP with all the amino acids or with a larger group of peptides is currently not available 1Bioseparation Engineering Group, Department of Mechanical Engineering, Technical University of Munich, 85748, Garching b. München, Germany. 2Institute of Nanotechnology, Karlsruhe Institute of Technology, 76344, Eggenstein- Leopoldshafen, Germany. Correspondence and requests for materials should be addressed to W.W. (email: wolfgang. wenzel@kit.edu) or S.B. (email: s.berensmeier@tum.de) SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 1 www.nature.com/scientificreports/ Figure 1. Schematic experimental setup of peptide array binding experiments and their evaluation. Peptides were immobilised on a cellulose membrane in spots. Each spot contained 5 to 10 nm of a certain peptide. The nanoparticles were suspended in buffer and added to the membrane. A higher load with particles resulted in a darker colour of the spot which was digitalised to yield a score as an indicator for their adsorption. Figure 1. Schematic experimental setup of peptide array binding experiments and their evaluation. Peptides were immobilised on a cellulose membrane in spots. Each spot contained 5 to 10 nm of a certain peptide. The nanoparticles were suspended in buffer and added to the membrane. A higher load with particles resulted in a darker colour of the spot which was digitalised to yield a score as an indicator for their adsorption. may stem from the fact that MNP not only interfere with fluorescence, but also degrade fluorophores commonly used for the quantification of biomolecules17. Binding patterns of homo-peptides on bare magnetic nanoparticles: insights into environmental dependence In this investigation we explore the influence of the buffer on MNP-peptide interactions and rationalise our findings by means of a model which quantitatively reproduces the change in the loading of the charged peptides on the surface for all the stud- ied conditions. These insights enable us not only to understand interaction rules but also to develop peptides that reversibly bind MNP as a function of the buffer. The interaction of peptides to surfaces in general is affected by many factors, such as structure26,27, stoichiome- try28, pH29, temperature and solvent conditions30,31, surface protonation state32, possible peptide conformation33,34 and composition35. Additionally, the conformation and binding states of short peptides to metal oxide nanoparti- cles are influenced by pH and buffer conditions36. Apart from this paper, the importance of the buffer species was only recently addressed in the context of biological systems37, where their impact on bio-nano interactions was, however, not discussed. In comparison to homogeneous surfaces, such as carefully grown gold and platinum, or a single crystalline surface, iron oxide MNP present a particularly challenging target concerning their representa- tion by an atomistic model, mostly due to the great heterogeneity present38–40. For this reason, it is difficult to utilise quantitative models based on ab-initio calculations that have been proven successful for other surfaces41,42. I hi k h fi i d i d f h h h f 20 l i id q p In this work, we present the first systematic adsorption study of the homo-hexamers of 20 natural amino acids with magnetic nanoparticles and show clear evidence for a strong influence of the environment (buffering sys- tem) on the electrostatic characteristics of the MNP surface. It in turn leads to an observable effect on the binding of the charged peptides, while the uncharged peptides binding affinity, although present, remains untouched. For the experiments, we choose buffers with different dissociation constants, and different sign and number of charged species. We find that for certain buffer conditions, the MNP surface is not homogeneous in charge and concurrently binds peptides with different physico-chemical characteristics. In literature, various models such as the Stern-Gouy-Chapman model43,44, the triple-layer model (TLM)45, physical surface-complexion model46 or the charge distribution-multi site complexation (CD-MUSIC) model47,48 have been employed to describe similar surface systems. These models assume heterogeneous surfaces interacting with the surrounding ion layers. Some studies also included the influence of buffer ions on the adsorption of metal cations49. Binding patterns of homo-peptides on bare magnetic nanoparticles: insights into environmental dependence In this investigation we explore the influence of the buffer on MNP-peptide interactions and rationalise our findings by means of a model which quantitatively reproduces the change in the loading of the charged peptides on the surface for all the stud- ied conditions. These insights enable us not only to understand interaction rules but also to develop peptides that reversibly bind MNP as a function of the buffer. Binding patterns of homo-peptides on bare magnetic nanoparticles: insights into environmental dependence In this study, we measure the interaction between bare MNP and homo-peptides using peptide arrays. Peptide arrays are commonly used in the biological context18–22 but have also found application in the tag development for metals and metal oxides23–25. This technique is particularly well suited, since iron oxide MNP stain distinctively, leaving dark spots when bound to peptides on a white cellulose membrane.hf The interaction of peptides to surfaces in general is affected by many factors, such as structure26,27, stoichiome- try28, pH29, temperature and solvent conditions30,31, surface protonation state32, possible peptide conformation33,34 and composition35. Additionally, the conformation and binding states of short peptides to metal oxide nanoparti- cles are influenced by pH and buffer conditions36. Apart from this paper, the importance of the buffer species was only recently addressed in the context of biological systems37, where their impact on bio-nano interactions was, however, not discussed. In comparison to homogeneous surfaces, such as carefully grown gold and platinum, or a single crystalline surface, iron oxide MNP present a particularly challenging target concerning their representa- tion by an atomistic model, mostly due to the great heterogeneity present38–40. For this reason, it is difficult to utilise quantitative models based on ab-initio calculations that have been proven successful for other surfaces41,42. In this work, we present the first systematic adsorption study of the homo-hexamers of 20 natural amino acids with magnetic nanoparticles and show clear evidence for a strong influence of the environment (buffering sys- tem) on the electrostatic characteristics of the MNP surface. It in turn leads to an observable effect on the binding of the charged peptides, while the uncharged peptides binding affinity, although present, remains untouched. For the experiments, we choose buffers with different dissociation constants, and different sign and number of charged species. We find that for certain buffer conditions, the MNP surface is not homogeneous in charge and concurrently binds peptides with different physico-chemical characteristics. In literature, various models such as the Stern-Gouy-Chapman model43,44, the triple-layer model (TLM)45, physical surface-complexion model46 or the charge distribution-multi site complexation (CD-MUSIC) model47,48 have been employed to describe similar surface systems. These models assume heterogeneous surfaces interacting with the surrounding ion layers. Some studies also included the influence of buffer ions on the adsorption of metal cations49. Results M ti Magnetic nanoparticles present a chimeric surface to peptides. We first investigated the binding behaviour of the nonfunctionalised magnetite nanoparticles (MNP) to peptides at a moderately basic pH of 7.4 in tris buffered saline (TBS) solution. These particles had been extensively characterised previously with respect to their shape and size distribution by TEM and XRD (see Fig. 2 and ref.17) and their chemical composition by XPS, Mössbauer (See supplementary information Figure S1 and ref.17) and ATR-IR spectroscopy50. The crystal- line particles present a spherical shape and a particle size of around 13.5 nm as shown in Fig. 2. The particles yield a typical spinel crystal pattern (JCPDS card no. 85-1436) with a Scherrer diameter of 8.3 nm which can also be observed in Fig. 2. g In TBS pH 7.4, zeta potential measurements indicate that the MNP surface is slightly positively charged (see Table 1). The point of zero charge was determined to be approximately 7.8 by acidometry measurements50. Under these conditions, one might expect that negatively charged peptides would bind to the surface with a high affinity. SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 2 www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 2. Transmission electron microscopy (TEM) image of magnetic nanoparticles used for peptide array binding experiments (top left) with corresponding particle size distribution (top right). X-ray diffractrogram of the freeze-dried nanoparticles (bottom left) and potentiometric titration (bottom right). Figure 2. Transmission electron microscopy (TEM) image of magnetic nanoparticles used for peptide array binding experiments (top left) with corresponding particle size distribution (top right). X-ray diffractrogram of the freeze-dried nanoparticles (bottom left) and potentiometric titration (bottom right). Buffer pH Zeta potential, mV TBS 7.4 + 3.7 ± 0.2 8 −5.6 ± 0.3 PBS 6 −27 ± 1.5 7.4 −25 ± 1.3 8 −27.8 ± 1.5 CBS 6 −34.7 ± 1.8 Table 1. Zeta potentials of magnetic nanoparticles in different buffers and at different pHs used in the peptide array experiments. The values are the average of 10 measurements and an experimental error of 5.3% was determined for static zeta potential measurements. Table 1. Zeta potentials of magnetic nanoparticles in different buffers and at different pHs used in the peptide array experiments. The values are the average of 10 measurements and an experimental error of 5.3% was determined for static zeta potential measurements. Table 1. Zeta potentials of magnetic nanoparticles in different buffers and at different pHs used in the peptide array experiments. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 3. Binding scores of magnetic nanoparticles on peptides at pH 7.4 and pH 8 in TBS. The dashed line indicates the noise level in this experiment and the error bars show the standard deviation of the membrane duplicates. Figure 3. Binding scores of magnetic nanoparticles on peptides at pH 7.4 and pH 8 in TBS. The dashed line indicates the noise level in this experiment and the error bars show the standard deviation of the membrane duplicates. Figure 4. Peptide array scores of MNP at different pHs in phosphate buffered saline (PBS). The noise level is denoted by the dashed line; n = 2. Figure 4. Peptide array scores of MNP at different pHs in phosphate buffered saline (PBS). The noise level is denoted by the dashed line; n = 2. When the pH is changed to 8, the surface charge of the iron oxide nanoparticles is expected to become more negative. In accordance with this expectation, the binding scores of the negative peptides also drop from their value at the pH 7.4, while the binding scores of positive peptides rise; nevertheless, binding of both species is visible, as before. The binding of the species with the same charge polarity suggests that the nanoparticles present sites with locally varying charge polarity. Regarding the uncharged peptides, the change in the pH from 7.4 to 8 shows no distinct trend observable with regard to the experimental accuracy (Fig. 3). The lack of change in the binding scores for uncharged peptides implies that their adsorption is governed by forces that are not influenced by a change in the pH. In fact, the uncharged peptides have the same binding score regardless of the buffer solu- tion and over the whole pH range for all the experiments performed, as seen in Figure S8 where a summary is given for the peptide interaction scores with MNP in all the different buffer conditions comparing the charged (Figure S8a) and the uncharged (Figure S8b) peptides. We clearly observe that a trend can be seen for the change in the scores of the charged peptides as a function of the buffer solution and the pH, while the uncharged peptides are unaffected with regard to these variables. Results M ti The values are the average of 10 measurements and an experimental error of 5.3% was determined for static zeta potential measurements. In agreement with this expectation, the colorimetric data in Fig. 3, which corresponds to the absorbed mass of MNP on spots of the same size and peptide density for all experiments, show that at a pH of 7.4, the nega- tively charged hexa-glutamic acid (6E) and hexa-aspartic acid (6D) peptides show the highest adsorption on the MNP. Surprisingly, however, the second highest set of scores at pH 7.4 is observed for the positively charged hexa-arginine (6R), hexa-lysine (6K), followed by 5RH and 5RE, as well as hexa-histidine (6H) peptides. Histidine has a low percentage of positively charged side chains at a pH of 7.4, as indicated by its theoretical pI of 7.21 deter- mined by the ProtParam Tool by ExPASy51. Histidine and carboxy groups are also known to form inner sphere complexes with metal ions52,53. The uncharged peptides come next although at a much lower score. Hence, the first observation is that the peptides can be categorised in their adsorption scores based on their charge. We also find that the nanoparticles do not bind strongly to the nonfunctionalised membrane seen from the low back- ground noise level shown in the graph 3; all colorimetric scores are reported relative to this background. It is known that hydrophobic amino acids may interact with the membrane; we took care to pre-treat the membrane with methanol to minimise this effect. Next we analysed the loading of the membrane in MES buffered saline at pH 6 (see Figure S6). A similar binding pattern was obtained. SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 3 SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 www.nature.com/scientificreports/ www.nature.com/scientificreports/ Substance pKa(1) pKa2 pKa3 Source Charge of dominant species Chemical structure at pH 7 At pH 6 At pH 7.4 At pH 8 Tris 8.07 55 + 1 + 1 + 1 Phosphate 2.16 7.21 12.33 68 −1 −2 −2 Citrate 3.13 4.76 6.40 69 −2 −3 −3 Table 2. Acid dissociation constants of buffer molecules used in the peptide array binding experiments at 25 °C in aqueous solutions. Table 2. Acid dissociation constants of buffer molecules used in the peptide array binding experiments at 25 °C in aqueous solutions. Figure 5. Binding scores of magnetic nanoparticles to peptides in citrate buffered saline pH 6 (CBS). The dashed line denotes the background noise level. Error bars were derived from the standard deviation of 4 measurements (two experiments were performed with membrane duplicates). Figure 5. Binding scores of magnetic nanoparticles to peptides in citrate buffered saline pH 6 (CBS). The dashed line denotes the background noise level. Error bars were derived from the standard deviation of 4 measurements (two experiments were performed with membrane duplicates). Buffer conditions, not only pH, control peptide affinity. Focusing on the charged peptides we con- sidered ways to manipulate the density of positively and negatively charged spots on the surface which are in the nanometer scale5. The surface composition of the MNP can be modulated during MNP synthesis or through a chemical post-synthetic modification which we wanted to avoid. An in-situ opportunity to modify the effec- tive concentration of patches on the surface arises through buffer manipulation. The effective binding surface the nanoparticles present to the peptide is defined by the charge concentration in the inner Helmholtz plane, which depends on the buffer conditions (buffer species and their concentrations)54. Hereby, ions from the buffer compete with the (charged) groups of the peptide for adsorption. As the pKa of the buffer, tris(hydroxymethyl) aminomethane (tris), is 8.07 at 25 °C55, a mixture of neutral and monovalent cationic tris molecules are present at the pHs of 7.4 and 8 in our experiments. Hence, to vary the number of available positive surface positions, further binding experiments were conducted in phosphate buffered saline (PBS, see Fig. 4), with phosphate present in the dissociated, anionic form at all of the aforementioned pH values (see Table 2). www.nature.com/scientificreports/ As a result, within the accuracy of the measurements, the choice of buffer and the pH cannot be adequately linked to the binding affinity of the uncharged peptides, or be used to further categorise them with regard to their physical properties. Only 6P (hexa-proline) and 6Y (hexa-tyrosine) show a slight change in the score as a function of pH, but an overall rationalisation of this effect is hindered by the lack of an experimentally significant change in the rest of the (13 other) uncharged peptides. SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 4 www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 6. Schematic representation of a nanoparticle surface with positive, negative and neutral ‘‘spots”, denoted by red, blue and white squares, respectively. Charged buffer species are represented by red and blue circles for positive and negative charges, respectively, peptides depicted by lines. The proportion of charged spots is exaggerated for the sake of illustration. Figure 6. Schematic representation of a nanoparticle surface with positive, negative and neutral ‘‘spots”, denoted by red, blue and white squares, respectively. Charged buffer species are represented by red and blue circles for positive and negative charges, respectively, peptides depicted by lines. The proportion of charged spots is exaggerated for the sake of illustration. with increasing pH, which is commensurate with an increase in negative charge on the MNP surface. We find that 6K scores are higher than 6R despite having a lower pKR of 10.53 compared to 12.48 (Table S1)56. It has been argued that the charge delocalisation in the guanidinium group of arginine leads to a reduction in its hydrogen bond donating capacity to phosphate groups compared to that of lysine57, which may account for this observation. In contrast to the results from TBS, there is a much lower MNP adsorption on negatively charged spots, suggest- ing that ions in the solution compete with the peptides for nanoparticle binding. Buffer conditions can thus be varied to manipulate the local density of positive and negative binding sites. Just as in TBS, however, uncharged peptides exhibit binding which is smaller than the charged ones and is indifferent to the change of the pH.ff p p g gf g p For the following tests, we chose a buffer with bivalent and trivalent anions. In citrate buffered saline (CBS) at pH 6 (Fig. 5, pka values shown in Table 2), positive peptides bind the highest quantity of magnetic nanoparticles while on negatively charged peptides no binding can be seen whatsoever. The scores for the uncharged peptides are also decreased, although remain constant from peptide to peptide. These findings are consistent with the fact that bivalent anionic buffer molecules are rather strongly adsorbed onto the iron oxide surface; this is also reflected by a change in the zeta potential from a positive value to a negative value (see Table 1 and ref.58). www.nature.com/scientificreports/ A closer look at the interaction score under these conditions reveals a clear trend of an increased binding of MNP to positively charged peptides SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 5 www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 7. Comparison of the theoretically and the experimentally obtained binding affinities for charged homo-peptides in oppositely charged buffers and different pH. All values were recalculated to be proportional to Gibbs free energy of binding. Figure 7. Comparison of the theoretically and the experimentally obtained binding affinities for charged homo-peptides in oppositely charged buffers and different pH. All values were recalculated to be proportional to Gibbs free energy of binding. adsorption models60, such as the multisite CD-MUSIC approach48. In these models, the surface is composed of a set of binding sites with a specific local chemistry that interact with ions and other adsorbates in solution. Because the loading of the surface correlates strongly with the charge of the peptides, in this investigation we perceive the bare MNP to present positive, negative and neutral binding “spots”52,61 (roughly the size of an amino acid5) that can competitively or cooperatively react with the components of the solution, as illustrated in Fig. 6. As the trends to be explained are found to be strongly related to the electrostatic nature of the species, only electrostatic interactions are considered in the model, in order to derive design rules regarding the adsorption of the charged peptides to the unfunctionalised MNP under various buffer conditions.f p pf Due to the fact that the concentration of buffer molecules is at least an order of magnitude larger than the concentration of the magnetite nanoparticle binding sites and many orders of magnitude larger than the con- centration of the peptide molecules (see Figure S4) we can use a hierarchical framework of chemical equilibria. Thereby, the buffer ions react first with the particle and the buffer composition and the pH determine the con- centration of the ionic species on the surface, which is then able to bind to the peptide (Fig. 6) independent of the concentration of the peptide itself. In this way the nanoparticle is assumed to present an effective surface to the peptides defined by the buffer solution. if The spots on the surface may be considered either as a homogeneous binding site in the sense of a multisite-adsorption model48 or may have a size smaller than that of the peptides. www.nature.com/scientificreports/ Recent MD simulations sup- port such a picture for peptide binding on titania and silica, where the peptide surface interactions are rational- ised by the interactions of individual amino acids with specific surface sites, either directly, or mediated by the solvent27,42,62–64. Under these conditions, peptides of increasing length would need to bind to multiple spots. As positive and negative spots are present on the surface and the amino acids are constrained by peptide bonds, the peptide is presented to a large set of arrangements of positive, negative and neutral spots on the surface with a concentration that varies with pH and the buffer. However, due to the MNP surface being in a great excess compared to the peptide and hence containing a large amount of randomly distributed adsorption sites, it can be assumed that suitable patterns on the MNP surface exist for the peptides to bind. If there were not enough suitable adsorbing site patterns on the surface to accommodate the peptides, the influence of the buffer and the pH would be hindered and the observation of the significant functionalities seen from Figures 3, 4 and 5 would not be possible. Figure S9 shows that there is a size dependence of the binding score, which indicates that multiple amino acids can participate in the interaction of the peptide with the surface. p p p p In order to model the loading on the homo-peptides, we have used the interactions of the individual amino acids with the MNP surface. We have computed the free energies of binding using all-atom peptide models within an implicit solvent/implicit surface approximation, where the interaction strength of an individual amino acid with the surface is parameterised heuristically (See SI). We note that this model does not assume any particular peptide conformation, but only that the free energy of binding of a charged peptide is proportional to the charge of the peptide and the concentration of the positive and negative spots on the MNP surface. As Figure S8 indicates for peptides of up to length eight, there is an almost linear relationship between the affinity of the peptide and its length, which is effectively the result of the lack of the influence of the conformational structure of the peptide. SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 www.nature.com/scientificreports/ The high score of positive peptides in CBS compared to TBS and PBS at similar or higher pHs may in addition contain the effect of ionic interactions between the lysine amino groups with citrate bound to the MNP surface59. A patchwork model for the observed peptide-surface interactions. In order to interpret the experi- mental results for the charged peptides, we investigate a theoretical approach to explain the diverse binding prop- erties of peptides with different physiochemical characteristics as a function of the environmental conditions. The experimental data suggest that there is a strong interplay between the buffer conditions and the interaction strength of the charged peptides with MNP. Our model is conceptually similar to the surface complexation and SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 6 www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 8. Binding scores of MNP on wet membrane after incubation in TBS pH 7.4 for 1 h and after membrane transfer to CBS pH 6 for 1 h. The noise level is represented by the dashed line. Two membrane replicates were used to obtain the data and the average and standard deviation was calculated from it. Figure 8. Binding scores of MNP on wet membrane after incubation in TBS pH 7.4 for 1 h and after membrane transfer to CBS pH 6 for 1 h. The noise level is represented by the dashed line. Two membrane replicates were used to obtain the data and the average and standard deviation was calculated from it. concentration of the positive and negative spots after reacting with the buffer were thus computed for all exper- imental situations. The fractional loading (proportional to the score of the measurements) is then correlated with the charge of the peptide. As the fractional loading is proportional to the interaction strength, the model assumes that the positive (K,R) and negative (E,D) amino acids are characterised each by one single interaction parameter. As Fig. 7 illustrates, the obtained result is consistent with the observed MNP loading of the membrane for all buffer conditions, including those with a high variety in the fractional loading. The good correlation found between the computed and the observed loadings indicates that the interaction of the peptide with the surface is largely determined by its charging interactions with the surface patches of the opposite sign. This is in line with the overall experimental finding that the buffer/pH conditions influence mostly the electrostatic interactions, thereby not influencing the uncharged peptides in a recognisable manner within the measurement accuracy. It is hence the concentration of the surface patches that is affected by the composition of the buffer. Reversible peptide binding. This analysis suggests that manipulation of the buffer may be exploited for reversible peptide binding. In order to demonstrate this effect, we first prepared MNP in TBS at pH 7.4, which is favourable to bind negatively charged peptides (see Fig. 8). After binding the peptides, we washed the peptide array membrane three times with TBS and an image of the wet membrane was taken. www.nature.com/scientificreports/ For the other experiments the membrane had been dried before the scanning in order to increase the contrast, but this would prolong the desorption of MNP due to the need to rehydrate the particles. The wet membrane was therefore transferred to fresh CBS with a pH of 6. Under these conditions, all positively charged MNP surface sites should be covered by bivalent anionic citrate ions, which are available in excess in solution. After an incubation time of 1 h in CBS, we found that the amount of MNP on the hexa-glutamate and hexa-aspartate spots were greatly reduced showing the reversibility of the interaction. On the other hand, the binding affinity towards positive peptides increased showing a reorganisation of the nanoparticles on the array. www.nature.com/scientificreports/ A similar, one-to-one relationship was observed experimentally for the tetra-to deca-homomers of aspartic and glutamic acid (Figure S9b).i g g Based on this analysis we now test a simplified model for the binding of charged peptides to MNP in var- ying buffer conditions (see model section in the SI for details). The fraction of charged buffer species and the SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 7 SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 Materials and Methods The degassed particle suspensions were adjusted to a con- centration of 2 g L−1 and equilibrated at a pH of 4 overnight. The whole titration was conducted under nitro- gen atmosphere at 298.5 K with HCl and NaOH as titrands and two different NaCl concentrations of 10 and 100 mmol L−1. Magnetic nanoparticle binding assay. In order to determine the binding of peptides to bare magnetic nanoparticles (MNP) CelluSpot peptide arrays from Intavis with 5 to 10 nmol of peptides per spot were used. The peptides were bound to the cellulose membrane via the C-terminus and the free N-terminus is acetylated. The spot diameter was 2 mm. A schematic picture of the setup is shown in Fig. 1. The membrane on which the peptides had been synthesised by the manufacturer was conditioned with 1 mL of methanol in order to rehydrate hydrophobic peptides67. Methanol is toxic and should be handled with care. The buffers used were 50 mM of cit- rate (CBS), phosphate (PBS), tris(hydroxymethyl) aminomethane (TBS) and 2-(N-morpholino)ethanesulfonic acid (MES-BS) in double distilled water and were supplemented with 137 mM NaCl and 2.7 mM KCl. Finally, Tween 20 was added to a concentration of 0.25% (v/v) to the buffers to reduce nonspecific binding. The orbital shaker used for incubation was MulitBio3D from Biosan.hf The array membrane was washed three times for 10 min each with 50 mL of buffer rotating of the orbital shaker at 30 rpm. After washing, the membrane was incubated for 60 minutes in a 0.4 g/L MNP suspension25 at the same rocking speed. The MNP suspension was freshly prepared before the experiment by adding buffer to lyophilised nanoparticles and sonicating for 15 minutes. Unbound particles were removed from the membrane by washing with buffer three times for 10 minutes. In order to test for reversible binding, the membrane was consecutively incubated in the buffer of interest for 1 h. The cellulose membrane was dried overnight at 4°C; then an image was taken using a GelDoku station. An example picture is shown in Figure S2. To quantify the staining of the spots, the microarray profile plugin for ImageJ was used. The output of this plugin is a mean value of the spot darkness, which is correlated to the amount of bound magnetic nanoparticles. An average value of the back- ground was determined from 32 spots on the membrane without any peptides. Materials and Methods All reagents used are commercially available and were used as received from the manufacturer without further purification. All reagents used are commercially available and were used as received from the manufacturer without further purification. Synthesis of bare magnetic nanoparticles. The bare iron oxide nanoparticles employed for this study were synthesised by co-precipitation of Fe2+ and Fe3+ in alkaline aqueous solutions according to our previously optimised procedure66. 21.2 g of FeCl3 × 6 H2O and 8.3 g of FeCl2 × 4 H2O were dissolved in 200 mL of deionised, degassed water resulting in a Fe(III): Fe(II) ratio of 1.9: 1. This iron chloride solution was added to 1 L of 1 M NaOH prepared with deionised, degassed water stirring at 250 rpm in a reaction vessel. The reaction mixture was kept under a nitrogen atmosphere at 25 °C and stirred for an additional 30 minutes before the resulting nanopar- ticles were washed with deionised water until the conductivity of the MNP solution was below 200 µS cm−1. In order to separate the particles, the mixture was placed on a NdFeB permanent magnet. Suspensions were lyoph- ilised with an ALPHA 1-2LD plus from Martin Christ Gefriertrocknungsanlagen GmbH, Germany in order to obtain solid particles. FeCl3 × 6 H2O and sodium hydroxide were purchased from AppliChem GmbH, Germany in the highest purity available. FeCl2 × 4 H2O extra pure was obtained from Merck KGaA, Germany. Transmission electron microscopy (TEM). TEM images were recorded using a JEM 100-CX (JEOL GmbH, Germany). For the TEM measurements the colloidal samples were diluted in degassed and deionised water, ultrasonicated to disperse any agglomerates and precipitated on carbon coated copper grids (Quantifoil Micro Tools GmbH, Germany). The pictures were manually processed in ImageJ. 30 particles were measured in random order. X-ray diffraction (XRD). Crystal structure and phase purity of the lyophilised samples were examined with powder X-ray diffraction (XRD). The measurements were performed with a Stadi-P diffractometer (STOE & Cie GmbH, Germany), equipped with a molybdenum source (λ = 0.7093 Å) and a Mythen 1 K detector (DECTRIS Ltd., Switzerland) in transmission geometry. Data was collected in the range from 2° to 50° (2θ). The software package STOE WinXPOW (STOE & Cie GmbH, Germany) was used for indexing and refinement purposes. Potentiometric titrations. Potentiometric titrations were accomplished in an OptiMax™ reactor (Mettler-Toledo GmbH, Germany) from pH 4 to 10. Conclusionsh The interaction of bare magnetic nanoparticles with all 20 natural amino acids has not previously been documented, as analytical methods struggle to quantify such affinities. By providing systematic data for homopeptides of all natu- rally occurring amino acids, we have shown the dependency of their interactions with non-functionalised magnetic iron oxide nanoparticles as a function of buffer solution and pH. We have demonstrated for bare MNP that chromo- phoric read-outs prove to be reliable for the characterisation of the interactions with short peptides. We find that both negative and positive peptides can bind the MNP while the uncharged peptides show relatively small binding propensity which is indifferent to the buffer and the pH. We therefore conclude that the electrostatic interactions play a major role in the adsorption of peptides on the MNP. Changing the buffer conditions can hence enable us to tune the interaction between the peptides and the magnetic nanoparticles. A patchwork electrostatic multisite model was used to describe the adsorption affinity (free energy) of the positive and negative peptides on oppositely charged spots on the MNP surface. The results of this analysis indicate that the binding scores of the positive and negative peptides can be fully explained by electrostatic interactions. It was further demonstrated that a reversible binding of the negative peptides to MNP can be achieved by changing the buffer conditions. These results provide a basis for the design of reversible or permanent binding tags for biomolecules that show an affinity to non-functionalised MNP. We note that this methodology is suitable to identify peptide tags for the binding of a protein to bare mag- netic nanoparticles and that tag interaction patterns found in peptide arrays are in accordance with the behaviour of tagged green fluorescent protein (GFP) variants in different buffers65. Such tags pave the way for fast and simple immobilisation procedures without the need for unstable13,14 and expensive affinity ligands currently in use15. They therefore have the potential to be used as both efficient and effective tools for the isolation of recombinant proteins. SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 8 www.nature.com/scientificreports/ Materials and Methods The difference between this back- ground value and the darkness of the peptide spots is dependent on the amount of magnetic nanoparticles that was adsorbed by the peptides and is therefore an indicator of the binding selectivity of the peptides towards the MNP. In order to determine the background noise, the standard deviation was calculated for the darkness values of these 32 spots without any peptides. For each experiment, this noise level is indicated as a dashed horizontal line in the respective figures. The membrane was regenerated in 100 mM oxalic acid for 40 minutes rotating on the orbital shaker after each use and washed for 10 minutes with double distilled water three times. Dried membranes were stored at −20°C in a sealed plastic bag. Zeta potential measurements. A buffered suspension of 0.4 g/L magnetic nanoparticles as used in the peptide array binding assays described above was sonicated for 15 minutes. The zeta potential was determined using the Smoluchowski equation in a Beckman Coulter Delsa Nano C at 25 °C. Each measurement was taken three times with 10 accumulations and a pinhole of 50 µm. SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 9 www.nature.com/scientificreports/ References References 1. Rittich, B. & Španová, A. SPE and Purification of DNA Using Magnetic Particles. J. Sep. Sci. 36, 2472–2485 (2013). 1. Rittich, B. & Španová, A. SPE and Purification of DNA Using Magnetic Particles. J. Sep. Sci. 36, 2472–2485 (2013). i 2. Hong, J. W., Studer, V., Hang, G., Anderson, W. F. & Quake, S. R. A Nanoliter-Scale Nucleic Acid Processor with Parallel Architecture Nat. Biotechnol. 22, 435–439 (2004). 3. Fraga García, P. et al. High-Gradient Magnetic Separation for Technical Scale Protein Recovery Using Low Cost Magnetic Nanoparticles. Sep. Purif. Technol. 150, 29–36 (2015). p p f 4. Colombo, M. et al. Biological Applications of Magnetic Nanoparticles. Chem. Soc. Rev. 41, 4306–4334 (2012). 5 R th H C S h i S P P F & B i S I bili ti f C ll l M ti N i Ch i t O lombo, M. et al. Biological Applications of Magnetic Nanoparticles 5. Roth, H.-C., Schwaminger, S. P., Peng, F. & Berensmeier, S. Immobilization of Cellulase on Magnetic Nanocarriers. ChemistryOpen 5, 183–187 (2016). 6. Misson, M., Zhang, H. & Jin, B. Nanobiocatalyst Advancements and Bioprocessing Applications. J. R. Soc., Interface 12, https://doi org/10.1098/rsif.2014.0891 (2015). g ( ) 7. Shinkai, M. Functional Magnetic Particles for Medical Application. J. Biosci. Bioeng. 94, 606–613 (2002). g 7. Shinkai, M. Functional Magnetic Particles for Medical Application. J. Biosci. Bioeng. 94, 606–613 (2002). . Senpan, A. et al. Conquering the Dark Side: Colloidal Iron Oxide p q g p 9. Boyer, C., Whittaker, M. R., Bulmus, V., Liu, J. & Davis, T. P. The Design and Utility of Polymer-Stabilized Iron-Oxide N for Nanomedicine Applications. NPG Asia Mater 2, 23–30 (2010). 9. Boyer, C., Whittaker, M. R., Bulmus, V., Liu, J. & Davis, T. P. The Design and Utility of Polymer Stabilized Iron Ox for Nanomedicine Applications. NPG Asia Mater 2, 23–30 (2010). pp V. et al. Magnetically Recoverable Nanocatalysts. Chemical Reviews 10. Polshettiwar, V. et al. Magnetically Recoverable Nanocatalysts. g y y 11. Block, H. et al. Chapter 27 Immobilized-Metal Affinity Chromatography (IMAC). Methods Enzymol. 463, 439–473 (2009). fi 2. Zhou, Y. et al. Synchronized purification and immobilization of his-tagged β-glucosidase via Fe3O4/PMG core/shell magneti nanoparticles. Sci Rep. 7, https://doi.org/10.1038/srep41741 (2017).fii i nanoparticles. Sci Rep. 7, https://doi.org/10.1038/srep41741 (2017)fi p p p g p 3. Hearn, M. T. W. & Acosta, D. References Applications of Novel Affinity Cassette Methods: Use of Peptide Fusion Handles for the Purification o Recombinant Proteins. J. Mol. Recognit. 14, 323–369 (2001).i 13. Hearn, M. T. W. & Acosta, D. Applications of Novel Affinity Cassette Methods: Use of Peptide Fusion Recombinant Proteins. J. Mol. Recognit. 14, 323–369 (2001).i 13. Hearn, M. T. W. & Acosta, D. Applications of Novel Affinity Ca fi Recombinant Proteins. J. Mol. Recognit. 14, 323–369 (2001). g 4. Gaberc-Porekar, V. & Menart, V. Potential for Using Histidine Tags in Purification of Proteins at Large Scale. Chem. Eng. Technol. 28 1306–1314 (2005).hi 15. Franzreb, M., Siemann-Herzberg, M., Hobley, T. J. & Thomas, O. R. T. Protein Purification Using Magnetic Adsorbent Particles. Appl. Microbiol. Biot. 70, 505–516 (2006). pp 16. Tamerler, C. & Sarikaya, M. Genetically Designed Peptide-Based Molecular Materials. ACS Nano 3, 1606–1615 (2009). 7. Schwaminger, S., Bauer, D., Fraga-García, P., Wagner, F. E. & Berensmeier, S. Oxidation of Magnetite Nanoparticles. Impact on Surface and Crystal Properties. CrystEngComm, 246–255 (2016). y p y g 8. Stroud, H. et al. Non-CG Methylation Patterns Shape the Epigenetic Landscape in Arabidopsis. Nat Struct Mol Biol 21, 64–72 (2014) 19. Price, J. V. et al. On Silico Peptide Microarrays for High-Resolution Mapping of Antibody Epitopes and Diverse Protein-Protein Interactions. Nat Med 18, 1434–1440 (2012). 0. Ahn, S. B. et al. Characterization of the Interaction between Heterodimeric αvβ6 Integrin and Urokinase Plasminogen Activato Receptor (uPAR) Using Functional Proteomics. J. Proteome Res. 13, 5956–5964 (2014).i 21. Ahmed, S., Mathews, A. S., Byeon, N., Lavasanifar, A. & Kaur, K. Peptide Arrays for Screening Cancer Specific Peptides. Ana 82, 7533–7541 (2010).h 22. Schutkowski, M., Thiele, A. & Koch, J. In Peptides as Drugs (Wiley-VCH Verlag GmbH & Co. KGaA2009), pp. 187–217.ih ki, M., Thiele, A. & Koch, J. In Peptides as Drugs (Wiley-VCH Verlag Schutkowski, M., Thiele, A. & Koch, J. In Peptides as Drugs (Wiley-V l b l ll l d d b 23. Kramer, A. et al. Combinatorial Cellulose-Bound Peptide Libraries: Screening Tools for the Identification of Peptides T Ligands with Predefined Specificity. Methods 6, 388–395 (1994).fi gi pi y 4. Okochi, M. et al. Screening of peptides with a high affinity for ZnO using spot-synthesized peptide arrays and computationa analysis. Acta Biomaterialia 6, 2301–2306 (2010). y 25. Kuboyama, M., Kato, R., Okochi, M. & Honda, H. Screening for Silver Nanoparticle-Binding Peptides by Using a Peptide Array. Biochem. Eng. J. References 66, 73–77 (2012). g 6. Shemetov, A. A., Nabiev, I. & Sukhanova, A. Molecular Interaction of Proteins and Peptides with Nanoparticles. ACS Nano 6 4585–4602 (2012).i ( ) 7. Schneider, J. & Ciacchi, L. C. Specific Material Recognition by Small Peptides Mediated by the Interfacial Solvent Structure. J. Am Chem Soc 134 2407 2413 (2012) 27. Schneider, J. & Ciacchi, L. C. Specific Material Recognition by Small Peptides Mediated by the Interfacial Solvent Structure. J. Am. Chem. Soc. 134, 2407–2413 (2012). 8. Gitelman, A. & Rapaport, H. Bifunctional Designed Peptides Induce Mineralization and Binding to TiO2. Langmuir 30, 4716–4724 (2014).fi 28. Gitelman, A. & Rapaport, H. Bifunctional Designed Peptides Induce Mineralization and Binding to TiO2. Langmuir 30, 4716–4724 (2014). Ch S l A l f d d Affi d O d l b A S A S 9. Chang, S. Y. et al. Analysis of Peptides and Proteins Affinity-Bound to Iron Oxide Nanoparticles by MALDIMS. J. Am. Soc. Mas Spectrom. 18, 910–918 (2007). p 30. Vallee, A., Humblot, V. & Pradier, C.-M. Peptide Interactions with Metal and Oxide Surfaces. Acc. Chem. Res. 43, 1297–1306 (2 30. Vallee, A., Humblot, V. & Pradier, C.-M. Peptide Interactions with Metal and Oxide Surfaces. Acc. Chem. Res. 43, 1297–1306 (2010). 31 Costa D Garrain P A & Baaden M Understanding Small Biomolecule Biomaterial Interactions: A Review of Fundamental 30. Vallee, A., Humblot, V. & Pradier, C.-M. Peptide Interactions with Metal and Oxide Surfaces. Acc. Chem. Res. 43, 1297–1306 (2010). 31. Costa, D., Garrain, P.-A. & Baaden, M. Understanding Small Biomolecule-Biomaterial Interactions: A Review of Fundamental Th l d E l A h f B l l I h I S f J B d M R P A 1. Costa, D., Garrain, P.-A. & Baaden, M. Understanding Small Biomolecule-Biomaterial Interactions: A Review of Fundamenta Theoretical and Experimental Approaches for Biomolecule Interactions with Inorganic Surfaces. J. Biomed. Mater. Res. Part A 101 1210–1222 (2013). 1210–1222 (2013). 32. Grasso, G. et al. Cell Penetrating Peptide Adsorption on Magnetite and Silica Surfaces: A Computational Investigation. J. Phys. Chem. B 119, 8239–8246 (2015). 2. Grasso, G. et al. Cell Penetrating Peptide Adsorption on Magnetite and Silica Surfaces: A Computational Investigation. J. Phys Chem. B 119, 8239–8246 (2015). 3. Rimola, A., Aschi, M., Orlando, R. & Ugliengo, P. Does Adsorption at Hydroxyapatite Surfaces Induce Peptide Folding? Insight from Large-Scale B3LYP Calculations. J. Am. Chem. Soc. 134, 10899–10910 (2012). www.nature.com/scientificreports/ p p g y J g g , ( ) 67. Brand, N. Protein-Protein Interactions: A Molecular Cloning Manual. By Erica A. Golemis and Peter D. Adams (Eds). Proteomics 6, 6119–6120 (2006). g g g ( y ) 69. Lide, D. R. CRC Handbook of Chemistry and Physics. A Ready-Reference Book of Chemical and Physical Data. 87th ed. (CRC Taylor & Francis, Boca Raton, Fla., 2006). Author Contributions S.A.B. planned and performed binding experiments and prepared the figures, S.P.S. synthesised and characterised the nanoparticles, M.B. and P.A. performed computational analysis, P.A. and P.Y. discussed and analysed the data, S.A.B. and W.W. analysed the data and wrote the manuscript. P.F.-G. and S.B. discussed the data. P.F.-G., S.P.S., S.A.B., P.A., M.B., K.F., P.Y., and W.W. edited the manuscript. M.B. and K.F. contributed with the theoretical model. S.B. conceived the idea and coordinated the project. Acknowledgementsh This work was funded by the Federal Ministry of Education and Research (Grant No. 031A173A and 031A173B). This work was supported by the German Research Foundation (DFG) and the Technical University of Munich within the funding programme Open Access Publishing. The Mössbauer spectroscopy measurements were performed and processed by Professor Fritz Wagner. The authors thank Dr. Isaac Shim for language editing the manuscript, and Johann Kaiser for help with the peptide array data analysis. www.nature.com/scientificreports/ Poly(L-Lysine)-Modified Iron Oxide Nanoparticles for Stem Cell Labeling. Bioconjugate Chem. 19, 740–750 (2008) 0 D i J A & L ki J O S f i i i d l i h id / i f 3 Ad i f i J C ll id I , y( y )i p g j g , ( ) 60. Davis, J. A. & Leckie, J. O. Surface ionization and complexation at the oxide/water interface. 3. Adsorption of anions. J Colloid Interf Sci 74, 32–43 (1980). 1. Garcell, Morales, Andres-Verges, Tartaj & Serna. Interfacial and Rheological Characteristics of Maghemite Aqueous Suspensions. J Colloid Interface Sci 205, 470–475 (1998).i f 2. Emami, F. S. et al. Prediction of Specific Biomolecule Adsorption on Silica Surfaces as a Function of pH and Particle Size. Chem Mater. 26, 5725–5734 (2014). 3. Ramakrishnan, S. K., Martin, M., Cloitre, T., Firlej, L. & Gergely, C. Molecular Mechanism of Selective Binding of Peptides to Silicon Surface. J. Chem. Inf. Model. 54, 2117–2126 (2014). f 64. Brandt, E. G. & Lyubartsev, A. P. Molecular Dynamics Simulations of Adsorption of Amino Acid Side Chain An Titanium Binding Peptide on the TiO2 (100) Surface. J. Phys. Chem. C 119, 18126–18139 (2015). f 4. Brandt, E. G. & Lyubartsev, A. P. Molecular Dynamics Simulations of Adsorption of Amino Acid Side Chain Analogues and Titanium Binding Peptide on the TiO2 (100) Surface J Phys Chem C 119 18126 18139 (2015) g p y 65. Schwaminger, S. P. et al. Bio-nano interactions. Cellulase on iron oxide nanoparticle surfaces. Adsorption 23, 281–292 (2017) 65. Schwaminger, S. P. et al. Bio nano interactions. Cellulase on iron oxide nanoparticle surfaces. Adsorption 23, 281 292 (2017). 66. Roth, H.-C., Schwaminger, S. P., Schindler, M., Wagner, F. E. & Berensmeier, S. Influencing Factors in the CO-Precipitation Process 65. Schwaminger, S. P. et al. Bio nano interactions. Cellulase on iron oxide nanoparticle surfaces. Adsorption 23, 281 292 (2017). 66. Roth, H.-C., Schwaminger, S. P., Schindler, M., Wagner, F. E. & Berensmeier, S. Influencing Factors in the CO-Precipitation Process of Superparamagnetic Iron Oxide Nano Particles: A Model Based Study J Magn Magn Mater 377 81 89 (2015) 6. Roth, H.-C., Schwaminger, S. P., Schindler, M., Wagner, F. E. & Berensmeier, S. Influencing Factors in the CO-Precipitation Proces of Superparamagnetic Iron Oxide Nano Particles: A Model Based Study. J. Magn. Magn. Mater. 377, 81–89 (2015). www.nature.com/scientificreports/ www.nature.com/scientificreports/ 8. Hiemstra, T. & van Riemsdijk, W. H. A Surface Structural Approach to Ion Adsorption: The Charge Distribution (CD) Model Journal of Colloid and Interface Science 179, 488–508 (1996). 49. Tiberg, C., Sjöstedt, C., Persson, I. & Gustafsson, J. P. Phosphate effects on copper(II) and lead(II) sorption to ferrihydrite. Geochimica et Cosmochimica Acta 120, 140–157 (2013). 50. Schwaminger, S. P. et al. Nature of Interactions of Amino Acids with Bare Magnetite Nanoparticles. J. Phys. Chem. C 119, 23032–23041 (2015).h ( ) 1. Bjellqvist, B. et al. The Focusing Positions of Polypeptides in Immobilized pH Gradients Can Be Predicted from their Amino Acid Sequences. Electrophoresis 14, 1023–1031 (1993).h q p ( ) 52. Cornell, R. M. & Schwertmann, U. The Iron Oxides (Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, FRG, 2003).i . Cornell, R. M. & Schwertmann, U. The Iron Oxides (Wiley-VCH V h 53. Kozlowski, H., Potocki, S., Remelli, M., Rowinska-Zyrek, M. & Valensin, D. Specific metal ion binding sites in unstructured re of proteins. Coordination Chemistry Reviews 257, 2625–2638 (2013). y 54. Lyklema, J. H. Fundamentals of Interface and Colloid Science. Vol. 2. Solid-Liquid Interfaces. Contr. A. de Keizer, B. H. Bijsterbosch etc. p. 3.61 (Academic Press, London - San Diego etc, 1995).h p. 3.61 (Academic Press, London - San Diego etc, 1995). p ( g ) 5. El-Harakany, A. A., Abdel Halim, F. M. & Barakat, A. O. Dissociation Constants and Related Thermodynamic Quantities of the Protonated Acid Form of Tris (Hydroxymethyl) Aminomethane in Mixtures of 2 Methoxyethanol and Water at Differen 55. El-Harakany, A. A., Abdel Halim, F. M. & Barakat, A. O. Dissociation Constants and Related Thermodynamic Quantities of the Protonated Acid Form of Tris-(Hydroxymethyl)-Aminomethane in Mixtures of 2-Methoxyethanol and Water at Different Temperatures. J. Electroanal. Chem. Interfacial Electrochem. 162, 285–305 (1984). Protonated Acid Form of Tris-(Hydroxymethyl)-Aminomethane in Mixtures of 2-Methoxyethanol and Water at D Temperatures. J. Electroanal. Chem. Interfacial Electrochem. 162, 285–305 (1984). p f ( ) 56. Nelson, D. L. & Cox, M. M. Lehninger Principles of Biochemistry. 5th ed. (W. H. Freeman & Company, New York, 2008).i 7. Kooijman, E. E. et al. An Electrostatic/Hydrogen Bond Switch as the Basis for the Specific Interaction of Phosphatidic Acid with Proteins. J. Biol. Chem. 282, 11356–11364 (2007). 58. Jing, J. et al. One-step reverse precipitation synthesis of water-dispersible superparamagnetic magnetite nanoparticles. J. Nanopart. Res. 14, https://doi.org/10.1007/s11051-012-0827-3 (2012).i p g 9. Babic, M. et al. References g 4. Capriotti, L. A., Beebe, T. P. Jr & Schneider, J. P. Hydroxyapatite Surface-Induced Peptide Folding. J. Am. Chem. Soc. 129, 5281–5287 (2007).fi 35. Okochi, M. et al. Peptide Array-Based Characterization and Design of Zno-High Affinity Peptides. Biotechnol. Bioeng. 106, 845–851 (2010). 6. Schwaminger, S., Blank-Shim, S. A., Scheifele, I., Fraga-García, P. & Berensmeier, S. Peptide binding to metal oxide nanoparticles Faraday Discuss, Advance Article (2017).iff 7. Salis, A. & Monduzzi, M. Not Only pH. Specific Buffer Effects in Biological Systems. Curr. Opin. Colloid Interface Sci. 23, 1–9 (2016) 8. Bliem, R. et al. Subsurface Cation Vacancy Stabilization of the Magnetite (001). Surface. Science 346, 1215–1218 (2014).h iff 38. Bliem, R. et al. Subsurface Cation Vacancy Stabilization of the Magnetite (001). Surface. Science 346, 1215–1218 (201 39. Navrotsky, A., Mazeina, L. & Majzlan, J. Size-Driven Structural and Thermodynamic Complexity in Iron Oxides. Science 319, 1635–1638 (2008). 40. Merte, L. R. et al. Water-Mediated Proton Hopping on an Iron Oxide. Surface. Science 336, 889–893 (2012). 41. Heinz, H. et al. Nature of Molecular Interactions of Peptides with Gold, Palladium, and Pd-Au Bimetal Surfaces in Aqueous Solution. J. Am. Chem. Soc. 131, 9704–9714 (2009). 42. Emami, F. S. et al. Force Field and a Surface Model Database for Silica to Simulate Interfacial Properties in Atomic Resolution. Chem. Mater. 26, 2647–2658 (2014). ( ) 43. Torrie, G. M. & Valleau, J. P. A Monte Carlo study of an electrical double layer. Chemical Physics Letters 65, 343–346 (1979).h 4. Stern, O. The theory of the electrolytic double-layer. Zeit. Elektrochem., 508–516 (1924). h 45. Davis, J. A., James, R. O. & Leckie, J. O. Surface ionization and complexation at the oxide/water interface. Journal of Colloid and Interface Science 63, 480–499 (1978). 6 S j k D A Ph i l f l i d l f i h i l i f N 36 6 80 (1993) f 6. Sverjensky, D. A. Physical surface-complexation models for sorption at the mineral–water interface. Nature 364, 776–780 (1993). j ky, y p p , ( ) 47. Goldberg, S. Application of surface complexation models to anion adsorption by natural materials. Environ Toxicol Chem 33, 2172–2180 (2014). SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 10 Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-017-13928-6. Competing Interests: The authors declare that they have no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2017 SCientifiC REPOrTS \| 7: 14047 \| DOI:10.1038/s41598-017-13928-6 11
https://openalex.org/W4200161146	https://journals.umcs.pl/al/article/download/13027/9244	es		Editorial	Anuario Latinoamericano	2,021	cc-by	228	Anuario Latinoamericano Ciencias Políticas y Relaciones Internacionales vol. 11, 2021 p. 9 Editorial DOI: 10.17951/al.2021.11.9 En el undécimo volumen de la revista Anuario Latinoamericano – Ciencias Políticas y Relaciones Internacionales, les recomendamos leer el contenido del Dossier titulado “Brasil: entre la democracia y el autoritarismo” y coordinado por Maria do Socorro Sousa Braga y Katarzyna Krzywicka. La inestable situación política, económica y social, así como las negligencias y amenazas en el campo de la seguridad sanitaria en el país, atraen la atención de la opinión pública, periodistas, científicos y analistas. Los problemas internos de Brasil tienen una dimensión internacional y son de particular importancia en el contexto de las elecciones que se celebrarán en 2022. Los autores de los artículos publicados en el Dossier hacen una valiosa e importante contribución a la discusión actual sobre el estado de la democracia brasileña. Tradicionalmente les recomendamos que presten atención a la sección “Artículos y ensayos”, así como los alentamos a leer reseñas de libros en la sección “Reseñas e informes”. Deseándoles una buena lectura, les invitamos a enviar artículos y reseñas de libros para los próximos volúmenes de Anuario Latinoamericano. Las informaciones sobre los temas del Dossier de los volúmenes previstos para la publicación y los plazos para el envío de contribuciones están disponibles en el sitio web de nuestra revista. Lublin, 18 de julio de 2021 Katarzyna Krzywicka 9
https://openalex.org/W2368877988	https://europepmc.org/articles/pmc5322833?pdf=render	English	null	Non-signalling energy use in the developing rat brain	Journal of cerebral blood flow and metabolism	2,016	cc-by	14,293	Original Article Journal of Cerebral Blood Flow & Metabolism 2017, Vol. 37(3) 951–966 ! Author(s) 2016 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0271678X16648710 journals.sagepub.com/home/jcbfm Keywords y ATP, brain development, brain slice, energy metabolism, lipids Received 12 January 2016; Revised 30 March 2016; Accepted 5 April 2016 synaptic function.19,20 To do this, the actin cytoskeleton treadmills continuously, by adding G-actin monomers to one end of strands of F-actin and releasing them at the other end.21 Along the F-actin polymer, ATP bound to G-actin gets hydrolysed to ADP.22 Bernstein and Bamburg23 claimed that actin cycling accounted for half of all energy use in neuronal cultures but, as their experimental method suppresses neuronal glutamate release, this fraction is likely to be overesti- mated.24 Conversely, modelling of actin and micro- tubule turnover suggested that less than 1% of total brain energy use was spent on actin treadmilling, and even less on microtubule turnover.24 Experimental data on the energetic cost of microtubule turnover are Non-signalling energy use in the developing rat brain Elisabeth Engl1, Renaud Jolivet1,2, Catherine N Hall3 and David Attwell1 Abstract Energy use in the brain constrains its information processing power, but only about half the brain’s energy consumption is directly related to information processing. Evidence for which non-signalling processes consume the rest of the brain’s energy has been scarce. For the first time, we investigated the energy use of the brain’s main non-signalling tasks with a single method. After blocking each non-signalling process, we measured oxygen level changes in juvenile rat brain slices with an oxygen-sensing microelectrode and calculated changes in oxygen consumption throughout the slice using a modified diffusion equation. We found that the turnover of the actin and microtubule cytoskeleton, followed by lipid synthesis, are significant energy drains, contributing 25%, 22% and 18%, respectively, to the rate of oxygen consumption. In contrast, protein synthesis is energetically inexpensive. We assess how these estimates of energy expenditure relate to brain energy use in vivo, and how they might differ in the mature brain. 1Department of Neuroscience, Physiology & Pharmacology, University College London, London, UK 2CERN, and De´partement de physique nucle´aire et corpusculaire (DPNC), University of Geneva, Geneva, Switzerland 3School of Psychology, University of Sussex, Falmer, UK Corresponding author: David Attwell, Department of Neuroscience, Physiology & Pharmacology, University College London, Gower St., London WC1E 6BT, UK. Email: d.attwell@ucl.ac.uk Introduction Energy availability limits information processing in the brain,1–3 which consumes energy disproportionately in relation to its fraction of total body mass.4,5 The brain’s most salient feature is neuronal communication, and the energetic cost of the diﬀerent cellular processes underlying signalling has been well described.1–3,6–11 Most energy use is on the removal of sodium ions that enter neurons to generate synaptic and action poten- tials.1,6,11 However, inhibiting the sodium–potassium ATPase, without which signalling activity ceases, has shown that around 45% of the brain’s baseline energy use in vivo is consumed on non-signalling processes,12 which is considerably more than is estimated or assumed in many models of brain energy use.1,11,13,14 Surprisingly, it is largely unknown which non-signalling processes con- sume the rest of the brain’s energy. 1Department of Neuroscience, Physiology & Pharmacology, University College London, London, UK 2CERN, and De´partement de physique nucle´aire et corpusculaire (DPNC), University of Geneva, Geneva, Switzerland 3School of Psychology, University of Sussex, Falmer, UK Often called ‘housekeeping’ processes, the many tasks that the brain performs in addition to signalling provide the scaﬀold on which signalling, plasticity and the encoding of memory can occur. The actin cytoskel- eton regulates the morphology of the mature neuron, as well as its growth in development,15–18 and modulates Corresponding author: David Attwell, Department of Neuroscience, Physiology & Pharmacology, University College London, Gower St., London WC1E 6BT, UK. Email: d.attwell@ucl.ac.uk Journal of Cerebral Blood Flow & Metabolism 37(3) 952 stage. The bath volume was 2 ml, and solution ﬂowed over the top, but not the bottom, of the slice. Slices and electrode placement were observed using a 25 lens. The aCSF pH was adjusted to 7.4 with NaOH, and con- tained (mM) 140 NaCl, 10 D-glucose, 2.5 KCl, 10 HEPES, 1 NaH2PO4, 1 MgCl2 and 2 CaCl2 (osmo- larity 300 mmol/kg, bubbled with 100% oxygen, and heated to 36–37C). lacking, even though microtubules also exist in a state of dynamic instability, growing and shrinking in an energy-dependent manner, and hydrolysing tubulin- bound GTP along the way.25–28 Alongside continuous actin and microtubule restruc- turing, phospholipids and proteins must be synthesised in the brain. Introduction Experimental data for the energy use of either process are scarce, but protein synthesis was esti- mated theoretically to account for no more than 2% of the total consumption of ATP in the brain1,29 and retina.14 While phospholipid synthesis was once thought to be similarly inexpensive,14,30 Purdon and Rapoport31 calculated that up to 25% of total brain ATP may be used on phospholipid metabolism, mainly on mainten- ance of the phosphorylation state of lipids (12%), with another 8% of total ATP spent on transport of phospholipids through the phospholipid bilayer to main- tain lipid asymmetry across the membrane,32 and 5% on recycling and incorporation of short-lived fatty acids inside phospholipids. Pharmacology To establish whether spontaneous signalling had a detect- able eﬀect on O2 consumption, postsynaptic currents were inhibited with 10 mM NBQX þ 50 mM D-AP5, pre- synaptic transmitter release and postsynaptic events were blocked with 250 mM cadmium, or action potentials and evoked synaptic events were inhibited with 1 mM TTX. Actin polymerisation was reversibly blocked with cyto- chalasin D, which binds to the plus end of F-actin and prevents G-actin monomer attachment.34 Cytochalasin does not aﬀect action potential propagation and duration in cardiac muscle cells.35 Cytochalasin D was dissolved in DMSO and made up in aCSF to a ﬁnal concentration of 10mM. When cytochalasin is applied extracellularly, max- imal inhibitory eﬀects are achieved with a concentration of 10mM, and the time needed for maximal eﬃcacy is three to four minutes.34 In order to verify that cytocha- lasin D does indeed inhibit actin treadmilling, we imaged isolectin B4 labelled microglia in hippocampal slices (see Supplementary Methods and Supplementary Figure 1). At rest, microglia constantly survey the brain by moving their processes.36 If actin treadmilling is inhibited, the continuous rearrangement of actin underlying micro- glial process movement should cease, and motility should decline. Here, we present the ﬁrst experimental evidence investigating, with a single approach, the relative con- tributions of the main non-signalling processes to the developing brain’s energy budget. We blocked actin treadmilling, microtubule turnover, or lipid or protein synthesis in developing rat brain slices. By measuring oxygen level changes with an oxygen-sensing microelec- trode as a proxy for energy expenditure, and modelling oxygen use throughout the slice, we could estimate the energy used on each process. Slice preparation Experimental protocols were approved by UCL’s Animal Welfare and Ethical Review Body. Procedures complied with the regulations of the Animals (scientiﬁc procedures) Act 1986 and reporting follows the ARRIVE (animal research: reporting of in vivo experiments) guidelines. Following cervical dislocation (no anaesthesia was required for animals of the age and size used), hippocam- pal slices (300 mm thick) from male and female P10 rats were cut on a Leica VT1200 S vibratome using ice- cold slicing solution containing (mM) 124 NaCl, 10 D-glucose, 26 NaHCO3, 2.5 KCl, 1 NaH2PO4, 2 MgCl2, 2.5 CaCl2 and 1 kynurenic acid (bubbled with 95% O2/5% CO2). The dissection followed the ‘magic cut’ method to maintain synaptic connectivity, using an angle of 10 for the ‘magic cut’ to preserve the CA1 region.33 Slices were used one to ﬁve hours after slicing. HEPES-buﬀered artiﬁcial cerebrospinal ﬂuid (aCSF), to which blockers were added, was bath-applied (at 2.5 ml/min), while submerged slices rested on the glass bottom of a bath on the Zeiss LSM 710 microscope Cytochalasin D, the most potent of the cytochala- sins, was chosen over jasplakinolide, another inhibitor of actin treadmilling, because the latter can promote F-actin polymerization as well as stabilization,37,38 and apoptosis,39 and was preferred over latrunculin because the latter drug distorts cell shape far more than cytocha- lasin D.40 Nevertheless, we also applied jasplakinolide to some slices, as Bernstein and Bamburg23 stated that it reduced ATP use more eﬀectively than cytochalasin D in neuronal cultures. In hippocampal slices, jasplakinolide has been used at concentrations from 0.1–10mM41,37 without changing the electrophysiological properties of neurons.41 We used 1 mM. Microtubule turnover was reversibly inhibited with 25mM nocodazole (dissolved in DMSO), a non-cytotoxic microtubule-depolymerizing agent binding to beta- tubulin.42 This concentration was shown to be eﬀective when bath-applied onto hippocampal slices without alter- ing neuronal electrophysiological characteristics,43 and nocodazole depolymerises microtubules within minutes.44 953 Engl et al. superfusing the slice, were ﬁtted with steady-state solu- tions of the following modiﬁed diﬀusion equation To arrest fatty acid synthesis, we used 60 mM 5-(tet- radecyloxy)-2 furoic acid (TOFA), which inhibits the acetyl-CoA carboxylase required to catalyse the carb- oxylation of acetyl-CoA.45 TOFA was dissolved in DMSO and 0.5% albumin to avoid precipitation. Modelling oxygen consumption through the slice The [O2] depth proﬁles obtained in each condition were used to model oxygen consumption throughout the slice, as in Hall et al.51 The depth proﬁle data points, obtained during continuous ﬂow of the solution Oxygen recordings A Unisense Clark-type oxygen microsensor (OX-10), which generates a current proportional to the oxygen concentration,49 was used to measure the oxygen level at the slice surface (in the CA1 region of the hippocam- pus) and at three diﬀerent depths in the slice before, during and after the application of a blocker of each non-signalling process (see Figure 1 and Supplementary Methods). Recordings were calibrated (Figure 1(d)) using solutions bubbled with known partial pressures of oxygen (converted to mM using Henry’s law and the solubility50 of O2: see Supplementary Methods). Individual depth proﬁles, and depth proﬁles aver- aged over all slices, including the unstirred layer above the slice surface, were ﬁtted with the non-linear least-square curve ﬁtting function lsqcurveﬁt in MATLAB. All ﬁts were evaluated by calculating the proportion of the sum of the squared residuals (diﬀer- ence between data and ﬁt) explained by the ﬁt (R2), and all ﬁts in this paper had R2 > 0.95. Since the electrode moved across the hippocampus when being lowered into the slice along its axis (starting from CA1 and usually not venturing beyond the stratum lacunosum-moleculare), we established that the baseline oxygen level was similar throughout the hippocampus by measuring the surface oxygen level at 28 points across 9 hippocampal regions (Figure 2(a) and (b)). We further determined the energy consumption on spontaneous electrical signalling in the hippocampal slice by blocking distinct signalling related processes (Figure 2(c)). Most analyses involved measuring the [O2] proﬁle through the depth of the slice; however, some experi- ments examined only the [O2] on the slice surface. Although the surface cell layer can be damaged by the slicing process33,56 up to a depth of 10 to 35 mm, this does not invalidate the use of surface [O2] measure- ments to assess the [O2] consumption of the slice, because the surface [O2] does not just reﬂect local meta- bolic activity, but reﬂects O2 use throughout the slice and thus changes when metabolic activity is altered (see plots in Figures 4(a) and (b), 5(a) and (b) and 6(a)). Slice preparation On adipocytes, 50 mM TOFA is eﬃcacious within 15 min.46 Application of 60 mM TOFA for up to one hour does not alter neuronal viability.45 D @2c @x2 ¼ cVmax c þ Km ð1Þ ð1Þ where D ¼ 1.54 109 m2/s is the diﬀusion coeﬃcient of O2 in brain at 37C,52 c is oxygen concentration, x is distance into the slice in mm, Vmax denotes the maximum rate of oxidative phosphorylation at saturating oxygen concentration in mM/min, and Km¼1 mM53 is the EC50 for oxygen activating oxidative phosphorylation. In equation (1), the left-hand side represents diﬀusion and the right-hand side represents the consumption of oxygen by mitochondria. This was solved using the pdepe function in MATLAB (the MathWorks; scripts available on request). At the bottom of the slice, we applied the boundary condition dc/dx ¼ 0. Anisomycin, an inhibitor of mRNA translation, was used at 20 mM (dissolved in DMSO) to inhibit protein synthesis. Anisomycin blocks protein synthesis within minutes at this concentration in hippocampal slices without aﬀecting basal synaptic transmission.47,48 To inhibit the sodium–potassium ATPase, 1 mM ouabain was applied for 10 min in external solutions containing either 2 mM Ca2þ or 2 mM EGTA. To block all oxidative phosphorylation, 25 mM antimycin was applied for 20 min. Above the surface of a slice, there is an unstirred layer of solution with no oxygen consumption, but only oxygen diﬀusion towards the slice from the bulk solution above51,54 with a diﬀusion coeﬃcient for oxygen in water at 37C of 2.68 109 m2/s.55 Consequently, the oxygen concentration in the unstirred layer gradually approaches that of the bulk solution with greater dis- tance from the membrane.54 In slices across all experi- mental conditions, the unstirred layer was measured after the experiment and incorporated into the ﬁts of slice oxygen data in order to more accurately estimate changes in O2 consumption in the slice from changes in the O2 depth proﬁles (see Supplementary Methods). Equal percentages (see Supplementary Methods) of solvents or carrier proteins used were added to all the extracellular solutions in any particular experiment in order to rule out any confounding eﬀects they may have. See Supplementary Methods for drug prepar- ation and purchasing information. Statistics The oxygen sensor was placed at the surface (touching the tissue) of the hippocampal brain slice at start of the experiment. After 10 min of baseline measurement, a depth profile of O2 concentration was obtained by moving the trode along its own axis to generate a vertical depth of 50, 100, and 150 mm (the midpoint of the slice) into the slice (see plementary Methods). After obtaining the [O2] depth profile, the electrode was returned to the surface. Depth profiles measured a greater spatial resolution, or with measurement points beyond the slice midpoint, did not affect the calculation of oxygen sumption through the slice (see Supplementary Figure 3). A specific blocker of an energy-consuming process was bath-perfused o the slice for 7–20 min (see Materials and Methods and Supplementary Methods). A second [O2] depth profile was obtained at the of the drug application, when the oxygen level had reached a plateau. After a 10–15 min recovery period and a third [O2] depth ile, 1 mM glutamate was applied to the slice and a final [O2] depth profile was obtained after oxygen levels stabilized after 3–5 min. ve the slice surface, oxygen diffuses through an unstirred layer but is not consumed. (c) Sample depth profile for [O2] through a ocampal slice (CA1 region). The [O2] at the slice surface is lower than at the top of the static unstirred layer, where, in turn, 0.6 (c) Time (min) Oxygen (mM) slice surface 0.3 0.4 0.5 0 0.5 1 1.5 2 (d) (d) d) Electrode reading (V) 0% O 2 20% O 2 95% O 2 0 200 400 600 800 1000 0 2 4 6 d) ) ) Time (min) (e) Oxygen (mM) 0.6 0 0.4 0.2 0 200 -200 Unstirred layer Slice (f) Calibrated oxygen levels normalised to surface (=1) -200 0 -100 -300 -400 1 1.2 2 2.2 2.4 1.4 1.6 1.8 Figure 1. Measuring oxygen and calculating energy use in a brain slice. (a) A Clark-type oxygen sensor was used to measure oxygen concentration during the experiment (electrode schematic adapted from Unisense:49 for sensor information see Supplementary Methods). (b) Experimental outline. The oxygen sensor was placed at the surface (touching the tissue) of the hippocampal brain slice at the start of the experiment. Statistics Data are shown as mean standard error of the mean (s.e.m.). After conﬁrming that the data were normally Journal of Cerebral Blood Flow & Metabolism 37(3) 954 0.6 (b) (a) Oxygen -0.8 V Gas-permeable membrane Sensing cathode Reduction Current Anode (c) Time (min) Oxygen (mM) slice surface 0.3 0.4 0.5 0 0.5 1 1.5 2 (d) Electrode reading (V) 0% O 2 20% O 2 95% O 2 0 200 400 600 800 1000 0 2 4 6 (f) (e) Calibrated oxygen levels normalised to surface (=1) -200 0 -100 -300 -400 1 1.2 2 2.2 2.4 1.4 1.6 1.8 Oxygen (mM) 0.6 0 0.4 0.2 0 200 -200 Unstirred layer Slice Unstirred layer (d) ) 50 µm 150 µm 100 µm Slice surface Bathing solution Bath Time re 1. Measuring oxygen and calculating energy use in a brain slice. (a) A Clark-type oxygen sensor was used to measure o ntration during the experiment (electrode schematic adapted from Unisense:49 for sensor information see Supplementa ods). (b) Experimental outline. The oxygen sensor was placed at the surface (touching the tissue) of the hippocampal brain s art of the experiment. After 10 min of baseline measurement, a depth profile of O2 concentration was obtained by movi d l it i t t ti l d th f 50 100 d 150 (th id i t f th li ) i t th li ( (b) Unstirred layer 50 µm 150 µm 100 µm Slice surface Bathing solution Bath Time (a) Oxygen -0.8 V Gas-permeable membrane Sensing cathode Reduction Current Anode 0.6 Oxygen -0.8 V Gas-permeable membrane Sensing cathode Reduction Anode (c) Time (min) Oxygen (mM) slice surface 0.3 0.4 0.5 0 0.5 1 1.5 2 (d) Electrode reading (V) 0% O 2 20% O 2 95% O 2 0 200 400 600 800 1000 0 2 4 6 (f) (e) Calibrated oxygen levels normalised to surface (=1) -200 0 -100 -300 -400 1 1.2 2 2.2 2.4 1.4 1.6 1.8 Oxygen (mM) 0.6 0 0.4 0.2 0 200 -200 Unstirred layer Slice (d) ) 50 µm 150 µm 100 µm Slice surface Bath Time ure 1. Measuring oxygen and calculating energy use in a brain slice. (a) A Clark-type oxygen sensor was used to measure oxygen centration during the experiment (electrode schematic adapted from Unisense:49 for sensor information see Supplementary hods). (b) Experimental outline. Statistics After 10 min of baseline measurement, a depth profile of O2 concentration was obtained by moving the electrode along its own axis to generate a vertical depth of 50, 100, and 150 mm (the midpoint of the slice) into the slice (see Supplementary Methods). After obtaining the [O2] depth profile, the electrode was returned to the surface. Depth profiles measured with a greater spatial resolution, or with measurement points beyond the slice midpoint, did not affect the calculation of oxygen consumption through the slice (see Supplementary Figure 3). A specific blocker of an energy-consuming process was bath-perfused onto the slice for 7–20 min (see Materials and Methods and Supplementary Methods). A second [O2] depth profile was obtained at the end of the drug application, when the oxygen level had reached a plateau. After a 10–15 min recovery period and a third [O2] depth profile, 1 mM glutamate was applied to the slice and a final [O2] depth profile was obtained after oxygen levels stabilized after 3–5 min. Above the slice surface, oxygen diffuses through an unstirred layer but is not consumed. (c) Sample depth profile for [O2] through a hippocampal slice (CA1 region). The [O2] at the slice surface is lower than at the top of the static unstirred layer, where, in turn, 955 Engl et al. distributed using the Kolmogorov–Smirnov test, one-way or two-way ANOVAs (repeated measures where appro- priate) or paired or one-sample t-tests were used to compare means, and data were corrected for multiple comparisons with Dunnett’s post hoc test or a modiﬁed Holm–Bonferroni correction (see Supplementary Methods). Degrees of freedom are reported in brackets after the t or F statistic, respectively. [Oxygen] change (µM) (b) (a) (c) 1 2 3 4 granule cells 6 5 25 26 dentate gyrus 7 8 9 CA4 10 11 12 13 14 CA3 pyramidal cells 15 16 CA2 pyr. cells 17 18 CA1 pyramidal cells 22 23 24 27 19 20 21 str. lac.-mol. mossy fibres 28 fimbria [Oxygen] (mM) DGG DG CA4 CA3P CA2P CA1P SL-M MFP F p=0.44 NBQX + D-AP5 TTX Cadmium Baseline 50 0 -100 -50 100 p=0.26 0.3 0.4 0.2 0 0.1 Figure 2. Baseline metabolic activity in a brain slice. (a) Constructing a hippocampal oxygen map of a P10 rat slice. The oxygen sensor was used to measure surface oxygen levels at each point on the schematic. Statistics mossy fibres 28 fimbria [Oxygen] (mM) DGG DG CA4 CA3P CA2P CA1P SL-M MFP F p=0.44 0.3 0.4 0.2 0 0.1 Figure 2. Baseline metabolic activity in a brain slice. (a) Constructing a hippocampal oxygen map of a P10 rat slice. The oxygen sensor was used to measure surface oxygen levels at each point on the schematic. The points were binned into nine distinct areas, colour-coded in (b). During the experiments reported subsequently in the paper, the electrode was placed in the CA1 region and advanced down through the slice into the stratum lacunosum-moleculare. (b) Mean ( s.e.m.) of oxygen concentration binned into the hippocampal regions shown in (a), n ¼ nine slices, N ¼ three animals. The following area measurements were binned: DGG (dentate gyrus granule cells) points 1–6, DG (dentate gyrus) 7–8 and 25–26, CA4 9, CA3P (CA3 pyramidal cells) 10–14, CA2P (CA2 pyramidal cells) 15–16, CA1P (CA1 pyramidal cells) 17–18, SL-M (stratum lacunosum-moleculare) 19–22, MFP (mossy fibre pathway) 23–24, and F (fimbria hippocampus) 27–28. No significant difference in oxygen level between hippocampal regions was found (p ¼ 0.44). (c) Signalling related energy expenditure is negligible in a resting slice. Specific blockers of postsynaptic currents (and thus postsynaptic action potentials, 10 mM NBQX þ 50 mM D-AP5, n ¼ nine slices, N ¼ four animals), presynaptic transmitter release and postsynaptic events (250 mM cadmium, n ¼ 6, N ¼ 2) or action potentials and synaptic events (1 mM TTX, n ¼ 9, N ¼ 9) were applied to different slices. Changes in oxygen level between the start of drug application and 15 min later were measured in those blockers and in a no-drug baseline condition (n ¼ 9, N ¼ 9). No blocker changed oxygen levels relative to control (p¼0.26). ( ) the [O2] is lower than in the reservoir bubbled with 100% O2 (see (d)), as oxygen is lost to the air above the reservoir and through the perfusion tube walls (see Results). (d) Oxygen sensor calibration. Three bottles of distilled water were heated up to 37C and bubbled for at least 15 min with 0%, 20%, and 95% oxygen. The electrode was then inserted into the three solutions consecutively for a few seconds until a stable reading was obtained. Statistics The points were binned into nine distinct areas, colour-coded in (b). During the experiments reported subsequently in the paper, the electrode was placed in the CA1 region and advanced down through the slice into the stratum lacunosum-moleculare. (b) Mean ( s.e.m.) of oxygen concentration binned into the hippocampal regions shown in (a), n ¼ nine slices, N ¼ three animals. The following area measurements were binned: DGG (dentate gyrus granule cells) points 1–6, DG (dentate gyrus) 7–8 and 25–26, CA4 9, CA3P (CA3 pyramidal cells) 10–14, CA2P (CA2 pyramidal cells) 15–16, CA1P (CA1 pyramidal cells) 17–18, SL-M (stratum lacunosum-moleculare) 19–22, MFP (mossy fibre pathway) 23–24, and F (fimbria hippocampus) 27–28. No significant difference in oxygen level between hippocampal regions was found (p ¼ 0.44). (c) Signalling related energy expenditure is negligible in a resting slice. Specific blockers of postsynaptic currents (and thus postsynaptic action potentials, 10 mM NBQX þ 50 mM D-AP5, n ¼ nine slices, N ¼ four animals), presynaptic transmitter release and postsynaptic events (250 mM cadmium, n ¼ 6, N ¼ 2) or action potentials and synaptic events (1 mM TTX, n ¼ 9, N ¼ 9) were applied to different slices. Changes in oxygen level between the start of drug application and 15 min later were measured in those blockers and in a no-drug baseline condition (n ¼ 9, N ¼ 9). No blocker changed oxygen levels relative to control (p¼0.26). [Oxygen] change (µM) (b) (a) (c) 1 2 3 4 granule cells 6 5 25 26 dentate gyrus 7 8 9 CA4 10 11 12 13 14 CA3 pyramidal cells 15 16 CA2 pyr. cells 17 18 CA1 pyramidal cells 22 23 24 27 19 20 21 str. lac.-mol. mossy fibres 28 fimbria [Oxygen] (mM) DGG DG CA4 CA3P CA2P CA1P SL-M MFP F p=0.44 NBQX + D-AP5 TTX Cadmium Baseline 50 0 -100 -50 100 p=0.26 0.3 0.4 0.2 0 0.1 (b) [Oxygen] (mM) DGG DG CA4 CA3P CA2P CA1P SL-M MFP F p=0.44 0.3 0.4 0.2 0 0.1 (b) a) 1 2 3 4 granule cells 6 5 25 26 dentate gyrus 7 8 9 CA4 10 11 12 13 14 CA3 pyramidal cells 15 16 CA2 pyr. cells 17 18 CA1 pyramidal cells 22 23 24 27 19 20 21 str. lac.-mol. Statistics Those readings were plotted against the dissolved oxygen concentrations corresponding to the different percentage saturation values for oxygen in water at 37C, which were obtained from Henry’s law as 0, 208, and 991 mM, respectively.50 Using the slope and intercept from the linear fit through these three points, a linear conversion to mM was then applied to the raw electrode output. (e) Oxygen measurements were taken at the end point of each depth step (blue dots). Measurements were fitted (see Materials and methods and Supplementary methods) with a modified diffusion equation (equation (1)) from the top of the unstirred layer to give Vmax, the maximum rate of oxidative phosphorylation at saturating [O2]. (f) The width of the unstirred layer was determined by moving the oxygen electrode upwards from the slice surface and calculating the break point of the [O2] profile between the unstirred layer and the bulk solution above it (see Supplementary Methods). Measurements were fitted (see Materials and methods and Supplementary methods) with a modified diffusion equation (equation (1)) from the top of the unstirred layer to give Vmax, the maximum rate of oxidative phosphorylation at saturating [O2]. (f) The width of the unstirred layer was determined by moving the oxygen electrode upwards from the slice surface and calculating the break point of the [O2] profile between the unstirred layer and the bulk solution above it (see Supplementary Methods). Journal of Cerebral Blood Flow & Metabolism 37(3) 956 Antimycin 25 µM 0.1 mM oxygen 5 min Nocodazole 25 µM Cytochalasin D 10 µM (a) (b) (c) (d) (e) (f) Glutamate Glutamate Anisomycin 20 µM TOFA + Anisomycin 60 µM + 20 µM Glutamate Glutamate Ouabain 1 mM Glutamate Blocker application Depth profile Figure 3. Sample traces for blockers of energy-consuming processes. Traces show oxygen levels on the slice surface interspersed with [O2] depth profiles (light blue bars). Oxygen levels rise when less oxygen is being consumed and fall when more oxygen is consumed (the grey dotted line is placed at the initial [O2] level at the slice surface for easier comparison). After 10 min of baseline (here only 5 min is shown before drug onset), the specific blocker of a non-signalling process was bath-perfused onto the slice (black bar), followed by 10–15 min of recovery and wash-in of 1 mM glutamate (open bar). Blocker application times are given in the Supplementary Methods. Baseline hippocampal oxygen levels are uniform in hippocampal slices Baseline hippocampal oxygen levels are uniform in hippocampal slices The hippocampus comprises diﬀerent functional areas which might have diﬀerent baseline energy uses. The oxygen electrode moved laterally through the slice when lowered along its axis so, although it mostly remained in the CA1 region, we investigated whether regional diﬀerences in baseline [O2] might aﬀect our results. We tested this by measuring the oxygen level at the slice surface at 28 points across 9 hippocampal regions (Figure 2(a) and (b)). No signiﬁcant diﬀerence Statistics (a) antimycin stops all oxidative phosphorylation: near the end of the trace the variation of [O2] with depth is abolished, (b) cytochalasin D blocks actin treadmilling, (c) nocodazole inhibits microtubule turnover, (d), TOFA arrests lipid synthesis, (e) anisomycin blocks protein synthesis and (f) ouabain inhibits the sodium–potassium ATPase (no calcium in the external solution). (d) (e) (f) Anisomycin 20 µM TOFA + Anisomycin 60 µM + 20 µM Glutamate Glutamate Ouabain 1 mM Glutamate Blocker application Depth profile Antimycin 25 µM 0.1 mM oxygen 5 min Nocodazole 25 µM Cytochalasin D 10 µM (a) (b) (c) Glutamate Glutamate Cytochalasin D (b) Glutamate Glutamate Glutamate Figure 3. Sample traces for blockers of energy-consuming processes. Traces show oxygen levels on the slice surface interspersed with [O2] depth profiles (light blue bars). Oxygen levels rise when less oxygen is being consumed and fall when more oxygen is consumed (the grey dotted line is placed at the initial [O2] level at the slice surface for easier comparison). After 10 min of baseline (here only 5 min is shown before drug onset), the specific blocker of a non-signalling process was bath-perfused onto the slice (black bar), followed by 10–15 min of recovery and wash-in of 1 mM glutamate (open bar). Blocker application times are given in the Supplementary Methods. (a) antimycin stops all oxidative phosphorylation: near the end of the trace the variation of [O2] with depth is abolished, (b) cytochalasin D blocks actin treadmilling, (c) nocodazole inhibits microtubule turnover, (d), TOFA arrests lipid synthesis, (e) anisomycin blocks protein synthesis and (f) ouabain inhibits the sodium–potassium ATPase (no calcium in the external solution). in [O2] across regions was detected (F(8,72)¼0.44, p ¼ 0.89, n ¼ nine slices), suggesting no signiﬁcant dif- ference in O2 use between hippocampal areas. This implies that moving the oxygen electrode across the slice during the process of lowering it into the slice would not have an eﬀect on the measured [O2]. No O2 use associated with spontaneous activity can be detected in brain slices p=0.63 p=9.1x10-9 p=3.9x10-12 Position above and in slice (µm), 0 = slice surface Oxygen (mM) Slice Position above and in slice (µm), 0 = slice surface p=0.29 p=2.9x10-8 p=2.6x10-7 Oxygen (mM) Slice (b) (a) Cytochalasin D Baseline Recovery Glutamate Nocodazole Baseline Recovery Glutamate 0.3 0.4 0.5 0.6 0.2 0 0.1 0.3 0.4 0.5 0.6 0.2 0 0.1 -300 -200 -100 0 200 100 -300 -200 -100 0 200 100 p=0.63 =9.1x10-9 p=3.9x10-12 Position above and in slice (µm), 0 = slice surface p=0.29 p=2.9x10-8 p=2.6x10-7 Oxygen (mM) Slice (b) Nocodazole Baseline Recovery Glutamate 0.3 0.4 0.5 0.6 0.2 0 0.1 -300 -200 -100 0 200 100 p=0.63 p=9.1x10-9 p=3.9x10-12 Position above and in slice (µm), 0 = slice surface Oxygen (mM) Slice (b) (a) Cytochalasin D Baseline Recovery Glutamate 0.3 0.4 0.5 0.6 0.2 0 0.1 -300 -200 -100 0 200 100 Rate of oxygen use (Vmax, mM/min) normalised to Baseline Baseline Recovery p=0.005 p=0.41 p=0.007 (c) Cytochal. D Glutamate 0.5 0 1 1.5 2 2.5 3 Baseline Recovery p=0.002 p=0.09 p=0.005 Rate of oxygen use (Vmax, mM/min) normalised to Baseline (d) Nocodazole Glutamate 0.5 0 1 1.5 2 2.5 3 skeleton treadmilling accounts for about a quarter of resting energy use, and microtubule turnover uses a b i ’ ( ) (b) A i ( M ) d h fil f h di i Figure 4. Actin cytoskeleton treadmilling accounts for about a quarter of resting energy use, and microtubule turnover uses a similar fraction of the brain’s energy. (a), (b) Average oxygen concentration (mM s.e.m.) depth profiles for each condition (black ¼ baseline, red ¼ 10 mM cytochalasin D in (a) and 25 mM nocodazole in (b), green ¼ recovery, blue ¼ 1 mM glutamate), for block of actin ((a), n ¼ eight slices, N ¼ six animals) and microtubule ((b), n ¼ 8, N ¼ 4) turnover. Data were fitted with equation (1) from the surface to the bottom of the slice, and with equation (1) but without the oxygen consumption term across the unstirred layer to the slice surface. The fit gives Vmax, the maximum rate of oxygen use. (c), (d) Averaged Vmax s.e.m. (red dots) and individual Vmax values (black dots, normalised to baseline Vmax (¼1)) for block of actin ((c), n ¼ 8) and microtubule ((d), n ¼ 8) turnover. No O2 use associated with spontaneous activity can be detected in brain slices The average fractional rate of energy consumption relative to baseline was calculated from the average of individual fits as being 75% during actin treadmilling inhibition and 78% during microtubule turnover block. is less in brain slices than in the brain, because long-range connections are disrupted by the slicing. We investigated oxygen use evoked by spontaneous electrical activity by blocking either postsynaptic cur- rents (using 10 mM NBQX þ 50 mM D-AP5, n ¼ nine slices), presynaptic transmitter release and postsynaptic events (using 250 mM cadmium, n ¼ six slices), or action potentials and synaptic events (using 1 mM TTX, n ¼ nine slices). None of these manipulations signiﬁ- cantly changed the oxygen level at the slice surface (F(3,29)¼1.4, p ¼ 0.26, Figure 2(c)), suggesting that ongoing electrical activity in the slice is too weak to be detected from its O2 consumption. The absence of spontaneous signalling activity facilitated selective measurement of the O2 use of non-signalling processes, but also implies that the percentage of O2 use that we measure for non-signalling tasks (below) would be a smaller percentage of total brain O2 consumption in vivo, where synaptic and action potentials consume more energy (see Discussion). is less in brain slices than in the brain, because long-range connections are disrupted by the slicing. We investigated oxygen use evoked by spontaneous electrical activity by blocking either postsynaptic cur- rents (using 10 mM NBQX þ 50 mM D-AP5, n ¼ nine slices), presynaptic transmitter release and postsynaptic events (using 250 mM cadmium, n ¼ six slices), or action potentials and synaptic events (using 1 mM TTX, n ¼ nine slices). None of these manipulations signiﬁ- cantly changed the oxygen level at the slice surface (F(3,29)¼1.4, p ¼ 0.26, Figure 2(c)), suggesting that ongoing electrical activity in the slice is too weak to be detected from its O2 consumption. The absence of spontaneous signalling activity facilitated selective measurement of the O2 use of non-signalling processes, but also implies that the percentage of O2 use that we measure for non-signalling tasks (below) would be a smaller percentage of total brain O2 consumption in vivo, where synaptic and action potentials consume more energy (see Discussion). No O2 use associated with spontaneous activity can be detected in brain slices Most brain energy is used on synaptic and action potentials,1 and this can be detected as changes of [O2] level in response to electrical stimulation in hippo- campal slices.51 However, endogenous synaptic activity 957 Engl et al. p=0.63 p=9.1x10-9 p=3.9x10-12 Position above and in slice (µm), 0 = slice surface Oxygen (mM) Slice Baseline Recovery p=0.002 p=0.09 p=0.005 Position above and in slice (µm), 0 = slice surface p=0.29 p=2.9x10-8 p=2.6x10-7 Oxygen (mM) Slice (b) (a) Rate of oxygen use (Vmax, mM/min) normalised to Baseline Rate of oxygen use (Vmax, mM/min) normalised to Baseline Baseline Recovery p=0.005 p=0.41 p=0.007 (d) (c) Nocodazole Glutamate Cytochal. D Glutamate Cytochalasin D Baseline Recovery Glutamate Nocodazole Baseline Recovery Glutamate 0.3 0.4 0.5 0.6 0.2 0 0.1 0.3 0.4 0.5 0.6 0.2 0 0.1 -300 -200 -100 0 200 100 -300 -200 -100 0 200 100 0.5 0 1 1.5 2 2.5 3 0.5 0 1 1.5 2 2.5 3 Figure 4. Actin cytoskeleton treadmilling accounts for about a quarter of resting energy use, and microtubule turnover uses a similar fraction of the brain’s energy. (a), (b) Average oxygen concentration (mM s.e.m.) depth profiles for each condition (black ¼ baseline, red ¼ 10 mM cytochalasin D in (a) and 25 mM nocodazole in (b), green ¼ recovery, blue ¼ 1 mM glutamate), for block of actin ((a), n ¼ eight slices, N ¼ six animals) and microtubule ((b), n ¼ 8, N ¼ 4) turnover. Data were fitted with equation (1) from the surface to the bottom of the slice, and with equation (1) but without the oxygen consumption term across the unstirred layer to the slice surface. The fit gives Vmax, the maximum rate of oxygen use. (c), (d) Averaged Vmax s.e.m. (red dots) and individual Vmax values (black dots, normalised to baseline Vmax (¼1)) for block of actin ((c), n ¼ 8) and microtubule ((d), n ¼ 8) turnover. The average fractional rate of energy consumption relative to baseline was calculated from the average of individual fits as being 75% during actin treadmilling inhibition and 78% during microtubule turnover block. Unstirred layer parameters In the following experiments, we ﬁt measurements of O2 concentration throughout the slice and the unstirred layer above the slice to obtain a value for the rate of oxygen consumption in the slice. We incorporated the value of [O2] at the top of the unstirred layer (which corresponds to the oxygen level in the bulk solution) in order to more accurately estimate changes in O2 consumption in the slice from changes in the O2 depth proﬁles (see Materials and Methods and Supplementary Methods). The unstirred layer width averaged across the cytochalasin D, nocoda- zole, ouabain, and anisomycin conditions in 17 slices was 235 10mm, and the [O2] in the bulk solution at the top of the unstirred layer was 0.57 0.05mM (less than the 1.04mM in the solution reservoir bubbled with 100% O2 due to O2 loss to the air above the reservoir and through the perfusion tube walls). Neither the width of the unstirred layer nor the [O2] at the top of the unstirred layer diﬀered signiﬁcantly between these conditions Journal of Cerebral Blood Flow & Metabolism 37(3) 958 Position above and in slice (µm), 0 = slice surface Baseline TOFA + anisomycin Recovery Glutamate p=0.04 p=0.54 p=0.02 Baseline Anisomycin Recovery Glutamate (b) (a) (d) (c) p=0.33 p=0.30 p=0.02 Slice p=2.1x10-7 p=0.85 p=0.67 Anisomycin Baseline Recovery Glutamate Oxygen (mM) 0.3 0.4 0.5 0.6 0.2 0 0.1 -300 -200 -100 0 200 100 Rate of oxygen use (Vmax, mM/min) normalised to Baseline 0.5 0 1 1.5 2 2.5 3 Slice Position above and in slice (µm), 0 = slice surface p=4.4x10-16 p=0.15 p=7.1x10-5 TOFA + anisomycin Baseline Recovery Glutamate Oxygen (mM) 0.3 0.4 0.5 0.6 0.2 0 0.1 -200 -100 0 100 200 0.7 0.8 0.9 Rate of oxygen use (Vmax, mM/min) normalised to Baseline 0.5 0 1 1.5 2 2.5 3 Figure 5. Lipid and protein synthesis together account for about 18% of O2 use, but O2 use on protein synthesis alone is too small to be measured. (a), (b) Average oxygen concentration s.e.m. across depth profiles per condition (black ¼ baseline, red ¼ 60 mM TOFA þ 20 mM anisomycin in (a) or 20 mM anisomycin in (b), green ¼ recovery, blue ¼ 1 mM glutamate) for block of lipid and protein synthesis ((a), n ¼ five slices, N ¼ two animals) and protein synthesis alone ((b), n ¼ 4, N ¼ 4). Unstirred layer parameters (c), (d) Average Vmax s.e.m. (red dots) and individual Vmax (black dots, normalised to baseline Vmax (¼1)) for block of lipid and protein synthesis ((a), n ¼ 5) and protein synthesis alone ((b), n ¼ 4). The average energy consumption was 82% of the control value when lipid and protein synthesis were blocked. No change could be detected when protein synthesis alone was inhibited. Position above and in slice (µm), 0 = slice surface (b) Slice p=2.1x10-7 p=0.85 p=0.67 Anisomycin Baseline Recovery Glutamate Oxygen (mM) 0.3 0.4 0.5 0.6 0.2 0 0.1 -300 -200 -100 0 200 100 -16 -5 (a) Slice Position above and in slice (µm), 0 = slice surface p=4.4x1 p=0.15 p=7.1x1 TOFA + anisomycin Baseline Recovery Glutamate Oxygen (mM) 0.3 0.4 0.5 0.6 0.2 0 0.1 -200 -100 0 100 200 0.7 0.8 0.9 (a) (c) Baseline Anisomycin Recovery Glutamate (d) p=0.33 p=0.30 p=0.02 Rate of oxygen use (Vmax, mM/min) normalised to Baseline 0.5 0 1 1.5 2 2.5 3 Baseline TOFA + anisomycin Recovery Glutamate p=0.04 p=0.54 p=0.02 (c) Rate of oxygen use (Vmax, mM/min) normalised to Baseline 0.5 0 1 1.5 2 2.5 3 Figure 5. Lipid and protein synthesis together account for about 18% of O2 use, but O2 use on protein synthesis alone is too small to be measured. (a), (b) Average oxygen concentration s.e.m. across depth profiles per condition (black ¼ baseline, red ¼ 60 mM TOFA þ 20 mM anisomycin in (a) or 20 mM anisomycin in (b), green ¼ recovery, blue ¼ 1 mM glutamate) for block of lipid and protein synthesis ((a), n ¼ five slices, N ¼ two animals) and protein synthesis alone ((b), n ¼ 4, N ¼ 4). (c), (d) Average Vmax s.e.m. (red dots) and individual Vmax (black dots, normalised to baseline Vmax (¼1)) for block of lipid and protein synthesis ((a), n ¼ 5) and protein synthesis alone ((b), n ¼ 4). The average energy consumption was 82% of the control value when lipid and protein synthesis were blocked. No change could be detected when protein synthesis alone was inhibited. (F(3,13)¼2.8, p¼ 0.08 and F(3,13)¼2.2, p¼ 0.13, respect- ively, n¼ 17 slices). (F(3,13)¼2.8, p¼ 0.08 and F(3,13)¼2.2, p¼ 0.13, respect- ively, n¼ 17 slices). Blocking all oxidative phosphorylation raises the [O2] to the bulk solution level For experiments containing albumin as a carrier pro- tein for the lipid synthesis blocker TOFA, the unstirred layer parameters were signiﬁcantly diﬀerent. The unstirred layer width in four slices was 165 21 mm, and the [O2] at the top of the unstirred layer was 0.86 0.45 mM. In the bulk solution outside the unstirred layer (unaﬀected by the slice’s oxygen con- sumption), the [O2] above 17 slices superfused with solution not containing albumin was 0.59 0.04 mM, but was 0.92 0.02 mM when external solutions con- taining albumin were used. Therefore, the raised base- line [O2] in experiments using solutions containing albumin (described below) is an eﬀect of a greater sat- uration of the bulk solution with oxygen, which may come about because the albumin foam formed on the surface of solution bubbled with gas impedes the loss of the bubbled O2 to the air above. In order to conﬁrm that our method of measuring [O2] in a brain slice with an oxygen electrode does indeed reﬂect changes in cellular respiration, we veriﬁed that blocking all oxidative phosphorylation in a brain slice elevated [O2] at the slice surface to bulk solution levels. Application of 25 mM antimycin, a respiratory chain inhibitor, increased the oxygen concentration at the slice surface within one to two minutes of application. After 15 min, the oxygen concentration at the surface of the slice (0.60 0.02 mM) was not signiﬁcantly diﬀerent from that in the bulk solution of the bath (which in these experiments was 0.66 0.01 mM, t(3)¼1.88, p ¼ 0.16, n ¼ slices slices) and was much larger than the surface oxygen level before drug application (0.24 0.06 mM). Consistent with this, the variation of [O2] with depth was abolished within minutes when 959 Engl et al. p=3x10-4 Slice Position above and in slice (µm), 0 = slice surface p=2.7x10-12 Slice Position above and in slice (µm), 0 = slice surface p=0.35 (a) (b) (d) (c) Ouabain (no ext. calcium) Baseline Baseline p=0.23 Ouabain (+ ext. calcium) Ouabain (+ ext. calcium) Baseline Ouabain (no ext. Blocking all oxidative phosphorylation raises the [O2] to the bulk solution level calcium) Baseline Rate of oxygen use (Vmax, mM/min) normalised to Baseline 0.5 0 1 1.5 2 2.5 3 (d) Figure 6. Block of the sodium–potassium ATPase nearly halves oxygen use in the absence of external calcium. (a), (b) Averaged oxygen concentration s.e.m. across depth profiles per condition (black ¼ baseline, red ¼ 1 mM ouabain) for inhibition of the sodium–potassium pump in the absence ((a), n ¼ seven slices, N ¼ two animals) and presence ((b), n ¼ 4, N ¼ 4) of external calcium. Slices did not recover after ouabain application and did not react to glutamate (data not shown). (c), (d) Averaged Vmax s.e.m. (red dots) and individual Vmax values (black dots, normalised to baseline Vmax (¼1)) for inhibition of the sodium–potassium pump in the absence ((c), n ¼ 7) and presence ((d), n ¼ 4) of external calcium. In ouabain, the average energy consumption was 50% of the control value in the absence of external calcium, an effect masked by the presence of calcium (see Results and Discussion). modiﬁed diﬀusion equation (see Materials and methods and Supplementary methods). These rates of oxygen consumption are quantiﬁed in Figures 4 to 6. oxygen consumption throughout the whole slice ceased (Figure 3(a)). Blocking all oxidative phosphorylation raises the [O2] to the bulk solution level calcium) Baseline Oxygen (mM) 0.3 0.4 0.5 0.6 0.2 0 0.1 -300 -200 -100 0 200 100 Oxygen (mM) 0.3 0.4 0.5 0.6 0.2 0 0.1 -300 -200 -100 0 200 100 Rate of oxygen use (Vmax, mM/min) normalised to Baseline 0.5 0 1 1.5 2 2.5 3 Rate of oxygen use (Vmax, mM/min) normalised to Baseline 0.5 0 1 1.5 2 2.5 3 Figure 6. Block of the sodium–potassium ATPase nearly halves oxygen use in the absence of external calcium. (a), (b) Averaged oxygen concentration s.e.m. across depth profiles per condition (black ¼ baseline, red ¼ 1 mM ouabain) for inhibition of the sodium–potassium pump in the absence ((a), n ¼ seven slices, N ¼ two animals) and presence ((b), n ¼ 4, N ¼ 4) of external calcium. Slices did not recover after ouabain application and did not react to glutamate (data not shown). (c), (d) Averaged Vmax s.e.m. (red dots) and individual Vmax values (black dots, normalised to baseline Vmax (¼1)) for inhibition of the sodium–potassium pump in the absence ((c), n ¼ 7) and presence ((d), n ¼ 4) of external calcium. In ouabain, the average energy consumption was 50% of the control value in the absence of external calcium, an effect masked by the presence of calcium (see Results and Discussion). Slice Position above and in slice (µm), 0 = slice surface p=2.7x10-12 Slice Position above and in slice (µm), 0 = slice surface p=0.35 (a) (b) Ouabain (+ ext. calcium) Baseline Ouabain (no ext. calcium) Baseline Oxygen (mM) 0.3 0.4 0.5 0.6 0.2 0 0.1 -300 -200 -100 0 200 100 Oxygen (mM) 0.3 0.4 0.5 0.6 0.2 0 0.1 -300 -200 -100 0 200 100 Position above and in slice (µm), 0 = slice surface p=2.7x10-12 Slice (a) (b) Ouabain (no ext. calcium) Baseline Oxygen (mM) 0.3 0.4 0.5 0.6 0.2 0 0.1 -300 -200 -100 0 200 100 Oxygen (mM) Slice -12 Position above and in slice (µm), 0 = slice surface p=0.35 b) Ouabain (+ ext. calcium) Baseline Oxygen (mM) 0.3 0.4 0.5 0.6 0.2 0 0.1 -300 -200 -100 0 200 100 (c) (d) Baseline p=0.23 Ouabain (+ ext. calcium) Rate of oxygen use (Vmax, mM/min) normalised to Baseline 0.5 0 1 1.5 2 2.5 3 p=3x10-4 (c) Ouabain (no ext. Blocking actin or microtubule cycling reduces oxygen consumption Microtubule turnover also con- tributes to resting brain slice oxygen use, as applying nocodazole (Figure 3(c)) raised the oxygen level at the slice surface from 0.28 0.02 mM to 0.36 0.03 mM. This recovered to 0.30 0.03 mM on removing the nocodazole (Figure 4(b)), while glutamate lowered the surface oxygen level to 0.22 0.02 mM (n ¼ 8). A two- way repeated measures ANOVA again showed a main eﬀect of condition on oxygen level across all depths in the slice (F(3,21)¼37.8, p ¼ 1.2 108, n ¼ eight slices). Oxygen levels across depth in nocodazole diﬀered sig- niﬁcantly from baseline (Figure 4(b), t(21)¼8.03, p ¼ 2.9 108) and the surface oxygen level after recov- ery from nocodazole application was indistinguishable ( ( ) p ) We then modelled the rate of oxygen consumption through the slice. After the [O2] depth proﬁles were ﬁt (Figure 4(a)) with the modiﬁed diﬀusion equation (equation (1)) incorporating the unstirred layer, the resulting values of Vmax, the maximum rate of oxidative phosphorylation at saturating [O2], were normalised to the initial control value for each slice. On average, the Vmax in cytochalasin D was 0.75 0.05 of that in the initial control condition (Figure 4(c), t(7)¼4.8, p ¼ 0.005, n ¼ eight slices). On removing cytochalasin D, Vmax recovered to 1.03 0.04 of the initial control value (t(7)¼0.88, p ¼ 0.41). In glutamate, the average Vmax rose to 1.70 0.1 of the initial control value (Figure 4(c), t(7)¼4.29, p ¼ 0.007, n ¼ eight slices). These results suggest that a signiﬁcant fraction of the slice’s resting energy budget, 25%, is spent on actin cycling. Two-photon imaging of ongoing microglial motility before and after the application of cytochalasin D conﬁrmed that the drug rapidly and eﬀectively inhib- ited microglial movement and therefore actin cycling (Supplementary Figure 1(c), n ¼ 15 cells from ﬁve slices). Since excitatory synaptic currents are thought to con- sume most ATP in the brain,9 we were concerned that our estimates of the energy use on cytoskeletal recycling could be confounded if the drugs used aﬀected the frequency of excitatory synaptic currents (see Supplementary Figure 1(a) and (b)). We therefore whole-cell patch-clamped area CA1 pyramidal cells and monitored spontaneous EPSCs while applying either cytochalasin D or nocodazole (see Supplementary Methods). The energy used on lipid and protein synthesis We next investigated the contributions of phospholipid and protein synthesis to the slice’s energy expenditure. Previous estimates of the energy consumption of protein synthesis are unanimously very low, at 2% of the total energy budget,1,14,29 but vary between 2% and 25% for lipid metabolism.14,30,31,57 We ﬁrst applied blockers of both phospholipid synthesis (60mm TOFA) and protein synthesis (20 mM anisomycin) together. In the experi- ments containing TOFA, serum albumin had to be pre- sent in the external solutions to prevent precipitation of the drug. Albumin increased the concentration of oxygen in the solutions (see above), which altered oxygen levels at the slice surface (F(2,16)¼12.74, p ¼ 0.006, n ¼ four slices). Application of 0.5% albumin (together with 0.55% DMSO) reversibly raised the oxygen level at the surface of the resting slice from We next investigated the contributions of phospholipid and protein synthesis to the slice’s energy expenditure. Previous estimates of the energy consumption of protein synthesis are unanimously very low, at 2% of the total energy budget,1,14,29 but vary between 2% and 25% for lipid metabolism.14,30,31,57 We ﬁrst applied blockers of In order to inhibit microtubule turnover, 25 mM nocodazole, a microtubule-depolymerising agent was perfused onto the slice. Microtubule turnover also con- tributes to resting brain slice oxygen use, as applying nocodazole (Figure 3(c)) raised the oxygen level at the slice surface from 0.28 0.02 mM to 0.36 0.03 mM. This recovered to 0.30 0.03 mM on removing the nocodazole (Figure 4(b)), while glutamate lowered the surface oxygen level to 0.22 0.02 mM (n ¼ 8). A two- way repeated measures ANOVA again showed a main eﬀect of condition on oxygen level across all depths in the slice (F(3,21)¼37.8, p ¼ 1.2 108, n ¼ eight slices). Oxygen levels across depth in nocodazole diﬀered sig- niﬁcantly from baseline (Figure 4(b), t(21)¼8.03, p ¼ 2.9 108) and the surface oxygen level after recov- ery from nocodazole application was indistinguishable both phospholipid synthesis (60mm TOFA) and protein synthesis (20 mM anisomycin) together. In the experi- ments containing TOFA, serum albumin had to be pre- sent in the external solutions to prevent precipitation of the drug. Albumin increased the concentration of oxygen in the solutions (see above), which altered oxygen levels at the slice surface (F(2,16)¼12.74, p ¼ 0.006, n ¼ four slices). Blocking actin or microtubule cycling reduces oxygen consumption Neither drug signiﬁcantly altered the number of EPSCs occurring during the last 6min in each condition (F(2,14)¼1.69, p¼ 0.23 for cyto- chalasin D and F(2,8)¼0.56, p¼ 0.59 for nocodazole, both n¼ eight slices). Blocking actin or microtubule cycling reduces oxygen consumption Similarly, glutam- ate application signiﬁcantly lowered [O2] relative to the control condition (t(21) ¼ 7.28, p ¼ 2.6 107, n ¼ eight slices). As in the previous experiment, these data were used to model oxygen consumption through the slice. When microtubule turnover was blocked with nocodazole (Figure 4(d)), the Vmax derived from the [O2] depth proﬁles was reduced to 0.78 0.04 (t(7) ¼ 5.6, p ¼ 0.002, n ¼ eight slices) of the control value, and recov- ered to 0.96 0.03 of the control value on removing nocodazole (t(7) ¼ 1.9, p ¼ 0.09). In glutamate, Vmax increased to 1.48 0.10 of the baseline value (t(7)¼4.6, p ¼ 0.005, n ¼ eight slices). Surprisingly, therefore, microtubule turnover also accounts for a substantial 22% of the slice’s baseline oxygen use. After the slices recovered from either actin or microtubule turnover inhibition, activating glutamate receptors throughout the slice by superfusing glutamate increased the energy use by 50–70%. The magnitude of the Vmax change after glutamate application was not signiﬁcantly diﬀerent in the cytochalasin D versus the nocodazole condition (t(14) ¼ 0.4, p ¼ 0.67). 0.28 0.02mM (t(6)¼5.56, p ¼ 0.004) in three slices. We then modelled the rate of oxygen consumption through the slice. After the [O2] depth proﬁles were ﬁt (Figure 4(a)) with the modiﬁed diﬀusion equation (equation (1)) incorporating the unstirred layer, the resulting values of Vmax, the maximum rate of oxidative phosphorylation at saturating [O2], were normalised to the initial control value for each slice. On average, the Vmax in cytochalasin D was 0.75 0.05 of that in the initial control condition (Figure 4(c), t(7)¼4.8, p ¼ 0.005, n ¼ eight slices). On removing cytochalasin D, Vmax recovered to 1.03 0.04 of the initial control value (t(7)¼0.88, p ¼ 0.41). In glutamate, the average Vmax rose to 1.70 0.1 of the initial control value (Figure 4(c), t(7)¼4.29, p ¼ 0.007, n ¼ eight slices). These results suggest that a signiﬁcant fraction of the slice’s resting energy budget, 25%, is spent on actin cycling. Two-photon imaging of ongoing microglial motility before and after the application of cytochalasin D conﬁrmed that the drug rapidly and eﬀectively inhib- ited microglial movement and therefore actin cycling (Supplementary Figure 1(c), n ¼ 15 cells from ﬁve slices). In order to inhibit microtubule turnover, 25 mM nocodazole, a microtubule-depolymerising agent was perfused onto the slice. Blocking actin or microtubule cycling reduces oxygen consumption Figure 3(b) to (e) shows sample traces for oxygen level changes at the slice surface for each non-signalling pro- cess that was blocked and the recovery of the oxygen level after the blocker was removed, while Figure 3(e) shows oxygen level changes after block of the sodium pump. At the end of each experiment, to check the health of the slice, 1 mM glutamate was applied to acti- vate a cation inﬂux through ionotropic receptors and thus increase oxygen consumption. For all drugs, oxygen level was measured as a function of depth in the slice, and these depth proﬁles were used to calculate the rate of oxygen consumption through the slice, by solving a To investigate the eﬀect on energy consumption of blocking actin treadmilling, 10mM cytochalasin D was applied to the slice to arrest actin polymerisation. At the end of the cytochalasin D application (Figure 3(b)), the oxygen level at the slice surface was elevated from 0.33 0.02mM to 0.43 0.02mM (n ¼ 8), suggesting that some slice oxygen consumption is due to actin treadmilling. On removing cytochalasin D, the oxygen level recovered to 0.32 0.02mM (Figure 4(a)). Subsequently, applying glutamate lowered the surface oxygen level to 0.23 0.02mM, reﬂecting oxygen Journal of Cerebral Blood Flow & Metabolism 37(3) 960 7(3) from baseline (t(21)¼1.59, p ¼ 0.29). Similarly, glutam- ate application signiﬁcantly lowered [O2] relative to the control condition (t(21) ¼ 7.28, p ¼ 2.6 107, n ¼ eight slices). consumption to fuel synaptic depolarisation. A two-way repeated measures ANOVA for treatment (baseline, cytochalasin D, recovery, glutamate) and position in the slice (surface, 50mm, 100 mm, 150 mm) showed a main eﬀect of condition on oxygen level across all depths used in the depth proﬁle (F(3,21)¼33.9, p ¼ 2.1 108, n ¼ eight slices). With Dunnett’s post-hoc test, the oxygen levels in cytochalasin diﬀered signiﬁ- cantly from control (t(21)¼8.42, p ¼ 9.1 109), and there was no diﬀerence in the oxygen levels between the baseline and recovery conditions (t(21)¼1.01, p ¼ 0.63). Glutamate application also signiﬁcantly chan- ged the oxygen levels relative to baseline (t(21)¼10.81, p ¼ 3.9 1012, n ¼ eight slices). Similarly, when 1 mM jasplakinolide, another actin cycling blocker that can, however, also promote polymerization,38 was applied, the surface oxygen level rose from 0.23 0.03mM to 0.28 0.02mM (t(6)¼5.56, p ¼ 0.004) in three slices. from baseline (t(21)¼1.59, p ¼ 0.29). Without external calcium, the Naþ/Kþ pump accounts for 50% of energy use Most brain ATP use is thought to be on the pumping out of ions that enter neurons to generate synaptic or action potentials, or that enter at the resting potential.9 This pumping is mainly carried out by the sodium pump. To examine the fraction of energy expended on sodium pumping in brain slices, we applied the pump blocker ouabain (1 mM). These experiments were performed both in the presence and absence of external calcium, because blocking the pump is expected to lead indirectly to a rise of [Ca2þ]i which can increase energy consumption (see Discussion). For example, while Shibuki58 found that ouabain decreased oxygen consumption in unstimulated neuro- hypophysis slices whether or not calcium was present, Rusˇ cˇ a´ k and Whittam59 found that blocking the Naþ/ Kþ pump only decreased O2 consumption in cortical slices when using Ca2þ-free external solutions, and instead found an increase in O2 consumption when the external solution contained Ca2þ. Because of these varying results, we measured the eﬀect of blocking the sodium pump on the [O2] proﬁle both in the pres- ence and absence of external calcium. ( ) p ) After lipid and protein synthesis were blocked using 60 mM TOFA and 20 mM anisomycin (Figure 5(a)), the oxygen level at the slice surface rose from 0.63 0.04 mM to 0.68 0.04 mM (t(12)¼6.2, p ¼ 7.0 105). On removal of the drugs, the [O2] recovered to 0.64 0.03 mM (t(12)¼2.06, p ¼ 0.14), and on applying glutamate it fell to 0.49 0.05 mM (t(12)¼17.01, p ¼ 4.4 1016). An overall two-way repeated measures ANOVA showed a main eﬀect of condition on oxygen levels across depths (F(3,12)¼79.1, p ¼ 3.75 108, n ¼ ﬁve slices). ) We then blocked protein synthesis alone with 20mM anisomycin (Figure 5(b)) to isolate its contribution to the slice’s oxygen consumption. There was no signiﬁcant change in oxygen level at the slice surface between base- line (0.32 0.05mM), anisomycin (0.32 0.05mM, t(9) ¼0.66, p ¼ 0.85) and recovery (0.33 0.04mM, t(9)¼ 0.96, p ¼ 0.67). However, [O2] again fell signiﬁcantly in glutamate to 0.23 0.04mM (t(9)¼10.98, p ¼ 2.14 107). The energy used on lipid and protein synthesis Application of 0.5% albumin (together with 0.55% DMSO) reversibly raised the oxygen level at the surface of the resting slice from 961 Engl et al. 0.34 0.03mM to 0.56 0.06mM (t(6)¼10.1, p ¼ 0.8 105). The surface oxygen level recovered to 0.34 0.05mM after albumin was washed oﬀ(t(6)¼41, p ¼ 0.88, n ¼ four slices). As albumin was added to all solutions in the lipid synthesis block experiments, the baseline oxygen level was uniformly raised throughout these experiments. To test whether, by altering the oxygen level in the superperfused solution, albumin changed the oxygen consumption of the slice in control conditions, we assessed the total oxygen consumption as being proportional to the amount of O2 diﬀusing through the unstirred layer towards the slice. This ﬂux is proportional to the diﬀerence between the oxygen level at the top of the unstirred layer (a mean value for which was averaged over all slices) and at the slice surface (measured for each individual slice), divided by the width of the unstirred layer (averaged over all slices). This parameter did not diﬀer between experiments with- out albumin in the external solution (1.22 0.07, n ¼ 31 slices) and those using external albumin (1.36 0.26, n ¼ ﬁve slices, t(34) ¼ 0.7, p ¼ 0.49). (t(3)¼6.39, p ¼ 0.02). From the data above, it can there- fore be estimated that lipid and protein synthesis together require 18% of the brain’s resting O2 use. Lipid synthesis likely account for most of this ﬁgure, as protein synthesis uses too little of the slice’s resting energy to be detected with the oxygen electrode. Discussion Here, we have presented the ﬁrst experimental data obtained with a single method investigating the relative contributions of the main non-signalling processes to the brain’s energy budget. This was done by inhibiting these processes in young rat hippocampal slices and measuring changes in oxygen consumption with a Clark-type oxygen sensor. Surprisingly, the actin and microtubule cytoskeletons contribute almost equally to a large (47%) fraction of the brain slice’s resting energy budget (Figure 4). The O2 use on actin turnover (25%) is considerably lower than previous sugges- tions23 that half of the energy use was spent on actin turnover alone, but is much higher than our previous modelling suggested.24 Second, there is a limit to the resolution of oxygen measurements using Clark-type oxygen sensors. Baseline recordings of brain oxygen levels show ﬂuctu- ations of 25 mM oxygen over 15 min (Figure 2(c), baseline). The noise that this introduces, together with block of the sodium–potassium pump being performed in the absence of external calcium, could account for the sum of the contributions of all processes exceed- ing 100%. Lipid synthesis accounts for most of the remaining non-signalling energy use (18% when blocked together with protein synthesis), but blocking protein synthesis did not detectably change energy use (Figure 5). These results are consistent with previous estimates that protein synthesis accounts for only 1.3% of the brain’s total energy use.1,29 Purdon and Rapoport31,57 calculated that up to 25% of the brain’s energy use might be spent on phospholipid metabolism, although only 5% of the total energy budget was estimated to be spent on the turnover of fatty acids within phospholipids, with the rest allo- cated to maintaining the phosphorylation state of phospholipids (12%) and maintaining asymmetries in the phospholipid bilayer (7.7%). g How can our estimates of non-signalling energy use in brain slices be extrapolated to the awake brain? Oxidative phosphorylation is lower in unstimulated brain slices than in vivo, since there is little spontaneous neuronal activity in slices (Figure 2(c)). Without external calcium, the Naþ/Kþ pump accounts for 50% of energy use Without external calcium, blocking the sodium pump therefore approximately halved the resting slice’s energy use, whereas no signiﬁ- cant change in oxygen consumption could be detected when ouabain was applied in the presence of external calcium. baseline (0.85 0.10 of the control value, t(3) ¼ 1.5, p ¼ 0.23, n ¼ four slices). Without external calcium, blocking the sodium pump therefore approximately halved the resting slice’s energy use, whereas no signiﬁ- cant change in oxygen consumption could be detected when ouabain was applied in the presence of external calcium. The contributions to O2 consumption of the sodium pump (50%), cytoskeletal turnover (47%), and protein and lipid synthesis (18%) that we have estimated sum to slightly more than 100%. However, the sodium pump contribution was calculated in the absence of calcium, and furthermore the experimental results pre- sented here need to be considered with several caveats. First, we are assuming that oxygen consumption is proportional to energy use, because under normal con- ditions glucose is essentially completely oxidised in the brain by the sequential operation of glycolysis and oxi- dative phosphorylation.4,5 However, we cannot exclude the possibility that, in our brain slices, glycolysis accounts for a larger proportion of ATP generation than in the normal brain. For example, microglia, the brain’s immune cells, switch their energy production mechanism from oxidative phosphorylation to glycoly- sis in hypoxia62, and the relative rates of these processes in astrocytes and neurons63 could diﬀer in slices and in vivo. However, in these experiments, the slices were superfused with 100% oxygen and thus were not lim- ited by oxygen availability. Without external calcium, the Naþ/Kþ pump accounts for 50% of energy use After ouabain application in the absence of external calcium, the oxygen level at the slice surface rose from 0.26 0.05 mM to 0.38 0.05 mM (Figure 6(a); recov- ery and eﬀect of glutamate are not shown as the eﬀect of ouabain was irreversible). The main eﬀect of ouabain on oxygen levels was signiﬁcant at all depths in the slice (F(3,18)¼17.86, p ¼ 1.2 105, n ¼ seven slices). Oxygen levels in the presence of ouabain diﬀered signiﬁcantly from those in control conditions (t(18)¼11.55, p ¼ 2.7 1012). In contrast, with exter- nal calcium present, the oxygen level at the slice surface rose from 0.17 0.02 mM to 0.20 0.01 mM (Figure 6(b)) when ouabain was applied, which was not signiﬁcant (F(3,6)¼1.32, p¼ 0.35, n¼ four slices). This is presumably because the pump block raises [Ca2þ]i which increases energy consumption (see Discussion). When modelling the rate of oxygen consumption through the slice, the Vmax for O2 usage after inhibition of both lipid and protein synthesis (Figure 5(c)) fell to 0.82 0.05 of the control value (t(4)¼4.1, p ¼ 0.04), recovered to 0.96 0.07 of the control value on remov- ing the blockers (t(4) ¼ 0.66, p ¼ 0.54), and increased to 1.63 0.15 of the control value (t(4)¼4.61, p ¼ 0.02) in glutamate. For protein synthesis block alone (Figure 5(d)), the Vmax for O2 usage in anisomycin rela- tive to baseline was 1.03 0.03, not signiﬁcantly diﬀer- ent from the control value (t(3)¼1.15, p ¼ 0.33). After removing the drug, the recovery value was 0.97 0.02 of the control value (t(3) ¼ 1.9, p ¼ 0.30), while Vmax in glutamate rose to 1.54 0.08 of the control value When modelling the rate of oxygen consumption through the slice, the average Vmax for oxygen usage in ouabain without external calcium (Figure 6(c)) was 0.50 0.07 of the control value without calcium (t(6) ¼ 7.4, p ¼ 3104, n ¼ seven slices). With exter- nal calcium present (Figure 6(d)), the Vmax after oua- bain application was not signiﬁcantly diﬀerent from 962 Journal of Cerebral Blood Flow & Metabolism 37(3) baseline (0.85 0.10 of the control value, t(3) ¼ 1.5, p ¼ 0.23, n ¼ four slices). Discussion The resting oxygen consumption in our experiments (in P10 rats) was 0.47 0.03 mM/min at 37C, while 0.7 mM/min was found previously in coronal brain slices of P21 rats at 35C51, 0.8 mM/min in cultured cerebellar slices of P8 rats, and 1.3 mM/min in acute cerebellar slices of P10 rats, both at 37C.64 However, oxygen consumption is higher in vivo: in anaesthetized rats, the CMRO2 at 37C is approximately 1.8 mM/min,65 while Sokoloﬀet al.66 estimated glucose consump- tion in the grey matter of conscious adult rats to be 1 mM/min, implying an oxygen consumption of 6 mM/min. In awake humans, the whole-brain averaged CMRO2 is around 1.3 mM/min.67 The much lower oxygen consumption in our data (0.47 0.03mM/min under baseline conditions) presumably reﬂects the young age of the animal and the presence of little neur- onal activity in the slice (Figure 2(c)) which is a major contributor to oxygen use.51,63,68–71 Interestingly, unlike our data in Figure 2(c), Huchzermeyer et al.68 found a p p p y ( ) As in previous experiments in vivo12, inhibiting the sodium–potassium pump approximately halved the total energy expenditure in our brain slices. However, as in cortical slices,58,59 this pronounced reduction in oxygen consumption was seen only when external calcium was absent, and the decrease in oxygen consumption was inhibited when calcium was present in the external solu- tion (Figure 6). It is known that, with extracellular cal- cium present, blocking the sodium pump with ouabain induces a rise of intracellular [Ca2þ]i, due to Naþ gradi- ent rundown and subsequent reversal of the Naþ/Ca2þ exchanger60. This may, in turn, increase the activity of oxygen-consuming processes (such as Ca2þ-ATPase activity to remove the extra Ca2þ) as well as increasing oxidative phosphorylation,61 which could occlude the decrease in oxygen consumption produced by sodium– potassium pump inhibition. 963 Engl et al. rise in oxygen level in organotypic slices following the application of TTX, implying that suﬃcient spontaneous activity was occurring to generate detectable O2 use. Whether this reﬂects the use of organotypic slices or of an interface chamber69 by Huchzermeyer et al.,68 as opposed to the submerged acute slices used in this study, is unclear. In summary, contributions of non-signalling pro- cesses to the brain’s energy budget can be signiﬁcant, but are probably dynamic and likely to change during development or with changes in the brain’s functional state. Discussion Pathology will also alter non-signalling processes and their energetic consumption. Alzheimer’s disease (AD), in which energy supply is impaired, aﬀects actin and microtubule turnover.79,80 Coﬁlin, an actin-binding protein, may link cytoskeletal aberrations to mitochon- drial impairments characteristic not only of AD, but also of related pathologies such as Parkinson’s and Huntington’s disease.81 Understanding how non- signalling processes contribute to the brain’s energy budget is therefore important. As a ﬁrst step, we have found that both actin and microtubule cytoskeleton turnover are surprisingly signiﬁcant energy drains in the healthy developing brain, with lipid synthesis close behind. In contrast, protein synthesis is energetically inexpensive. rise in oxygen level in organotypic slices following the application of TTX, implying that suﬃcient spontaneous activity was occurring to generate detectable O2 use. Whether this reﬂects the use of organotypic slices or of an interface chamber69 by Huchzermeyer et al.,68 as opposed to the submerged acute slices used in this study, is unclear. If the non-signalling energy use were the same inde- pendent of age, brain location or level of neuronal activ- ity, then the percentage contributions of non-signalling energy processes to the total energy budget would be signiﬁcantly smaller in vivo than we calculate above. However, non-signalling processes are not completely uncoupled from signalling activity: for instance, actin turnover is regulated by neuronal activity, with faster treadmilling after block of neuronal activity and reduced motility due to actin stabilisation after NMDA applica- tion.19,72 This would imply that the energetic cost of actin cycling in brain slices, with little neuronal activity, is higher than in the intact brain. It is also likely that non-signalling processes in more mature animals have a diﬀerent energy demand from those at the early developmental stage used here (P10) when cell processes are still being extended. Funding The author(s) disclosed receipt of the following ﬁnancial sup- port for the research, authorship, and/or publication of this article: This work was supported by the Wellcome Trust (PhD studentship 097265/Z/11/A to EE and Senior Investigator Award 099222/Z/12/Z to DA), the ERC (Advanced Investigator Award 249670 BRAINPOWER to DA), and a Marie Curie Fellowship to RJ. Discussion Energy metabolism as a whole changes: at P10, rodents may lack full capabilities for oxidative phosphorylation, which might explain the lower baseline Vmax in our slices compared to those from older animals,50 and neuronal function is more resistant to oxygen depriv- ation than in mature rats.73 In the case of the actin cytoskeleton, developmental changes occur: motility in highly dynamic dendritic spines decreases signiﬁ- cantly between P10 (a peak time for synaptogenesis) and P20 in cortical mouse slices.74 Total actin levels in the brain also peak at an early stage of development and later decline.75 This would again imply that actin recycling uses less energy in vivo than in our brain slices. The pattern of microtubule spatial organisation also shifts during early development: uniformly high- polymerisation rates throughout neuronal axons and dendrites in the ﬁrst days of development give way to more stable proximal regions in older cultured neurons, while more dynamic distal regions maintain high poly- merisation rates.76 Furthermore, de novo phospholipid and protein synthesis both decline with age.77 Non-sig- nalling processes are thus developmentally regulated, and it would be valuable to repeat the experiments pre- sented here at diﬀerent stages of development. Here, we have presented data at P10, an age of rapid synapse restructuring and synaptogenesis.78 It would be inter- esting to investigate whether the energetic cost of actin cycling decreases after P20, when synapses have sta- bilised and less cytoskeletal rearrangement is needed. In addition, it would be interesting to test for regional diﬀerences in the non-signalling energy budget. Supplementary material Supplementary material for this paper can be found at http:// jcbfm.sagepub.com/content/by/supplemental-data Authors’ contributions EE, DA, RJ, and CH designed research, EE and CH per- formed experiments, EE, RJ, and CH analysed data, and EE, DA, CH, and RJ wrote the manuscript. Declaration of conflicting interests The author(s) declared no potential conﬂicts of interest with respect to the research, authorship, and/or publication of this article. Acknowledgements We thank Dr Julia Harris and Oliver Gauld for valuable help with experiments and their interpretation. References 1. Attwell D and Laughlin SB. An energy budget for signal- ing in the grey matter of the brain. J Cereb Blood Flow Metab 2001; 21: 1133–1145. 2. Laughlin SB and Sejnowski T. Communication in neur- onal networks. Science 2003; 301: 1870–1874. 2. Laughlin SB and Sejnowski T. Communication in neur- onal networks. Science 2003; 301: 1870–1874. Journal of Cerebral Blood Flow & Metabolism 37(3) 964 7(3) 3. Attwell D and Gibb A. Neuroenergetics and the kinetic design of excitatory synapses. Nat Rev Neurosci 2005; 6: 841–849. 25. Margolis RL. Role of GTP hydrolysis in microtubule treadmilling and assembly. Proc Natl Acad Sci 1981; 78: 1586–1590. 25. Margolis RL. Role of GTP hydrolysis in microtubule treadmilling and assembly. Proc Natl Acad Sci 1981; 78: 1586–1590. 4. Kety S. The general metabolism of the brain in vivo. In: Richter D (ed.) Metabolism of the nervous system. London: Pergamon, 1957, pp.221–237. 26. Mitchison T and Kirschner M. Dynamic instability of microtubule growth. Nature 1984; 312: 237–242. 27. Vorobjev IA, Rodionov VI, Maly IV, et al. Contribution of plus and minus end pathways to microtubule turnover. J Cell Sci 1999; 112: 2277–2289. 5. Sokoloff L. The metabolism of the central nervous sys- tem in vivo. In: Field J, Magoun HW and Hall VE (eds) Handbook of physiology-neurophysiology. Washington: American Physiological Society, 1960, pp.1843–1864. 28. Kline-Smith SL and Walczak CE. Mitotic spindle assem- bly and chromosome segregation. Mol Cell 2004; 15: 317–327. 6. Lennie P. The cost of cortical computation. Curr Biol 2003; 13: 493–497. 29. Rolfe DFS and Brown GC. Cellular energy utilization and molecular origin of standard metabolic rate in mam- mals. Physiol Rev 1997; 77: 731–758. 7. Niven JE and Laughlin SB. Energy limitation as a select- ive pressure on the evolution of sensory systems. J Exp Biol 2008; 211: 1792–1804. 30. McKenna M, Dienel G, Sonnewald U, et al. Energy metabolism of the brain. In: Siegel GJ, Agranoff BW, Albers RW and Molinoff MD (eds) Basic neurochemistry: Molecular, cellular and medical aspects. 8th edn. New York: Raven Press, 2012, p.202. 8. Jolivet R, Magistretti PJ and Weber B. Deciphering neuron-glia compartmentalization in cortical energy metabolism. Front Neuroenergetics 2009; 1: 4. 9. Harris JJ, Jolivet R and Attwell D. Synaptic energy use and supply. Neuron 2012; 75: 762–777. 31. Purdon A and Rapoport S. Energy consumption by phospholipid metabolism in mammalian brain. References In: Lajtha A, Gibson GE and Dienel GA (eds) Handbook of neurochemistry and molecular neurobiology: Brain ener- getics. Integration of molecular and cellular processes. New York: Springer, 2007, pp.402–427. 10. Rangaraju V, Calloway N and Ryan TA. Activity-driven local ATP synthesis is required for synaptic function. Cell 2014; 156: 825–835. 11. Harris JJ and Attwell D. The energetics of CNS white matter. J Neurosci 2012; 32: 356–371. 32. Beleznay Z, Zachowski A, Devaux PF, et al. ATP- dependent aminophospholipid translocation in erythro- cyte vesicles: stoichiometry of transport. Biochemistry (Mosc) 1993; 32: 3146–3152. 12. Astrup J, Sorensen PM and Sorensen HR. Oxygen and glucose consumption related to NaþK þ transport in canine brain. Stroke 1981; 12: 726–730. 13. Sibson NR, Dhankhar A, Mason GF, et al. Stoichiometric coupling of brain glucose metabolism and glutamatergic neuronal activity. Proc Natl Acad Sci 1998; 95: 316–321. 33. Bischofberger J, Engel D, Li L, et al. Patch-clamp rec- ording from mossy fiber terminals in hippocampal slices. Nat Protoc 2006; 1: 2075–2081. 34. Casella JF, Flanagan MD and Lin S. Cytochalasin D inhibits actin polymerization and induces depolymeriza- tion of actin filaments formed during platelet shape change. Nature 1981; 293: 302–305. 14. Ames A. CNS energy metabolism as related to function. Brain Res Rev 2000; 34: 42–68. 15. Condeelis J. Life at the leading edge: the formation of cell protrusions. Annu Rev Cell Biol 1993; 9: 411–444. g 35. Wu J, Biermann M, Rubart M, et al. Cytochalasin D as excitation-contraction uncoupler for optically mapping action potentials in wedges of ventricular myocardium. J Cardiovasc Electrophysiol 1998; 9: 1336–1347. 16. Carlier M-F and Pantaloni D. Control of actin dynamics in cell motility. J Mol Biol 1997; 269: 459–467. 17. Hall A. Rho GTPases and the actin cytoskeleton. Science 1998; 279: 509–514. 36. Davalos D, Grutzendler J, Yang G, et al. ATP mediates rapid microglial response to local brain injury in vivo. Nat Neurosci 2005; 8: 752–758. 18. Pollard TD and Borisy GG. Cellular motility driven by assembly and disassembly of actin filaments. Cell 2003; 112: 453–465. 37. Okamoto K-I and Hayashi Y. Visualization of F-actin and G-actin equilibrium using fluorescence resonance energy transfer (FRET) in cultured cells and neurons in slices. Nat Protoc 2006; 1: 911–919. 19. Star EN, Kwiatkowski DJ and Murthy VN. Rapid turn- over of actin in dendritic spines and its regulation by activity. Nat Neurosci 2002; 5: 239–246. 20. References Cingolani LA and Goda Y. Actin in action: the interplay between the actin cytoskeleton and synaptic efficacy. Nat Rev Neurosci 2008; 9: 344–356. 38. Holzinger A. Jasplakinolide: an actin-specific reagent that promotes actin polymerization. In: Gavin RH (ed.) Cytoskeleton methods and protocols. Totowa, NJ: Humana Press, 2009, pp.71–87. 21. Wegner A. Head to tail polymerization of actin. J Mol Biol 1976; 108: 139–150. 39. Odaka C, Sanders ML and Crews P. Jasplakinolide induces apoptosis in various transformed cell lines by a caspase-3-like protease-dependent pathway. Clin Diagn Lab Immunol 2000; 7: 947–952. 22. Carlier M-F, Pantaloni D, Evans JA, et al. The hydrolysis of ATP that accompanies actin polymerization is essen- tially irreversible. FEBS Lett 1988; 235: 211–214. 23. Bernstein BW and Bamburg JR. Actin-ATP hydrolysis is a major energy drain for neurons. J Neurosci 2003; 23: 1–6. 40. Spector I, Shochet NR, Blasberger D, et al. Latrunculins – novel marine macrolides that disrupt microfilament organization and affect cell growth: I. Comparison with cytochalasin D. Cell Motil Cytoskeleton 1989; 13: 127–144. 24. Engl E and Attwell D. Non-signalling energy use in the brain. J Physiol 2015; 593: 3417–3429. 965 Engl et al. 41. Fonseca R. Activity-dependent actin dynamics are required for the maintenance of long-term plasticity and for synaptic capture: role of actin dynamics in synaptic- capture. Eur J Neurosci 2012; 35: 195–206. 57. Purdon A and Rapoport S. Energy requirements for two aspects of phospholipid metabolism in mammalian brain. Biochem J 1998; 335: 313–318. 58. Shibuki K. Calcium-dependent and ouabain-resistant oxygen consumption in the rat neurohypophysis. Brain Res 1989; 487: 96–104. 42. Vasquez RJ, Howell B, Yvon AM, et al. Nanomolar concentrations of nocodazole alter microtubule dynamic instability in vivo and in vitro. Mol Biol Cell 1997; 8: 973. 59. Rusˇ cˇ a´ k M and Whittam M. The metabolic response of brain slices to agents affecting the sodium pump. J Physiol 1967; 190: 595–610. 43. Barnes SJ, Opitz T, Merkens M, et al. Stable mossy fiber long-term potentiation requires calcium influx at the granule cell soma protein synthesis, and microtubule- dependent axonal transport. J Neurosci 2010; 30: 12996–13004. 60. Tower DB. Ouabain and the distribution of calcium and magnesium in cerebral tissues in vitro. Exp Brain Res 1968; 6: 273–283. 61. Llorente-Folch I, Rueda CB, Pardo B, et al. The regula- tion of neuronal mitochondrial metabolism by calcium: regulation of neuronal mitochondrial metabolism. J Physiol 2015; 593: 3447–3462. 44. References Spurck TP, Pickett-Heaps JD and Klymkowsky MW. Metabolic inhibitors and mitosis: II. Effects of dinitro- phenol/deoxyglucose and nocodazole on the microtubule cytoskeleton. Protoplasma 1986; 131: 60–74. 45. Landree LE. C75, a fatty acid synthase inhibitor, modu- lates AMP-activated protein kinase to alter neuronal energy metabolism. J Biol Chem 2003; 279: 3817–3827. 62. Orihuela R, McPherson CA and Harry GJ. Microglial M1/M2 polarization and metabolic states: microglia bio- energetics with acute polarization. Br J Pharmacol 2015; 173: 649–665. 46. Halvorson DL and McCune SA. Inhibition of fatty acid synthesis in isolated adipocytes by 5-(tetradecyloxy)-2- furoic acid. Lipids 1984; 19: 851–856. 63. Ivanov AI, Malkov AE, Waseem T, et al. Glycolysis and oxidative phosphorylation in neurons and astrocytes during network activity in hippocampal slices. J Cereb Blood Flow Metab 2014; 34: 397–407. 47. Karpova A. Involvement of protein synthesis and degrad- ation in long-term potentiation of schaffer collateral CA1 synapses. J Neurosci 2006; 26: 4949–4955. 64. Hall CN and Attwell D. Assessing the physiological con- centration and targets of nitric oxide in brain tissue. J Physiol 2008; 586: 3597–3615. 48. Xiong W, Kojic LZ, Zhang L, et al. Anisomycin activates p38 MAP kinase to induce LTD in mouse primary visual cortex. Brain Res 2006; 1085: 68–76. 65. Zhu X-H, Zhang Y, Zhang N, et al. Noninvasive and three-dimensional imaging of CMRO2 in rats at 9.4 T: reproducibility test and normothermia/hypothermia com- parison study. J Cereb Blood Flow Metab 2007; 27: 1225–1234. 49. Unisense. Oxygen sensor user manual, 2012, http://www. unisense.com/files/PDF/Manualer/Oxygen%20Sensor% 20Manual.pdf (accessed 20 September 2015). 50. Sander R. Compilation of Henry’s law constants for inor- ganic and organic species of potential importance in envir- onmental chemistry. Max-Planck Institute of Chemistry, Air Chemistry Department Mainz, Germany, 1999, http://enviromap.utah.gov/businesses/E/EnSolutions/ depleteduranium/performassess/compliancereport/docs/ 2014/07Jul/supinfo/appreferences/Sander1999.pdf (accessed 1 October 2015). 66. Sokoloff L, Reivich M, Kennedy C, et al. The [14 c]deox- yglucose method for the measurement of local cerebral glucose utilization: theory, procedure, and normal values in the conscious and anesthetized albino rat. J Neurochem 1977; 28: 897–916. 67. Xu F, Ge Y and Lu H. Noninvasive quantification of whole-brain cerebral metabolic rate of oxygen (CMRO 2) by MRI. Magn Reson Med 2009; 62: 141–148. 51. Hall CN, Klein-Flu¨ gge MC, Howarth C, et al. Oxidative phosphorylation not glycolysis, powers presynaptic and postsynaptic mechanisms underlying brain information processing. J Neurosci 2012; 32: 8940–8951. 68. Huchzermeyer C, Berndt N, Holzhu¨ tter H-G, et al. References Oxygen consumption rates during three different neur- onal activity states in the hippocampal CA3 network. J Cereb Blood Flow Metab 2013; 33: 263–271. 52. Ganfield RA, Nair P and Whalen WJ. Mass transfer, storage, and utilization of O2 in cat cerebral cortex. Am J Physiol Content 1970; 219: 814–821. 69. Ha´ jos N, Ellender TJ, Zemankovics R, et al. Maintaining network activity in submerged hippocampal slices: importance of oxygen supply. Eur J Neurosci 2009; 29: 319–327. 53. Cooper C. Competitive, reversible, physiological? Inhibition of mitochondrial cytochrome oxidase by nitric oxide. IUBMB Life 2003; 55: 591–597. 70. Ivanov A and Zilberter Y. Critical state of energy metab- olism in brain slices: the principal role of oxygen delivery and energy substrates in shaping neuronal activity. Front Neuroenergetics 2011; 3: 9. 54. Barry PH and Diamond JM. Effects of unstirred layers on membrane phenomena. Physiol Rev 1984; 64: 763–872. 55. Han P and Bartels DM. Temperature dependence of oxygen diffusion in H2O and D2O. J Phys Chem 1996; 100: 5597–5602. 71. Kann O, Huchzermeyer C, Kovacs R, et al. Gamma oscillations in the hippocampus require high complex I gene expression and strong functional performance of mitochondria. Brain 2011; 134: 345–358. 56. Frenguelli BG, Llaudet E and Dale N. High-resolution real-time recording with microelectrode biosensors reveals novel aspects of adenosine release during hypoxia in rat hippocampal slices: adenosine release during hyp- oxia. J Neurochem 2003; 86: 1506–1515. 72. Urban NT, Willig KI, Hell SW, et al. STED nanoscopy of actin dynamics in synapses deep inside living brain slices. Biophys J 2011; 101: 1277–1284. Journal of Cerebral Blood Flow & Metabolism 37(3) 966 37(3) in different brain areas during aging. Lipids 1982; 17: 291–296. 73. Nabetani M, Okada Y, Kawai S, et al. Neural activity and the levels of high energy phosphates during depriv- ation of oxygen and/or glucose in hippocampal slices of immature and adult rats. Int J Dev Neurosci 1995; 13: 3–12. 78. Blue ME and Parnavelas JG. The formation and matur- ation of synapses in the visual cortex of the rat. II Quantitative analysis. J Neurocytol 1983; 12: 697–712. 79. Lee VM-Y and Trojanowski JQ. The disordered neuronal cytoskeleton in Alzheimer’s disease. Curr Opin Neurobiol 1992; 2: 653–656. 74. Dunaevsky A, Tashiro A, Majewska A, et al. Developmental regulation of spine motility in the mam- malian central nervous system. Proc Natl Acad Sci 1999; 96: 13438–13443. 80. Penzes P and VanLeeuwen J-E. References Impaired regulation of synaptic actin cytoskeleton in Alzheimer’s disease. Brain Res Rev 2011; 67: 184–192. 75. Santerre RF and Rich A. Actin accumulation in develop- ing chick brain and other tissues. Dev Biol 1976; 54: 1–12. 76. Kollins KM, Bell RL, Butts M, et al. Dendrites differ from axons in patterns of microtubule stability and polymeriza- tion during development. Neural Develop 2009; 4: 26. 81. Kang DE, Roh SE, Woo JA, et al. The interface between cytoskeletal aberrations and mitochondrial dysfunction in Alzheimer’s disease and related disorders. Exp Neurobiol 2011; 20: 67. 77. Gaiti A, Brunetti M, Piccinin GL, et al. The synthesis in vivo of choline and ethanolamine phosphoglycerides
https://openalex.org/W2256564424	http://www.scielo.br/pdf/floram/v22n3/2179-8087-floram-22-3-377.pdf	Portuguese	null	Detecção da Infraestrutura para Exploração Florestal em Rondônia Utilizando Dados de Sensoriamento Remoto	Floresta e Ambiente	2,015	cc-by	6,712	Floresta e Ambiente 2015; 22(3): 377-390 http://dx.doi.org/10.1590/2179-8087.064013 ISSN 1415-0980 (impresso) ISSN 2179-8087 (online) Artigo Original Floresta e Ambiente 2015; 22(3): 377-390 http://dx.doi.org/10.1590/2179-8087.064013 ISSN 1415-0980 (impresso) ISSN 2179-8087 (online) Artigo Original Floresta e Ambiente 2015; 22(3): 377-390 http://dx.doi.org/10.1590/2179-8087.064013 ISSN 1415-0980 (impresso) ISSN 2179-8087 (online) Floresta e Ambiente 2015; 22(3): 377-390 http://dx.doi.org/10.1590/2179-8087.064013 ISSN 1415-0980 (impresso) ISSN 2179-8087 (online) Artigo Original RESUMO A infraestrutura para exploração florestal é fundamental nas atividades de exploração madeireira, mas responde por grande parte dos impactos ambientais dessa atividade. O presente estudo aplicou e testou cinco técnicas de processamento digital de imagens orbitais com três resoluções espaciais distintas para a detecção de estradas e pátios florestais construídos em áreas sob manejo florestal sustentável na Floresta Nacional do Jamari, estado de Rondônia. Os resultados mostraram que o NDVI e a Análise de componentes principais apresentaram a melhor acurácia global nas resoluções espaciais de 5 e 10 metros e na de 30 metros, respectivamente. De maneira geral, a acurácia do produtor para a classe de interesse não foi alta, alcançando no máximo 39,2%, e índice Kappa de 0,38. A baixa performance das técnicas de geoprocessamento utilizadas na detecção de florestas exploradas seletivamente está relacionada às alterações ocorridas no dossel da floresta manejada, pouco perceptíveis em imagens de satélite. Palavras-chave: sensoriamento remoto, manejo florestal sustentável, processamento digital de imagens. Detecção da Infraestrutura para Exploração Florestal em Rondônia Utilizando Dados de Sensoriamento Remoto Ekena Rangel Pinagé1,2,Eraldo Aparecido Trondoli Matricardi2 1Embrapa Monitoramento por Satélite, Campinas/SP, Brasil 2Departamento de Engenharia Florestal, Universidade de Brasília – UnB, Brasília/DF, Brasil 2.1. Área de estudo A Flona do Jamari é uma Unidade de Conservação de Uso Sustentável criada em 25 de setembro de 1984. Situa-se nos municípios de Itapuã do Oeste, Cujubim e Candeias do Jamari, no estado de Rondônia, com área de 220 mil hectares (Figura 1). Segundo o plano de manejo da unidade, a tipologia vegetal predominante é a Floresta Ombrófila Densa com porções de Floresta Ombrófila Aberta, que podem se apresentar com predominância de palmeiras ou com cipós, de acordo com a classificação fisionômica-ecológica do IBGE (2012). A infraestrutura florestal é o principal elemento de contexto do padrão espacial do corte seletivo de madeira, constituída por estradas primárias que dão acesso à área de manejo e estradas secundárias ligando pátios de estocagem. Geralmente, esses elementos podem ser detectados em imagens de satélite, possibilitando o mapeamento da exploração madeireira por sensoriamento remoto. Em 2008, a Flona do Jamari destinou uma área de 96 mil hectares para concessão florestal, dividida em três unidades de manejo, o que a tornou a primeira Floresta Nacional a passar por esse tipo de licitação no Brasil. Uma vez que há limitações para a detecção do impacto de outras operações da exploração florestal com o uso de imagens ópticas, tais como abate e arraste de árvores (Coops et al., 2007), a infraestrutura foi selecionada como objeto da avaliação da qualidade do mapeamento do corte seletivo de madeira a partir de dados orbitais. Nesse trabalho foram consideradas como área de estudo as Unidades de Produção Anual (UPA) 1 das Unidades de Manejo Florestal (UMF) 1 e 2, áreas delimitadas em amarelo na Figura 1. A área da UPA 1 na UMF 1 tem 594 hectares e a área da UPA 1 na UMF 2, 1.068 hectares. O volume médio de madeira retirado foi de 14,8 m3.ha–1 na área da UMF 1 e de 9,6 m3.ha–1 na área da UMF 2. O objetivo desta análise é aplicar e testar a detecção da infraestrutura florestal utilizando diferentes técnicas de processamento de imagens de satélite com diferentes resoluções espaciais. 2.1. Área de estudo Os testes de detecção da infraestrutura foram realizados com a aplicação de algumas técnicas de processamento digital de imagens em três resoluções espaciais distintas: o Modelo Linear de Mistura Espectral (MLME), técnica mais utilizada atualmente nesse tipo de detecção, com imagens de média resolução espacial (Souza et al., 2005; Asner et al., 2005; Matricardi et al., 2010); alguns índices de vegetação: o NDVI, por ser o índice mais utilizado (Rouse et al., 1973), o MSAVI, por ter correção para os efeitos do solo (Qi et al., 1994), e o GEMI, por apresentar correção ABSTRACT Logging lands (forest roads and log decks) are an underlying issue during selective logging activities, but they are responsible for most impacts on the forest. This study aimed to apply and assess the performance of five geoprocessing techniques on remotely sensed data using three different spatial resolutions to detect logging lands under forest management at the Jamari National Forest, state of Rondônia, Brazil. The research results showed that Normalized Difference Vegetation Index (NDVI) and Principal Components Analysis (PCA) presented the best overall accuracy using spatial resolutions of 5 and 10 meters, and 30 meters, respectively. Generally, the overall accuracy and Kappa statistics for the selectively logged forest classifications were not good (39.2% or lower, and 0.38 or lower, respectively). The low performance of the geoprocessing techniques is related to the subtle changes on the forest canopy cover under selective logging activities. Keywords: remote sensing, sustainable forest management, digital image processing. Floresta e Ambiente 2015; 22(3): 377-390 378 Pinagé ER, Matricardi EAT 1. INTRODUÇÃO E OBJETIVO para os efeitos atmosféricos (Pinty & Verstraete, 1992); e a Análise de componentes principais (ACP). Para que o manejo florestal seja praticado, é necessária a abertura da infraestrutura florestal permanente, constituída de pátios de estocagem, estradas primárias e secundárias, além das pontes. Porém, a sustentabilidade da atividade demanda que as vias de acesso gerem o menor impacto possível aos ecossistemas. De acordo com um levantamento do IFT (2011), obras de infraestrutura como estradas e pátios podem impactar diretamente até 10% da área onde se pratica o manejo florestal. Quando as estradas não são bem construídas, podem provocar prejuízos aos recursos naturais, afetando qualidade da água, migração e deslocamento de espécies, ocasionando ainda fragmentação de habitats e erosão do solo (Uhl & Vieira, 1989). As técnicas e testes foram aplicados na área de floresta pública sob concessão federal na Floresta Nacional (Flona) do Jamari, Estado de Rondônia, onde existem planos de manejo florestal sustentável em execução desde 2010. 2.2. Imagens de satélite e técnicas de processamento Foram utilizadas imagens multiespectrais do satélite Landsat 5, sensor TM, largamente utilizadas no monitoramento da Amazônia (INPE, 2006) e atualmente distribuídas gratuitamente. Além desse motivo, o Landsat 5 é o único satélite da série Landsat operacional nas datas escolhidas para a análise. Foram 379 Detecção da Infraestrutura para Exploração Florestal... Floresta e Ambiente 2015; 22(3): 377-390 Figura 1. Localização da Floresta Nacional do Jamari, de suas Unidades de Manejo Florestal (UMF) e das Unidade de Produção Anual (UPA) objetos deste estudo. Figure 1. Study site location at the Jamari National Forest, its Forest Management Plots (FMP), and the Annua Production Plots (APP) in this study. Figura 1. Localização da Floresta Nacional do Jamari, de suas Unidades de Manejo Florestal (UMF) e das Unidades de Produção Anual (UPA) objetos deste estudo. Figure 1. Study site location at the Jamari National Forest, its Forest Management Plots (FMP), and the Annual Production Plots (APP) in this study. testadas também ortoimagens da constelação de satélites RapideEye, no nível de processamento 3A. Tabela 1. Dados de sensoriamento remoto utilizado neste estudo. Table 1. Remotely sensed data used in this study. Table 1. Remotely sensed data used in this study. As imagens de satélite que cobrem a área de estudo foram adquiridas em duas coberturas distintas: uma antes do início da exploração florestal na área, que ocorreu em setembro de 2010, e outra após o início, em maio de 2011. Três resoluções espaciais foram utilizadas: as imagens RapidEye em sua resolução de ortoimagens (5 metros) e reamostradas para 10 metros, e a imagem Landsat 5 TM com 30 metros. A Tabela 1 mostra as informações das imagens Landsat e RapidEye utilizadas. Satélite/ Constelação Órbita-ponto/ Tile Data de aquisição Landsat5 232/066 15/5/2010 Landsat5 232/066 5/7/2011 RapidEye 2034814 14/5/2010 RapidEye 2034815 14/5/2010 RapidEye 2034915 14/5/2010 RapidEye 2034916 10/5/2010 RapidEye 2034814 31/5/2011 RapidEye 2034815 31/5/2011 RapidEye 2034915 31/5/2011 RapidEye 2034916 31/5/2011 As imagens Landsat passaram por calibração radiométrica – conversão de valores digitais para níveis de radiância e, posteriormente, para valores de reflectância no topo da atmosfera e, em seguida, por correção geométrica. As imagens RapidEye passaram apenas pela correção radiométrica, por serem disponibilizadas já com a correção geométrica. A partir das imagens calibradas para reflectância no topo da atmosfera, as técnicas de processamento digital de imagens detalhadas a seguir foram aplicadas. 2.2.5. Análise de componentes principais (ACP) 2.2.5. Análise de componentes principais (ACP) em que ρIVP = Reflectância na faixa espectral do infravermelho próximo; e ρV = Reflectância na faixa espectral do vermelho; e em que ρIVP = Reflectância na faixa espectral do infravermelho próximo; e ρV = Reflectância na faixa espectral do vermelho; e A correlação entre bandas é um problema frequentemente encontrado em imagens multiespectrais. A transformação por componentes principais é uma técnica utilizada para reduzir essa redundância entre os dados presentes nas diferentes faixas espectrais. O propósito dessa técnica é comprimir toda a informação contida em um conjunto original de dados com n bandas para um dado com menos do que n bandas, que pode então ser utilizado no lugar do dado original (Lillesand et al., 2008). Eliminar a alta correlação das imagens tem a vantagem de reduzir o volume de dados a ser analisado e de redistribuir a informação espectral entre as novas principais componentes (Meneses & Almeida, 2012). ( ) ( ) L ( IVP V s 1 IVP V ² 8.0s* IVP V = ρ −ρ + + ρ + ρ − ρ −ρ     (3) em que s = Declividade da linha de solos calculada a partir da reflectância do solo exposto na imagem RapidEye. 2.2.1. Normalized Difference Vegetation Index (NDVI) O NDVI foi um dos primeiros índices de vegetação desenvolvidos e é amplamente utilizado até os dias atuais, tendo sido explorado em diferentes abordagens em Floresta e Ambiente 2015; 22(3): 377-390 80 Pinagé ER, Matricardi EAT 380 2.2.3. Global Environment Monitoring Index (GEMI) Considerando que a influência da atmosfera é maior na faixa espectral do vermelho do que na do infravermelho próximo, Pinty & Verstraete (1992) propuseram para o monitoramento global da vegetação esse novo índice, o GEMI, criado para minimizar a influência dos efeitos atmosféricos no valor final do índice. Seu cálculo é dado pelas Equações 4 e 5: (MLME) Dentro de um único pixel podem estar incluídos diferentes objetos ou elementos da cobertura da superfície terrestre. Isso gera uma mistura espectral, ou seja, a resposta espectral de um pixel da imagem é resultante da combinação da resposta espectral dos diferentes componentes que formam esse pixel. ( ) ( ) IVP V / IVP V NDVI = ρ −ρ ρ + ρ (1) (1) em que ρIVP = Reflectância na faixa espectral do infravermelho próximo e ρV = Reflectância na faixa espectral do vermelho. Com o uso do Modelo linear de mistura espectral é possível estimar a proporção de cada um desses alvos no pixel a partir da aplicação da técnica de otimização dos mínimos quadrados (Shimabukuro & Smith, 1991). Com a aplicação desse modelo é gerada uma imagem fração para cada um dos componentes considerados, conhecidos como endmembers. Essas imagens representam as proporções (abundância) de cada um dos respectivos componentes da cena selecionados para o modelo de mistura espectral (Ponzoni & Shimabukuro, 2009). 2.2.2. Modified Soil Adjusted Vegetation Index (MSAVI) Desenvolvido por Qi et al. (1994), o MSAVI oferece uma fórmula com a inclusão de um fator de correção para a influência dos solos, L, que depende da cobertura vegetal da área de trabalho. Sua fórmula é mostrada nas Equações 2 e 3: ( ) ( ) ( ) ( ) MSAVI IVP V / IVP V L * 1 L = ρ −ρ ρ + ρ + + (2) 2.2.4. Modelo linear de mistura espectral (MLME) estudos agrícolas, florestais e climáticos. Desenvolvido por Rouse et al. (1973), deu origem a vários outros índices, que são transformações do NDVI, na tentativa de reduzir os efeitos do solo e da atmosfera e melhorar o desempenho dos índices de vegetação na avaliação da vegetação. Sua fórmula é descrita na Equação 1: 2.2.6. Análise de acurácia A análise estatística da detecção da infraestrutura florestal consistiu na avaliação de acurácia das classificações geradas com as cinco técnicas de processamento digital de imagens descritas anteriormente. Nessa avaliação é gerada para cada classificação uma matriz de confusão (também conhecida como tabela de contingência) com dois dados: aqueles gerados pela classificação e pontos de GPS coletados em campo com um receptor GPS de navegação, adotados como dado de referência para a ( ) ( ) ( ) = 1 0,25 ( 0,125 / 1 , GEMI V V η η ρ ρ − − − (4) (4) em que ρV = Reflectância na faixa espectral do vermelho; e em que ρV = Reflectância na faixa espectral do vermelho; e ( ) ( ) ( ) ( ) 2 2 2 1 ,5 0,5 / 0,5 , IVP V IVP V IVP V η ρ ρ ρ ρ ρ ρ = − + + + + (5) onde ρIVP = Reflectância na faixa espectral do infravermelho próximo; e ρV = Reflectância na faixa espectral do vermelho. ( ) ( ) ( ) ( ) 2 2 2* 1 ,5 0,5 / 0,5 , IVP V IVP V IVP V η ρ ρ ρ ρ ρ ρ = − + + + + (5) onde ρIVP = Reflectância na faixa espectral do infravermelho próximo; e ρV = Reflectância na faixa espectral do vermelho. Floresta e Ambiente 2015; 22(3): 377-390 38 Detecção da Infraestrutura para Exploração Florestal... 381 validação. Esse procedimento é descrito em detalhes por Congalton & Green (2008). Um fluxograma resumindo as etapas seguidas para a detecção da infraestrutura florestal é apresentado na Figura 2. Foram analisadas as estatísticas fornecidas pela matriz de confusão: índice Kappa, erros de omissão e comissão e acurácia global, do produtor e do usuário. O índice Kappa representa a medida da concordância entre os dados de referência e os dados da classificação (indicados pela diagonal da matriz de confusão) e da concordância esperada ao acaso (indicada pelas marginais da matriz de confusão). A acurácia global representa a proporção dos pontos de referência corretamente classificados (Congalton & Green, 2008). 3. RESULTADOS E DISCUSSÃO A configuração das estradas e pátios de estocagem bem como as árvores abatidas até a data de aquisição das imagens auxiliam na interpretação dos resultados das classificações geradas (Figura 3). As classificações geradas a partir da aplicação das cinco técnicas nos três conjuntos de imagens mostraram resultados bastante Figura 2. Fluxo utilizado na detecção da infraestrutura florestal utilizando dados de sensoriamento remoto. Figure 2. Flowchart for the logging infrastructure detection using remotely sensed data. Figura 2. Fluxo utilizado na detecção da infraestrutura florestal utilizando dados de sensoriamento remoto. Figure 2. Flowchart for the logging infrastructure detection using remotely sensed data. Figura 3. Infraestrutura florestal existente nas áreas de estudo e árvores abatidas até a data das imagens de satélite utilizadas. Figure 3. Existing logging infrastructure in the study sites and harvested trees by the acquisition date of the satellite images. Figura 3. Infraestrutura florestal existente nas áreas de estudo e árvores abatidas até a data das imagen utilizadas. Figura 3. Infraestrutura florestal existente nas áreas de estudo e árvores abatidas até a data das imagens de satélite utilizadas. Figure 3. Existing logging infrastructure in the study sites and harvested trees by the acquisition date of the satellite images. Figura 3. Infraestrutura florestal existente nas áreas de estudo e árvores abatidas até a data das imagens de satélite utilizadas. Figure 3. Existing logging infrastructure in the study sites and harvested trees by the acquisition date of the satellite images. utilizadas. Figure 3. Existing logging infrastructure in the study sites and harvested trees by the acquisition date of the satellite images. 382 Pinagé ER, Matricardi EAT Floresta e Ambiente 2015; 22(3): 377-390 As outras técnicas apresentaram resultados intermediários entre o NDVI e o GEMI, apresentando acurácia global e índice Kappa de: 63,34% e 0,24 (MLME); 62,75% e 0,23 (ACP); e 62,17% e 0,22 (MSAVI). As demais estatísticas geradas apresentadas na matriz de confusão também seguiram esse padrão. discrepantes para as diferentes resoluções espaciais, especialmente entre as imagens RapidEye e Landsat, conforme mostram as matrizes de confusão dos dados gerados a partir das imagens RapidEye na resolução espacial de 5 metros e o mapa das classificações (Tabela 2 e Figura 3, respectivamente). Para essa resolução houve uma variação de até 13% entre os resultados das cinco técnicas empregadas. O NDVI apresentou os melhores resultados de acurácia global e índice Kappa (70,01% e 0,38, respectivamente). 3. RESULTADOS E DISCUSSÃO explorada 3 41 44 25,15% 93,18% 6,82% 74,85% Total 178 163 341 -- -- -- -- Floresta e Ambiente 2015; 22(3): 377-390 3 Detecção da Infraestrutura para Exploração Florestal... 383 Esses valores de acurácia do produtor e do usuário indicam que o mapeamento da infraestrutura florestal foi conservador, com altas taxas de omissão para a classe área explorada, concentradas nas áreas de estradas secundárias. Essas feições são de difícil detecção, especialmente numa exploração de impacto reduzido como a que ocorre nas concessões florestais. Em muitos trechos elas não apresentam largura suficiente para abrir o dossel florestal e expor porções de solo, como acontece nas estradas primárias e pátios de estocagem. Entretanto, dessa vez o MSAVI se mostrou a melhor técnica para a detecção da classe área explorada, com acurácia do produtor cerca de 5% maior que a do NDVI. A imagem resultante da aplicação dessa técnica é a que detém maior quantidade de pequenos polígonos da classe de interesse, então a chance de ocorrerem coincidências com os pontos de validação são maiores. Inversamente, mas pelo mesmo motivo, a acurácia do produtor da classe floresta para essa técnica foi a menor de todas. Além da análise estatística obtida com a matriz de confusão, é pertinente também analisar visualmente o resultado das classificações. Comparando-se o NDVI com o GEMI, por exemplo, percebe-se que o NDVI mapeou melhor a infraestrutura e apresentou menos polígonos espúrios, sugerindo que ele foi mais eficiente na detecção da exploração florestal. Comparando-se os resultados obtidos com a resolução de 10 metros com os da de 5 metros, observa‑se que apenas o NDVI apresentou resultado superior na resolução de 5 metros, as demais técnicas tiveram desempenho melhor na resolução de 10 metros. Esses resultados gerais indicam que não há ganho representativo em aumentar a resolução espacial de 10 para 5 metros. Visualmente, os resultados das duas resoluções também são bastante similares. De maneira geral, as técnicas baseadas em índices de vegetação apresentaram mais polígonos classificados como infraestrutura florestal do que o MLME e a ACP. Mas, em contrapartida, essas técnicas apresentaram melhor configuração das estradas primárias e alguma sinalização de estradas secundárias, mesmo que incipiente e descontínua, especialmente o MSAVI. Apresentam ainda alguma indicação de impacto no dossel ocasionado pelas operações de derrubada e arraste das toras, especialmente o NDVI. 3. RESULTADOS E DISCUSSÃO O GEMI apresentou o pior desempenho entre os resultados obtidos para essa resolução, com 57,47% e 0,12 de acurácia global e índice Kappa, respectivamente. Para a classe de interesse do mapeamento (área explorada), os valores de acurácia do produtor foram bem mais baixos, variando de 39,26% (NDVI) a 14,72% (GEMI). Já para a acurácia do usuário, os valores para a classe de interesse foram bem mais altos, sendo que somente os índices GEMI e MSAVI apresentaram valor inferior a 90%. Tabela 2. Matrizes de confusão das cinco técnicas aplicadas nas imagens RapidEye com resolução espacial de 5 metros. Table 2. Confusion matrices of the five processing techniques applied in RapidEye imagery with 5 meters spatial resolution. resolution. Normalized Difference Vegetation Index (NDVI) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 175 99 274 98,31% 63,87% 36,13% 1,69% 70,01% 0,39 A. explorada 3 64 67 39,26% 95,52% 4,48% 60,74% Total 178 163 341 -- -- -- -- Global Environmental Monitoring Index (GEMI) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 172 139 311 96,63% 55,31% 44,69% 3,37% 57,47% 0,12 A. explorada 6 24 30 14,72% 80,00% 20,00% 85,38% Total 178 163 341 -- -- -- -- Modified Soil Adjusted Vegetation Index (MSAVI) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 160 111 271 89,89% 59,04% 40,96% 10,11% 62,17% 0,22 A. explorada 18 52 70 31,90% 74,29% 25,71% 68,10% Total 178 163 341 -- -- -- -- Análise de componentes principais (ACP) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 175 124 299 98,31% 58,53% 41,47% 1,69% 62,75% 0,23 A. explorada 3 39 42 23,93% 92,86% 7,14% 76,07% Total 178 163 341 -- -- -- -- Modelo linear de mistura espectral (MLME) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 175 122 297 98,31% 58,92% 41,08% 1,69% 63,34% 0,24 A. 3. RESULTADOS E DISCUSSÃO A comparação do mapa da Figura 3 com a Figura 4 permite a identificação do impacto no dossel relacionado ao abate das árvores até a data de aquisição das imagens, que foi melhor mapeado nas classificações geradas com os índices de vegetação. Na resolução de 30 metros, os resultados de acurácia global das cinco técnicas foram muito similares, com variação de aproximadamente 5% (Tabela 4). A técnica que apresentou o maior valor dessa estatística foi a ACP (57,72%); o NDVI e o MSAVI apresentaram resultados idênticos (53,64%). Entretanto, quando se analisa a acurácia do produtor para a classe área explorada, observa-se que os valores foram bem baixos (variando de 0,61% para o GEMI a 15,15% para a ACP). A acurácia global ficou em torno dos 50%, porque o mapeamento da classe floresta foi excelente para todas as técnicas, como pode ser verificado na matriz de confusão, dado que é uma classificação binária. Somente o resultado da Análise de componentes principais apresentou valor da estatística Kappa superior a 0,1, valor já considerado de baixa concordância (Congalton & Green, 2008). Para a resolução espacial de 10 metros, a acurácia global das diferentes técnicas apresentaram no máximo 5,2% de variação, menor do que a variação encontrada na resolução de 5 metros, como pode ser observado nas matrizes de confusão (Tabela 3). Na análise visual da Figura 6 observa-se que apenas os resultados do MSAVI, da ACP e do MLME mostram alguma configuração das estradas primárias existentes, mesmo que descontínuas. Corroborando a maior acurácia global obtida entre as cinco técnicas, a ACP apresenta o resultado com maior significado real (no campo), principalmente na área da UMF 2. Silva et al. (2009) também identificaram a ACP, em comparação com o MLME e a transformação IHS com razão de bandas, como a melhor técnica na detecção O mapa das classificações geradas a partir das imagens RapidEye reamostradas para a resolução espacial de 10 metros mostra que as técnicas que não envolveram detecção de mudanças apresentaram resultados idênticos e, visualmente, seus resultados também são muito semelhantes (Figuras 4 e 5). O NDVI continuou apresentando melhor acurácia global e índice Kappa, assim como o GEMI, seguiu apresentando o pior desempenho. 384 Pinagé ER, Matricardi EAT Floresta e Ambiente 2015; 22(3): 377-390 ura 4. Resultados das classificações geradas com as cinco técnicas aplicadas nas imagens RapidEye com re acial de 5 metros. ure 4. 3. RESULTADOS E DISCUSSÃO Results of the binary classifications derived from the five processing techniques applied in R agery with 5 meters spatial resolution. Figura 4. Resultados das classificações geradas com as cinco técnicas aplicadas nas imagens RapidEye com resolução espacial de 5 metros. Figure 4. Results of the binary classifications derived from the five processing techniques applied in RapidEye imagery with 5 meters spatial resolution. Figura 4. Resultados das classificações geradas com as cinco técnicas aplicadas nas imagens RapidEye com resolução espacial de 5 metros. Figure 4. Results of the binary classifications derived from the five processing techniques applied in RapidEye imagery with 5 meters spatial resolution. Figura 4. Resultados das classificações geradas com as cinco técnicas aplicadas nas imagens RapidEye com resolução espacial de 5 metros. Figure 4. Results of the binary classifications derived from the five processing techniques applied in RapidEye imagery with 5 meters spatial resolution. 385 Detecção da Infraestrutura para Exploração Florestal... Floresta e Ambiente 2015; 22(3): 377-390 Figura 5. Resultados das classificações geradas com as cinco técnicas aplicadas nas imagens RapidEye com re spacial de 10 metros. Figure 5. Results of the binary classifications derived from the five processing techniques applied in R magery with 10 meters spatial resolution Figura 5. Resultados das classificações geradas com as cinco técnicas aplicadas nas imagens RapidEye com resolução espacial de 10 metros. Figure 5. Results of the binary classifications derived from the five processing techniques applied in RapidEye imagery with 10 meters spatial resolution. Figura 5. Resultados das classificações geradas com as cinco técnicas aplicadas nas imagens RapidEye com resolução espacial de 10 metros. Figure 5. Results of the binary classifications derived from the five processing techniques applied in RapidEye imagery with 10 meters spatial resolution. 386 Pinagé ER, Matricardi EAT Floresta e Ambiente 2015; 22(3): 377-390 ura 6. Resultados das classificações geradas com as cinco técnicas aplicadas nas imagens Landsat 5 T lução espacial de 30 metros. ure 6. Results of the binary classifications derived from the five processing techniques applied in Lands h l l Figura 6. Resultados das classificações geradas com as cinco técnicas aplicadas nas imagens Landsat 5 TM com resolução espacial de 30 metros. Figure 6. Results of the binary classifications derived from the five processing techniques applied in Landsat 5 TM imagery with 30 meters spatial resolution. Figura 6. p Results of the binary classifications derived from the five processing techniques applied in Landsat 5 TM with 30 meters spatial resolution. 3. RESULTADOS E DISCUSSÃO Resultados das classificações geradas com as cinco técnicas aplicadas nas imagens Landsat 5 TM com resolução espacial de 30 metros. Figure 6. Results of the binary classifications derived from the five processing techniques applied in Landsat 5 TM imagery with 30 meters spatial resolution. Figure 6. Results of the binary classifications derived from the five processing techniques applied in Landsat 5 TM imagery with 30 meters spatial resolution. Floresta e Ambiente 2015; 22(3): 377-390 387 Detecção da Infraestrutura para Exploração Florestal... 387 Tabela 3. Matrizes de confusão das cinco técnicas aplicadas nas imagens RapidEye com resolução espacial de 10 metros.i Table 3. Confusion matrices of the five processing techniques applied in RapidEye imagery with 10 meters spatial resolution. Normalized Difference Vegetation Index (NDVI) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 171 109 280 96,07% 61,07% 38,93% 3,93% 65,98% 0,30 A. explorada 7 54 61 33,13% 88,52% 11,48% 66,87% Total 178 163 341 -- -- -- -- Global Environmental Monitoring Index (GEMI) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 168 124 292 94,38% 57,53% 42,47% 5,62% 60,70% 0,19 A. explorada 10 39 49 23,93% 79,59% 20,41% 76,07% Total 178 163 341 -- -- Modified Soil Adjusted Vegetation Index (MSAVI) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 157 100 257 88,20% 61,09% 38,91% 11,80% 64,52% 0,27 A. explorada 21 63 84 38,65% 75,00% 25,00% 61,35% Total 178 163 341 -- -- -- -- Análise de componentes principais (ACP) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 176 121 297 98,88% 59,26% 40,74% 1,12% 63,93% 0,25 A. explorada 2 42 44 25,77% 95,45% 4,55% 74,23% Total 178 163 341 -- -- -- -- Modelo linear de mistura espectral (MLME) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 176 121 297 98,88% 59,26% 40,74% 1,12% 63,93% 0,25 A. 3. RESULTADOS E DISCUSSÃO explorada 2 42 44 25,77% 95,45% 4,55% 74,23% Total 178 163 341 -- -- -- -- do corte seletivo com imagens Landsat 5 TM numa área do Mato Grosso. melhor desempenho na classificação do que o NDVI, contudo isso não foi constatado. Uma possível causa para o melhor desempenho do NDVI é a data das imagens utilizadas na detecção da infraestrutura. As imagens RapidEye pós-exploração foram adquiridas em maio, já no início da estação seca. Se elas tivessem sido adquiridas no auge da estação chuvosa, a vegetação provavelmente estaria no seu vigor máximo e a saturação com vegetação densa que ocorre no NDVI poderia ter diminuído seu desempenho. Os baixos valores do índice Kappa sugerem que a detecção da infraestrutura florestal nas imagens Landsat tem pouca aderência aos dados de referência, que representam a realidade de campo. A extensão dos alvos em relação à resolução espacial das imagens explica esses resultados: muitos alvos apresentam dimensões menores que 30 metros, resolução espacial do sensor TM. A acurácia da classe floresta elevou fortemente os valores da acurácia global para todas as classificações. Esse fato é evidenciado pela diferença entre os valores obtidos na acurácia global (dada em porcentagem) e o índice Kappa (apresentado em proporção). O índice De maneira geral, considerando-se as três resoluções espaciais analisadas, esperava-se que os índices de vegetação com correção para os fatores atmosféricos (GEMI) e de influência do solo (MSAVI) apresentassem Floresta e Ambiente 2015; 22(3): 377-390 388 Pinagé ER, Matricardi EAT Tabela 4. Matrizes de confusão das cinco técnicas aplicadas nas imagens Landsat 5 TM com resolução espacial de 30 metros. Table 4. Confusion matrices of the five processing techniques applied in Landsat 5 TM imagery with 30 meters spatial resolution. spatial resolution. Normalized Difference Vegetation Index (NDVI) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 176 157 333 98,88% 52,85% 47,15% 1,12% 53,64% 0,04 A. explorada 2 8 10 4,85% 80,00% 20,00% 95,15% Total 178 165 343 -- -- -- -- Global Environmental Monitoring Index (GEMI) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 178 164 342 100,00% 52,05% 47,95% 0,00% 52,19% 0,01 A. 3. RESULTADOS E DISCUSSÃO explorada 0 1 1 0,61% 100,00% 0,00% 100,00% Total 178 165 343 -- -- -- -- Modified Soil Adjusted Vegetation Index (MSAVI) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 176 157 333 98,88% 52,85% 47,15% 1,12% 53,64% 0,04 A. explorada 2 8 10 4,85% 80,00% 20,00% 95,15% Total 178 165 343 -- -- -- -- Análise de componentes principais (ACP) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 173 140 313 97,19% 55,27% 44,73% 2,81% 57,73% 0,13 A. explorada 5 25 30 15,15% 83,33% 16,67% 84,85% Total 178 165 343 -- -- -- -- Modelo linear de mistura espectral (MLME) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 177 154 331 99,44% 53,47% 46,53% 0,56% 54,81% 0,06 A. explorada 1 11 12 6,67% 91,67% 8,33% 93,33% Total 178 165 343 -- -- -- -- Normalized Difference Vegetation Index (NDVI) Classificação Dado de referência Acurácia do produtor Acurácia do usuário Erro de comissão Erro de omissão Acurácia global Índice Kappa Floresta A. explorada Total Floresta 176 157 333 98,88% 52,85% 47,15% 1,12% 53,64% 0,04 A. explorada 2 8 10 4,85% 80,00% 20,00% 95,15% Total 178 165 343 -- -- -- -- Gl b l E i t l M it i I d (GEMI) Kappa compara os resultados obtidos a uma classificação aleatória, considerando a classificação de forma global e não fazendo uma média entre as acurácias das classes consideradas, como acontece no cálculo da acurácia global. pela exploração florestal de impacto reduzido está dentro da variabilidade natural encontrada em áreas de floresta primária, quando o dossel é visto de cima pelos sensores ópticos (no nível do solo as mudanças são muito mais evidentes). A grande quantidade de pixels atribuídos à classe de interesse nas classificações geradas a partir dos índices de vegetação, inclusive nas áreas não exploradas na UPA, sugere que o grau de mudança ocasionado pela exploração florestal é similar às mudanças naturais que ocorrem na floresta de um ano para o outro. Esse resultado corrobora a afirmação defendida por D’Oliveira et al. (2012) de que grande parte das alterações ocasionadas 4. CONCLUSÕES E CONSIDERAÇÕES FINAIS Estimating forest biomass and identifying low-intensity logging areas using airborne scanning lidar in Antimary State Forest, Acre State, Western Brazilian Amazon. Remote Sensing of Environment 2012; 124: 479-491. http://dx.doi. org/10.1016/j.rse.2012.05.014. logging areas using airborne scanning lidar in Antimary State Forest, Acre State, Western Brazilian Amazon. Remote Sensing of Environment 2012; 124: 479-491. http://dx.doi. org/10.1016/j.rse.2012.05.014. Numa área com exploração florestal de impacto reduzido como é o caso das concessões florestais é fundamental o uso de imagens ópticas com maior resolução espacial ou de sensores ativos, especialmente se o objetivo é avaliar e monitorar impactos que ocorrem em pequena escala (remoção de uma árvore, abertura de ramais de arraste). Os resultados encontrados com as imagens Landsat neste estudo apontam claramente essa necessidade. O uso de dados de aerolevantamento com tecnologia de perfilhamento a laser (LiDAR) tem mostrado grande potencial e resultados promissores para esse fim (Andersen et al., 2014; Kent et al., 2015). Instituto Brasileiro de Gografia e Estatística - IBGE. Manual técnico da vegetação brasileira. 2. ed. Rio de Janeiro: IBGE; 2012. Instituto Floresta Tropical – IFT. [online]. [citado em 2011 set. 9]. Disponível em: http://www.ift.org.br/ Instituto Nacional de Pesquisas Espaciais - INPE. Metodologia para o cálculo da taxa anual de desmatamento na Amazônia Legal [online]. São José dos Campos: INPE; 2006. [citado em 2011 jul. 21]. Disponível em: http://www.obt.inpe.br/ prodes/metodologia.pdf Kent R, Lindsell JA, Laurin GV, Valentini R, Coomes DA. Airborne LiDAR detects selectively logged tropical forest even in an advanced stage of recovery. Remote Sensing 2015; 7: 8348-8367. http://dx.doi.org/10.3390/rs70708348. Recebido: 21 out., 2013 Aceito: 30 mar., 2015 AUTOR(ES) PARA CORRESPONDÊNCIA Recebido: 21 out., 2013 Aceito: 30 mar., 2015 Recebido: 21 out., 2013 Aceito: 30 mar., 2015 Matricardi EAT, Skole DL, Pedlowski MA, Chomentowski W, Fernandes LC. Assessment of tropical forest degradation by selective logging and fire using Landsat imagery. Remote Sensing of Environment 2010; 114(5): 1117-1129. http:// dx.doi.org/10.1016/j.rse.2010.01.001. 4. CONCLUSÕES E CONSIDERAÇÕES FINAIS A acurácia global máxima obtida na detecção da infraestrutura florestal obtida nesta análise foi de aproximadamente 70%, com a utilização do índice de vegetação NDVI na resolução espacial de 5 metros. Floresta e Ambiente 2015; 22(3): 377-390 389 Detecção da Infraestrutura para Exploração Florestal... 389 Valores de acurácia menores ainda foram obtidos para a classe de interesse (área explorada). As imagens ópticas são normalmente utilizadas no monitoramento do corte seletivo em florestas tropicais, possibilitando a detecção de sua ocorrência baseada em uma série de elementos de contexto que constituem seu padrão espacial, mas são limitadas em sua capacidade de distinguir as mudanças estruturais que ocorrem abaixo do dossel (Coops et al., 2007), como é o caso de grande parte das alterações que ocorrem devido à execução do manejo florestal. REFERÊNCIAS Andersen HE, Reutebuch SE, McGaughey RJ, d’Oliveira MVN, Keller M. Monitoring selective logging in western Amazonia with repeat lidar flights. Remote Sensing of Environment 2014; 151: 157-165. http://dx.doi.org/10.1016/j. rse.2013.08.049. Asner GP, Knapp DE, Broadbent EN, Oliveira PJC, Keller M, Silva JN. Selective logging in the Brazilian Amazon. Science 2005; 310(5747): 480-482. http://dx.doi.org/10.1126/ science.1118051. PMid:16239474. Congalton RG, Green K. Assessing the accuracy of remotely sensed data. 2. ed. Boca Raton: CRC Press; 2008. http:// dx.doi.org/10.1201/9781420055139. O GEMI apresentou o pior desempenho dentre as técnicas utilizadas nas três resoluções espaciais consideradas. Esse índice foi concebido para uso em monitoramento global e essa pode ser uma razão pela qual o teste em pequenas áreas como as UPA da Flona do Jamari não obteve bons resultados. Ele também já foi aplicado em imagens de média resolução espacial em escala regional (Matricardi et al., 2010; Wang et al., 2005), mas em nenhum desses trabalhos foi indicado como melhor opção em relação aos outros índices e/ou técnicas utilizados. Coops NC, Hilker T, Wulder MA, St-Onge B, Newnham G, Siggins A et al. Estimating canopy structure of Douglas-fir forest stands from discrete-return LIDAR. Trees - Structure and Function 2007; 21(3): 295-310. http://dx.doi.org/10.1007/ s00468-006-0119-6. Coops NC, Hilker T, Wulder MA, St-Onge B, Newnham G, Siggins A et al. Estimating canopy structure of Douglas-fir i forest stands from discrete-return LIDAR. Trees - Structure and Function 2007; 21(3): 295-310. http://dx.doi.org/10.1007/ s00468-006-0119-6. D’Oliveira MVN, Reutebuch SE, McGaughey RJ, Andersen HE. Estimating forest biomass and identifying low-intensity D’Oliveira MVN, Reutebuch SE, McGaughey RJ, Andersen HE. STATUS DA SUBMISSÃO Lillesand TM, Kiefer RW, Chipman JW. Remote sensing and image interpretation. 6. ed. Hoboken: John Willey & Sons; 2008. Ekena Rangel Pinagé Ekena Rangel Pinagé Embrapa Monitoramento por Satélite, CEP 13070-115, Campinas, SP, Brasil e-mail: ekenapinage@hotmail.com Ekena Rangel Pinagé Embrapa Monitoramento por Satélite, CEP 13070-115, Campinas, SP, Brasil e-mail: ekenapinage@hotmail.com Meneses PR, Almeida T. Transformação por rotação espectral. In: Meneses PR, Almeida T, editores. Introdução ao processamento de imagens de sensoriamento remoto. Brasília: UnB:CNPq; 2012. Floresta e Ambiente 2015; 22(3): 377-390 390 Pinagé ER, Matricardi EAT Pinty B, Verstraete MM. GEMI: a non-linear index to monitor global vegetation from satellites. Vegetatio 1992; 101(1): 15-20. http://dx.doi.org/10.1007/BF00031911. Silva FB, Fonseca LMG, Valeriano DM, Shimabukuro YE. Identificação de processos de desmatamento seletivo na região amazônica utilizando dados orbitais e técnicas de processamento digital de imagens. In: Anais XIV Simpósio Brasileiro de Sensoriamento Remoto; 2009; Natal. São José dos Campos: INPE; 2009. p. 3043-3049. Silva FB, Fonseca LMG, Valeriano DM, Shimabukuro YE. Identificação de processos de desmatamento seletivo na região amazônica utilizando dados orbitais e técnicas de processamento digital de imagens. In: Anais XIV Simpósio Brasileiro de Sensoriamento Remoto; 2009; Natal. São José dos Campos: INPE; 2009. p. 3043-3049. Ponzoni FJ, Shimabukuro YE. Sensoriamento remoto no estudo da vegetação. 2. ed. São José dos Campos: Parêntese; 2009. Souza C Jr, Roberts D, Cochrane MA. Combining spectral and spatial information to map canopy damage from selective logging and forest fires. Remote Sensing of Environment 2005; 98(2-3): 329-343. http://dx.doi.org/10.1016/j. rse.2005.07.013. Souza C Jr, Roberts D, Cochrane MA. Combining spectral and spatial information to map canopy damage from selective logging and forest fires. Remote Sensing of Environment 2005; 98(2-3): 329-343. http://dx.doi.org/10.1016/j. rse.2005.07.013. Qi J, Chehbouni A, Huete AR, Kerr YH, Sorooshian S. A modified soil adjusted vegetation index. Remote Sensing of Environment 1994; 48(2): 119-126. http://dx.doi. org/10.1016/0034-4257(94)90134-1. Rouse JW, Haas RH, Schell JA, Deering DW. Monitoring vegetation systems in the great plains with ERTS. In: Third ERTS Symposium Abstracts; 1973; Washington. Greenbelt, MD: NASA; 1973. p. 309-317. Rouse JW, Haas RH, Schell JA, Deering DW. Monitoring vegetation systems in the great plains with ERTS. In: Third Uhl C, Vieira ICG. Ecological impacts of selective logging in the Brazilian Amazon - a case study from the Paragominas Region of the State of Pará. Biotropica 1989; 21(2): 98-106. http://dx.doi.org/10.2307/2388700. Uhl C, Vieira ICG. Ecological impacts of selective logging in the Brazilian Amazon - a case study from the Paragominas Region of the State of Pará. Biotropica 1989; 21(2): 98-106. http://dx.doi.org/10.2307/2388700. ERTS Symposium Abstracts; 1973; Washington. Greenbelt, MD: NASA; 1973. p. 309-317. Ekena Rangel Pinagé Shimabukuro YE, Smith JA. The least-squares mixing models to generate fraction images derived from remote sensing multispectral data. International Journal of Remote Sensing 1991; 29: 16-20. Wang C, Qi J, Cochrane M. Assessment of Tropical Forest degradation with canopy fractional cover from landsat ETM+ and IKONOS imagery. Earth Interactions 2005; 9(22): 1-18. http://dx.doi.org/10.1175/EI133.1.
https://openalex.org/W3012845130	https://europepmc.org/articles/pmc7082934?pdf=render	English	null	Comparison of RNA isolation methods on RNA-Seq: implications for differential expression and meta-analyses	BMC genomics	2,020	cc-by	7,487	© The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Scholes and Lewis BMC Genomics (2020) 21:249 https://doi.org/10.1186/s12864-020-6673-2 Scholes and Lewis BMC Genomics (2020) 21:249 https://doi.org/10.1186/s12864-020-6673-2 Open Access Abstract Background: The increasing number of transcriptomic datasets has allowed for meta-analyses, which can be valuable due to their increased statistical power. However, meta-analyses can be confounded by so-called “batch effects,” where technical variation across different batches of RNA-seq experiments can clearly produce spurious signals of differential expression and reduce our power to detect true differences. While batch effects can sometimes be accounted for, albeit with caveats, a better strategy is to understand their sources to better avoid them. In this study, we examined the effects of RNA isolation method as a possible source of batch effects in RNA- seq design. Results: Based on the different chemistries of “classic” hot phenol extraction of RNA compared to common commercial RNA isolation kits, we hypothesized that specific mRNAs may be preferentially extracted depending upon method, which could masquerade as differential expression in downstream RNA-seq analyses. We tested this hypothesis using the Saccharomyces cerevisiae heat shock response as a well-validated environmental response. Comparing technical replicates that only differed in RNA isolation method, we found over one thousand transcripts that appeared “differentially” expressed when comparing hot phenol extraction with the two kits. Strikingly, transcripts with higher abundance in the phenol-extracted samples were enriched for membrane proteins, suggesting that indeed the chemistry of hot phenol extraction better solubilizes those species of mRNA. Conclusions: Within a self-contained experimental batch (e.g. control versus treatment), the method of RNA isolation had little effect on the ability to identify differentially expressed transcripts. However, we suggest that researchers performing meta-analyses across different experimental batches strongly consider the RNA isolation methods for each experiment. Keywords: RNA-seq, RNA isolation, Batch effects, Meta-analysis, Transcriptomics Comparison of RNA isolation methods on RNA-Seq: implications for differential expression and meta-analyses Amanda N. Scholes1,2 and Jeffrey A. Lewis1* Background transcriptomic datasets need to be cautious in their use and aware of so-called “batch effects,” where technical differences between experimental batches can clearly produce spurious signals of differential expression and reduce our power to detect true differences. The decreasing cost of massively parallel sequencing has led to an explosion of transcriptomic datasets. This large number of datasets has allowed for meta-analyses, which can be valuable due to their increase in statistical power. However, researchers performing meta-analyses on In some cases the sources of batch effects are known and can be avoided. Some well-known batch effects in- clude sequencing lane effects, library construction proto- col, and RNA quality [1–3]. Other sources of batch effects clearly exist but remain unknown. While batch USA Full list of author information is available at the end of the article * Correspondence: lewisja@uark.edu 1Department of Biological Sciences, University of Arkansas, Fayetteville, AR, p j @ 1Department of Biological Sciences, University of Arkansas, Fayetteville, AR, USA Full list of author information is available at the end of the article Scholes and Lewis BMC Genomics (2020) 21:249 Page 2 of 9 Page 2 of 9 zol methods. Based on the combined chemistries of so- dium dodecyl sulfate (SDS) and phenol on cellular mem- branes [8, 9], we hypothesized that the Phenol method would better solubilize membrane-associated mRNAs. To test this hypothesis, and whether the choice of RNA isolation method had downstream effects on our ability to detect differentially expression transcripts, we collected four biological replicates of the model yeast Saccharomy- ces cerevisiae before and after a 20-min heat shock. Im- portantly, each biological sample was split into three identical technical replicates that differed only in their mode of RNA isolation. This allowed us to systematically test whether the RNA isolation method affects relative transcript abundance between technical replicates, and whether that matters for differential expression analysis. effects can sometimes be accounted for, this comes with some major caveats. If the batch effect completely confounds the experimental design, for example with different sequencing lanes being used for controls and treatments, statistically accounting for the batch effect will remove any “real” signal [4]. Even in the case where the batch effect is not a complete confounder, account- ing for batch can reduce our power to detect true bio- logical signal [5]. Thus, a better understanding of the sources of batch effects can help us to avoid them. In this study, we examined the effects of RNA isolation method as a possible source of batch effects in RNA-seq design. It is well known that the RNA distribution within cells is not uniform. Newly synthesized pre-mRNAs are processed in the nucleus before being exported. Once exported, mRNAs are frequently trafficked to specific subcellular sites as a mechanism for spatially controlling protein synthesis. Indeed, perhaps the most widespread example of mRNA localization is that used for spatial control of protein synthesis, where mRNAs encoding se- creted and membrane proteins are translated at the ER membrane allowing for proper protein localization and folding [6]. Our analysis found a striking number of transcripts (nearly 1/3 of the genome) that appeared “differentially” expressed when comparing the Phenol method to either Kit method, and a small number of differences when comparing the Kit methods to each other. Transcripts over-represented by Phenol extraction compared to ei- ther Kit were enriched for membrane proteins, suggest- ing that indeed the combination of SDS plus phenol better extracts those species of mRNA. Importantly, there were virtually no differences when comparing dif- ferential expression for the heat shock response within samples where RNA was isolated via same method. Based on these results, we strongly recommend that meta-analyses be performed on groups of experiments with common RNA isolation methods. Despite the widespread acknowledgement that mRNAs are differentially localized within the cell, there has been a paucity of studies examining whether “common” RNA extraction methods are equivalent in their abilities to ex- tract differentially localized RNA species, and whether the method of RNA isolation affects our ability to detect differentially expressed transcripts. Sultan and colleagues compared two RNA isolation methods (Qiagen RNeasy kit and guanidinium-phenol (TRIzol) extraction) and two library selection schemes (poly-A enrichment and rRNA depletion) on downstream transcript abundance estimates, and found that rRNA depletion was particu- larly sensitive to the RNA extraction method [2]. How- ever, their comparisons were done using only two biological replicates, and they only examined transcript abundance across technical replicates and not whether the method of extraction affects the ability to detect dif- ferential expression in the types of sample comparisons that biologists frequently care about (e.g. wild-type ver- sus mutant or treatment versus control). Experimental setup To test whether RNA extraction methods impact between- sample comparisons and the power to identify differentially expressed genes, we used the well-characterized yeast heat shock response as an environmental perturbation. We col- lected four biological replicates for comparison. For each biological replicate, three “technical replicate” samples were collected to understand the impact of RNA extraction method. The only difference was that each technical repli- cate had their RNA extracted by one of three methods: classic hot acid phenol (Phenol Method), a silica-based col- umn kit (RNeasy Method) and a guanidinium-phenol (TRI- zol)-based kit (Direct-zol Method) (Fig. 1). RNA isolated via the Phenol method was subsequently “cleaned” with a Qiagen RNeasy Kit using the optional on-column DNase treatment, thus controlling for both DNase treatment and potential differential binding of different RNA species to the column. To minimize against batch effects other than RNA extraction method, all RNA-seq libraries were con- structed on the same day using an automated robotic plat- form, and all libraries were multiplexed and sequenced on a single lane of an Illumina HiSeq4000 instrument. Thus, we sought to systematically examine whether three common RNA isolation methods led to differences in transcript abundance and/or our ability to detect dif- ferential expression between two experimental condi- tions in the form of the Saccharomyces cerevisiae heat shock response. The different RNA isolation methods were the classic “hot acid phenol” method, and the two most commonly-used types of kits [7]—a silica-based column kit (Qiagen RNeasy Kit) and a guanidinium- phenol (TRIzol)-based kit (Zymo Research Direct-zol), hereafter referred to as the Phenol, RNeasy, and Direct- Scholes and Lewis BMC Genomics (2020) 21:249 Page 3 of 9 Page 3 of 9 Scholes and Lewis BMC Genomics Fig. 1 Schematic of the experimental design. Yeast cells were grown to mid-exponential phase at 30 °C, unstressed control samples were collected, and then cells were shifted to a 37 °C heat shock with samples collected after 20 min. For both unstressed and stressed cells, we collected three identical samples (technical replicates), and RNA was isolated using either hot acid phenol extraction, a Qiagen RNeasy Kit, or a Zymo Research Direct-zol RNA Miniprep Kit. Experimental setup Libraries were constructed in a single batch using a liquid handling robot, and then were pooled and sequenced on a single Illumina HiSeq4000 lane We were particular interested in whether differences in the RNA isolation method could masquerade as “dif- ferential” expression due to differences in transcript quantification. We first performed principal component analysis (PCA) (Fig. 2). Not surprisingly, a substantial proportion of the variance (50.5%) was explained by treatment (unstressed versus heat shock). The second principal component corresponded to RNA isolation method and explained 26.9% of the variation. Samples with RNA isolated by the two different kit methods clus- tered together, with the Phenol-isolated samples forming a separate cluster. It could seem counterintuitive that Direct-zol and Phenol methods would be so dissimilar, considering that both methods use phenol. However, the Direct-zol method uses a milder detergent than SDS (sarkosyl), is performed at room temperatures instead of 65 °C, and samples are exposed to phenol for 10 min in- stead of 45 min. We speculate that these differences with the Phenol method result in both silica-column-based kits behaving similarly (see Discussion and conclusions). To visualize differences in transcript abundance across RNA isolation methods, we performed hierarchical clus- tering on the TPMs of the unstressed samples (Fig. 3a). Hierarchical clustering of the samples largely recapitu- lated the patterns of PCA—again, the Phenol-isolated samples formed a discreet cluster distinct from the two kits. The RNeasy- and Direct-zol-isolated samples also had far fewer visible differences. To quantify these differ- ences, we used edgeR to identify transcripts with signifi- cantly differential abundance in pairwise comparisons of each RNA isolation method (FDR < 0.01, see Methods). Pairwise comparisons of the Phenol method with each Kit method identified a large number of transcripts with differential abundance: 2430 transcripts (Phenol vs. RNeasy) and 2512 transcripts (Phenol vs. Direct-zol), which we validated with qPCR for representative tran- scripts with differential abundance (Supplementary Fig. 1). Of those transcripts with differential abundance in both comparisons, 1917 overlapped, which was highly significant (P = 1 × 10−520, Fisher’s exact test) (Fig. 3c). In contrast, only 230 transcripts had differential abun- dance when comparing the kits to each other, suggesting only slight differences. Fig. 1 Schematic of the experimental design. Yeast cells were grown to mid-exponential phase at 30 °C, unstressed control samples were collected, and then cells were shifted to a 37 °C heat shock with samples collected after 20 min. Experimental setup For both unstressed and stressed cells, we collected three identical samples (technical replicates), and RNA was isolated using either hot acid phenol extraction, a Qiagen RNeasy Kit, or a Zymo Research Direct-zol RNA Miniprep Kit. Libraries were constructed in a single batch using a liquid handling robot, and then were pooled and sequenced on a single Illumina HiSeq4000 lane Differences in relative transcript abundance between phenol-extracted RNA and kit-extracted RNA Looking more closely at the cellular component GO enrichments, transcripts with higher abundance in the phenol samples were strongly enriched for those encoding intrinsic membrane proteins (P < 4 × 10−191), as well as proteins localized to the endoplasmic reticulum (P < 6 × 10−84), cell periphery (P < 3 × 10−80), and the vacuole (P < 3 × 10−53). In contrast, mRNAs with lower relative abundance in the phenol samples were enriched for nuclear in localization (P < 3 × 10−60), and in- cluded those encoding functions related to nucleic acid me- tabolism (P < 1 × 10−38), RNA metabolism (P < 6 × 10−28), chromosome organization (P < 4 × 10−17), and gene expres- sion (P < 8 × 10−17). Notably, the transcripts that appeared “repressed” in the Phenol samples compared to both Kits also had significantly lower expression relative to the genomic average (60.7 TPMs vs. 170.3 TPMs; P < 3 × 10−24, t-test). We hypothesize that lowly expressed transcripts are more sensitive to appearing spuriously “repressed” (see Discussion and conclusions). also examined the relationship between differential abundance and direct estimates of in vivo transcript sta- bility (half-lives) from Neymotin and colleagues [10]. We did find a significant difference in the Phenol vs. Direct- zol comparison, but not for the Phenol vs. RNeasy com- parison. To determine how much of the variation was explained by GC content, gene length, and transcript half-life, we performed linear regression of those param- eters on the average fold changes for phenol-extracted samples vs. the kits. Both GC content and transcript length showed weak to moderate correlation (r = 0.06– 0.32) with log2 fold changes, depending upon the com- parison group, while estimated in vivo half-life weakly correlated with log2 fold changes in either comparison (Supplementary Tables 3 and 4). Because differences in GC content and length are associated with differences in transcript degradation rates in vitro [3], we repeated the edgeR analysis using RIN as a factor. We expected that because the RIN values for the Direct-zol samples were all lower than the others, using RIN as a covariate would eliminate most of the signal for differential expression. This turned out to be correct—we identified 788 “differ- entially” expressed genes in the Phenol vs. Direct-zol comparison compared to 2513 when RIN was not in- cluded as a factor. Differences in relative transcript abundance between phenol-extracted RNA and kit-extracted RNA All of the RNA isolation methods yielded generally high quality RNA, as defined by a RIN of 9.0 or above, though the phenol extracted RNA averaged significantly higher RIN values than those isolated from the Direct-zol kit (9.96 vs. 9.33; P = 2 × 10−6, t-test) or the RNeasy kit (9.96 vs. 9.79; P = 0.01, t-test) (Supplementary Table 1). The percentage of total mapped reads was similar across sam- ples, with slight (though significant) differences (Supple- mentary Table 2). There were larger differences in the percentage of uniquely mapped reads across RNA isola- tion methods (Supplementary Table 2). These differences did not correlate with RNA integrity, as the Direct-zol samples had the lowest RIN values and highest uniquely and total mapped reads. Overall, we feel that both the RNA quality and read mapping would not raise any red flags in laboratories performing RNA-seq on either their own samples, or conducting a meta-analysis, though those values can be used a factor to be controlled for in differen- tial expression analysis [3]. To better visualize these differences, we performed hierarchical clustering on all 3127 transcripts with sig- nificantly differential abundance (FDR < 0.01) in any pairwise comparison of RNA isolation method (Fig. 3b). We found striking functional gene ontology (GO) en- richments for transcripts with higher or lower abun- dance in the phenol-extracted samples compared to both kits. Transcripts with higher abundance in phenol- extracted RNA in comparison to both kits were strongly enriched for transmembrane transport (P < 4 × 10−68), establishment of localization (P < 9 × 10−54), lipid Scholes and Lewis BMC Genomics (2020) 21:249 Page 4 of 9 Fig. 2 Principal component analysis (PCA) strongly implicates RNA isolation method as a batch effect. PCA on TPMs for each sample (see Methods) shows clear separation on both treatment (PC1) and RNA isolation method (PC2). Kit samples were more similar to each other than they were to the Phenol samples Fig. 2 Principal component analysis (PCA) strongly implicates RNA isolation method as a batch effect. PCA on TPMs for each sample (see Methods) shows clear separation on both treatment (PC1) and RNA isolation method (PC2). Kit samples were more similar to each other than they were to the Phenol samples metabolism (P < 1 × 10−27), and cell wall organization (P < 1 × 10−18). Differences in relative transcript abundance between phenol-extracted RNA and kit-extracted RNA Because the RNeasy-isolated samples had relatively high RIN values relative to the Direct-zol- isolated samples, the vast majority of transcripts with differential expression were retained as significant when accounting for RIN in the edgeR QL model (2362 / 2430). With RIN as a covariate, we repeated the analyses looking at relationships between “differential” expression and GC content, gene length, and in vivo transcript sta- bility (Supplementary Fig. 3 and Supplementary Ta- bles 3). The correlations between log2 fold changes and GC content or in vivo transcript stability were essentially Properties of transcripts with spurious differential expression That Phenol-isolated samples have higher transcript abundance for mRNAs encoding membrane proteins fits with the hypothesis that the Phenol method better solu- bilizes that species of mRNA. Another possibility is that differences in transcript degradation rates are respon- sible for the spurious patterns of differential expression. Because GC content and transcript length correlate with in vivo mRNA degradation rates [3], we examined those relationships in our data. Transcripts with significantly higher or lower abundance in Phenol-extracted samples compared to each Kit method had significantly higher GC content and gene length (Supplementary Fig. 2). We Page 5 of 9 Scholes and Lewis BMC Genomics (2020) 21:249 Fig. 3 Phenol preferentially extracts mRNAs that encode for membrane proteins. a Hierarchical clustering of unstressed samples (P = Phenol, R = RNeasy, D = Direct-zol). Clustering on relative transcript abundance (TPMs) reveals differences depending upon RNA isolation method, while clustering on sample identity shows that the Phenol method diverges from both Kits. Red indicates higher than average transcript abundance within a sample, and blue indicates lower than average transcript abundance. b Hierarchical clustering of 3127 transcripts with significantly differential abundance (FDR < 0.01) in any pairwise comparisons between each RNA isolation method. Brown indicates higher expression than the comparison group (e.g. Phenol in the P v. R column) and violet indicates lower expression than the comparison group (e.g. RNeasy in the P v. R column). Enriched Gene Ontology (GO) categories (Bonferroni-corrected P < 0.01) are shown on the right. Complete GO enrichments for each cluster can be found in Supplementary File 3. c Overlap between transcripts with significantly differential abundance (FDR < 0.01) in the Phenol v. RNeasy and Phenol v. Direct-zol comparisons Fig. 3 Phenol preferentially extracts mRNAs that encode for membrane proteins. a Hierarchical clustering of unstressed samples (P = Phenol, R = RNeasy, D = Direct-zol). Clustering on relative transcript abundance (TPMs) reveals differences depending upon RNA isolation method, while clustering on sample identity shows that the Phenol method diverges from both Kits. Red indicates higher than average transcript abundance within a sample, and blue indicates lower than average transcript abundance. b Hierarchical clustering of 3127 transcripts with significantly differential abundance (FDR < 0.01) in any pairwise comparisons between each RNA isolation method. Brown indicates higher expression than the comparison group (e.g. Phenol in the P v. R column) and violet indicates lower expression than the comparison group (e.g. Properties of transcripts with spurious differential expression RNeasy in the P v. R column). Enriched Gene Ontology (GO) categories (Bonferroni-corrected P < 0.01) are shown on the right. Complete GO enrichments for each cluster can be found in Supplementary File 3. c Overlap between transcripts with significantly differential abundance (FDR < 0.01) in the Phenol v. RNeasy and Phenol v. Direct-zol comparisons extraction are responsible for the batch effect, and not RNA degradation (see Discussion and conclusions). eliminated for the Phenol vs. Direct-zol comparison, while the correlation with length was slightly reduced. Thus, it is possible that some subset of spurious differ- ential expression is due to differences in RNA degrad- ation rates. However, the surviving differentially expressed transcripts with higher expression in the Phenol-isolated samples relative to the Direct-zol iso- lated samples were still strongly enriched for those en- coding intrinsic membrane proteins (P < 3 × 10−100). Moreover, because of the substantial overlap between genes called as differentially expressed in the Phenol vs. RNeasy and Phenol vs. Direct-zol comparisons, we hypothesize that the differing chemistries in the Differences in RNA isolation method have little effect on the ability to detect differential expression with a batch Discussion and conclusions In this study, we tested whether differences in RNA iso- lation method affect relative transcript abundance be- tween samples, and whether the RNA isolation method impacts our ability to detect differential expression. Our results suggest that differences in RNA isolation method can substantially affect relative transcript abundance, and we saw thousands of differences in transcript abun- dance when comparing hot acid phenol extraction with an RNeasy or Direct-zol kit. It is well established that mRNAs encoding membrane and secreted proteins are anchored to the membrane during translation [11]. That transcripts with higher abundance in the Phenol-isolated samples are strongly enriched for encoding membrane proteins suggests the Phenol method better solubilizes those mRNAs. Because relatively more membrane- associated mRNAs are being extracted, there must be relatively less abundance of other mRNAs. Thus, we see decreased abundance of certain nuclear transcripts, which were already more lowly expressed, and thus likely more sensitive to appearing “repressed.” Fig. 4 The method of RNA extraction has little effect on differential expression analysis. Hierarchical clustering of median-centered log2- fold TPM changes for 4232 transcripts that were differentially expressed in response to heat (FDR < 0.01) in at least one set of samples (P = phenol, R = RNeasy, D = Direct-zol). The left portion of the heat map displays gene expression changes during heat shock across the four biological replicates, with red indicating genes induced by heat shock, and blue indicating genes repressed by heat shock. The right portion shows differences in abundance in pairwise comparisons between each RNA isolation method, with brown indicating higher expression than the comparison group, and violet indicating lower expression than the comparison group. The Venn Diagram depicts overlap between differentially expressed genes in the Phenol, RNeasy, and Direct-zol isolated samples We disfavor the alternative hypothesis that we are cap- turing differences in transcript degradation rates for a number of reasons. First, while we do see differences in RIN values across the different RNA isolation methods, the differences are relatively small, and our RIN values are all much higher than the points where other studies identified them as confounding RNA-seq analysis [3, 12]. Differences in RNA isolation method have little effect on the ability to detect differential expression with a batch The striking differences in transcript abundance depend- ing on RNA isolation could conceivably affect the ability to detect differential expression. To test this, we exam- ined our ability to detect differential expression in cells shifted from 30 °C to 37 °C for 20 min—the classic yeast heat shock response. We identified ~ 3800 differentially expressed transcripts for all three RNA isolation methods, with substantial overlap for all three (Fig. 4). Page 6 of 9 Scholes and Lewis BMC Genomics (2020) 21:249 Page 6 of 9 Scholes and Lewis BMC Genomics Fig. 4 The method of RNA extraction has little effect on differential expression analysis. Hierarchical clustering of median-centered log2- fold TPM changes for 4232 transcripts that were differentially expressed in response to heat (FDR < 0.01) in at least one set of samples (P = phenol, R = RNeasy, D = Direct-zol). The left portion of the heat map displays gene expression changes during heat shock across the four biological replicates, with red indicating genes induced by heat shock, and blue indicating genes repressed by heat shock. The right portion shows differences in abundance in pairwise comparisons between each RNA isolation method, with brown indicating higher expression than the comparison group, and violet indicating lower expression than the comparison group. The Venn Diagram depicts overlap between differentially expressed genes in h h l d l l d l One possibility is that the transcripts encoding the processes most affected by differences in the extraction methods (i.e. membrane-associated proteins) are lowly represented during the heat shock response, and thus the method of RNA isolation could affect differential expression under different conditions. To test this, we examined whether genes with the GO component term of “membrane” were less likely to be differentially expressed during heat shock. Genes encoding membrane proteins comprise 28.7% of the yeast genome, which turned out to be statistically equivalent (via TOST, see Methods) to their proportion within the heat shock re- sponse (28.4, 28.7, and 28.3% for RNAs extracted by Phenol, RNeasy, and Direct-zol, respectively). Thus, we hypothesize that at sufficient sequencing depth, the abil- ity to detect differential expression is robust to the mod- est differences in transcript counts caused by differences in RNA isolation method. Discussion and conclusions Second, it is likely that any degradation that is occurring in our samples is happening in vitro during RNA isola- tion, and Opitz and colleagues have found that in vitro RNA degradation rates are likely relatively equal across transcripts and thus have little effect on differential ex- pression analysis [13]. And while RNA degradation rates in vivo are strongly biased and can lead to spurious Hierarchical clustering yielded no clear pattern among differentially expressed transcripts that were missed in sample set over another (Fig. 4). We also detected zero transcripts that had significant fold change differences in their heat shock response in any pairwise comparison between RNA isolation methods (Supplementary File 2). Page 7 of 9 Page 7 of 9 Page 7 of 9 Page 7 of 9 Scholes and Lewis BMC Genomics (2020) 21:249 Scholes and Lewis BMC Genomics (2020) 21:249 functional enrichments in downstream analysis, we found little relationship between estimated mRNA half- lives from [10] and fold-changes in comparisons between kits. Only one of the Phenol vs. Kit comparisons showed a significant difference in half-lives, but the correlation was still rather poor (r2 = 0.02). And while transcripts with higher relative abundance in the phenol-extracted samples versus the kits had higher GC content and gene length, which both correlate with higher in vivo degrad- ation rates [3], the correlation between those parameters and fold-change differences was not strong (Supplemen- tary Table 3). Notably, GC content and gene length are not random, and membrane proteins tend to be longer and have higher GC content than average [14, 15]. Fi- nally, if RNA degradation is responsible, it is somewhat hard to reconcile that we see similar patterns of “differ- ential” expression when comparing the Phenol vs. Direct-zol or RNeasy kits, even though the RNeasy kits have quite a bit higher RIN values. RNA isolation with two different Miniprep kits RNA isolation with two different Miniprep kits RNA was extracted using two different kits: the Qiagen RNeasy Mini Kit (Cat. 74,104) and the Zymo Research Direct-zol RNA Miniprep Kit (Cat. R2050). Cell concen- trations were all below the maximum recommendation of 5 × 107 cells from both manufacturers (ranging from 2.5 × 107–4.5 × 107 cells). For both kits, we mechanically lysed cells with a Beadbeater-24 (3500 oscillations/mi- nute, 45 s on ice between cycles). Mechanical lysis was performed in 2-ml screw-capped tubes containing an equal volume (600 μl) of lysis buffer (RLT for RNeasy or TRI reagent for Direct-zol) and acid-washed glass beads (425–600 μm, Sigma-Aldrich). Yeast growth and sampling procedures RNA was then purified according to each manufac- turer’s protocol for yeast, including the optional on- column DNase digestion. For all samples, RNA was quantitated using a Qubit RNA HS Assay kit and Qubit fluorometer according to the manufacturer’s instruc- tions. The RNA integrity number (RIN) for each sample was measured using an Agilent 2200 TapeStation. RNA concentrations and RIN values for each sample can be found in Supplementary Table 1. g g All experiments were performed using yeast strain BY4741 (S288c background; MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0), obtained from Open Biosystems. To compare RNA isolation methods, we collected three identical 10-ml ‘technical’ replicates for each biological replicate (4 biological replicates in total). Cells were grown > 8 generations in 100-ml synthetic complete medium (SC) [16] at 30 °C with orbital shaking (270 rpm) to mid-exponential phase (OD600 of 0.3–0.6), and 10-ml samples were removed representing the un- stressed control. For heat shock treatment, one volume of 55 °C medium was added to the remaining culture, immediately bringing the final temperature to 37 °C, and the culture was incubated at 37 °C for another 20 min before removing 10-ml samples. Both unstressed and heat shocked cells were collected by centrifugation at 1500 x g for 3 min, and cell pellets were flash frozen in liquid nitrogen and stored at −80 °C until processing. RNA isolation methods Hot phenol isolation RNA isolation methods Hot phenol isolation Cells were lysed and RNA was isolated using a standard hot phenol method as described [17], and a detailed proto- col can be found on the protocols.io repository under DOI dx.doi.org/10.17504/protocols.io.inwcdfe. Briefly, 1 volume of acid saturated phenol and 1 volume of lysis buffer (10 mM Tris-HCl pH 7.4, 10 mM EDTA, 0.5% SDS) were added to frozen cell pellets, vortexed, and then placed in a 65 °C preheated Multi-Therm incubated vortexer (Bench- mark Scientific) at 1500 rpm for 45 min. Samples were cen- trifuged for 10 min at 4 °C at maximum speed in a microcentrifuge, extracted once more with phenol, once with chloroform, and then precipitated overnight at − 20 °C with 0.1 volumes of sodium acetate (pH 5.2) and 2.5 volumes of 100% ethanol. Precipitated RNA was washed once with 70% ethanol and then resuspended in TE (10 mM Tris-HCl pH 8.0, 1 mM EDTA). The phenol extracted RNA was then ‘cleaned’ using an RNeasy Miniprep Kit with optional on-column DNase treatment according to the manufacturer’s instructions. Regardless of the cause of these differences between hot-phenol extracted samples and kits, it clear that this can represent a large source of batch-effect variation be- tween samples whose RNA has been isolated via differ- ent methods. Within an individual lab, we are largely agnostic. The method of RNA isolation had little effect on the ability to identify differentially expressed tran- scripts in our heat shock test case. Thus, experiments within a single lab are unlikely to be affected by the choice of RNA isolation method as long as the same method is used throughout an experiment. For meta- analyses however, we recommend that researchers avoid comparing experiments where the RNA isolation methods differ. Quantitative PCR h Six genes with either significantly higher (LAS17, SED1, PRY3) or lower (JNM1, EAF7, RRP36) abundance in the Phenol vs. Kit RNA-seq data were validated by real-time quantitative (q) PCR using the Maxima SYBR-qPCR Mas- ter Mix (Thermo Fisher Scientific) and a Bio-Rad CFX96 Touch Real-Time PCR Detection System as described [25]. Briefly, cDNA was synthesized from the same RNA samples used for RNA-seq using 10 μg total RNA, 3 μg an- chored oligo-dT (T20VN), and SuperScript III (Thermo Fisher Scientific) according to the manufacturer’s instruc- tions. One ng of cDNA was used as template, and qPCR was performed using the following thermal cycling param- eters: 95 °C for 3 min, 40 cycles 95 °C for 15 s, 55 °C for 1 min for 40 cycles, followed by a melt curve analysis to val- idate the presence of only a single amplicon. Cq values were determined using regression analysis, with baseline subtraction via curve fit. Relative abundance between sam- ples was determined using the ΔΔCt method [26], by nor- malizing to ERV25 as control gene whose expression is unaffected by various stresses [27]. A detailed protocol is available on protocols.io under DOI dx.doi.org/10.17504/ protocols.io.bbgpijvn. Reads were trimmed of low-quality reads and adapter sequence (KAPA v1 indices) using Trimmomatic (version 0.32) [18], with the following commands: ILLUMINA- CLIP:Kapa_indices.fa:2:30:10 LEADING:3 TRAILING:3 MAXINFO:40:0.4 MINLEN:40. Reads were mapped to the S288c genome (version Scer3), using STAR (version 020201) [19]. Mapping statistics can be found in Supple- mentary Table 2. Transcripts per million (TPM) and ex- pected counts for each gene were calculated using RSEM (version 1.3.1) [20]. The RSEM output can be found in Additional File 2. Differential expression analysis was conducted using the Bioconductor package edgeR (version 3.22.3) using the quasi-likelihood (QL) framework. For the QL model, sample type (i.e. Phenol unstressed, Phenol heat shock, RNeasy unstressed …) and biological replicate were used as factors. To account for differences in RIN across sam- ples, we also performed a separate analysis that included sample type, replicate, and RIN as factors in the model. To control for differences in sequencing depth across samples, the edgeR function thincounts was used to ran- domly subsample counts across all samples to be equal to the sample with the lowest number of total counts (8, 678,188). Only genes with at least 1 count per million (CPM) in at least one condition were included for TMM normalization and differential expression analysis. RNA sequencing and analysis RNA-seq libraries were prepared from polyA-enriched RNA using the KAPA Biosystems mRNA HyperPrep Kit (KK8581) and KAPA Single-Indexed Adapter Set A + B (KK8700), according to manufacturer’s instructions. We started with 500 ng total RNA, fragmentation time (6 min) was optimized to generate 200–300-nt RNA frag- ments, and the libraries were amplified with 9 cycles of PCR. All libraries were constructed in a single batch Scholes and Lewis BMC Genomics (2020) 21:249 Page 8 of 9 equivalency bounds of +/−0.029 (10% of the background proportion). through an automated Eppendorf epMotion 5075 liquid handling robot, and a detailed a protocol can be found on protocols.io under DOI dx.doi.org/10.17504/proto- cols.io.uueewte. cDNA libraries were sequenced on a HiSeq4000 at the University of Chicago Genomics Facil- ity, generating single-end 50-bp reads. Abbreviations bp: Base pairs; cDNA: Complementary DNA; Ct: Critical threshold; DNase: deoxyribonuclease; DOI: Digital object identifier; edgeR: Empirical analysis of Digital Gene Expression in R; ER: Endoplasmic reticulum; FDR: False discovery rate; GEO: Gene Expression Omnibus; GO: Gene Ontology; PCA: Principal component analysis; poly-A: polyadenylation; QL: Quasi-likelihood; qPCR: Quantitative polymerase chain reaction; RIN: RNA integrity number; RNA-seq: RNA sequencing; RSEM: RNA-Seq by Expectation- Maximization; SC: Synthetic complete media; SDS: Sodium dodecyl sulfate; STAR: Spliced Transcripts Alignment to a Reference; TOST: Two One-Sided Tests; TOSTER: Two One-Sided Tests (TOST) Equivalence Testing; TPM: Transcripts per million Supplementary information l f pp y Supplementary information accompanies this paper at https://doi.org/10. 1186/s12864-020-6673-2. Additional file 1: Figure S1. qPCR validation of representative differentially abundant transcripts. Figure S2. Properties of transcripts with differential abundance depending upon RNA isolation method. Figure S3. Properties of transcripts with differential abundance when RIN is included as a factor. Table S1. RNA concentrations and integrity (RIN) values. Table S2. Summary of mapping statistics. Table S3. Correlation coefficient (r) for log2 fold changes versus each factor. Table 4. Log2 fold-abundance changes versus each factor. Additional file 2. RSEM Output. Additional file 3. EdgeR Output. Additional file 4. GO enrichments for clusters in Fig. 3. Additional file 1: Figure S1. qPCR validation of representative differentially abundant transcripts. Figure S2. Properties of transcripts with differential abundance depending upon RNA isolation method. Figure S3. Properties of transcripts with differential abundance when RIN is included as a factor. Table S1. RNA concentrations and integrity (RIN) values. Table S2. Summary of mapping statistics. Table S3. Correlation coefficient (r) for log2 fold changes versus each factor. Table 4. Log2 fold-abundance changes versus each factor. Additional file 2. RSEM Output. Additional file 3. EdgeR Output. Additional file 4. GO enrichments for clusters in Fig. 3. p Principal component analysis (PCA) was performed using ClustVis [21] on ln-transformed TPM values for all transcripts included in the differential expression ana- lysis, using unit variance scaling and singular value de- composition. Hierarchical clustering was performed with Cluster 3.0 (http://bonsai.hgc.jp/~mdehoon/software/ cluster/software.htm) using uncentered Pearson correl- ation and centroid linkage as the metric [22]. RNA-seq samples were weighted using a cutoff value of 0.4 and an exponent value of 1. Functional enrichments of gene ontology (GO) categories were performed using GO- TermFinder (https://go.princeton.edu/cgi-bin/GOTerm- Finder) [23], with Bonferroni-corrected P-values < 0.01 taken as significant. Complete lists of enriched categories can be found in Additional File 4. Equivalency testing on proportions of gene sets was performed using two one- sided tests (TOST) through the TOSTER R package (version 0.3.4) [24], using an alpha level of 0.05, and Additional file 4. GO enrichments for clusters in Fig. 3. Quantitative PCR h All RNA-seq data are available through the National Insti- tutes of Health Gene Expression Omnibus (GEO) data- base under accession no. GSE135430, and the edgeR outputs can be found in Additional File 3. References 1. Li S, Labaj PP, Zumbo P, Sykacek P, Shi W, Shi L, et al. Detecting and correcting systematic variation in large-scale RNA sequencing data. Nat Biotechnol. 2014;32(9):888–95. https://doi.org/10.1038/nbt.3000 PMID: 25150837. 24. Lakens D. Equivalence Tests. Soc Psychol Personal Sci. 2017;8(4):355–62. https://doi.org/10.1177/1948550617697177 PMID. 24. Lakens D. Equivalence Tests. Soc Psychol Personal Sci. 2017;8(4):355–62. https://doi.org/10.1177/1948550617697177 PMID. 2. Sultan M, Amstislavskiy V, Risch T, Schuette M, Dokel S, Ralser M, et al. Influence of RNA extraction methods and library selection schemes on RNA- seq data. BMC Genomics. 2014;15:675. https://doi.org/10.1186/1471-2164-15- 675 PMID: 25113896. 2. Sultan M, Amstislavskiy V, Risch T, Schuette M, Dokel S, Ralser M, et al. Influence of RNA extraction methods and library selection schemes on RNA- seq data. BMC Genomics. 2014;15:675. https://doi.org/10.1186/1471-2164-15- 675 PMID: 25113896. 25. Stuecker TN, Scholes AN, Lewis JA. Linkage mapping of yeast cross protection connects gene expression variation to a higher-order organismal trait. PLoS Genet. 2018;14(4):e1007335. https://doi.org/10.1371/journal.pgen. 1007335 PMID: 29649251. 3. Gallego Romero I, Pai AA, Tung J, Gilad Y. RNA-seq: impact of RNA degradation on transcript quantification. BMC Biol. 2014;12:42. https://doi. org/10.1186/1741-7007-12-42 PMID: 24885439. 3. Gallego Romero I, Pai AA, Tung J, Gilad Y. RNA-seq: impact of RNA degradation on transcript quantification. BMC Biol. 2014;12:42. https://doi. org/10.1186/1741-7007-12-42 PMID: 24885439. 26. Livak KJ, Schmittgen TD. Analysis of Relative Gene Expression Data Using Real-Time Quantitative PCR and the 2−ΔΔCT Method. Methods. 2001;25(4): 402–8. https://doi.org/10.1006/meth.2001.1262 PMID: 11846609. 26. Livak KJ, Schmittgen TD. Analysis of Relative Gene Expression Data Using Real-Time Quantitative PCR and the 2−ΔΔCT Method. Methods. 2001;25(4): 402–8. https://doi.org/10.1006/meth.2001.1262 PMID: 11846609. 4. Gilad Y, Mizrahi-Man O. A reanalysis of mouse ENCODE comparative gene expression data. F1000Res. 2015;4:121. https://doi.org/10.12688/ f1000research.6536.1 PMID: 26236466. 4. Gilad Y, Mizrahi-Man O. A reanalysis of mouse ENCODE comparative gene expression data. F1000Res. 2015;4:121. https://doi.org/10.12688/ f1000research.6536.1 PMID: 26236466. 27. Berry DB, Gasch AP. Stress-activated genomic expression changes serve a preparative role for impending stress in yeast. Mol Biol Cell. 2008;19(11): 4580–7. https://doi.org/10.1091/mbc E07-07-0680 PMID: 18753408. 27. Berry DB, Gasch AP. Stress-activated genomic expression changes serve a preparative role for impending stress in yeast. Mol Biol Cell. 2008;19(11): 4580–7. https://doi.org/10.1091/mbc E07-07-0680 PMID: 18753408. 5. Goh WWB, Wang W, Wong L. Why Batch Effects Matter in Omics Data, and How to Avoid Them. Trends Biotechnol. 2017;35(6):498–507. https://doi.org/ 10.1016/j.tibtech.2017.02.012 PMID: 28351613. 5. Goh WWB, Wang W, Wong L. Why Batch Effects Matter in Omics Data, and How to Avoid Them. Trends Biotechnol. 2017;35(6):498–507. https://doi.org/ 10.1016/j.tibtech.2017.02.012 PMID: 28351613. Acknowledgements We thank Zymo Research for providing a free sample of the Direct-zol RNA Kit, Dr. Elizabeth Ruck for running samples on the TapeStation, and Dr. An- drew Alverson for the use of equipment. Page 9 of 9 Page 9 of 9 Page 9 of 9 Scholes and Lewis BMC Genomics (2020) 21:249 Scholes and Lewis BMC Genomics (2020) 21:249 Funding h k This work was supported in part by National Science Foundation Grant No. IOS-1656602 (JAL), the Arkansas Biosciences Institute (Arkansas Settlement Proceeds Act of 2000) (JAL), and a Research Assistantship provided through the University of Arkansas Cell and Molecular Biology Graduate Program (ANS). The funders had no role in study design, data collection and interpret- ation, or the decision to submit the work for publication. 13. Opitz L, Salinas-Riester G, Grade M, Jung K, Jo P, Emons G, et al. Impact of RNA degradation on gene expression profiling. BMC Med Genomics. 2010;3: 36. https://doi.org/10.1186/1755-8794-3-36 PMID: 20696062. 14. D'Onofrio G, Jabbari K, Musto H, Bernardi G. The correlation of protein hydropathy with the base composition of coding sequences. Gene. 1999; 238(1):3–14. https://doi.org/10.1016/s0378-1119(99)00257-7 PMID: 10570978. 15. Young MD, Wakefield MJ, Smyth GK, Oshlack A. Gene ontology analysis for RNA-seq: accounting for selection bias. Genome Biol. 2010;11(2):R14. https:// doi.org/10.1186/gb-2010-11-2-r14 PMID: 20132535. Received: 17 September 2019 Accepted: 13 March 2020 23. Boyle EI, Weng S, Gollub J, Jin H, Botstein D, Cherry JM, et al. GO:: TermFinder--open source software for accessing Gene Ontology information and finding significantly enriched Gene Ontology terms associated with a list of genes. Bioinformatics. 2004;20(18):3710–5. https:// doi.org/10.1093/bioinformatics/bth456 PMID: 15297299. Authors’ contributions 11. Chartron JW, Hunt KC, Frydman J. Cotranslational signal-independent SRP preloading during membrane targeting. Nature. 2016;536(7615):224–8. https://doi.org/10.1038/nature19309 PMID: 27487213. A.N.S. performed experiments and analyses. J.A.L. conceived and designed the project, and performed experiments and analyses. A.N.S. and J.A.L. wrote the paper. The author(s) have read and approved this manuscript. 12. Chen EA, Souaiaia T, Herstein JS, Evgrafov OV, Spitsyna VN, Rebolini DF, et al. Effect of RNA integrity on uniquely mapped reads in RNA-Seq. BMC Res Notes. 2014;7:753. https://doi.org/10.1186/1756-0500-7-753 PMID: 25339126. Competing interests The authors declare that they have no competing interests. 21. Metsalu T, Vilo J. ClustVis: a web tool for visualizing clustering of multivariate data using Principal Component Analysis and heatmap. Nucleic Acids Res. 2015;43(W1):W566–70. https://doi.org/10.1093/nar/gkv468 PMID: 25969447. Author details 1 1Department of Biological Sciences, University of Arkansas, Fayetteville, AR, USA. 2Interdisciplinary Graduate Program in Cell and Molecular Biology, University of Arkansas, Fayetteville, AR, USA. 1Department of Biological Sciences, University of Arkansas, Fayetteville, AR, USA. 2Interdisciplinary Graduate Program in Cell and Molecular Biology, University of Arkansas, Fayetteville, AR, USA. 22. Eisen MB, Spellman PT, Brown PO, Botstein D. Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci U S A. 1998;95:14863– 8 PMID: 9843981. Received: 17 September 2019 Accepted: 13 March 2020 Ethics approval and consent to participate Not applicable. Ethics approval and consent to participate Not applicable. 18. Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014;30(15):2114–20. https://doi.org/10.1093/ bioinformatics/btu170 PMID: 24695404. Consent for publication Not applicable. 19. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29(1):15–21. https:// doi.org/10.1093/bioinformatics/bts635 PMID: 23104886. 20. Li B, Dewey CN. RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics. 2011;12:323. https://doi.org/10.1186/1471-2105-12-323 PMID: 21816040. Competing interests h h d l h Competing interests The authors declare that they have no competing interests. Availability of data and materials ll d l bl h All RNA-seq data are available through the National Institutes of Health Gene Expression Omnibus (GEO) database under accession no. GSE135430. The analyses generated during this study are included in the supplementary in- formation files. 16. Sherman F. Getting started with yeast. Methods Enzymol. 2002;350:3–41 PMID: 12073320. 17. Gasch AP. Yeast genomic expression studies using DNA microarrays. Methods Enzymol. 2002;350:393–414 doi: PMID: 12073326. Methods Enzymol. 2002;350:393–414 doi: PMID: 12073326. Publisher’s Note 6. Lerner RS, Seiser RM, Zheng T, Lager PJ, Reedy MC, Keene JD, et al. Partitioning and translation of mRNAs encoding soluble proteins on membrane-bound ribosomes. RNA. 2003;9(9):1123–237. https://doi.org/10. 1261/rna.5610403 PMID: 12923260. 6. Lerner RS, Seiser RM, Zheng T, Lager PJ, Reedy MC, Keene JD, et al. Partitioning and translation of mRNAs encoding soluble proteins on membrane-bound ribosomes. RNA. 2003;9(9):1123–237. https://doi.org/10. 1261/rna.5610403 PMID: 12923260. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 7. Johnson M. Kits for RNA Extraction, Isolation, and Purification. Materials Methods. 2012;2:201. https://doi.org/10.13070/mm.en.2.201 PMID:. 8. Kragh-Hansen U, le Maire M, Moller JV. The mechanism of detergent solubilization of liposomes and protein-containing membranes. Biophys J. 1998;75(6):2932–46. https://doi.org/10.1016/S0006-3495(98)77735-5 PMID: 9826614. 9. Davidson PM, Branden AL. Antimicrobial Activity of Non-Halogenated Phenolic Compounds. J Food Prot. 1981;44(8):623–32. https://doi.org/10. 4315/0362-028X-44.8.623 PMID: 30836539. 10. Neymotin B, Athanasiadou R, Gresham D. Determination of in vivo RNA kinetics using RATE-seq. RNA. 2014;20(10):1645–52. https://doi.org/10.1261/ rna.045104.114 PMID: 25161313.
https://openalex.org/W1544797161	https://hal.inria.fr/hal-01402042/document	English	null	Parametric LTL on Markov Chains	Lecture notes in computer science	2,014	cc-by	9,290	To cite this version: Souymodip Chakraborty, Joost-Pieter Katoen. Parametric LTL on Markov Chains. 8th IFIP Inter- national Conference on Theoretical Computer Science (TCS), Sep 2014, Rome, Italy. pp.207-221, 10.1007/978-3-662-44602-7_17. hal-01402042 Distributed under a Creative Commons Attribution 4.0 International License Souymodip Chakraborty and Joost-Pieter Katoen⋆ RWTH Aachen University, Ahornstraße 55, D-52074 Aachen, Germany Abstract. This paper is concerned with the veriﬁcation of ﬁnite Markov chains against parametrized LTL (pLTL) formulas. In pLTL, the until- modality is equipped with a bound that contains variables; e.g., ♦⩽x ϕ asserts that ϕ holds within x time steps, where x is a variable on natural numbers. The central problem studied in this paper is to determine the set of parameter valuations V≺p(ϕ) for which the probability to satisfy pLTL-formula ϕ in a Markov chain meets a given threshold ≺p, where ≺is a comparison on reals and p a probability. As for pLTL determin- ing the emptiness of V>0(ϕ) is undecidable, we consider several logic fragments. We consider parametric reachability properties, a sub-logic of pLTL restricted to next and ♦⩽x, parametric B¨uchi properties and ﬁnally, a maximal subclass of pLTL for which emptiness of V>0(ϕ) is decidable. ⋆Currently on sabbatical leave at the University of Oxford, United Kingdom. HAL Id: hal-01402042 https://inria.hal.science/hal-01402042v1 Submitted on 24 Nov 2016 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License 1 Introduction Verifying a ﬁnite Markov chain (MC, for short) M against an LTL-formula ϕ amounts to determining the probability that M satisﬁes ϕ, i.e., the likelihood of the set of inﬁnite paths of M satisfying ϕ. Vardi [1] considered the qualita- tive version of this problem, that is, does M almost surely satisfy ϕ, or with positive probability. Together with Wolper, he showed that the qualitative LTL model-checking problem for MCs is PSPACE-complete. The quantitative veri- ﬁcation problem – what is the probability of satisfying ϕ? – has been treated by Courcoubetis and Yannakakis [2]. An alternative algorithm that has a time complexity which is polynomial in the size of the MC and exponential in \|ϕ\| is by Couvreur et al. [3]. Recently, practical improvements have been obtained by Chatterjee et al. for verifying the LTL(F,G)-fragment on MCs using generalized deterministic Rabin automata [4]. This paper considers the veriﬁcation of MCs against parametric LTL formu- las. In parametric LTL [5] (pLTL, for short), temporal operators can be sub- scripted by a variable ranging over the natural numbers. The formula ♦⩽x a means that in at most x steps a occurs, and □♦⩽y a means that at every index a occurs within y steps. Note that x and y are variables whose value is not ﬁxed in advance. The central question is now to determine the values of x and y such that the probability of a given MC satisfying the pLTL-formula ϕ meets a cer- tain threshold p. This is referred to as the valuation set V≺p(ϕ) for comparison 2 operator ≺. This problem has both a qualitative (threshold > 0 and = 1) and a quantitative variant (0 < p < 1). operator ≺. This problem has both a qualitative (threshold > 0 and = 1) and a quantitative variant (0 < p < 1). The main results of this paper are as follows. Just as for the setting with Kripke structures [5], it is shown that checking the emptiness of V>0(ϕ) in gen- eral is undecidable. We therefore resort to fragments of pLTL. We show that determining V⩾p(♦⩽x a) can be done by searching in a range deﬁned by the precision of the input, whereas polynomial time graph algorithms suﬃce for its qualitative variant. The same applies to formulas of the form □♦⩽x a. 1 Introduction We pro- vide necessary and suﬃcient criteria for checking the emptiness of V>0(ϕ) (and V=1(ϕ)) for the fragments pLTL(F,X) and pLTL♦, and prove that checking these criteria are NP-complete and PSPACE-complete, respectively. We also deﬁne a representation of these sets and provide algorithms to construct them. Related work. The veriﬁcation of parametric probabilistic models in which cer- tain transition probabilities are given as parameters (or functions thereof) has recently received considerable attention. Most of these works are focused on pa- rameter synthesis: for which parameter instances does a given (LTL or PCTL) formula hold? To mention a few, Han et al. [6] considered this problem for timed reachability in continuous-time MCs, Hahn et al. [7] and Pugelli et al. [8] for Markov decision processes (MDPs), and Benedikt et al. [9] for ω-regular prop- erties of interval MCs. Hahn et al. [10] provide an algorithm for computing the rational function expressing the probability of reaching a given set of states in a parametric (reward) MDP based on exploiting regular expressions as initially proposed by Daws [11]. Other related work includes the synthesis of loop invari- ants for parametric probabilistic programs [12]. To the best of our knowledge, verifying parametric properties on MCs has not been considered so far. The closest related works are on combining two-variable FO with LTL for MDPs by Benedikt et al. [13] and the computation of quantiles by Ummels and Baier [14]. Organization of the paper. Section 2 presents pLTL and MCs and a ﬁrst un- decidability result. Section 3 considers parametric reachability. Section 4 treats the fragment pLTL(F,X) and Section 5 parametric B¨uchi properties. Section 6 treats the bounded always-free fragment of pLTL. Section 7 concludes the paper. Full version of the paper can be found in the archive. 2 Preliminaries Parametric LTL. Parametric LTL extends propositional LTL with bounded tem- poral modalities, for which the bound is either a constant or a variable. Let Var be a ﬁnite set of variables ranged over by x, y, and AP be a ﬁnite set of propo- sitions ranged over by a and b. Let c ∈IN. Parametric LTL formulas adhere to the following syntax: ϕ ::= a \| ¬ϕ \| ϕ ∧ϕ \| ⃝ϕ \| ϕ U ϕ \| ♦≺x ϕ \| ♦≺c ϕ where ≺∈{ =, ⩽, <, >, ⩾}. A pLTL structure is a triple (w, i, v) where w ∈Σω with Σ = 2AP is an inﬁnite word over sets of propositions, i ∈IN is an index, and 3 v : Var →IN is a variable valuation. Analogously, we consider a valuation v as a vector in Nd, where d for pLTL formula ϕ is the number of variables occurring in ϕ. E.g. for d = 1, the valuation is just a number v. We compare valuations v and v ′ as v ⩽v ′ iﬀv(x) ⩽v ′(x) for all x. Let w[i] denote the i-th element of w. The satisfaction relation \|= is deﬁned by structural induction over ϕ as follows: (w, i, v) \|= a iﬀa ∈w[i] (w, i, v) \|= ¬ ϕ iﬀ(w, i, v) ̸\|= ϕ (w, i, v) \|= ϕ1 ∧ϕ2 iﬀ(w, i, v) \|= ϕ1 and (w, i, v) \|= ϕ2 (w, i, v) \|= ♦≺x ϕ iﬀ(w, j, v) \|= ϕ for some j ≺v(x)+i. (w, i, v) \|= a iﬀa ∈w[i] (w, i, v) \|= ¬ ϕ iﬀ(w, i, v) ̸\|= ϕ (w, i, v) \|= ϕ1 ∧ϕ2 iﬀ(w, i, v) \|= ϕ1 and (w, i, v) \|= ϕ2 (w, i, v) \|= ♦≺x ϕ iﬀ(w, j, v) \|= ϕ for some j ≺v(x)+i. For the sake of brevity, we have omitted the semantics of the standard LTL modalities. As usual, ϕ1 R ϕ2 ≡¬(¬ϕ1 U ¬ϕ2), ♦ϕ ≡true U ϕ and □ϕ ≡¬♦¬ϕ. The language of ϕ is deﬁned by L(ϕ) = {(w, v) \| (w, 0, v) \|= ϕ}. Alur et al. [5] have shown that other modalities such as U⩽x, ♦>x, □>x, U>x, R ⩽x and R >x, can all be encoded in our syntax. 2 Preliminaries In particular, we will focus on the decidability and complexity of the emptiness problem for V≺p(ϕ), i.e., the decision problem whether V≺p(ϕ) = ∅ or not, on algorithms (if any) determining the set V≺p(ϕ), and on the size of the minimal representation of V≺p(ϕ). In the qualitative setting, the bound ≺p is either > 0, or = 1. Proposition 1. For ϕ ∈pLTL, the problem if V>0(ϕ) = ∅is undecidable. Proposition 1. For ϕ ∈pLTL, the problem if V>0(ϕ) = ∅is undecidable. Proof. The proof is based on [5, Th. 4.1], see the archived version for details. ■ It follows that deciding whether V=1(ϕ) = ∅is undecidable, as V>0(ϕ) = ∅iﬀ V=1(¬ϕ) ̸= ∅. As a combination of ♦⩽x and □⩽x modalities can encode U=x, e.g., ¬a ∧⃝(¬a U=x a) ≡⃝(¬a U⩽x a) ∧(¬a U>x a), we will restrict ourselves to fragments of pLTL where each formula is in negative normal form and the only parametrized operator is ♦⩽x ϕ. We refer to this fragment as pLTL♦: we will restrict ourselves to fragments of pLTL where each formula is in negative normal form and the only parametrized operator is ♦⩽x ϕ. We refer to this fragment as pLTL♦: ϕ ::= a \| ¬a \| ϕ∧ϕ \| ϕ∨ϕ \| ⃝ϕ \| ϕ U ϕ \| ϕ R ϕ \| □ϕ \| ♦⩽x ϕ \| ♦⩽c ϕ \| □⩽c ϕ. (2) ϕ ::= a \| ¬a \| ϕ∧ϕ \| ϕ∨ϕ \| ⃝ϕ \| ϕ U ϕ \| ϕ R ϕ \| □ϕ \| ♦⩽x ϕ \| ♦⩽c ϕ \| □⩽c ϕ. (2) We show it is a sub-logic of pLTL for which the emptiness problem for V>0(ϕ) is decidable. The logic has a favourable monotonicity property, i.e., Remark 1. For every pLTL♦-formula ϕ, inﬁnite word w and valuations v, v ′, v ⩽v ′ implies (w, v) \|= ϕ =⇒(w, v ′) \|= ϕ. Here (w, v) \|= ϕ is s shorthand for (w, 0, v) \|= ϕ. We start oﬀwith brieﬂy consid- ering (only) parametric eventualities and then consider the sub-logic pLTL(F,X) restricted to next and ♦x. Later on, we also consider parametric B¨uchi formulas, and ﬁnally, pLTL♦. 2 Preliminaries For instance, the following equivalences hold: ♦>x ϕ ≡□⩽x ♦⃝ϕ, □>x ϕ ≡♦⩽x □⃝ϕ, ϕ U⩽x ψ ≡(ϕ U ψ) ∧♦⩽x ψ, ϕ U>x ψ ≡□⩽x (ϕ ∧⃝(ϕ U ψ)) (1) (1) In the remainder of this paper, we focus on bounded always and eventualities where all bounds are upper bounds. We abbreviate ♦⩽x by ♦x and do similar for the other modalities. For valuation v and pLTL-formula ϕ, let v(ϕ) denote the LTL formula obtained from ϕ by replacing variable x by its valuation v(x); e.g., v(♦x ϕ) equals ♦v(x) v(ϕ). Markov chains. A discrete-time Markov chain M is a quadruple (S, P, s0, L) where S is a ﬁnite set of states with m = \|S\|, P : S × S →[0, 1] is a stochastic matrix, s0 ∈S an initial state, and L : S →2AP a state-labeling function. P(u, v) denotes the one-step probability of moving from state u to v. A trajectory (or path) of a Markov chain (MC, for short) M is a sequence { si }i⩾0 such that P(si, si+1) > 0 for all i ⩾0. A trajectory π = s0s1s2 . . . induces the trace trace(π) = L(s0)L(s1)L(s2) . . .. Let Paths(M) denote the set of paths of MC M. A path π satisﬁes the pLTL-formula ϕ under the valuation v, denoted π \|= v(ϕ), whenever (trace(π), 0, v) \|= ϕ (or equivalently, (trace(π), v) ∈L(ϕ)). A ﬁnite path (or path fragment) satisﬁes a formula under a valuation if any inﬁnite extension of it also satisﬁes the formula. Let Pr be the probability measure on sets of paths, deﬁned by a standard cylinder construction [1]. The probability of satisfying ϕ by M under valuation v is given by Pr{ π ∈Paths(M) \| π \|= v(ϕ) }, generally abbreviated as Pr(M \|= v(ϕ)). Valuation set. The central problem addressed in this paper is to determine the valuation set of a pLTL formula ϕ. Let M be an MC, p ∈[0, 1] a probability bound, and ≺∈{ =, ⩽, <, >, ⩾}. Then we are interested in determining: V≺p(ϕ) = { v \| Pr(M \|= v(ϕ)) ≺p }, 4 i.e., the set of valuations under which the probability of satisfying ϕ meets the bound ≺p. 3 Parametric Reachability In this section, we consider pLTL-formulas of the form ♦x a for proposition a, or equivalently, ♦x T for the set of target states T = { s ∈S \| a ∈L(s) }. We consider bounds of the form ⩾p with 0 < p < 1. The valuation set of interest is thus V⩾p(♦x a). Let µi be the probability of reaching T within i steps; the sequence {µi} is ascending. There can be two cases: (a) the sequence reaches a constant value in m steps (m being the size of Markov chain) or (b) the sequence monotonically increases and converges to µ∞. This makes the emptiness problem for V⩾p(♦x a) decidable. In the ﬁrst case, we check µm ⩾p. In the second case, emptiness is decidable in time polynomial in m, by determining µ∞= Pr(♦a) which can be done by solving a system of linear equations with at most m variables. Then, V⩾p(♦x a) ̸= ∅iﬀp < µ∞. 5 Assume in the sequel that T is non-empty. Let min V⩾p(♦x a) = n0. The valuation set can thus be represented by n0 (this gives a minimal representation of the set). Membership queries, i.e., does n ∈V⩾p(♦x a), then simply boil down to checking whether n0 ⩽n, which can be done in constant time (modulo the size of n0). The only catch is that n0 can be very large if p is close to µ∞. A simple example elucidates this fact. Example 1. Consider the MC M with S = { s0, t }, L(t) = { a }, L(s0) = ∅, P(s0, s0) = 1 2 = P(s0, t) and P(t, t) = 1. Then Pr(M \|= ♦n a) = 1 − 1 2 n. It follows that min V⩾p(♦x a) goes to inﬁnity when p approaches one. The following bound on n0 can nonetheless be provided. This bound allows for obtaining the minimum value n0 by a binary search. Proposition 2. For MC M, min V⩾p(♦xa) ⩽logγ(1−(1−γ) p b ), where 0 < γ < 1 and b > 0. Proof. Collapse all a-states into a single state t and make it absorbing (i.e., replace all outgoing transitions by a self-loop with probability one). Let t be the only bottom strongly connected component (BSCC) of M (other BSCCs can be safely ignored). Let {1, . . . Proof. The problem is the same as reachability in directed graphs. Proposition 4. The sets V>0(♦x a) and V=1(♦x a) can be determined in poly- nomial time by a graph analysis of MC M. Proposition 4. The sets V>0(♦x a) and V=1(♦x a) can be determined in poly- nomial time by a graph analysis of MC M. Proof. Collapse all the a-states into a target state t and make t absorbing. If V>0(♦x a) is non-empty, it suﬃces to determine min V>0(♦x a) which equals the length of a shortest path from s0 to t. To determine whether V=1(♦x a) is empty or not, we proceed as follows. If a cycle without t is reachable from s0, then no ﬁnite n exists for which the probability of reaching t within n steps equals one. Thus, V=1(♦x a) = ∅. If this is not the case, then the graph of M is a DAG (apart from the self-loop at t), and min V=1(♦x a) equals the length of a longest path from s0 to t. ■ 4 The Fragment pLTL(F,X) This section considers the fragment pLTL(F,X) which is deﬁned by: ϕ ::= a \| ¬a \| ϕ ∧ϕ \| ϕ ∨ϕ \| ⃝ϕ \| ♦ϕ \| ♦⩽x ϕ \| ♦⩽c ϕ Our ﬁrst result is a necessary and suﬃcient condition for the emptiness of V>0(ϕ). Theorem 1. For ϕ ∈pLTL(F,X) and MC M with m states, V>0(ϕ) ̸= ∅iﬀ¯v ∈ V>0(ϕ) with ¯v(x) = m·\|ϕ\|. Theorem 1. For ϕ ∈pLTL(F,X) and MC M with m states, V>0(ϕ) ̸= ∅iﬀ¯v ∈ V>0(ϕ) with ¯v(x) = m·\|ϕ\|. Proof. Let ϕ be a pLTL(F,X)-formula and assume V>0(ϕ) ̸= ∅. By monotonicity, it suﬃces to prove that v ∈V>0(ϕ) with v ̸⩽¯v implies ¯v ∈V>0(ϕ). The proof proceeds in a number of steps. (1) We show that it suﬃces to consider formulas without disjunction. (2) We show that if path fragment π[0..l] \|= ¯ϕ, (where LTL(F,X)-formula ¯ϕ is obtained from ϕ by omitting all parameters from ϕ) then π[0..l] \|= v l(ϕ) with v l(x) = l for every x. (3) We construct a deterministic B¨uchi automaton (DBA) A ¯ϕ for ¯ϕ such that its initial and ﬁnal state are at most \| ¯ϕ\| transitions apart. (4) We show that reachability of a ﬁnal state in the product of MC M and DBA A ¯ϕ implies the existence of a ﬁnite path in M of length at most m·\|ϕ\| satisfying ¯ϕ. 1. As disjunction distributes over ∧,⃝ ⃝ ⃝, ♦, and ♦x, each formula can be written in disjunctive normal form. Let ϕ ≡ϕ1∨. . .∨ϕk, where each ϕi is disjunction- free. Evidently, \|ϕi\| ⩽\|ϕ\|. Assume v ∈V>0(ϕ). Then, v ∈V>0(ϕi) for some 0 < i ⩽k. Assuming the theorem holds for ϕi (this will be proven below), ¯v i ∈V>0(ϕi) with ¯v i(x) = \|ϕi\|·m. Since ¯v ⩾¯v i, it follows by monotonicity that ¯v ∈V>0(ϕi), and hence, ¯v ∈V>0(ϕ). It thus suﬃces in the remainder of the proof to consider disjunction-free formulas. 1. As disjunction distributes over ∧,⃝ ⃝ ⃝, ♦, and ♦x, each formula can be written in disjunctive normal form. Let ϕ ≡ϕ1∨. . .∨ϕk, where each ϕi is disjunction- free. Evidently, \|ϕi\| ⩽\|ϕ\|. Assume v ∈V>0(ϕ). Then, v ∈V>0(ϕi) for some 0 < i ⩽k. Assuming the theorem holds for ϕi (this will be proven below), ¯v i ∈V>0(ϕi) with ¯v i(x) = \|ϕi\|·m. Since ¯v ⩾¯v i, it follows by monotonicity that ¯v ∈V>0(ϕi), and hence, ¯v ∈V>0(ϕ). 3 Parametric Reachability , m} be the states of the modiﬁed MC M, with the initial state s0 and the target state t represented by 1 and m, respectively. Let Q be the (m−1) × (m−1) transition matrix of the modiﬁed MC without the state t. That is, Q(i, j) = P(i, j) iﬀj ̸= m where P is the transition probability matrix of M. We have the following observation: 1. Let the coeﬃcient of ergodicity τ(Q) of Q deﬁned as τ(Q) = 1 −min i,j X k min{Q(i, k), Q(j, k)} ! . As Q is sub-stochastic and no row of Q is zero, it follows 0 < τ(Q) < 1. T As Q is sub-stochastic and no row of Q is zero, it follows 0 < τ(Q) < 1. T 2. Let vector rT = (r1, . . . , rm−1) with ri = P(i, m), rmax be the maximum element in r and iT be (1, 0, . . . , 0). The probability of reaching the state m from the state 1 in at most n+1 steps is the probability of being in some state i < m within n steps and taking the next transition to m: µn+1 = n+1 X j=0 iT Qjr ⩽ n+1 X j=0 τ(Q)jrmax. Let τ(Q) = γ and rmax = b. The integer n0 is the smallest integer such that µn0 ⩾p, which implies that b· 1−γn0 1−γ ⩾p. This yields n0 ⩽logγ(1 −(1 −γ) p b ). ■ Let τ(Q) = γ and rmax = b. The integer n0 is the smallest integer such that µn0 ⩾p, which implies that b· 1−γn0 1−γ ⩾p. This yields n0 ⩽logγ(1 −(1 −γ) p b ). ■ As in the non-parametric setting, it follows that (for ﬁnite MCs) the valuation sets V>0(♦x a) and V=1(♦x a) can be determined by a graph analysis, i.e. no inspection of the transition probabilities is necessary for qualitative parametric reachability properties. 6 Proposition 3. The problem V>0(♦x a) = ∅is NL-complete. Proposition 3. The problem V>0(♦x a) = ∅is NL-complete. Proposition 3. The problem V>0(♦x a) = ∅is NL-complete. ■ Proof. The problem is the same as reachability in directed graphs. 4 The Fragment pLTL(F,X) It thus suﬃces in the remainder of the proof to consider disjunction-free formulas. 7 2. For pLTL(F,X)-formula ϕ, let ¯ϕ be the LTL(F,X)-formula obtained from ϕ by replacing all occurrences of ♦x by ♦, e.g., for ϕ = ♦x(a ∧♦yb), ¯ϕ = ♦(a ∧♦b). We claim that π[0...l] \|= ¯ϕ implies π[0...l] \|= v l(ϕ) with v l(x) = l for all x. This is proven by induction on the structure of ϕ. The base cases a and ¬a are obvious. For the induction step, conjunctions, ⃝ϕ and ♦ϕ are straightforward. It remains to consider ♦x ϕ. Assume π[0...l] \|= ♦¯ϕ. Thus, for some i ⩽l, π[i...l] \|= ¯ϕ. By induction hypothesis, π[i...] \|= v il(ϕ) with v il(y) = l−i for each variable y in ϕ. Thus, π[0..l] \|= v l(♦x ϕ) with v l(x) = l and for all y in ϕ, v l(y) = l. y ϕ, l(y) 3. We provide a DBA A ¯ϕ = ⟨Q, Σ, δ, q0, F⟩with Σ = 2AP for each LTL(F,X)- formula ¯ϕ using the construction from [15]. We ﬁrst treat ¯ϕ = a and ¯ϕ = ♦a. As every LTL(F,X)-formula can be obtained from ♦(a∧ϕ), ϕ1∧ϕ2 and ⃝ϕ, we then treat these inductive cases. (Negations are treated similarly.) For ¯ϕ = a, Aa = ⟨{ q0, q1 }, Σ, δ, q0, { q1 }⟩with δ(q0, a) = q1 and δ(q1, true) = q1. For ¯ϕ = ♦a , the DBA A♦a = ⟨{ q0, q1 }, Σ, δ, q0, { q1 }⟩, where δ(q0, a) = q1, δ(q0, ¬a) = q0 and δ(q1, true) = q1. This completes the base cases. For the three inductive cases, the DBA is constructed as follows. (a) Let A ¯ϕ = ⟨Q, Σ, δ, q0, F⟩. A♦(a∧¯ϕ) = ⟨Q ∪{ q′ 0 }, Σ, δ′, q′ 0, F⟩where q′ 0 is fresh, δ′(q, ·) = δ(q, ·) if q ∈Q, δ′(q′ 0, a) = δ(q0, a), and δ′(q′ 0, ¬a) = q′ 0. (b) F ¯ ∧¯ th DBA i t d d h d t f th DBA (a) Let A ¯ϕ = ⟨Q, Σ, δ, q0, F⟩. A♦(a∧¯ϕ) = ⟨Q ∪{ q′ 0 }, Σ, δ′, q′ 0, F⟩where q′ 0 is fresh, δ′(q, ·) = δ(q, ·) if q ∈Q, δ′(q′ 0, a) = δ(q0, a), and δ′(q′ 0, ¬a) = q′ 0. 4 The Fragment pLTL(F,X) (b) For ¯ϕ1 ∧¯ϕ2, the DBA is a standard synchronous product of the DBA for ¯ϕ1 and ¯ϕ2 fresh, δ′(q, ·) = δ(q, ·) if q ∈Q, δ′(q′ 0, a) = δ(q0, a), and δ′(q′ 0, ¬a) = q′ 0. (b) For ¯ϕ1 ∧¯ϕ2, the DBA is a standard synchronous product of the DBA for ¯ϕ1 and ¯ϕ2. (c) Let A ¯ϕ = ⟨Q, Σ, δ, q0, F⟩. A⃝¯ϕ = ⟨Q ∪{ q′ 0 }, Σ, δ′, q′ 0, F⟩where q′ 0 is fresh, δ′(q′ 0, a) = q0 for all a ∈Σ and δ′(q, a) = δ(q, a) for every q ∈Q. A few remarks are in order. The resulting DBA have a single ﬁnal state. In addition, the DBA enjoy the property that the reﬂexive and transitive closure of the transition relation is a partial order [15]. Formally, q ⪯q′ iﬀ q′ ∈δ∗(q, w) for some w ∈Σω. The diameter of A ¯ϕ is the length of a longest simple path from the initial to the ﬁnal state. This implies that the diameter of A♦(a∧¯ϕ) and A⃝¯ϕ is n+1 where n is this diameter of A ¯ϕ, and the diameter of A ¯ϕ1∧¯ϕ2 is n1 + n2 where ni is the diameter of A ¯ϕi, i ∈{ 1, 2 }. ϕ1 ϕ2 ϕi { } 4. Let ϕ ≡ϕ1 ∨. . . ∨ϕk, where each ϕi is disjunction-free, with DBA A ¯ϕi. Evidently, V>0(ϕ) ̸= ∅iﬀV>0(ϕi) ̸= ∅for some disjunct ϕi. Consider the product of MC M and DBA A ¯ϕi, denoted M ⊗A ¯ϕi; see, e.g., [16, Def. 10.50]. By construction, M ⊗A ¯ϕi is partially ordered and has diameter at most m·\|ϕi\|. We have that Pr(M \|= ¯ϕi) > 0 iﬀan accepting state in M ⊗ A ¯ϕi is reachable. Thus, there exists a ﬁnite path π[0..m·\|ϕi\|] in M with π[0..m·\|ϕi\|] \|= ¯ϕ, or, π[0..m·\|ϕ\|] \|= ¯v(ϕ). This concludes the proof. M ⊗A ¯ϕi can also be used to show that, if we have a valuation v such that v(x) > m·\|ϕ\| and for all other variables y ̸= x, v(x) ⩽m·\|ϕ\| and v ∈V>0(ϕ) then v ′ ∈V>0(ϕ), where v ′(x) = m·\|ϕ\| and for y ̸= x, v ′(y) = v(y). The argument proceeds as induction on ¯ϕi. ■ The above Theorem 1 leads to the following proposition. The above Theorem 1 leads to the following proposition. Proposition 5. 4 The Fragment pLTL(F,X) For ϕ ∈pLTL(F,X), deciding if V>0(ϕ) = ∅is NP-complete. For almost sure properties, a similar approach as for V>0(ϕ) suﬃces. For almost sure properties, a similar approach as for V>0(ϕ) suﬃces. 8 Theorem 2.For ϕ ∈pLTL(F,X) and MC M with m states, V=1(ϕ) ̸= ∅iﬀ¯v ∈ V=1( ¯ϕ) with ¯v(x) = m·\|ϕ\|. Theorem 2.For ϕ ∈pLTL(F,X) and MC M with m states, V=1(ϕ) ̸= ∅iﬀ¯v ∈ V=1( ¯ϕ) with ¯v(x) = m·\|ϕ\|. Proof. Consider the direction from left to right. The argument goes along similar lines as the proof of Theorem 1. We build the DBA A ¯ϕ for ¯ϕ and take the cross product with Markov chain M. There are m·\|ϕ\| state in the cross product. If Pr(M \|= ¯v(ϕ)) < 1 then there is some cycle in the cross product that does not contain the ﬁnal state. Thus, V=1(ϕ) is empty. ■ Theorem 1 suggests that min V>0(ϕ) lies in the hyper-cube H = { 0, . . . , N }d, where N = m·\|ϕ\|. A possible way to ﬁnd min V>0(ϕ) is to apply the bisection method in d-dimensions. We recursively choose a middle point of the cube, say v ∈H —in the ﬁrst iteration v(x) = N/2— and divide H in 2d equally sized hypercubes. If v ∈V>0(ϕ), then the hypercube whose points exceed v is dis- carded, else the cube whose points are below v is discarded. The asymptotic time-complexity of this procedure is given by the recurrence relation: T(k) = (2d −1) · T(k·2−d) + F (3) (3) where k is the number of points in the hypercube and F is the complexity of checking v ∈V>0(ϕ) where \|v\| ⩽N. Section 6 presents an algorithm working in O(m·N d·2\|ϕ\|) for a somewhat more expressive logic. From (3), this yields a complexity of O(m·N d·2\|ϕ\|· log N). The size of a set of minimal points can be exponential in the number of variables, as shown below. Proposition 6. \| min V>0(ϕ)\| ⩽(N·d)d−1. Proposition 6. \| min V>0(ϕ)\| ⩽(N·d)d−1. Proposition 6. \| min V>0(ϕ)\| ⩽(N·d)d−1. 5 Qualitative Parametric B¨uchi In this section, we consider pLTL-formulas of the form ϕ = □♦x a, for proposi- tion a. We are interested in V>0(ϕ), i.e., does the set of inﬁnite paths visiting a-states that are maximally x apart inﬁnitely often, have a positive measure? Let MC M = (S, P, s0, L). A bottom strongly-connected component (BSCC) B ⊆S of M is a set of mutually reachable states with no edge leaving B. For BSCC B, let na,B = max{ \|π\| \| ∀i ⩽\|π\|, π[i] ∈B ∧a /∈L(π[i]) }. Proposition 8. Let B be a BSCC and s ∈B. Then, ∀n ∈N, n > na,B ⇔ Pr(s \|= □♦n a) = 1 and n ⩽na,B ⇔Pr(s \|= □♦n a) = 0. Proof. If n > na,B, then each path π from any state s ∈B will have at least one a-state in ﬁnite path fragment π[i, . . . , i+n] for all i. Hence, Pr(s \|= □♦n a) = 1. If n ⩽na,B, then there exists a ﬁnite path fragment ρ of B, such that, for all i ⩽n, a /∈L(ρ[i]). Consider an inﬁnite path π starting from any arbitrary s ∈B. As s ∈B, π will almost surely inﬁnitely often visit the initial state of ρ. Therefore, by [16, Th.10.25], π will almost surely visit every ﬁnite path fragment starting in that state, in particular ρ. Path π thus almost surely refutes □♦n a, i.e. Pr(s \|= □♦n a) = 0. ■ Proof. If n > na,B, then each path π from any state s ∈B will have at least one a-state in ﬁnite path fragment π[i, . . . , i+n] for all i. Hence, Pr(s \|= □♦n a) = 1. If n ⩽na,B, then there exists a ﬁnite path fragment ρ of B, such that, for all i ⩽n, a /∈L(ρ[i]). Consider an inﬁnite path π starting from any arbitrary s ∈B. As s ∈B, π will almost surely inﬁnitely often visit the initial state of ρ. Therefore, by [16, Th.10.25], π will almost surely visit every ﬁnite path fragment starting in that state, in particular ρ. Path π thus almost surely refutes □♦n a, i.e. Pr(s \|= □♦n a) = 0. ■ For any BSCC B and □♦x a, na,B < ∞iﬀevery cycle in B has at least one a-state. Hence, na,B can be obtained by analysing the digraph of B (in O(m2), the number of edges). 4 The Fragment pLTL(F,X) r b r b r b r b g x1 x2 x3 5 10 14 5 9 15 5 8 16 5 7 17 4 11 15 4 10 16 4 9 17 4 8 18 x1 x2 x3 3 10 16 3 11 17 3 10 18 3 9 19 2 13 17 2 12 18 2 11 19 2 10 20 Fig. 1. MC and min V>0(ϕ) for pLTL(F,X)-formula ϕ = ♦x1 r ∧♦x2 b ∧♦x3 g Fig. 1. MC and min V>0(ϕ) for pLTL(F,X)-formula ϕ = ♦x1 r ∧♦x2 b ∧♦x3 g Example 2. There exist MCs for which \| min V>0(ϕ)\| grows exponentially in d, the number of parameters in ϕ, whereas the number m of states in the MC grows linearly in d. For instance, consider the MC M in Fig. 1 and ϕ = ♦x1 r ∧♦x2 b ∧ ♦x3 g, i.e., d=3. We have \| min V>0(ϕ)\| = 42 as indicated in the table. Example 2. There exist MCs for which \| min V>0(ϕ)\| grows exponentially in d, the number of parameters in ϕ, whereas the number m of states in the MC grows linearly in d. For instance, consider the MC M in Fig. 1 and ϕ = ♦x1 r ∧♦x2 b ∧ ♦x3 g, i.e., d=3. We have \| min V>0(ϕ)\| = 42 as indicated in the table. We conclude this section by brieﬂy considering the membership query: does v ∈V>0(ϕ) for pLTL(F,X)-formula ϕ with d parameters? Checking membership We conclude this section by brieﬂy considering the membership query: does v ∈V>0(ϕ) for pLTL(F,X)-formula ϕ with d parameters? Checking membership 9 of a valuation v ∈V>0(ϕ) boils down to deciding whether there exists a v ′ ∈ min V>0(ϕ) such that v ⩾v ′. A representation of min V>0(ϕ) facilitating an eﬃcient membership test can be obtained by putting all elements in this set in lexicographical order. This involves sorting over all d coordinates. A membership query then amounts to a recursive binary search over d dimensions. This yields: Proposition 7. For pLTL(F,X)-formula ϕ, v ∈V>0(ϕ)? takes O(d· log N·d) time, provided a representation of min V>0(ϕ) is given. 1 The modalities ♦⩽c and □⩽c can be removed with only quadratic blow up. 5 Qualitative Parametric B¨uchi BSCC B is called accepting for □♦x a if na,B < ∞and B is reachable from the initial state s0. Note that this may diﬀer from being an accepting BSCC for □♦a. Evidently, V>0(□□□♦x a) ̸= ∅iﬀna,B < ∞. This result can be extended to generalized B¨uchi formula ϕ = □♦x1 a1 ∧. . . ∧□♦xd ad, by checking nai,B < ∞for each ai. , As a next problem, we determine min V>0(□♦x a). For the sake of simplicity, let MS M have a single accepting BSCC B. For states s and t in MC M, let d(s, t) be the distance from s to t in the graph of M. (Recall, the distance between state s and t is the length of the shortest path from s to t.) For BSCC B, let da,B(s) = mint∈B,a∈L(t) d(s, t), i.e., the minimal distance from s to an a-state in B. Let the proposition aB hold in state s iﬀs ∈B and a ∈L(s). Let Ga = (V, E) be the digraph deﬁned as follows: V contains all a-states of M and the initial state s0 and (s, s′) ∈E iﬀthere is path from s to s′ in M. Let c be a cost function deﬁned on a ﬁnite path s0 . . . sn in graph Ga as: c(s0 . . . sn) = maxi d(si, si+1), (d is deﬁned on the graph of M). Using these auxiliary notions we obtain the following characterization for min V>0(□♦x a): 10 π=s0...sn,sn\|=aB if na,B < da,B(s0) and n0 = na,B otherwise. Proof. We show for n ⩾n0, Pr(□♦n a) > 0, and for n < n0, Pr(□♦n a) = 0. Distinguish: if na,B < da,B(s0) and n0 = na,B otherwise. if na,B < da,B(s0) and n0 = na,B otherwise. if na,B < da,B(s0) and n0 = na,B otherwise. Proof. We show for n ⩾n0, Pr(□♦n a) > 0, and for n < n0, Pr(□♦n a) = 0. Distinguish: if na,B < da,B(s0) and n0 na,B otherwise. Proof. We show for n ⩾n0, Pr(□♦n a) > 0, and for n < n0, Pr(□♦n a) = 0. Distinguish: Proof. We show for n ⩾n0, Pr(□♦n a) > 0, and for n < n0, Pr(□♦n a) = 0. Distinguish: 1. na,B ⩾da,B(s0). Then, from s0 an a-state in B can be reached within na,B steps, i.e., Pr(s0 \|= ♦na,B aB) > 0. 5 Qualitative Parametric B¨uchi In the former case, it follows (as in the ﬁrst case) by Proposition 8 that Pr(s0 \|= □♦n a) = 0 for all n ⩾n0. Consider now n0 = cmin ⩾na,B. Let n < n0. By contra-position. Assume Pr(s0 \|= □♦n a) > 0. Let π = s0 . . . s1,a . . . s2,a . . . . . . sk,a be a ﬁnite path fragment in M where si,a \|= a and sk,a is the ﬁrst a-state along π which belongs to B. Then, by deﬁnition of the digraph Ga, the sequence π = s0s1,as2,a . . . sk,a is a path in Ga satisfying c(si,a, si+1,a) ⩽n for all 0 ⩽k < n. But then cmin ⩽n. Contradiction. ■ If MC M has more than one accepting BSCC, say { B1, . . . , Bk } with k > 1, then n0 = mini n0,Bi, where n0,Bi for 0 < i ⩽k is obtained as in Theorem 3. If MC M has more than one accepting BSCC, say { B1, . . . , Bk } with k > 1, then n0 = mini n0,Bi, where n0,Bi for 0 < i ⩽k is obtained as in Theorem 3. Proposition 9. The sets V>0(□♦x a) and V=1(□♦x a) can be determined in polynomial time by a graph analysis of MC M. Determining min V⩾p(□♦x a) for arbitrary p reduces to reachability of accepting BSCCs. In a similar way as for parametric reachability (cf. Section 3), this can be done searching. For generalized B¨uchi formula ϕ = □♦xi ai ∧. . . ∧□♦xd ad and BSCC B, naiB is at most m. Thus, min V>0(ϕ) ∈{ 0, . . . , m·d }d and can be found by the bisection method, similar to the procedure described in Section 4. 5 Qualitative Parametric B¨uchi For this aB-state, s, say, by Proposition 8 it follows Pr(s \|= □♦na,B a) = 1. Together this yields Pr(s0 \|= □♦n a) > 0 for each n ⩾na,B = n0. For n < n0 = na,B, it follows by Proposition 8 that Pr(s \|= □♦n a) = 0 for every aB-state s. Thus, Pr(s0 \|= □♦n a) = 0. ( \| ) ( \| ) 2. na,B < da,B(s0). As B is accepting, da,B(s0) ̸= ∞. Consider a simple path π from s0 to an a-state in B. Let c(π) be the maximal distance between two consecutive a-states along this path. Then it follows Pr(s0 \|= □♦k a) > 0 where k = max(c(π), na,B). By taking the minimum cmin over all simple paths between s0 and B, it follows Pr(s0 \|= □♦n a) > 0 for each n ⩾ n0 = max(na,B, cmin) with cmin = minπ∈Paths(s0,♦aB) c(π). For n < n0, distinguish between n0 = na,B and n0 = cmin. In the former case, it follows (as in the ﬁrst case) by Proposition 8 that Pr(s0 \|= □♦n a) = 0 for all n ⩾n0. Consider now n0 = cmin ⩾na,B. Let n < n0. By contra-position. Assume Pr(s0 \|= □♦n a) > 0. Let π = s0 . . . s1,a . . . s2,a . . . . . . sk,a be a ﬁnite path fragment in M where si,a \|= a and sk,a is the ﬁrst a-state along π which belongs to B. Then, by deﬁnition of the digraph Ga, the sequence π = s0s1,as2,a . . . sk,a is a path in Ga satisfying c(si,a, si+1,a) ⩽n for all 0 ⩽k < n. But then cmin ⩽n. Contradiction. ■ 2. na,B < da,B(s0). As B is accepting, da,B(s0) ̸= ∞. Consider a simple path π from s0 to an a-state in B. Let c(π) be the maximal distance between two consecutive a-states along this path. Then it follows Pr(s0 \|= □♦k a) > 0 where k = max(c(π), na,B). By taking the minimum cmin over all simple paths between s0 and B, it follows Pr(s0 \|= □♦n a) > 0 for each n ⩾ n0 = max(na,B, cmin) with cmin = minπ∈Paths(s0,♦aB) c(π). For n < n0, distinguish between n0 = na,B and n0 = cmin. 6 The Fragment pLTL♦ This section is concerned with the logical fragment pLTL♦, as deﬁned in (2): This section is concerned with the logical fragment pLTL♦, as deﬁned in (2): This section is concerned with the logical fragment pLTL♦, as deﬁned in (2): \| \| \| \| ⃝ \| U \| R \| □ \| ♦ 1 ϕ ::= a \| ¬a \| ϕ ∧ϕ \| ϕ ∨ϕ \| ⃝ϕ \| ϕ U ϕ \| ϕ R ϕ \| □ϕ \| ♦⩽x ϕ.1 ϕ ::= a \| ¬a \| ϕ ∧ϕ \| ϕ ∨ϕ \| ⃝ϕ \| ϕ U ϕ \| ϕ R ϕ \| □ϕ \| ♦⩽x ϕ.1 We will focus on the emptiness problem: is V>0(ϕ) = ∅. The decision problem whether V=1(ϕ) is very similar. Similar as for pLTL(F,X), we obtain necessary and suﬃcient criteria for both cases. The proofs for these criteria depend on an algorithm that checks whether v ∈V>0(ϕ). This algorithm is presented ﬁrst. 11 11 Automata constructions. Let ϕ be a pLTL♦-formula, and v a variable valuation. W.l.o.g. we assume that each variable occurs once in ϕ. We will extend the classical automaton-based approach for LTL by constructing a nondeterministic B¨uchi automaton for ϕ that is amenable to treat the variables occurring in ϕ. To that end, inspired by [17], we proceed in a number of steps: Automata constructions. Let ϕ be a pLTL♦-formula, and v a variable valuation. W.l.o.g. we assume that each variable occurs once in ϕ. We will extend the classical automaton-based approach for LTL by constructing a nondeterministic B¨uchi automaton for ϕ that is amenable to treat the variables occurring in ϕ. To that end, inspired by [17], we proceed in a number of steps: 1. Construct an automaton Gϕ for ϕ, independent from the valuation v, with two types of acceptance sets, one for treating until and release-modalities (as standard for LTL [18]), and one for treating the parameter constraints. 2. Establish how for a given valuation v, a B¨uchi automaton Bϕ(v) can be obtained from Gϕ such that for inﬁnite word w, (w, v) ∈L(ϕ) iﬀw is an accepting run of Bϕ(v). 3. Exploit the technique advocated by Couvreur et al. [3] to verify MC M versus Bϕ(v). We start with constructing Gϕ. Like for the LTL-approach, the ﬁrst step is to consider consistent sets of sub-formulas of ϕ. Let cl(ϕ) be the set of all sub- formulas of ϕ. Proposition 10 ([17]). For ϕ ∈pLTL♦, the automaton Gϕ is unambiguous and L(Gϕ) = L(ϕ). 6 The Fragment pLTL♦ Set H ⊆cl(ϕ) is consistent, when: – a ∈H iﬀ¬a ̸∈H, – ϕ2 ∈H implies ϕ1 U ϕ2 ∈H, – a ∈H iﬀ¬a ̸∈H, – ϕ1 ∧ϕ2 ∈H iﬀϕ1 ∈H and ϕ2 ∈H, – ϕ1, ϕ2 ∈H implies ϕ1 R ϕ2 ∈H, – ϕ1 ∨ϕ2 ∈H iﬀϕ1 ∈H or ϕ2 ∈H, – ϕ1 ∈H implies ♦x ϕ1 ∈H. We are now in a position to deﬁne Gϕ, an automaton with two acceptance sets. For ϕ ∈pLTL♦, let Gϕ = (Q, 2AP , Q0, δ, AccB, AccP ) where – Q is the set of all consistent sub-sets of cl(ϕ) and Q0 = { H ∈Q \| ϕ ∈H }. A – (H, a, H′) ∈δ, where a ∈2AP whenever: • H ∩AP = { a }, • ⃝ϕ1 ∈H ⇐⇒ϕ1 ∈H′, • ⃝ϕ1 ∈H ⇐⇒ϕ1 ∈H′, • ϕ1 U ϕ2 ∈H ⇐⇒ϕ2 ∈H or (ϕ1 ∈H and ϕ1 U ϕ2 ∈H′), ( ′) • ϕ1 U ϕ2 ∈H ⇐⇒ϕ2 ∈H or (ϕ1 ∈H and ϕ1 U ϕ2 ∈H′), • ♦x ϕ1 ∈H ⇐⇒ϕ1 ∈H or ♦x ϕ1 ∈H′, • ♦x ϕ1 ∈H ⇐⇒ϕ1 ∈H or ♦x ϕ1 ∈H′, • ♦x ϕ1 ∈H ⇐⇒ϕ1 ∈H or ♦x ϕ1 ∈H′, – (generalized) B¨uchi acceptance AccB and parametric acceptance AccP : • AccB = { Fϕ′ \| ϕ′ ∈cl(ϕ) ∧(ϕ′ = ϕ1 U ϕ2 ∨ϕ′ = ϕ1 R ϕ2) } where ϕ ∗Fϕ′ = { H \| ϕ2 ∈H ⇒ϕ′ ∈H } if ϕ′ = ϕ1 R ϕ • AccP = { Fxi \| ♦xi ϕi ∈cl(ϕ) } with Fxi = { H \| ♦xi ϕi ∈H ⇒ϕi ∈H }. A run ρ ∈Qω of Gϕ is accepting under valuation v if it visits each set in AccB inﬁnitely often and each Fxi ∈AccP in every inﬁx of length v(xi). L(Gϕ) contains all pairs (w, v) such that there is an accepting run of w under the valuation v. Gϕ is unambiguous if q a−→q′ and q a−→q′′ implies L(q′)∩L(q′′) = ∅, where L(q) is the language starting from the state q. 12 The automaton Gϕ can be constructed in O(2\|ϕ\|). Apart from the parametric acceptance condition, Gϕ behaves as a generalized B¨uchi automaton (GNBA) with accepting set AccB = { F1, . . . , Fk }. In order to obtain a non-deterministic automaton, we ﬁrst apply a similar transformation as for GNBA to NBA [16]. 6 The Fragment pLTL♦ We convert Gϕ to Uϕ = (Q′, 2AP, Q′ 0, δ′, Acc′ B, Acc′ P ) where Q′ = Q×{ 1, . . . , k }, Q′ 0 = Q0×{ 1 }. If (q, a, q′) ∈δ, then ((q, i), a, (q′, i′)) ∈δ′ with i=i′ if q ̸∈Fi else i′ = (i mod k)+1. AccB = F1 × { 1 } and Acc′ P = { F ′ xi \| Fxi ∈AccP }, where F ′ xi = Fxi × { 1, . . . , k }. Note that the construction preserves unambiguity and the size of Uϕ is in O(\|ϕ\|·2\|ϕ\|). For a given valuation v, Uϕ can be converted into an NBA Bϕ(v). This is done as follows. Let Uϕ = (Q′, 2AP , Q′ 0, δ′, Acc′ B, Acc′ P ) and v a valuation of ϕ with d parameters. Then Bϕ(v) = (Q′′, 2AP , Q′′ 0, δ′′, Acc) with: – Q′′ ⊆Q′ × {0, . . . , v(x1)} × . . . × {0, . . . , v(xd)}, – ((q, n), a, (q′, n′)) ∈δ′′ if (q, a, q′) ∈δ′ and for all xi: • if q′ ∈F ′ xi and n(xi) < v(xi) then n′(xi) = 0, • if q′ /∈F ′ xi and n(xi) < v(xi) then n′(xi) = n(xi) + 1. – Q′′ 0 = Q′ 0 × 0d and Acc = Acc′ B × {0, . . . , v(x1)} × . . . × {0, . . . , v(xd)}. – Q′′ ⊆Q′ × {0, . . . , v(x1)} × . . . × {0, . . . , v(xd)}, • if q′ ∈F ′ xi and n(xi) < v(xi) then n′(xi) = 0, i • if q′ /∈F ′ xi and n(xi) < v(xi) then n′(xi) = n(xi) + 1. – Q′′ 0 = Q′ 0 × 0d and Acc = Acc′ B × {0, . . . , v(x1)} × . . . × {0, . . . , v(xd)}. It follows that Bϕ(v) is unambiguous for any valuation v. Furthermore, every run of Bϕ(v) is either ﬁnite or satisﬁes the parametric acceptance condition for valuation v. Thus we have: Proposition 11. An inﬁnite word w ∈L(Bϕ(v)) if and only if (w, v) ∈L(ϕ). Proposition 11. An inﬁnite word w ∈L(Bϕ(v)) if and only if (w, v) ∈L(ϕ). The size of Bϕ(v) is in O(cv·\|ϕ\|·2\|ϕ\|) where cv = Q xi (v(xi) + 1). i As a next step, we exploit the fact that Bϕ(v) is unambiguous, and apply the technique by Couvreur et al. [3] for verifying MC M against Bϕ(v). Let M ⊗Bϕ(v) be the synchronous product of M and Bϕ(v) [16], Π1 the projection to M and Π2 the projection to Bϕ(v). Let L(s, q) = { π ∈Paths(s) \| trace(π) ∈ L(q) } and Pr(s, q) = Pr(L(s, q)). Let Pr(M ⊗Bϕ(v)) = P q0∈Q0 Pr(s0, q0). As Bϕ(v) is unambiguous, we have for any (s, q): Pr(s, q) = X (t,q′)∈δ(s,q) P(s, t) · Pr(t, q′), Pr(s, q) = X (t,q′)∈δ(s,q) P(s, t) · Pr(t, q′), where δ is the transition relation of M ⊗Bϕ(v) and P(s, t) is the one-step transition probability from s to t in MC M. A (maximal) strongly connected component (SCC, for short) C ⊆S is complete if for any s ∈Π1(C) : Paths(s) = [ (s,q)∈C LC(s, q) where LC(s, q) restricts runs to C (runs only visits states from C). The SCC C is accepting if Acc ∩Π2(C) ̸= ∅(where Acc is the set of accepting states in Bϕ(v)). 13 Proposition 12 ([3]). Let C be a complete and accepting SCC in M ⊗Bϕ(v). Then for all s ∈Π1(C): Proposition 12 ([3]). Let C be a complete and accepting SCC in M ⊗Bϕ(v). Then for all s ∈Π1(C): Pr [ (s,q)∈C LC(s, q) = 1. Moreover, since Bϕ(v) is unambiguous, Pr(M ⊗Bϕ(v)) > 0 implies there exists a reachable, complete and accepting SCC. Moreover, since Bϕ(v) is unambiguous, Pr(M ⊗Bϕ(v)) > 0 implies there exists a reachable, complete and accepting SCC. Finding complete and accepting SCC in M ⊗Bϕ(v) is done by standard graph analysis. Altogether, v ∈V>0(ϕ) is decided in O(m·cv·\|ϕ\|·2\|ϕ\|). The space complexity is polynomial in the size of the input (including the valuation), as M ⊗Bϕ(v) can be stored in O(log m+\|ϕ\|+log cv) bits. In the sequel, we exploit these results to obtain a necessary and suﬃcient criterion for the emptiness of V>0(ϕ) for ϕ in pLTL♦. Theorem 4. For ϕ ∈pLTL♦, V>0(ϕ) ̸= ∅iﬀ¯v ∈V>0(ϕ) s.t. ¯v(x) = m·\|ϕ\|·2\|ϕ\|. Theorem 4. For ϕ ∈pLTL♦, V>0(ϕ) ̸= ∅iﬀ¯v ∈V>0(ϕ) s.t. ¯v(x) = m·\|ϕ\|·2\|ϕ\|. Proof. Consider the direction from left to right. The only non-trivial case is when there exists a valuation v ̸⩽¯v such that v ∈V>0(ϕ) implies ¯v ∈V>0(ϕ). In the model checking algorithm described above, we ﬁrst construct Gϕ, and then Uϕ with a single B¨uchi accepting set Acc′ B and d parametric accepting sets F ′ xi, one for each variable xi in ϕ. For the sake of clarity, assume d = 1, i.e., we consider valuation v. The explanation extends to the general case in a straightforward manner. For valuation v, consider M ⊗Bϕ(v). We show that, for r < v, Pr(M ⊗Bϕ(v)) > 0 implies Pr(M ⊗Bϕ(r)) > 0, where r = m·\|Uϕ\|, which is in O(m·\|ϕ\|·2\|ϕ\|). Note that every cycle in M ⊗Bϕ(r) contains a state (s, q, i) with i = 0. (b) Suppose C′ is accepting. Then there exists (s′, q′, i′) with q′ ∈Acc. Since C′ is an SCC and C ⊆C′, there is a path from (s, q, 0) ∈C to (s′, q′, i′). If the length of the path is less than r, then we are done. If i′ > r, then some (s′′, q′′) pair in the path must be repeated. Thus, we can ﬁnd another path of length less For almost sure properties, a similar approach as for V>0(ϕ) suﬃces. Theorem 5.For ϕ ∈pLTL♦, V=1(ϕ) ̸= ∅iﬀ¯v ∈V=1( ¯ϕ) with ¯v(x) = m·\|ϕ\|·2\|ϕ\|. Let N M = m \|ϕ\| 2\|ϕ\| Note that c equals (N M)d Thus we have: Theorem 5.For ϕ ∈pLTL♦, V=1(ϕ) ̸= ∅iﬀ¯v ∈V=1( ¯ϕ) with ¯v(x) = m·\|ϕ\|·2\|ϕ\|. Let NϕM = m·\|ϕ\|·2\|ϕ\|. Note that c¯v equals (NϕM)d. Thus, we have: Proposition 13. For ϕ ∈pLTL♦, deciding if V>0(ϕ) = ∅is PSPACE-complete. Proof. Theorem 4 gives an algorithm in PSPACE, as M ⊗Bϕ(¯v) can be stored in O(log m + \|ϕ\| + d log NϕM) bits. PSPACE hardness follows trivially, as for LTL formula ϕ and MC M, deciding Pr(M \|= ϕ) > 0 (which is known to be a PSPACE complete problem) is the same as checking the emptiness of V>0(ϕ). ■ Just as for pLTL(F,X), we can use the bisection method to ﬁnd min V>0(ϕ). The search procedure invokes the model checking algorithm multiple times. We can reuse the space each time we check Pr(M \|= v(ϕ)) > 0. Hence, min V>0(ϕ) can be found in polynomial space. The time complexity of ﬁnding min V>0(ϕ) is O(m·(NϕM)d·2\|ϕ\|· log NϕM). Membership can also be similarly solved. Proposition 14. For pLTL♦-formula ϕ, v ∈V>0(ϕ)? takes O(d· log NϕM d ) time, provided a representation of V>0(ϕ) is given. 7 Concluding Remarks This paper considered the veriﬁcation of ﬁnite MCs against parametric LTL. We obtained several results on the emptiness problem for qualitative veriﬁcation problems, including necessary and suﬃcient conditions as well as some complex- ity results. Future work consists of devising more eﬃcient algorithms for the quantitative veriﬁcation problems, and lifting the results to extended temporal logics [19] and stochastic games, possibly exploiting [17]. Acknowledgement. This work was partially supported by the EU FP7 projects MoVeS and Sensation, the EU Marie Curie project MEALS and the Excellence initiative of the German federal government. Pr(s, q) = X (t,q′)∈δ(s,q) P(s, t) · Pr(t, q′), Moreover, the graph of M ⊗Bϕ(r) is a sub-graph of M ⊗Bϕ(v). We now prove that, if a (maximal) SCC C of M ⊗Bϕ(r) is not complete (or accepting) then any SCC C′ of M ⊗Bϕ(v) containing C is also not complete (or accepting, respectively). (a) Suppose C is not complete. Then there exists a ﬁnite path σ = s s1 . . . sk of M, such that from any q, with (s, q, 0) ∈C, the run ρ = (s, q, 0)(s1, q1, 1) . . . (sj, qj, j) leads to a deadlock state. This can have two causes: either (sj, qj, j) has no successor for any j. Then, C′ is not complete. Or, the path ρ terminates in (sj, qj, j) where j = r. This means, for all (s′, q′, j+1) ∈δ(sj, qj, j) in C′, q′ ̸∈Fx. As the length of ρ exceeds r, there are states in the run whose ﬁrst and second component appear multiple times. Thus, we can ﬁnd another path σ′ (possibly longer than σ) for C′ which goes through states where the ﬁrst and the second component of some of its states are repeated suﬃciently many times to have a run (s, q, 0)(s1, q1, 1) . . . (sj, qj, v) which is a deadlock state. Thus, C′ is not complete. 14 than r to a state (s′, q′, i), where i ⩽r. Therefore, C is accepting. The rest of the proof follows from Proposition 12. ■ than r to a state (s′, q′, i), where i ⩽r. Therefore, C is accepting. The rest of the proof follows from Proposition 12. ■ For almost sure properties, a similar approach as for V>0(ϕ) suﬃces. References 1. Vardi, M.Y.: Automatic veriﬁcation of probabilistic concurrent ﬁnite-state pro- grams. In: FOCS, IEEE Computer Society (1985) 327–338 ( ) 2. Courcoubetis, C., Yannakakis, M.: The complexity of probabilistic veriﬁcation. J. ACM 42(4) (1995) 857–907 3. Couvreur, J.M., Saheb, N., Sutre, G.: An optimal automata approach to LTL model checking of probabilistic systems. In: LPAR. Volume 2850 of LNCS., Springer (2003) 361–375 15 4. Chatterjee, K., Gaiser, A., Kret´ınsk´y, J.: Automata with generalized Rabin pairs for probabilistic model checking and LTL synthesis. In: CAV. Volume 8044 of LNCS., Springer (2013) 559–575 5. Alur, R., Etessami, K., La Torre, S., Peled, D.: Parametric temporal logic for ”model measuring”. ACM Trans. Comput. Log. 2(3) (2001) 388–407 6. Han, T., Katoen, J.P., Mereacre, A.: Approximate parameter synthesis for prob- abilistic time-bounded reachability. In: IEEE Real-Time Systems Symposium (RTSS), IEEE Computer Society (2008) 173–182 ( ) ( ) 7. Hahn, E.M., Han, T., Zhang, L.: Synthesis for PCTL in parametric Markov decision processes. In: NFM. Volume 6617 of LNCS., Springer (2011) 146–161 8. Puggelli, A., Li, W., Sangiovanni-Vincentelli, A., Seshia, S.: Polynomial-time veri- ﬁcation of PCTL properties of MDPs with convex uncertainties. In: CAV. Volume 8044 of LNCS., Springer (2013) 527–542 ( ) 9. Benedikt, M., Lenhardt, R., Worrell, J.: LTL model checking of interval Markov chains. In: TACAS. Volume 7795 of LNCS., Springer (2013) 32–46 10. Hahn, E.M., Hermanns, H., Zhang, L.: Probabilistic reachability for parametric Markov models. STTT 13(1) (2011) 3–19 11. Daws, C.: Symbolic and parametric model checking of discrete-time Markov chains. In: ICTAC. Volume 3407 of LNCS., Springer (2005) 280–294 12. Katoen, J.P., McIver, A., Meinicke, L., Morgan, C.C.: Linear-invariant generation for probabilistic programs. In: Static Analysis Symposium (SAS). Volume 6337 of LNCS., Springer (2010) 390–406 13. Benedikt, M., Lenhardt, R., Worrell, J.: Two variable vs. linear temporal logic in model checking and games. Logical Methods in Computer Science 9(2) (2013) 14. Ummels, M., Baier, C.: Computing quantiles in Markov reward models. In: FoS- SaCS. Volume 7794 of LNCS., Springer (2013) 353–368 ( ) 15. Alur, R., La Torre, S.: Deterministic generators and games for LTL fragments. ACM Trans. Comput. Log. 5(1) (2004) 1–25 16. Baier, C., Katoen, J.P.: Principles of Model Checking. MIT Press (2008) 17. Zimmermann, M.: Optimal bounds in parametric LTL games. Theor. Comput. Sci. 493 (2013) 30–45 18. Vardi, M.Y.: An automata-theoretic approach to linear temporal logic. In: Logics for Concurrency: Structure versus Automata. References Volume 1043 of LNCS., Springer (1996) 238–266 ( ) 19. Vardi, M.Y., Wolper, P.: Reasoning about inﬁnite computations. Information and Computation 115 (1994) 1–37
https://openalex.org/W2411709723	https://research.rug.nl/files/48407608/ncomms10023.pdf	English	null	Genetic associations at 53 loci highlight cell types and biological pathways relevant for kidney function	Nature communications	2,016	cc-by	28,196	ARTICLE Received 12 Mar 2015 \| Accepted 27 Oct 2015 \| Published 21 Jan 2016 Received 12 Mar 2015 \| Accepted 27 Oct 2015 \| Published 21 Jan 2016 University of Groningen University of Groningen Genetic associations at 53 loci highlight cell types and biological pathways relevant for kidney function Pattaro, Cristian; Teumer, Alexander; Gorski, Mathias; Chu, Audrey Y; Li, Man; Mijatovic, Vladan; Garnaas, Maija; Tin, Adrienne; Sorice, Rossella; Li, Yong Published in: Nature Communications DOI: 10.1038/ncomms10023 IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2016 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Pattaro, C., Teumer, A., Gorski, M., Chu, A. Y., Li, M., Mijatovic, V., Garnaas, M., Tin, A., Sorice, R., Li, Y., Taliun, D., Olden, M., Foster, M., Yang, Q., Chen, M.-H., Pers, T. H., Johnson, A. D., Ko, Y.-A., Fuchsberger, C., ... ICBP Consortium (2016). Genetic associations at 53 loci highlight cell types and biological pathways relevant for kidney function. Nature Communications, 7(10023). https://doi.org/10.1038/ncomms10023 Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Results S 1 7 9 Exploratory comparison of the association effect sizes in subjects with and without hypertension based on our previous work7 showed that novel and known loci are also similarly associated with eGFRcrea among individuals with and without hypertension (Supplementary Fig. 6). Tests for SNP associations with related phenotypes. We tested for overlap with traits that are known to be associated with kidney function in the epidemiologic literature by investigating SNP associations with systolic and diastolic blood pressure17, myocardial infarction18, left ventricular mass19, heart failure20, fasting glucose21 and urinary albumin excretion (CKDGen Consortium, personal communication). We observed little association of the 24 novel SNPs with other kidney function-related traits, with only 2 out of 165 tests reaching the Bonferroni signiﬁcance level of 0.0003 (see Methods and Supplementary Table 6). In addition to conﬁrming 29 previously identiﬁed loci7–9 (Fig. 1a; Supplementary Table 2), we identiﬁed 48 independent novel loci (Supplementary Table 3) where the index SNP, deﬁned as the variant with the lowest P value in the region, had an association P value o1.0 10 6. Of these 48 novel SNPs, 21 were genome-wide signiﬁcant with P values o5.0 10 8. Overall, 43 SNPs were identiﬁed in association with eGFRcrea (nine in the non-diabetes sample), one with eGFRcys and four with CKD, as reported in Supplementary Table 3. Manhattan plots for CKD, eGFRcys and eGFRcrea in diabetes are shown in Fig. 1b,c and Supplementary Fig. 2, respectively. y To investigate whether additional traits are associated with the 24 new eGFR loci, we queried the NHGRI GWAS catalog (www.genome.gov). Overall, nine loci were previously identiﬁed in association with other traits at a P value of 5.0 10 8 or lower (Supplementary Table 7), including body mass index (ETV5) and serum urate (INHBC, A1CF and AP5B1). Stage-2 replication. Novel loci were tested for replication in up to 42,166 additional European ancestry individuals from 15 studies (Supplementary Table 1). Of the 48 novel candidate SNPs sub- mitted to replication, 24 SNPs demonstrated a genome-wide signiﬁcant combined stage 1 and 2 P value o5.0 10 8 (Table 1). Of these, 23 fulﬁlled additional replication criteria (q-value o0.05 in stage 2). Only rs6795744 at the WNT7A locus demonstrated suggestive replication (P value o5.0 10 8, q-value 40.05). ARTICLE ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 C hronic kidney disease (CKD) is a global public health problem1–3, and is associated with an increased risk for cardiovascular disease, all-cause mortality and end-stage renal disease4,5. Few new therapies have been developed to prevent or treat CKD over the past two decades1,6, underscoring the need to identify and understand the underlying mechanisms of CKD. creatinine-independent biomarker of GFR (Supplementary Fig. 3; Supplementary Table 4). The majority of loci (22/24) demonstrated consistent effect directions of their association with both eGFRcrea and eGFRcys. Association plots of the 24 newly identiﬁed genomic regions that contain a replicated or suggestive index SNP appear in Supplementary Fig. 4. The odds ratio for CKD for each of the novel loci ranged from 0.93 to 1.06 (Supplementary Table 4). As evidenced by the relatively small effect sizes, the proportion of phenotypic variance of eGFRcrea explained by all new and known loci was 3.22%: 0.81% for the newly uncovered loci and 2.41% for the already known loci. Prior genome-wide association studies (GWAS) have identiﬁed multiple genetic loci associated with CKD and estimated glomerular ﬁltration rate (eGFR), a measure of the kidney’s ﬁltration ability that is used to diagnose and stage CKD7–15. Subsequent functional investigations point towards clinically relevant novel mechanisms in CKD that were derived from initial GWAS ﬁndings16, providing proof of principle that locus discovery through large-scale GWAS efforts can translate to new insights into CKD pathogenesis. Associations stratiﬁed by diabetes and hypertension status. The effects of the 53 known and novel loci in individuals with (stage 1 þ stage 2 n ¼ 16,477) and without (stage 1 þ stage 2 n ¼ 154,881) diabetes were highly correlated (correlation coefﬁ- cient: 0.80; 95% conﬁdence interval: 0.67, 0.88; Supplementary Fig. 5) and of similar magnitude (Fig. 2; Supplementary Table 5), suggesting that identiﬁcation of genetic loci in the overall popu- lation may also provide insights into loci with potential impor- tance among individuals with diabetes. The previously identiﬁed UMOD locus showed genome-wide signiﬁcant association with eGFRcrea among those with diabetes (Supplementary Fig. 2; rs12917707, P value ¼ 2.5 10 8), and six loci (NFKB1, UNCX, TSPAN9, AP5B1, SIPA1L3 and PTPRO) had nominally signiﬁcant associations with eGFRcrea among those with diabetes. Of the previously identiﬁed loci, 13 demonstrated nominal associations among those with diabetes, for a total of 19 loci associated with eGFRcrea in diabetes. Results S 1 Stage 1 discovery analysis. We analysed associations of eGFR based on serum creatinine (eGFRcrea), cystatin C (eGFRcys, an additional, complementary biomarker of renal function) and CKD (deﬁned as eGFRcrea o60 ml min 1 per 1.73 m2) with B2.5 million autosomal single-nucleotide polymorphisms (SNPs) in up to 133,413 individuals of European ancestry from 49 pre- dominantly population-based studies (Supplementary Table 1). Results from discovery GWAS meta-analysis are publicly avail- able at http://fox.nhlbi.nih.gov/CKDGen/. We performed ana- lyses in each study sample in the overall population and stratiﬁed by diabetes status, since genetic susceptibility to CKD may differ in the presence of this strong clinical CKD risk factor. Population stratiﬁcation did not impact our results as evidenced by low genomic inﬂation factors in our meta-analyses, which ranged from 1.00 to 1.04 across all our analyses (Supplementary Fig. 1). In addition to conﬁrming 29 previously identiﬁed loci7–9 (Fig. 1a; Supplementary Table 2), we identiﬁed 48 independent novel loci (Supplementary Table 3) where the index SNP, deﬁned as the variant with the lowest P value in the region, had an association P value o1.0 10 6. Of these 48 novel SNPs, 21 were genome-wide signiﬁcant with P values o5.0 10 8. Overall, 43 SNPs were identiﬁed in association with eGFRcrea (nine in the non-diabetes sample), one with eGFRcys and four with CKD, as reported in Supplementary Table 3. Manhattan plots for CKD, eGFRcys and eGFRcrea in diabetes are shown in Fig. 1b,c and Supplementary Fig. 2, respectively. Stage 1 discovery analysis. We analysed associations of eGFR based on serum creatinine (eGFRcrea), cystatin C (eGFRcys, an additional, complementary biomarker of renal function) and CKD (deﬁned as eGFRcrea o60 ml min 1 per 1.73 m2) with B2.5 million autosomal single-nucleotide polymorphisms (SNPs) in up to 133,413 individuals of European ancestry from 49 pre- dominantly population-based studies (Supplementary Table 1). Results from discovery GWAS meta-analysis are publicly avail- able at http://fox.nhlbi.nih.gov/CKDGen/. We performed ana- lyses in each study sample in the overall population and stratiﬁed by diabetes status, since genetic susceptibility to CKD may differ in the presence of this strong clinical CKD risk factor. Population stratiﬁcation did not impact our results as evidenced by low genomic inﬂation factors in our meta-analyses, which ranged from 1.00 to 1.04 across all our analyses (Supplementary Fig. 1). ARTICLE To identify additional genetic variants associated with eGFR and guide future experimental studies of CKD-related mechan- isms, we have now performed GWAS meta-analyses in up to 133,413 individuals, more than double the sample size of previous studies. Here we describe multiple novel genomic loci associated with kidney function traits and provide extensive locus characterization and bioinformatics analyses, further delineating the physiologic basis of kidney function. NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications Genetic associations at 53 loci highlight cell types and biological pathways relevant for kidney function Cristian Pattaro et al.# Reduced glomerular ﬁltration rate deﬁnes chronic kidney disease and is associated with cardiovascular and all-cause mortality. We conducted a meta-analysis of genome-wide association studies for estimated glomerular ﬁltration rate (eGFR), combining data across 133,413 individuals with replication in up to 42,166 individuals. We identify 24 new and conﬁrm 29 previously identiﬁed loci. Of these 53 loci, 19 associate with eGFR among indi- viduals with diabetes. Using bioinformatics, we show that identiﬁed genes at eGFR loci are enriched for expression in kidney tissues and in pathways relevant for kidney development and transmembrane transporter activity, kidney structure, and regulation of glucose metabolism. Chromatin state mapping and DNase I hypersensitivity analyses across adult tissues demonstrate preferential mapping of associated variants to regulatory regions in kidney but not extra-renal tissues. These ﬁndings suggest that genetic determinants of eGFR are mediated largely through direct effects within the kidney and highlight important cell types and biological pathways. terials should be addressed to C.P. (email: cristian.pattaro@eurac.edu) or to A.K. (email: anna.koettgen@uniklinik-freiburg.de) ov). Correspondence and requests for materials should be addressed to C.P. (email: cristian.pattaro@eurac.edu) or to A.K. (email: anna.koettgen@uniklinik-freiburg.de) or to C.S.F. (email: foxca@nhlbi.nih.gov). #A full list of authors and their afﬁliations appears at the end of the paper. 1 NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 21 19 17 15 14 13 12 11 10 9 8 7 6 5 4 4 3 3 2 2 1 1 0 eGFRcrea overall/nonDM –log10 (P value) 10 9 8 7 6 P value) Overall (n=133,413) NonDM (n=118,448) Chromosome n=32,834 eGFRcys overall c a SHROOM3 CPS1 NAT8 GCKR ANXA9 DDX1 TFDP2 DAB2 VEGFA TMEM60 STC1 PIP5K1B MPPED2 BCAS3 GATM UMOD SLC6A13 DACH1 INO80 WDR72 UBE2Q2 SLC47A1 CDK12 SLC7A9 WDR37 SLC22A2 SLC34A1 PRKAG2 CASP9 CACNA1S IGFBP5 ETV5 KCNQ1 DPEP1 NFATC1 A1CF SKIL SYPL2 SDCCAG8 LRP2 WNT7A NFKB1 ZNF204 UNCX RNF32 AP5B1 TSPAN9 PTPRO INHBC BCAS1 TP53INP2 SIPA1L3 KBTBD2 NAT8 VEGFA PRKAG2 STC1 ATXN2 UBE2Q2 UMOD CST3 SHROOM3 KCNQ1 10 9 8 7 6 5 4 3 2 1 0 –log10 (P value) 21 19 17 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Chromosome n=117,165 CKD overall b CPS1 ANXA9 GCKR DDX1 CASP9 NAT8 TFDP2 SHROOM3 DAB2 VEGFA SLC22A2 SLC34A1 TMEM60 STC1 WDR37 PIP5K1B MPPED2 GATM UMOD WDR72 SLC47A1 CDK12 BCAS3 SLC7A9 DACH1 INO80 UBE2Q2 SLC6A13 PRKAG2 SDCCAG8 CACNA1S SYPL2 LRP2 IGFBP5 WNT7A SKIL ETV5 NFKB1 ZNF204 RNF32 UNCX KBTBD2 A1CF AP5B1 KCNQ1 INHBC TSPAN9 PTPRO DPEP1 NFATC1 TP53INP2 BCAS1 SIPA1L3 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 21 19 17 15 14 13 12 11 10 9 8 7 6 5 4 4 3 3 2 2 1 1 0 eGFRcrea overall/nonDM –log10 (P value) Overall (n=133,413) NonDM (n=118,448) Chromosome a SHROOM3 CPS1 NAT8 GCKR ANXA9 DDX1 TFDP2 DAB2 VEGFA TMEM60 STC1 PIP5K1B MPPED2 BCAS3 GATM UMOD SLC6A13 DACH1 INO80 WDR72 UBE2Q2 SLC47A1 CDK12 SLC7A9 WDR37 SLC22A2 SLC34A1 PRKAG2 CASP9 CACNA1S IGFBP5 ETV5 KCNQ1 DPEP1 NFATC1 A1CF SKIL SYPL2 SDCCAG8 LRP2 WNT7A NFKB1 ZNF204 UNCX RNF32 AP5B1 TSPAN9 PTPRO INHBC BCAS1 TP53INP2 SIPA1L3 KBTBD2 a eGFRcrea overall/nonDM 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 21 19 17 15 14 13 12 11 10 9 8 7 6 5 4 4 3 3 2 2 1 1 0 –log10 (P value) 10 9 8 7 6 5 4 3 2 1 0 –log10 (P value) Chromosome 21 19 17 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Chromosome n=32,834 eGFRcys overall c CPS1 NAT8 GCKR ANXA9 DDX1 TFDP2 DAB2 VEGFA TMEM60 STC1 PIP5K1B MPPED2 BCAS3 SLC6A13 DACH1 INO80 WDR72 UBE2Q2 SLC47A1 CDK12 SLC7A9 WDR37 SLC22A2 SLC34A1 CASP9 CACNA1S IGFBP5 ETV5 KCNQ1 DPEP1 NFATC1 A1CF SKIL SYPL2 SDCCAG8 LRP2 WNT7A NFKB1 ZNF204 UNCX RNF32 AP5B1 TSPAN9 PTPRO INHBC BCAS1 TP53INP2 SIPA1L3 KBTBD2 NAT8 DDX1 GCKR CASP9 ANXA9 TFDP2 CPS1 DAB2 SLC34A1 VEGFA PRKAG2 STC1 SLC22A2 TMEM60 PIP5K1B WDR37 ATXN2 UBE2Q2 WDR72 BCAS3SLC7A9 DACH1 GATM INO80 SLC47A1 CDK12 UMOD CST3 MPPED2 SLC6A13 SHROOM3 SYPL2 SDCCAG8 CACNA1S LRP2 WNT7A ETV5 NFKB1 ZNF204 UNCX KBTBD2 KCNQ1 AP5B1 PTPBO TSPAN9 DPEP1 TP53INP2 BCAS1 NFATC1 SIPA1L3 INHBC SKIL IGFBP5 RNF32 A1CF 10 9 8 7 6 5 4 3 2 1 0 –log10 (P value) 21 19 17 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Chromosome n=117,165 CKD overall b CPS1 ANXA9 GCKR DDX1 CASP9 NAT8 TFDP2 SHROOM3 DAB2 VEGFA SLC22A2 SLC34A1 TMEM60 STC1 WDR37 PIP5K1B MPPED2 GATM UMOD WDR72 SLC47A1 CDK12 BCAS3 SLC7A9 DACH1 INO80 UBE2Q2 SLC6A13 PRKAG2 SDCCAG8 CACNA1S SYPL2 LRP2 IGFBP5 WNT7A SKIL ETV5 NFKB1 ZNF204 RNF32 UNCX KBTBD2 A1CF AP5B1 KCNQ1 INHBC TSPAN9 PTPRO DPEP1 NFATC1 TP53INP2 BCAS1 SIPA1L3 ery stage genome-wide association analysis. Results S 1 Because serum creatinine is used to estimate eGFRcrea, associated genetic loci may be relevant to creatinine production or metabolism rather than kidney function per se. For this reason, we contrasted associations of eGFRcrea versus eGFRcys, the latter estimated from an alternative and Trans-ethnic analyses. To assess the generalizability of our ﬁndings across ethnicities, we evaluated the association of the 24 newly identiﬁed loci with eGFRcrea in 16,840 participants of 12 African ancestry population studies (Supplementary Table 8) and in up to 42,296 Asians from the AGEN consortium11 (Supplementary Table 9). Seven SNPs achieved nominal direction-consistent signiﬁcance (Po0.05) in AGEN, and one SNP was signiﬁcant in the African ancestry meta-analysis (Supplementary Table 9). Random-effect meta-analysis showed that 12 loci (SDCCAG8, LRP2, IGFBP5, SKIL, UNCX, KBTBD2, A1CF, KCNQ1, AP5B1, PTPRO, TP53INP2 and BCAS1) had fully consistent effect direction across the three ethnic groups NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications 2 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 Table 1 \| The 24 novel SNPs associated with eGFRcrea in European ancestry individuals. SNP ID* Chr. Position (bp)w Locus namez Effect/Non effect allele (EAF) SNP functiony Stage 1 (discovery)\|\| Stage 2 (replication) Combined analysisz Beta P value Beta q-value Beta P value# I2 %** The eight loci whose smallest P value was observed in the ‘no diabetes’ group rs3850625 1 201,016,296 CACNA1S A/G (0.12) Exonic, nonsyn. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 SNV 0.0080 2.55E 09 0.0071 5.46E 03 0.0083 6.82E 11 0 rs2712184 2 217,682,779 IGFBP5 A/C (0.58) Intergenic 0.0049 1.65E 08 0.0055 2.06E 03 0.0053 1.33E 10 0 rs9682041 3 170,091,902 SKIL T/C (0.87) Intronic 0.0067 1.36E 07 0.0046 2.33E 02 0.0068 2.58E 08 2 rs10513801 3 185,822,353 ETV5 T/G (0.87) Intronic 0.0070 3.80E 09 0.0046 1.79E 02 0.0072 1.03E 09 0 rs10994860 10 52,645,424 A1CF T/C (0.19) UTR5 0.0075 1.00E 11 0.0061 5.46E 03 0.0077 1.07E 12 2 rs163160 11 2,789,955 KCNQ1 A/G (0.82) Intronic 0.0067 9.02E 09 0.0050 9.89E 03 0.0065 2.26E 09 14 rs164748 16 89,708,292 DPEP1 C/G (0.53) Intergenic 0.0047 9.92E 09 0.0019 4.19E 02 0.0046 1.95E 08 17 rs8091180 18 77,164,243 NFATC1 A/G (0.56) Intronic 0.0054 1.43E 08 0.0052 5.46E 03 0.0060 1.28E 09 0 The 16 loci whose smallest P value was observed in the ‘overall’ group rs12136063 1 110,014,170 SYPL2 A/G (0.70) Intronic 0.0049 2.33E 07 0.0028 2.31E 02 0.0045 4.71E 08 0 rs2802729 1 243,501,763 SDCCAG8 A/C (0.43) Intronic 0.0050 7.37E 08 0.0029 2.05E 02 0.0046 2.20E 08 9 rs4667594 2 170,008,506 LRP2 A/T (0.53) Intronic 0.0045 2.37E 07 0.0043 5.62E 03 0.0044 3.52E 08 4 rs6795744ww 3 13,906,850 WNT7A A/G (0.15) Intronic 0.0071 9.60E 09 0.0019 5.15E 02 0.0060 3.33E 08 18 rs228611 4 103,561,709 NFKB1 A/G (0.47) Intronic 0.0055 4.66E 10 0.0060 8.91E 04 0.0056 3.58E 12 3 rs7759001 6 27,341,409 ZNF204 A/G (0.76) ncRNA intronic 0.0053 2.64E 07 0.0045 9.10E 03 0.0051 1.75E 08 0 rs10277115 7 1,285,195 UNCX A/T (0.23) Intergenic 0.0095 1.05E 10 0.0079 9.03E 04 0.0090 8.72E 14 0 rs3750082 7 32,919,927 KBTBD2 A/T (0.33) Intronic 0.0049 2.52E 07 0.0031 1.96E 02 0.0045 3.22E 08 2 rs6459680 7 156,258,568 RNF32 T/G (0.74) Intergenic 0.0065 1.96E 10 0.0019 4.62E 02 0.0055 1.07E 09 0 rs4014195 11 65,506,822 AP5B1 C/G (0.64) Intergenic 0.0061 2.19E 11 0.0034 1.42E 02 0.0055 1.10E 11 0 rs10491967 12 3,368,093 TSPAN9 A/G (0.10) Intronic 0.0092 3.03E 10 0.0106 3.93E 04 0.0095 5.18E 14 0 rs7956634 12 15,321,194 PTPRO T/C (0.81) Intronic 0.0068 2.46E 09 0.0069 1.51E 03 0.0068 7.17E 12 0 rs1106766 12 57,809,456 INHBC T/C (0.22) Intergenic 0.0062 4.67E 08 0.0058 8.79E 03 0.0061 2.41E 09 9 rs11666497 19 38,464,262 SIPA1L3 T/C (0.18) Intronic 0.0064 8.58E 08 0.0041 1.53E 02 0.0058 4.25E 08 24 rs6088580 20 33,285,053 TP53INP2 C/G (0.47) Intergenic 0.0055 7.17E 10 0.0027 2.31E 02 0.0049 1.79E 09 0 rs17216707 20 52,732,362 BCAS1 T/C (0.79) Intergenic 0.0084 5.96E 13 0.0051 6.69E 03 0.0077 8.83E 15 1 bp, basepairs; Chr, chromosome; EAF, effect allele frequency; eGFRcrea, eGFR based on serum creatinine; GWAS, genome-wide association studies; SNP, single-nucleotide polymorphism; UTR, untranslated region. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 SNPs are grouped by the stratum where the smallest P value in the discovery and combined analysis was observed. In the ‘no diabetes’ group, sample size/number of studies were equal to 118,448/45, 36,433/13 and 154,881/58, in the discovery, replication and combined analyses, respectively. In the ‘overall’ group, the numbers for the three analyses were equal to 133,413/48, 42,116/14 and 175,579/62, respectively. wOn the basis of RefSeq genes (build 37). zConventional locus name based on relevant genes in the region as identiﬁed by bioinformatic investigation (Supplementary Table 12) or closest gene. A complete overview of the genes in each locus is given in the regional association plots (Supplementary Fig. 4). ySNP function is derived from NCBI RefSeq genes and may not correspond to the named gene. \|\|Twice genomic-control (GC) corrected P value from discovery GWAS meta-analysis: at the individual study level and after the meta-analysis. zFor random-effect estimate, see Supplementary Table 4. #P value of the meta-analysis of the twice GC-corrected discovery meta-analysis results and replication studies. Between-study heterogeneity, as assessed by the I2. Q statistic P value 40.05 for all SNPs, except rs11666497 (SIPA1L3, P value ¼ 0.04). wwFor this SNP, the conditions for replication were not all met (q-value 40.05 in the replication stage). Table 1 \| The 24 novel SNPs associated with eGFRcrea in European ancestry individuals. connected novel SNPs to transcript abundance of SYPL2, SDCCAG8, MANBA, KBTBD2, PTPRO and SPATA33 (C16orf55), thereby supporting these as potential candidate genes in the respective associated regions (Supplementary Table 11). (Supplementary Fig. 7), suggesting that our ﬁndings can likely be generalized beyond the European ancestry group. g y p y g p To identify additional potentially associated variants and more formally evaluate trans-ethnic heterogeneity of the loci identiﬁed through meta-analysis in European ancestry populations, we performed a trans-ethnic meta-analysis22, combining the 12 African ancestry studies with the 48 European Ancestry studies used in the discovery analysis of eGFRcrea. Of the 24 new loci uncovered for eGFRcrea, 15 were also genome-wide signiﬁcant in the trans-ethnic meta-analysis (deﬁned as log10 Bayes Factor 46, Supplementary Table 10), indicating that for most of these loci, there is little to no allelic effect heterogeneity across the two ethnic groups. No additional loci were signiﬁcantly associated with log10 Bayes Factor 46. Pathway analyses. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 Manhattan plots for eGFRcrea, CKD and eGFRcys. Previously labels). (a) Novel loci uncovered for eGFRcrea in the overall and in the non-diabetes groups are highligh esults from CKD analysis with highlighted known and novel loci for eGFRcrea. (c) Results from eGFRcys with and known eGFRcys loci. CATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications 10 9 8 7 6 5 4 3 2 1 0 –log10 (P value) 21 19 17 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Chromosome n=117,165 CKD overall b CPS1 ANXA9 GCKR DDX1 CASP9 NAT8 TFDP2 SHROOM3 DAB2 VEGFA SLC22A2 SLC34A1 TMEM60 STC1 WDR37 PIP5K1B MPPED2 GATM UMOD WDR72 SLC47A1 CDK12 BCAS3 SLC7A9 DACH1 INO80 UBE2Q2 SLC6A13 PRKAG2 SDCCAG8 CACNA1S SYPL2 LRP2 IGFBP5 WNT7A SKIL ETV5 NFKB1 ZNF204 RNF32 UNCX KBTBD2 A1CF AP5B1 KCNQ1 INHBC TSPAN9 PTPRO DPEP1 NFATC1 TP53INP2 BCAS1 SIPA1L3 b 10 9 8 7 6 5 4 3 2 1 0 –log10 (P value) 21 19 17 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Chromosome n=32,834 eGFRcys overall c NAT8 DDX1 GCKR CASP9 ANXA9 TFDP2 CPS1 DAB2 SLC34A1 VEGFA PRKAG2 STC1 SLC22A2 TMEM60 PIP5K1B WDR37 ATXN2 UBE2Q2 WDR72 BCAS3SLC7A9 DACH1 GATM INO80 SLC47A1 CDK12 UMOD CST3 MPPED2 SLC6A13 SHROOM3 SYPL2 SDCCAG8 CACNA1S LRP2 WNT7A ETV5 NFKB1 ZNF204 UNCX KBTBD2 KCNQ1 AP5B1 PTPBO TSPAN9 DPEP1 TP53INP2 BCAS1 NFATC1 SIPA1L3 INHBC SKIL IGFBP5 RNF32 A1CF c Figure 1 \| Discovery stage genome-wide association analysis. Manhattan plots for eGFRcrea, CKD and eGFRcys. Previously reported loci are highlighted in light blue (grey labels). (a) Novel loci uncovered for eGFRcrea in the overall and in the non-diabetes groups are highlighted in blue and green, respectively. (b) Results from CKD analysis with highlighted known and novel loci for eGFRcrea. (c) Results from eGFRcys with highlighted known and novel loci for eGFRcrea and known eGFRcys loci. NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 Displayed are effects and their 95% conﬁdence intervals on ln(eGFRcrea). Results are sorted by increasing effects in the diabetes group. The majority of loci demonstrated similar effect sizes in the diabetes as compared with non-diabetes strata. SNP-speciﬁc information and detailed sample sizes are reported in Supplementary Table 5. b a Effect on log(eGFRcrea) Effect on log(eGFRcrea) Figure 2 \| Association eGFRcrea loci in subjects with and without diabetes. Novel (a) and known (b) loci were considered. Displayed are effects and their 95% conﬁdence intervals on ln(eGFRcrea). Results are sorted by increasing effects in the diabetes group. The majority of loci demonstrated similar effect sizes in the diabetes as compared with non-diabetes strata. SNP-speciﬁc information and detailed sample sizes are reported in Supplementary Table 5. DNase I hypersensitivity and H3K4m3 chromatin mark analyses. To evaluate whether eGFRcrea-associated SNPs map into gene regulatory regions and to thereby gain insight into their potential function, we evaluated the overlap of independent eGFRcrea- associated SNPs with P values o10 4 (or their proxies) with DHSs using publicly available data from the Epigenomics Road- map Project and ENCODE for 123 cell types (see Methods). DHSs mark accessible chromatin regions where transcription may occur. Compared with a set of control SNPs (see Methods), eGFRcrea-associated SNPs were signiﬁcantly more likely to map to DHS in six speciﬁc tissues or cell types (Fig. 3b), including adult human renal cortical epithelial cells, adult renal proximal tubule epithelial cells, H7 embryonic stem cells (differentiated 2 days), adult human renal epithelial cells, adult small airway epi- thelial cells and amniotic epithelial cells. No signiﬁcant enrich- ment was observed for adult renal glomerular endothelial cells, the only other kidney tissue evaluated. Chromatin annotation maps. In addition to assessing individual regulatory marks separately, we annotated the known and repli- cated novel SNPs, as well as their perfect proxies in a com- plementary approach. Chromatin annotation maps were generated integrating 410 epigenetic marks from cells derived from adult human kidney tissue and a variety of non-renal tissues from the ENCODE project (see Methods). The proportion of variants to which a function could be assigned was signiﬁcantly higher when using chromatin annotation maps from renal tissue compared with using maps that investigated the same epigenetic marks in other non-renal tissues (Fig. 3c), again indicating that eGFRcrea associated SNPs are, or tag, kidney-speciﬁc regulatory variants. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 rs10491967 (A) - TSPAN9 rs491567 (A)-WDR72 No diabetes Diabetes rs10794720 (T)- WDR37 rs7422339 (A)-CPS1 rs267734 (T)- LASS2 rs3925584 (T)- MPPED2 rs1394125 (A)- UBE2Q2 rs7805747 (A)- PRKAG2 rs2453533 (A)-GATM rs11959928 (A) -DAB2 rs13538 (A)-ALMS1 rs17319721 (A) -SHROOM3 rs626277 (A)- DACH1 rs6431731 (T)- DDX1 rs10109414 (T)- STC1 rs881858 (A)- SLC34A1 rs12460876 (T)- SLC7A9 rs4744712 (A)- PIP5K1B rs2453580 (T)-SLC47A1 rs10774021 (T)-SLC6A13 rs7208487 (T)- CDK12 rs347685 (A)-TFDP2 rs9895661 (T)- BCAS3 rs2928148 (A)- INO80 rs848490 (C)- TMEM60 rs12124078 (A) -CASP9 rs2279463 (A)-VEGFA rs6465825 (T)-SLC22A2 rs1260326 (T)- GCKR rs12917707 (T)- UMOD rs11666497 (T) - SIPA1L3 rs228611 (A) -NFKB1 rs7956634 (T) -PTPRO rs17216707 (T) - BCAS1 rs2802729 (A) - SDCCAG8 rs6459680 (T) -RNF32 rs2712184 (A) - IGFBP5 rs4667594 (A) - LRP2 rs6088580 (C) -TP53INP2 rs7759001 (A) -ZNF204 rs9682041 (T) -SKIL rs12136063 (A) - SYPL2 rs1106766 (T) - INHBC rs10513801 (T) - ETV5 rs164748 (C) - DPEP1 rs8091180 (A) - NFATC1 rs10994860 (T) - A1CF rs3850625 (A) - CACNA1S rs3750082 (A) - KBTBD2 rs163160 (A) -KCNQ1 rs6795744 (A) - WNT7A rs4014195 (C) -AP5B1 rs10277115 (A) -UNCX –0.02 –0.02 –0.01 0 0.01 0.02 –0.01 0 0.01 0.02 Effect on log(eGFRcrea) Effect on log(eGFRcrea) a b rs10491967 (A) - TSPAN9 rs491567 (A)-WDR72 No diabetes Diabetes rs10794720 (T)- WDR37 rs7422339 (A)-CPS1 rs267734 (T)- LASS2 rs3925584 (T)- MPPED2 rs1394125 (A)- UBE2Q2 rs7805747 (A)- PRKAG2 rs2453533 (A)-GATM rs11959928 (A) -DAB2 rs13538 (A)-ALMS1 rs17319721 (A) -SHROOM3 rs626277 (A)- DACH1 rs6431731 (T)- DDX1 rs10109414 (T)- STC1 rs881858 (A)- SLC34A1 rs12460876 (T)- SLC7A9 rs4744712 (A)- PIP5K1B rs2453580 (T)-SLC47A1 rs10774021 (T)-SLC6A13 rs7208487 (T)- CDK12 rs347685 (A)-TFDP2 rs9895661 (T)- BCAS3 rs2928148 (A)- INO80 rs848490 (C)- TMEM60 rs12124078 (A) -CASP9 rs2279463 (A)-VEGFA rs6465825 (T)-SLC22A2 rs1260326 (T)- GCKR rs12917707 (T)- UMOD rs11666497 (T) - SIPA1L3 rs228611 (A) -NFKB1 rs7956634 (T) -PTPRO rs17216707 (T) - BCAS1 rs2802729 (A) - SDCCAG8 rs6459680 (T) -RNF32 rs2712184 (A) - IGFBP5 rs4667594 (A) - LRP2 rs6088580 (C) -TP53INP2 rs7759001 (A) -ZNF204 rs9682041 (T) -SKIL rs12136063 (A) - SYPL2 rs1106766 (T) - INHBC rs10513801 (T) - ETV5 rs164748 (C) - DPEP1 rs8091180 (A) - NFATC1 rs10994860 (T) - A1CF rs3850625 (A) - CACNA1S rs3750082 (A) - KBTBD2 rs163160 (A) -KCNQ1 rs6795744 (A) - WNT7A rs4014195 (C) -AP5B1 rs10277115 (A) -UNCX –0.02 –0.02 –0.01 0 0.01 0.02 –0.01 0 0.01 0.02 Effect on log(eGFRcrea) Effect on log(eGFRcrea) a b Figure 2 \| Association eGFRcrea loci in subjects with and without diabetes. Novel (a) and known (b) loci were considered. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 We used a novel method, Data-driven Expression Prioritized Integration for Complex Traits (DEPICT)23, to prioritize genes at associated loci, to test whether genes at associated loci are highly expressed in speciﬁc tissues or cell types and to test whether speciﬁc biological pathways and gene sets are enriched for genes in associated loci. On the basis of all SNPs with eGFRcrea association P values o10 5 in the discovery meta- analysis, representing 124 independent regions, we identiﬁed at least one signiﬁcantly prioritized gene in 49 regions, including in 9 of the 24 novel genome-wide signiﬁcant regions (Supplementary Table 12). Five tissue and cell type annotations were enriched for expression of genes from the associated regions, including the kidney and urinary tract, as well as hepatocytes and adrenal glands and cortex (Fig. 3a; Supplementary Table 13). Nineteen reconstituted gene sets showed enrichment of genes mapping into the associated regions at a permutation P value o10 5 (Supplementary Table 14; Fig. 4), highlighting processes related to renal development, kidney transmembrane transporter activity, kidney and urogenital system morphology, regulation of glucose metabolism, as well as speciﬁc protein complexes important in renal development. Bioinformatic and functional characterization of new loci. We used several techniques to prioritize and characterize genes underlying the identiﬁed associations, uncover connections between associated regions, detect relevant tissues and assign functional annotations to associated variants. These included expression quantitative trait loci (eQTL) analyses, pathway analyses, DNAse I hypersensitivity site (DHS) mapping, chromatin mapping, manual curation of genes in each region and zebraﬁsh knockdown. eQTL analysis. We performed eQTL analysis using publically available eQTL databases (see Methods). These analyses NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications 4 4 NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 ARTICLE NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 6×10 8×10 2×10 9×10 3×10 2×10 8×10 3×10 8×10 ure 3 \| Bioinformatic analysis of eGFR-associated SNPs. Connection of eGFR-associated SNPs to gene expression and Figure 3 \| Bioinformatic analysis of eGFR-associated SNPs. Connection of eGFR-associated SNPs to gene expression and variant function across a variety of tissues, pathways and regulatory marks was considered. (a) The DEPICT method shows that implicated eGFR-associated genes are highly expressed in particular tissues, including kidney and urinary tract. Shown are permutation test P values (see Methods). (b) Enrichment of eGFRcrea-associated SNPs in DHS according to discovery P value threshold. SNPs from the eGFR discovery genome-wide scan meeting a series of P value thresholds in the range 10 4– 10 16 preferentially map to DHSs, when compared with a set of control SNPs, in 6 of 123 cell types. Represented are main effects odds ratios from a logistic mixed effect model. Cell types indicated with coloured lines had nominally signiﬁcant enrichment ( indicate P values o0.05) at the P value o10 16 threshold and/or were derived from renal tissues (H7esDiffa2d: H7 embryonic stem cells, differentiated 2 days with BMP4, activin A and bFGF; Hae, amniotic epithelial cells; Hrce, renal cortical epithelial cells; Hre, renal epithelial cells; Hrgec, renal glomerular endothelial cells; Rptec, renal proximal tubule epithelial cells; Saec, small airway epithelial cells). (c) ENCODE/Chromatin ChIP-seq mapping: known and replicated novel eGFRcrea-associated SNPs and their perfect proxies were annotated based on genomic location using chromatin annotation maps from different tissues including adult kidney epithelial cells. P values from Fishers’ exact tests for 2 2 tables are reported (signiﬁcance level ¼ 5.6 10 3, see Methods). There is signiﬁcant enrichment of variants mapping to enhancer regions speciﬁcally in kidney but not other non-renal tissues. (to ensure relative homogeneity of the beta coefﬁcients); (3) nearest gene if the signal was located in a region containing a single gene. Using this approach, NFKB1, DPEP1, TSPAN9, NFATC1, WNT7A, PTPRO, SYPL2, UNCX, KBTBD2, SKIL and A1CF were prioritized as likely genes underlying effects at the new loci (Supplementary Table 12). complementary analytic, bioinformatic and functional approaches indicate enrichment of implicated gene products in kidney and urinary tract tissues. A greater proportion of the lead SNPs or their perfect proxies map into gene regulatory regions, speciﬁcally enhancers, in adult renal tissues compared with non- renal tissues. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 Endocrine a b c Transcription associated Poised promoter Active promoter (strong or weak) Enhancer (strong or weak) Proportion OR eGFRcrea SNP in DHS (relative to Eur. GWAS Catalog P<5×10–8) –log10 (P value) * * * * * * * * * * * * 0 1 2 3 4 Cardiovascular Digestive Hemic and immune Integumentary Musculoskeletal Nervous Respiratory Stomatognathic Urogenital Physiological system Liver Adrenal cortex Adrenal glands Urinary tract Kidney Repressed or not mapped region Insulator 4 5 6 3 2 1 0 1 0.8 0.6 0.4 0.2 0.0 –log10(eGFRcrea P value) threshold 6 4 8 10 12 14 16 Human renal glome- rular endothelial cells Small airway epithelial cells Human amniotic epithelial cells Human renal epithelial cells H7 embryonic stem cells, diff. 2 days Human renal cortical epithelial cells Renal proximal tubule epithelial cells Adult kidney Tissue type HepG2 H1 hESC HSMM NHLF K562 GM 12878 NHEK HMEC HUVEC P value comparing proportion of GWAS SNPs in enhancer regions between adult renal and other tissue types Ref. 6×10 –4 8×10 –6 2×10 –6 9×10 –5 3×10 –4 2×10 –5 8×10 –6 3×10 –3 8×10 –6 Figure 3 \| Bioinformatic analysis of eGFR-associated SNPs. Connection of eGFR-associated SNPs to gene expression and variant function across a variety of tissues, pathways and regulatory marks was considered. (a) The DEPICT method shows that implicated eGFR-associated genes are highly expressed in particular tissues, including kidney and urinary tract. Shown are permutation test P values (see Methods). (b) Enrichment of eGFRcrea-associated SNPs in DHS according to discovery P value threshold. SNPs from the eGFR discovery genome-wide scan meeting a series of P value thresholds in the range 10 4– 10 16 preferentially map to DHSs, when compared with a set of control SNPs, in 6 of 123 cell types. Represented are main effects odds ratios from a logistic mixed effect model. Cell types indicated with coloured lines had nominally signiﬁcant enrichment (* indicate P values o0.05) at the P value o10 16 threshold and/or were derived from renal tissues (H7esDiffa2d: H7 embryonic stem cells, differentiated 2 days with BMP4, activin A and bFGF; Hae, amniotic epithelial cells; Hrce, renal cortical epithelial cells; Hre, renal epithelial cells; Hrgec, renal glomerular endothelial cells; Rptec, renal proximal tubule epithelial cells; Saec, small airway epithelial cells). NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 In addition to the importance in the adult kidney, our results indicate a role for kidney function variants during development. (to ensure relative homogeneity of the beta coefﬁcients); (3) nearest gene if the signal was located in a region containing a single gene. Using this approach, NFKB1, DPEP1, TSPAN9, NFATC1, WNT7A, PTPRO, SYPL2, UNCX, KBTBD2, SKIL and A1CF were prioritized as likely genes underlying effects at the new loci (Supplementary Table 12). We investigated the role of these genes during vertebrate kidney development by examining the functional consequences of gene knockdown in zebraﬁsh embryos utilizing antisense morpholino oligonucleotide (MO) technology. After knockdown, we assessed the expression of established renal markers pax2a (global kidney), nephrin (podocytes) and slc20a1a (proximal tubule) at 48 hours post fertilization by in situ hybridization12. In all cases, morphant embryos did not display signiﬁcant gene expression defects compared with controls (Supplementary Table 15). We extend our previous ﬁndings, as well as those from other groups7–13 by identifying 450 genomic loci for kidney function, many of which were not previously known to be connected to kidney function and disease. Using a discovery data set that is nearly double in size to our prior effort7, we are now able to robustly link associated SNPs to kidney-speciﬁc gene regulatory function. Our work further exempliﬁes the continued value of increasing the sample size of GWAS meta-analyses to uncover additional loci and gain novel insights into the mechanisms underlying common phenotypes26. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 The difference between kidney and non-renal tissues was particularly evident for marks that deﬁne enhancers: the proportion of SNPs mapping to weak and strong enhancer regions in the kidney tissue was higher than in all non-kidney tissues (Fishers’ exact test P values from 3.1 10 3 to 7.9 10 6, multiple testing threshold a ¼ 5.6 10 3). Next, we analysed the overlap of the same set of SNPs with H3K4me3 chromatin marks, promoter-speciﬁc histone modiﬁca- tions associated with active transcription24, in order to gather more information about cell-type speciﬁc regulatory potential of eGFRcrea-associated SNPs. Comparing 33 available adult-derived cell types, we found that eGFRcrea-associated SNPs showed the most signiﬁcant overlap with H3K4me3 peaks in adult kidney (P value ¼ 0.0029), followed by liver (P value ¼ 0.0117), and rectal mucosa (P value ¼ 0.0445). Taken together, these ﬁndings are suggestive of cell-type-speciﬁc regulatory roles for eGFR loci, with greatest speciﬁcity for the kidney. Functional characterization of new loci. To prioritize genes for functional studies, we applied gene prioritization algorithms including GRAIL25, DEPICT and manual curation of selected genes in each region (Supplementary Table 12). For each region, gene selection criteria were as follows: (1) either GRAIL P value o0.05 or DEPICT false discovery rate (FDR) o0.05; (2) the effect of a given allele on eGFRcrea and on eGFRcys was direction-consistent and their ratio was between 0.2 and 5 5 NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 (c) ENCODE/Chromatin ChIP-seq mapping: known and replicated novel eGFRcrea-associated SNPs and their perfect proxies were annotated based on genomic location using chromatin annotation maps from different tissues including adult kidney epithelial cells. P values from Fishers’ exact tests for 2 2 tables are reported (signiﬁcance level ¼ 5.6 10 3, see Methods). There is signiﬁcant enrichment of variants mapping to enhancer regions speciﬁcally in kidney but not other non-renal tissues. b OR eGFRcrea SNP in DHS (relative to Eur. GWAS Catalog P<5×10–8) * * * * * * * * * * * * 4 5 6 3 2 1 0 –log10(eGFRcrea P value) threshold 6 4 8 10 12 14 16 Human renal glome- rular endothelial cells Small airway epithelial cells Human amniotic epithelial cells Human renal epithelial cells H7 embryonic stem cells, diff. 2 days Human renal cortical epithelial cells Renal proximal tubule epithelial cells Endocrine a –log10 (P value) 0 1 2 3 4 Cardiovascular Digestive Hemic and immune Integumentary Musculoskeletal Nervous Respiratory Stomatognathic Urogenital Liver Adrenal cortex Adrenal glands Urinary tract Kidney b 8 a Endocrine c Transcription associated Poised promoter Active promoter (strong or weak) Enhancer (strong or weak) Proportion OR (relative to * 0 Cardiovascular Digestive Hemic and immune Integumentary Musculoskeletal Nervous Respiratory Stomatognathic Urogenital Physiological system Repressed or not mapped region Insulator 1 0 1 0.8 0.6 0.4 0.2 0.0 –log10(eGFRcrea P value) threshold 6 4 8 10 12 14 16 Human renal glome- rular endothelial cells Adult kidney Tissue type HepG2 H1 hESC HSMM NHLF K562 GM 12878 NHEK HMEC HUVEC P value comparing proportion of GWAS SNPs in enhancer regions between adult renal and other tissue types Ref. 6×10 –4 8×10 –6 2×10 –6 9×10 –5 3×10 –4 2×10 –5 8×10 –6 3×10 –3 8×10 –6 c Transcription associated Poised promoter Active promoter (strong or weak) Enhancer (strong or weak) Proportion Physiological system Repressed or not mapped region Insulator 1 0.8 0.6 0.4 0.2 0.0 Adult kidney Tissue type HepG2 H1 hESC HSMM NHLF K562 GM 12878 NHEK HMEC HUVEC P value comparing proportion of GWAS SNPs in enhancer regions between adult renal and other tissue types Ref. 6×10 –4 8×10 –6 2×10 –6 9×10 –5 3×10 –4 2×10 –5 8×10 –6 3×10 –3 8×10 –6 c P value comparing proportion of GWAS SNPs in enhancer regions between adult renal and other tissue types Ref. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 BBS2 protein complex Decreased percent body fat Monosaccharide transmembrane transporter activity Hexose transmembrane transporter activity Glucose transmembrane transporter activity Renal tubule atrophy Kidney cortex atrophy Decreased kidney weight Abnormal kidney cortex morphology Urogenital system development Gene set P values Gene set overlap Low Medium High P < 10–5 P < 10–6 P < 10–7 Increased liver weight Organic anion transmembrane transporter activity Negative regulation of cysteine-type endopeptidase activity RXRA protein complex NCOA3 protein complex Decreased embryo size Renal system development Abnormal placenta morphology Partial embryonic lethality during organogenesis Figure 4 \| Gene set overlap analysis. The 19 reconstituted gene sets with P valueo10 5 were considered. Their overlap was estimated by computing the pairwise Pearson correlation coefﬁcient r between each pair of gene sets followed by discretization into one of three bins: 0.3rro0.5, low overlap; 0.5rro0.7, medium overlap; rZ0.7, high overlap. Overlap is shown by edges between gene set nodes and edges representing overlap corresponding to ro0.3 are not shown. The network was drawn with Cytoscape48. BBS2 protein complex Organic anion transmembrane transporter activity Negative regulation of cysteine-type endopeptidase activity Renal system development Figure 4 \| Gene set overlap analysis. The 19 reconstituted gene sets with P valueo10 5 were considered. Their overlap was estimated by computing the pairwise Pearson correlation coefﬁcient r between each pair of gene sets followed by discretization into one of three bins: 0.3rro0.5, low overlap; 0.5rro0.7, medium overlap; rZ0.7, high overlap. Overlap is shown by edges between gene set nodes and edges representing overlap corresponding to ro0.3 are not shown. The network was drawn with Cytoscape48. metabolically active organ that receives 20% of all cardiac output, contains an extensive endothelium-lined capillary network, and is sensitive to ischaemic and toxic injury. As a result, hypertension, cardiovascular diseases and diabetes each affect renal hemody- namics and contribute to kidney injury. However, many of the eGFR-associated SNPs in our GWAS could be assigned gene regulatory function speciﬁcally in the kidney and its epithelial cells, but not in human glomerular endothelial cells or the general vasculature. In addition, variants associated with eGFR were not associated with vascular traits, such as blood pressure or myocardial infarction. Taken together, these ﬁndings suggest that genetic determinants of eGFR may be mediated largely through direct effects within the kidney. NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 phenotype upon gene knockdown does not mean that the gene cannot be the one underlying the observed association signal in humans. Finally, our conclusions that eGFRcrea-associated SNPs regulate the expression of nearby genes speciﬁcally in kidney epithelial cells are based on DHSs, H3K4me3 chromatin marks and chromatin annotation maps. Since these analyses rely mostly on variant positions, additional functional investigation such as luciferase assay that assess transcriptional activity more directly are likely to gain additional insights into the variants’ mechanism of action. The kidney speciﬁcity for loci we identiﬁed may have important translational implications, particularly since our DHS and chromatin annotation analyses suggest that at least a set of gene regulatory mechanisms is important in the adult kidney. Kidney-speciﬁc pathways are important for the development of novel therapies to prevent and treat CKD and its progression with minimal risk of toxicity to other organs. Finally, the biologic insights provided by these new loci may help elucidate novel mechanisms and pathways implicated not only in CKD but also of kidney function in the physiological range. Second, despite the speciﬁcity related to renal processes, we also identiﬁed several SNPs that are associated with eGFR in diabetes, and our pathway analyses uncovered gene sets associated with glucose transporter activity and abnormal glucose homeostasis. Uncovering bona ﬁde genetic loci for diabetic CKD has been difﬁcult. We have now identiﬁed a total of 19 SNPs that demonstrate at least nominal association with eGFR in diabetes. The diabetes population is at particularly high risk of CKD, and identifying kidney injury pathways may help develop new treatments for diabetic CKD. y p y g g In conclusion, we have conﬁrmed 29 genomic loci and identiﬁed 24 new loci in association with kidney function that together highlight target organ-speciﬁc regulatory mechanisms related to kidney function. Finally, even though CKD is primarily a disease of the elderly, our pathway enrichment analyses highlight developmental processes relevant to the kidney and the urogenital tract. Kidney disease has been long thought to have developmental origins, in part related to early programming (Barker hypothesis)27, low birth weight, nephron endowment and early growth and early-life nutrition28. Our pathway enrichment analyses suggest that developmental pathways such as placental morphology, kidney weight and embryo size, as well as protein complexes of importance in renal development may in part contribute to the developmental origins of CKD. Discussion d ﬁ There are several messages from our work. First, many of the genetic variants associated with eGFR appear to affect processes speciﬁcally within the kidney. The kidney is a highly vascular and Discussion We identiﬁed 24 new loci in association with eGFR and conﬁrmed 29 previously identiﬁed loci. A variety of 6 ARTICLE ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 estimated to test for consistent effect direction with the discovery stage); (3) q-value o0.05 in the replication stage. Q-values were estimated using the package QVALUE36 in R. The tuning parameter lambda for the estimation of the overall proportion of true null hypotheses, p0, was estimated using the bootstrap method37. When the third criterion was not satisﬁed, the locus was declared ‘suggestive’. o15ml min per 1.73m2 were set to 15, and those 4200 were set to 200mlmin 1 per 1.73 m2. CKD was deﬁned as eGFRcrea o60ml min 1 per 1.73 m2. o15ml min per 1.73m2 were set to 15, and those 4200 were set to 200mlmin 1 per 1.73 m2. CKD was deﬁned as eGFRcrea o60ml min 1 per 1.73 m2. Diabetes was deﬁned as fasting glucose Z126 mg dl 1, pharmacologic treatment for diabetes or by self-report. In all studies, diabetes and kidney function were assessed at the same point in time. Genotypes. Genotyping was conducted in each study as speciﬁed in Supplementary Tables 18 and 19. After applying appropriate quality ﬁlters, 45 studies used markers of highest quality to impute B2.5 million SNPs, based on European-ancestry haplotype reference samples (HapMap II CEU). Four studies based their imputation on the 1000 Genomes Project data. Imputed genotypes were coded as the estimated number of copies of a speciﬁed allele (allelic dosage). Power analysis. With the sample size achieved in the combined analysis of stage 1 and stage 2 data, the power to assess replication at the canonical genome-wide signiﬁcance level of 5.0 10 8 was estimated with the software QUANTO38 version 1.2.4, assuming the same MAF and effect size observed in the discovery sample. Power to replicate associations ranged from 87 to 100% for eGFRcrea associated SNPs (median ¼ 98%), from 72 to 96% for the CKD-associated SNPs, and was equal to 59% for the eGFRcys-associated SNP (Supplementary Table 3). Genome-wide association analysis. By following a centralized analysis plan, each study regressed sex- and age-adjusted residuals of the logarithm of eGFRcrea or eGFRcys on SNP dosage levels. Logistic regression of CKD status was performed on SNP dosage levels adjusting for sex and age. For all traits, adjustment for appropriate study-speciﬁc features, including study site and genetic principal components was included in the regression and family-based studies appropriately accounted for relatedness. Associations stratiﬁed by diabetes and hypertension status. Proportion of phenotypic variance explained. The percent of phenotypic 53 Proportion of phenotypic variance explained. The percent of phenotypic variance explained by novel and known loci was estimated as P 53 i¼1 R2 i , where R2 i ¼ b2 i varðSNPiÞ=varðyÞ is the coefﬁcient of determination for each of the 53 individual SNPs associated with eGFRcrea uncovered to date (24 novel and 29 known ones), bi is the estimated effect of the ith SNP on y, y corresponds to the sex- and age-adjusted residuals of the logarithm of eGFRcrea and var(SNPi) ¼ 2 MAFSNPi (1 MAFSNPi)39. Var(y) was estimated in the ARIC study and all loci were assumed to have independent effects on the phenotype. R2 i ¼ b2 i varðSNPiÞ=varðyÞ is the coefﬁcient of determination for each of the 53 individual SNPs associated with eGFRcrea uncovered to date (24 novel and 29 known ones), bi is the estimated effect of the ith SNP on y, y corresponds to the sex- and age-adjusted residuals of the logarithm of eGFRcrea and var(SNPi) ¼ 2 MAFSNPi (1 MAFSNPi)39. Var(y) was estimated in the ARIC study and all loci were assumed to have independent effects on the phenotype. After removing SNPs with MAF of o0.05 and which were available in o50% of the studies, SNPs with a P value of r10 6 were selected and clustered into independent loci through LD pruning based on an r2 of r0.2 within a window of ±1 MB to each side of the index SNP. After removing loci containing variants that have been previously replicated at a P value of 5.0 10 8 (refs 7,8), the SNP with the lowest P value within each locus was selected for replication (‘index SNP’). If a SNP had an association P value of r10 6 with more than one trait, the trait where the SNP had the lowest P value was selected as discovery trait/stratum. Altogether, this resulted in 48 SNPs: 34 from eGFRcrea, 9 from eGFRcrea among those without diabetes, 4 from CKD and 1 from eGFRcys. Test for SNP associations with related traits. We performed evaluations of SNP association with results generated from consortia investigating other traits. Speci- ﬁcally, we evaluated systolic and diastolic blood pressure in ICBP17, myocardial infarction in CARDiOGRAM18, left ventricular mass19, heart failure20, the urinary albumin to creatinine ratio (CKDGen consortium, personal communication) and fasting plasma glucose in MAGIC21. ARTICLE For all the 24 novel and 29 known SNPs, the difference between the SNP effect on eGFRcrea in the diabetes versus the non-diabetes groups was assessed by means of a two-sample t-test for correlated data at a signiﬁcance level of 0.05. We used the following two- sample t-test for correlated data: bDM bnonDM ð Þ t ¼ bDM bnonDM ð Þ s:e: bDM ð Þ2 þ s:e: bnonDM ð Þ2 2r bDM; bnonDM ð Þs:e: bDM ð Þs:e: bnonDM ð Þ 0:5 ; t ¼ s:e: bDM ð Þ2 þ s:e: bnonDM ð Þ2 2r bDM; bnonDM ð Þs:e: bDM ð Þs:e Stage 1 discovery meta-analysis. GWAS of eGFRcrea were contributed by 48 studies (total sample size, N ¼ 133,413); 45 studies contributed GWAS data for the non-diabetes subgroup (N ¼ 118,448) and 39 for the diabetes group (N ¼ 11,522). GWAS of CKD were comprised by 43 studies, for a total sample size of 117,165, including 12,385 CKD cases. GWAS of eGFRcys were comprised by 16 studies for a total sample size of 32,834. All GWAS ﬁles underwent quality control using the GWAtoolbox package32 in R, before including them into the meta-analysis. Genome-wide meta-analysis was performed with the software METAL33, assuming ﬁxed effects and using inverse-variance weighting. The genomic inﬂation factor l was estimated for each study as the ratio between the median of all observed test statistics (b/s.e.)2 and the expected median of a w2 with 1 degree of freedom, with b and s.e. representing the effect of each SNP on the phenotype and its standard error, respectively34. Genomic-control (GC) correction was applied to P values and s.e.’s in case of l41 (ﬁrst GC correction). SNPs with an average minor allele frequency (MAF) of Z0.01 were used for the meta-analysis. To limit the possibility of false positives, after the meta-analysis, a second GC correction on the aggregated results was applied. Between-study heterogeneity was assessed through the I2 statistic. Stage 1 discovery meta-analysis. GWAS of eGFRcrea were contributed by 48 studies (total sample size, N ¼ 133,413); 45 studies contributed GWAS data for the non-diabetes subgroup (N ¼ 118,448) and 39 for the diabetes group (N ¼ 11,522). GWAS of CKD were comprised by 43 studies, for a total sample size of 117,165, including 12,385 CKD cases. GWAS of eGFRcys were comprised by 16 studies for a total sample size of 32,834. ARTICLE All GWAS ﬁles underwent quality control using the GWAtoolbox package32 in R, before including them into the meta-analysis. Genome-wide meta-analysis was performed with the software METAL33, assuming ﬁxed effects and using inverse-variance weighting. The genomic inﬂation factor l was estimated for each study as the ratio between the median of all observed test statistics (b/s.e.)2 and the expected median of a w2 with 1 degree of freedom, with b and s.e. representing the effect of each SNP on the phenotype and its standard error, respectively34. Genomic-control (GC) correction was applied to P values and s.e.’s in case of l41 (ﬁrst GC correction). SNPs with an average minor allele frequency (MAF) of Z0.01 were used for the meta-analysis. To limit the possibility of false positives, after the meta-analysis, a second GC correction on the aggregated results was applied. Between-study heterogeneity was assessed through the I2 statistic. Stage 1 discovery meta-analysis. GWAS of eGFRcrea were contributed by 48 studies (total sample size, N ¼ 133,413); 45 studies contributed GWAS data for the non-diabetes subgroup (N ¼ 118,448) and 39 for the diabetes group (N ¼ 11,522). GWAS of CKD were comprised by 43 studies, for a total sample size of 117,165, including 12,385 CKD cases. GWAS of eGFRcys were comprised by 16 studies for a total sample size of 32,834. All GWAS ﬁles underwent quality control using the GWAtoolbox package32 in R, before including them into the meta-analysis. 33 Stage 1 discovery meta-analysis. GWAS of eGFRcrea were contributed by 48 studies (total sample size, N ¼ 133,413); 45 studies contributed GWAS data for the non-diabetes subgroup (N ¼ 118,448) and 39 for the diabetes group (N ¼ 11,522). GWAS of CKD were comprised by 43 studies, for a total sample size of 117,165, including 12,385 CKD cases. GWAS of eGFRcys were comprised by 16 studies for a total sample size of 32,834. All GWAS ﬁles underwent quality control using the GWAtoolbox package32 in R, before including them into the meta-analysis. G id l i f d i h h f METAL33 i where bDM and bnonDM represent the SNP effects on log(eGFRcrea) in the two groups, s.e. is the standard error of the estimate and r(.) indicates the correlation between effects in the two groups, which was estimated as 0.044 by sampling 100,000 independent SNPs from our DM and nonDM GWAS, after removing known and novel loci associated with eGFRcrea. Proportion of phenotypic variance explained. The percent of phenotypic 53 In total, we performed 165 tests, corresponding to 7 traits tested for association against each of the 24 novel SNP, with the exception of myocardial infarction for which results from 3 SNPs were not available (Supplementary Table 6). Signiﬁcance was evaluated at the Bonferroni corrected level of 0.05/165 ¼ 0.0003. Stage 2 replication analysis. In silico replication analysis for any of the studied traits was carried out using eight independent studies whose genotyping platforms are provided in Supplementary Table 19. De novo genotyping was performed in seven additional studies (N ¼ 22,850 individuals) of European ancestry (Supplementary Table 20), including the Bus Sante´, ESTHER, KORA-F3 (subset of F3 without GWAS), KORA-F4 (subset of F4 without GWAS), Ogliastra Genetic Park (OGP, without Talana whose GWAS was included in the discovery analysis), SAPHIR and SKIPOGH studies (Supplementary Table 20). Summarizing all in silico and de novo replication studies (Supplementary Table 1), replication data for eGFRcrea were contributed by 14 studies (total sample size ¼ 42,166), which also contributed eGFRcrea results from non-diabetes (13 studies, N ¼ 36,433) and diabetes samples (13 studies, N ¼ 4,955). Thirteen studies contributed replication data on CKD (N ¼ 33,972; 4,245 CKD cases; studies with o50 CKD cases were excluded) and ﬁve on eGFRcys (N ¼ 14,930). Lookup of replicated loci in the NHGRI GWAS catalog. All replicated SNPs, as well as SNPs in LD (r240.2) within ±1 MB distance were checked for their association with other traits according to the NHGRI GWAS catalog40 (accessed April 14, 2014). SNP assessments in other ethnic groups. We performed cross-ethnicity SNP evaluations in participants of African ancestry from a meta-analysis of African ancestry individuals and from participants of Asian descent from the AGEN consortium11. SNP assessments in other ethnic groups. We performed cross-ethnicity SNP evaluations in participants of African ancestry from a meta-analysis of African ancestry individuals and from participants of Asian descent from the AGEN consortium11. ) y ( ) Association between eGFRcrea, CKD and eGFRcys and each of the 48 SNPs in the replication studies was assessed using the same analysis protocol detailed for the discovery studies above. Quality control of the replication ﬁles was performed with the same software as described above. African ancestry meta-analysis. We performed ﬁxed-effect meta-analysis of the genome-wide association data from 12 African ancestry studies (Supplementary Table 8) with imputation to HapMap reference panel, based on inverse-variance weighting using METAL. Methods Overview Methods Overview. This was a collaborative meta-analysis with a distributive data model. Brieﬂy, an analysis plan was created and circulated to all participating studies. Studies then uploaded study-speciﬁc data centrally; ﬁles were cleaned, and a speciﬁc meta-analysis for each trait was performed. Details regarding each step are provided below. All participants in all discovery and replication studies provided informed consent. Each study had its research protocol approved by the local ethics committee. Phenotype deﬁnitions. Serum creatinine was measured in each discovery and replication study as described in Supplementary Tables 16 and 17, and statistically calibrated to the US nationally representative National Health and Nutrition Examination Study data in all studies to account for between-laboratory varia- tion9,29,30. eGFRcrea was estimated using the four-variable Modiﬁcation of Diet in Renal Disease Study Equation. Cystatin C, an alternative biomarker for kidney function, was measured in a sub-set of participating studies. eGFRcys was estimated as 76.7 (serum cystatin C) 1.19 (ref. 31). eGFRcrea and eGFRcys values g A limitation of our work is that causal variants and precise molecular mechanisms underlying the observed associations were not identiﬁed and will require additional experimental follow-up projects. Our attempt to gain insights into potentially causal genes through knockdown in zebraﬁsh did not yield any clear CKD candidate gene, although the absence of a zebraﬁsh 7 ARTICLE For a large sample size, as in our case, t follows a standard normal distribution. A similar analysis was performed to compare results in subjects with and without hypertension, based on results from our previous work7. The correlation between the two strata was of 0.01. ARTICLE We tested whether the proportion of SNPs pointing to either strong or weak enhancers in the human kidney tissue cells was different from that of the other nine tissues by means of a Fishers’ exact test for 2 2 tables, contrasting each of the nine cell lines listed above against the reference kidney cell line, at a Bonferroni-corrected signiﬁcance level of 0.05/9 ¼ 5.6 10 3. Gene Relationships Across Implicated Loci (GRAIL). To prioritize the gene(s) most likely to give rise to association signals in a given region, the software GRAIL was used25. The index SNP of all previously known kidney function associated regions, as well as the novel SNPs identiﬁed here was used as input, using the CEU HapMap (hg18 assembly) and the functional datasource text_2009_03, established before the publication of kidney function-related GWAS. Results from GRAIL were used to prioritize genes for follow-up functional work. Expression quantitative trait loci analysis. We identiﬁed alias rsIDs and proxies (r240.8) for our index SNPs using SNAP software across 4 HapMap builds. SNP rsIDs and aliases were searched for primary SNPs and LD proxies against a collected database of expression SNP (eSNP) results. The collected eSNP results met criteria for statistical thresholds for association with gene transcript levels in their respective original analyses (for references see Supplementary Table 11). Correlation of selected eSNPs to the best eSNPs per transcript per expression quantitative trait loci (eQTL) data set were assessed by pairwise LD. All results are reported in Supplementary Table 11. DEPICT analysis. In this work, we ﬁrst used PLINK42 to identify independently associated SNPs using all SNPs with eGFRcrea association P values o10 5 (HapMap release 27 CEU data43; LD r2 threshold ¼ 0.01; physical kb threshold ¼ 1,000). We then used the DEPICT method23 to construct associated regions by mapping genes to independently associated SNPs if they overlapped or resided within LD (r240.5) distance of a given associated SNP. After merging overlapping regions and discarding regions that mapped within the major histocompatibility complex locus (chromosome 6, base pairs 20,000,000–40,000,000), 124 non- overlapping regions remained that covered a total of 320 genes. ARTICLE ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 of Trans-ethnic Association studies) software22. We combined the 48 European ancestry studies that contributed eGFRcrea, which were included in stage 1 discovery meta-analysis, and the 12 African ancestry studies mentioned above for a total sample size of 150,253 samples. We limited our analysis to biallelic SNPs with MAF Z0.01 and imputation quality r2Z0.3. Relatedness between the 60 studies was estimated using default settings from up to 5.9 million SNPs. Only SNPs that were present in more than 25 European ancestry studies and 6 African ancestry studies (total sample size Z120,000) were considered after meta-analysis. Genome- wide signiﬁcance was deﬁned as a log10 Bayes’ Factor (log10BF) Z6 (ref. 41). Interrogation of human kidney chromatin annotation maps. Different chro- matin modiﬁcation patterns can be used to generate tissue-speciﬁc chromatin-state annotation maps. These can serve as a valuable resource to discover regulatory regions and study their cell-type-speciﬁc distributions and activities, which may help with the interpretation especially of intergenic variants identiﬁed in associa- tion studies45. We therefore investigated the genomic mapping of the known and replicated novel index SNPs, as well as their perfect LD proxies (n ¼ 173, r2 ¼ 1 for proxies) using a variety of resources, including chromatin maps generated from human kidney tissue cells (HKC-E cells). Chromatin immune-precipitation sequencing (ChIP-seq) data from human kidney samples were generated by NIH Roadmap Epigenomics Mapping Consortium46. Brieﬂy, proximal tubule cells derived from an adult human kidney were collected and cross-linked with 1% formaldehyde. Subsequently, ChIP-seq was conducted using whole-cell extract from adult kidney tissue as the input (GSM621638) and assessing the following chromatin marks: H3K36me3 (GSM621634), H3K4me1 (GSM670025), H3K4me3 (GSM621648), H3K9ac (GSM772811) and H3K9me3 (GSM621651). The MACS version 1.4.1 (model-based analysis of ChIP-Seq) peak-ﬁnding algorithm was used to identify regions of ChIP-Seq enrichment47. A FDR threshold of enrichment of 0.01 was used for all data sets. The resulting genomic coordinates in bed format were further used in ChromHMM v1.06 for chromatin annotation45. For comparison, the same genomic coordinates were investigated in chromatin annotation maps of renal tissue, as well as across nine different cell lines from the ENCODE Project: umbilical vein endothelial cells (HUVEC), mammary epithelial cells (HMEC), normal epidermal keratinocytes (NHEK), B-lymphoblastoid cells (GM12878), erythrocytic leukemia cells (K562), normal lung ﬁbroblasts (NHLF), skeletal muscle myoblasts (HSMM), embryonic stem cells (H1 ES) and hepatocellular carcinoma cells (HepG2). ARTICLE Finally, we ran the DEPICT software program on the 124 regions to prioritize genes that may represent promising candidates for experimental follow up studies, identify reconstituted gene sets that are enriched in genes from associated regions and therefore may provide insight into general kidney function biology, and identify tissue and cell-type annotations in which genes from associated regions are highly expressed. Speciﬁcally, for each tissue, the DEPICT method performs a t-test comparing the tissue-speciﬁc expression of eGFRcrea-associated genes and all other genes. Next, for each tissue, empirical enrichment P values are computed by repeatedly sampling random sets of loci (matched to the actual eGFRcrea loci by gene density) from the entire genome to estimate the empirical mean and s.d. of the enrichment statistic’s null distribution. To visualize the nineteen reconstituted gene sets with P value o1e 5 (Fig. 4), we estimated their overlap by computing the pairwise Pearson correlation coefﬁcient r between each pair of gene sets followed by discretization into one of three bins; 0.3rro0.5, low overlap; 0.5rro0.7, medium overlap; rZ0.7, high overlap. Functional characterization of new loci. Replicated gene regions were prioritized for functional studies using the following criteria: (1) GRAIL identiﬁcation of a gene in each region of P valueo0.05 or DEPICT, FDR o0.05); (2) an eGFRcrea to eGFRcys ratio between 0.2 and 5 with direction consistency between the beta coefﬁcients; (3) nearest gene if the signal was located in a gene-poor region. The list of genes selected for functional work can be found in Supplementary Table 12. This same prioritization scheme was also used to assign locus names. Morpholino knockdowns were performed in zebraﬁsh. Zebraﬁsh (strain Tu¨bingen, TU) were maintained according to established Harvard Medical School Institutional Animal Care and Use Committee protocols (protocol # 04626). Male and female ﬁsh were mated (age 6–12 months) for embryo production. Embryos were injected at the one-cell stage with MOs (GeneTools) designed to block either the ATG start site or an exon–intron splice site of the target gene (Supplementary Table 21). In cases where human loci are duplicated in zebraﬁsh, both orthologues were knocked down simultaneously by combination MO injection. MOs were injected in escalating doses at concentrations up to 250 mM. Embryos were ﬁxed in 4% paraformaldehyde at 48 h post fertilization for in situ hybridization using published methods (http://zﬁn.org/ ZFIN/Methods/ThisseProtocol.html). ARTICLE Gene expression was visualized using established renal markers pax2a (global kidney), nephrin (podocytes) and slc20a1a (proximal tubule). The number of morphant embryos displaying abnormal gene expression was compared with control embryos by means of a Fisher’s exact test. DNase I hypersensitivity analysis. The overlap of SNPs associated with eGFR- crea at Po10 4 with DHSs was examined using publically available data from the Epigenomics Roadmap Project and ENCODE. In all, DHS mappings were available for 123 mostly adult cells and tissues44 (downloaded from http:// hgdownload.cse.ucsc.edu/goldenPath/hg19/encodeDCC/wgEncodeUwDnase/). The analysis here pertains to DHS’s deﬁned as ‘broad’ peaks, which were available as experimental replicates (typically duplicates) for the majority of cells and tissues. SNPs from our stage 1 eGFRcrea GWAS meta-analysis were ﬁrst clumped in PLINK42 in windows of 100 kb and maximum r2 of 0.1 using LD relationships from the 1,000 Genomes EUR panel (phase I, v3, 3/14/2012 haplotypes) using a series of P value thresholds (10 4, 10 6, 10 8, ... and 10 16). LD proxies of the index SNPs from the clumping procedure were then identiﬁed by LD tagging in PLINK with r2 ¼ 0.8 in windows of 100 kb, again using LD relationships in the 1000G EUR panel, restricted to SNPs with MAF 41% and also present in the HapMap2 CEU population. A reference set of control SNPs was constructed using the same clumping and tagging procedures applied to NHGRI GWAS catalog SNPs (available at http://www.genome.gov/gwastudies/, accessed 13 March 2013) with discovery P values o5.0 10 8 in European populations. In total, there were 1,204 such reference SNPs after LD pruning. A small number of reference SNPs or their proxies overlapping with the eGFRcrea SNPs or their proxies were excluded. For each cell-type and P value threshold, the enrichment of eGFR SNPs (or their LD proxies) mapping to DHSs relative to the GWAS catalog reference SNPs (or their LD proxies) was expressed as an odds ratio from logistic mixed effect models that treated the replicate peak determinations as random effects (lme4 package in R). Signiﬁcance for enrichment odds ratio was derived from the signiﬁcance of beta coefﬁcients for the main effects in the mixed models. NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications Proportion of phenotypic variance explained. The percent of phenotypic 53 Only SNPs with MAF Z0.01 and imputation quality r2Z0.3 were considered for the meta-analysis. After meta-analysis, we removed SNPs with MAF o0.05 and which were available in o50% of the studies. Sta- tistical signiﬁcance was assessed at the standard threshold of 5.0 10 8. Genomic control correction was applied at both the individual study level before meta- analysis and after the meta-analysis. African ancestry meta-analysis. We performed ﬁxed-effect meta-analysis of the genome-wide association data from 12 African ancestry studies (Supplementary Table 8) with imputation to HapMap reference panel, based on inverse-variance weighting using METAL. Only SNPs with MAF Z0.01 and imputation quality r2Z0.3 were considered for the meta-analysis. After meta-analysis, we removed SNPs with MAF o0.05 and which were available in o50% of the studies. Sta- tistical signiﬁcance was assessed at the standard threshold of 5.0 10 8. Genomic control correction was applied at both the individual study level before meta- analysis and after the meta-analysis. We performed a combined ﬁxed-effect meta-analysis of the double-GC corrected results from the discovery meta-analysis and the replication studies, based on inverse-variance weighting. The total sample size in the combined analysis of eGFRcrea was 175,579 subjects (154,881 in the non-diabetes stratum and 16,477 in the diabetes stratum; the sum of these two sample sizes is smaller than the sample size of the overall analysis because some studies did not contribute both strata), 151,137 samples for CKD (16,630 CKD cases) and 47,764 for eGFRcys. Three criteria were used to ensure validity of novel loci declared as signiﬁcant: (1) P value from the combined meta-analysis r5.0 10 8 in accordance with previously published guidelines35; (2) direction-consistent associations of the beta coefﬁcients in stage 1 and stage 2 (one-sided P values were Transethnic meta-analysis. We performed a trans-ethnic meta-analysis of GWAS data from cohorts of different ethnic backgrounds using MANTRA (Meta-Analysis Transethnic meta-analysis. We performed a trans-ethnic meta-analysis of GWAS data from cohorts of different ethnic backgrounds using MANTRA (Meta-Analysis 8 NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunica Author contributions Subject recruitment: C Helmer, P Hamet, TB Harris, T Aspelund, V Gudnason, AR Shuldiner, BD Mitchell, J Coresh, WHL Kao, M Cavalieri, R Schmidt, JB Whitﬁeld, NG Martin, L Ferrucci, P Mitchell, I Guessous, DS Siscovick, O Devuyst, A Metspalu, BA Oostra, CM van Duijn, BK Kra¨mer, ST Turner, S Kloiber, PS Wild, BI Freedman, MA McEvoy, RJ Scott, AB Zonderman, MK Evans, GC Curhan, A Adeyemo, CN Rotimi, D Cusi, A Lupo, G Gambaro, P d’Adamo, A Robino, S Ulivi, D Ruggiero, M Ciullo, R Sorice, D Toniolo, C Gieger, C Meisinger, CA Bo¨ger, HE Wichmann, T Illig, W Koenig, I Rudan, I Kolcic, M Boban, T Zemunik, PP Pramstaller, EP Bottinger, BW Penninx, Å Johansson, I Persico, M Pirastu, JF Wilson, SH Wild, A Franke, G Jacobs, GJ Navis, IM Leach, BM Buckley, I Ford, JW Jukema, N Probst-Hensch, B Paulweber, L Kedenko, F Kronenberg, R Rettig, R Biffar, S Stracke, H Vo¨lzke, P Muntner, JK Fernandes, MM Sale, B Ponte, D Ackermann, M Pruijm, GB Ehret, A To¨njes, ll l k k h k h k Subject recruitment: C Helmer, P Hamet, TB Harris, T Aspelund, V Gudnason, AR Shuldiner, BD Mitchell, J Coresh, WHL Kao, M Cavalieri, R Schmidt, JB Whitﬁeld, NG Martin, L Ferrucci, P Mitchell, I Guessous, DS Siscovick, O Devuyst, A Metspalu, BA Oostra, CM van Duijn, BK Kra¨mer, ST Turner, S Kloiber, PS Wild, BI Freedman, MA McEvoy, RJ Scott, AB Zonderman, MK Evans, GC Curhan, A Adeyemo, CN Rotimi, D Cusi, A Lupo, G Gambaro, P d’Adamo, A Robino, S Ulivi, D Ruggiero, M Ciullo, R Sorice, D Toniolo, C Gieger, C Meisinger, CA Bo¨ger, HE Wichmann, T Illig, W Koenig, I Rudan, I Kolcic, M Boban, T Zemunik, PP Pramstaller, EP Bottinger, BW Penninx, Å Johansson, I Persico, M Pirastu, JF Wilson, SH Wild, A Franke, G Jacobs, GJ Navis, IM Leach, BM Buckley, I Ford, JW Jukema, N Probst-Hensch, B Paulweber, L Kedenko, F Kronenberg, R Rettig, R Biffar, S Stracke, H Vo¨lzke, P Muntner, JK F d MM S l B P t D A k M P ij GB Eh t A T¨ j 33. Willer, C. J., Li, Y. & Abecasis, G. R. METAL: fast and efﬁcient meta-analysis of genomewide association scans. Bioinformatics 26, 2190–2191 (2010). 34. Devlin, B. & Roeder, K. Genomic control for association studies. Biometrics 55, 997–1004 (1999). 35. Skol, A. ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 11. Okada, Y. et al. Meta-analysis identiﬁes multiple loci associated with kidney function-related traits in east Asian populations. Nat. Genet. 44, 904–909 (2012). 41. Wang, X. et al. Comparing methods for performing trans-ethnic meta-analysis of genome-wide association studies. Hum. Mol. Genet. 22, 2303–2311 (2013). 42. Purcell, S. et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet. 81, 559–575 (2007). 12. Liu, C. T. et al. Genetic association for renal traits among participants of African ancestry reveals new loci for renal function. PLoS Genet. 7, e1002264 (2011). 43. International HapMap Consortium. The International HapMap Project. Nature 426, 789–796 (2003). 44. Maurano, M. T. et al. Systematic localization of common disease-associated variation in regulatory DNA. Science 337, 1190–1195 (2012). 13. Boger, C. A. et al. Association of eGFR-Related Loci Identiﬁed by GWAS with Incident CKD and ESRD. PLoS Genet. 7, e1002292 (2011). g y 45. ENCODE Project Consortium et al. An integrated encyclopedia of DNA 45. ENCODE Project Consortium et al. An integrated encyclopedia of DN elements in the human genome. Nature 489, 57–74 (2012). 14. Chasman, D. I. et al. Integration of genome-wide association studies with biological knowledge identiﬁes six novel genes related to kidney function. Hum. Mol. Genet. 21, 5329–5343 (2012). elements in the human genome. Nature 489, 57–74 (2012). 46. Ko, Y. A. et al. Cytosine methylation changes in enhancer regions of core pro- ﬁbrotic genes characterize kidney ﬁbrosis development. Genome Biol. 14, R108 (2013). 15. Pattaro, C. et al. A meta-analysis of genome-wide data from ﬁve European isolates reveals an association of COL22A1, SYT1, and GABRR2 with serum creatinine level. BMC Med. Genet. 11, 41 (2010). 47. Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137 (2008). 16. Trudu, M. et al. Common noncoding UMOD gene variants induce salt- sensitive hypertension and kidney damage by increasing uromodulin expression. Nat. Med. 19, 1655–1660 (2013). 48. Shannon, P. et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 13, 2498–2504 (2003). p 17. International Consortium for Blood Pressure Genome-Wide Association Studies et al. Genetic variants in novel pathways inﬂuence blood pressure and cardiovascular disease risk. Nature 478, 103–109 (2011). Author contributions D., Scott, L. J., Abecasis, G. R. & Boehnke, M. Joint analysis is more efﬁcient than replication-based analysis for two-stage genome-wide association studies. Nat. Genet. 38, 209–213 (2006). GJ Navis, IM Leach, BM Buckley, I Ford, JW Jukema, N Probst-Hensch, B Paulweber, L Kedenko, F Kronenberg, R Rettig, R Biffar, S Stracke, H Vo¨lzke, P Muntner, 36. Storey, J. D. A direct approach to false discovery rates. J. R. Stat. Soc. B 64, 479–498 (2002). JK Fernandes, MM Sale, B Ponte, D Ackermann, M Pruijm, GB Ehret, A To¨njes, JC Denny, RJ Carroll, O Polasek, J Viikari, M Ka¨ho¨nen, O Raitakari, T Lehtima¨ki. 37. Storey, J. D., Taylor, J. E. & Siegmund, D. Strong control, conservative point estimation, and simultaneous conservative consistency of false discovery rates: a uniﬁed approach. J. R. Stat. Soc. B 66, 187–205 (2004). Interpretation of results: C Helmer, JC Lambert, M Metzger, B Stengel, V Chouraki, J Tremblay, J Chalmers, M Woodward, P Hamet, AV Smith, A Parsa, JR O’Connell, A Tin, A Ko¨ttgen, M Li, WHL Kao, Y Li, EG Holliday, J Attia, I Guessous, CA Peralta, AC Morrison, JF Felix, C Pattaro, G Li, IH de Boer, O Devuyst, H Lin, A Isaacs, V Emilsson, AD Johnson, CS Fox, M Olden, Q Yang, EJ Atkinson, M de Andrade, ST Turner, T Zeller, J Ding, Y Liu, M Nalls, A Adeyemo, D Shriner, D Cusi, E Salvi, V Mijatovic, D Ruggiero, R Sorice, AW Dreisbach, AY Chu, DI Chasman, CA Bo¨ger, IM Heid, M Gorski, B Tayo, C Fuchsberger, H Snieder, IM Nolte, W Igl, K Susztak, Interpretation of results: C Helmer, JC Lambert, M Metzger, B Stengel, V Chouraki, J Tremblay, J Chalmers, M Woodward, P Hamet, AV Smith, A Parsa, JR O’Connell, A Tin, A Ko¨ttgen, M Li, WHL Kao, Y Li, EG Holliday, J Attia, I Guessous, CA Peralta, AC Morrison, JF Felix, C Pattaro, G Li, IH de Boer, O Devuyst, H Lin, A Isaacs, V Emilsson, AD Johnson, CS Fox, M Olden, Q Yang, EJ Atkinson, M de Andrade, ST Turner, T Zeller, J Ding, Y Liu, M Nalls, A Adeyemo, D Shriner, D Cusi, E Salvi, V Mijatovic, D Ruggiero, R Sorice, AW Dreisbach, AY Chu, DI Chasman, CA Bo¨ger, IM Heid, M Gorski, B Tayo, C Fuchsberger, H Snieder, IM Nolte, W Igl, K Susztak, pp 38. Gauderman, W. J. Acknowledgements Study-speciﬁc acknowledgements and funding sources for participating studies are reported in Supplementary Note. Zebraﬁsh work was supported by NIH R01DK090311 and R24OD017870 to W.G. 18. Schunkert, H. et al. Large-scale association analysis identiﬁes 13 new susceptibility loci for coronary artery disease. Nat. Genet. 43, 333–338 (2011). reported in Supplementary Note. Zebraﬁsh work was supported by NIH R01DK090311 and R24OD017870 to W.G. 19. Vasan, R. S. et al. Genetic variants associated with cardiac structure and function: a meta-analysis and replication of genome-wide association data. JAMA 302, 168–178 (2009). Author contributions Study management: C Helmer, M Metzger, J Tremblay, J Chalmers, M Woodward, P Hamet, G Eiriksdottir, TB Harris, T Aspelund, V Gudnason, A Parsa, AR Shuldiner, BD Mitchell, E Boerwinkle, J Coresh, WHL Kao, R Schmidt, L Ferrucci, E Rochtchina, JJ Wang, J Attia, P Mitchell, I Guessous, JM Gaspoz, M Bochud, DS Siscovick, O Devuyst, P Froguel, T Esko, BA Oostra, CM van Duijn, V Emilsson, AK Dieffenbach, H Brenner, I Borecki, CS Fox, M Rheinberger, ST Turner, S Kloiber, PS Wild, J Ding, Y Liu, SLR Kardia, AB Zonderman, MK Evans, MC Cornelis, A Adeyemo, CN Rotimi, D Cusi, E Salvi, PB Munroe, P Gasparini, M Ciullo, R Sorice, C Sala, D Toniolo, AW Dreisbach, DI Chasman, C Gieger, C Meisinger, M Waldenberger, HE Wichmann, T Illig, W Koenig, I Rudan, I Kolcic, M Boban, T Zemunik, W Ma¨rz, H Kramer, PP Pramstaller, EP Bottinger, O Gottesman, BW Penninx, H Snieder, JH Smit, AF Wright, H Campbell, JF Wilson, SH Wild, W Lieb, GJ Navis, BM Buckley, I Ford, JW Jukema, A Hofman, OH Franco, M Adam, M Imboden, N Probst-Hensch, B Paulweber, L Kedenko, F Kronenberg, S Coassin, M Haun, HK Kroemer, K Endlich, M Nauck, R Rettig, R Biffar, S Stracke, U Vo¨lker, H Wallaschofski, H Vo¨lzke, KL Keene, MM Sale, B Ponte, D Ackermann, M Pruijm, GB Ehret, A To¨njes, P Kovacs, JC Denny, RJ Carroll, C Hayward, O Polasek, V Vitart, PM Ridker, J Viikari, M Ka¨ho¨nen, O Raitakari, T Lehtima¨ki. Author contributions Study design: C Helmer, B Stengel, J Chalmers, M Woodward, P Hamet, G Eiriksdottir, LJ Launer, TB Harris, V Gudnason, JR O’Connell, A Ko¨ttgen, E Boerwinkle, WHL Kao, Study design: C Helmer, B Stengel, J Chalmers, M Woodward, P Hamet, G Eiriksdottir LJ Launer, TB Harris, V Gudnason, JR O’Connell, A Ko¨ttgen, E Boerwinkle, WHL Kao P Mitchell, I Guessous, JM Gaspoz, N Bouatia-Naji, P Froguel, A Metspalu, T Esko, BA Oostra CM van Duijn V Emilsson H Brenner I Borecki CS Fox Q Yang y g , g , J , , , LJ Launer, TB Harris, V Gudnason, JR O’Connell, A Ko¨ttgen, E Boerwinkle, WHL Ka P Mitchell, I Guessous, JM Gaspoz, N Bouatia-Naji, P Froguel, A Metspalu, T Esko, l k LJ Launer, TB Harris, V Gudnason, JR O Connell, A Kottgen, E Boerwinkle, WHL Ka P Mitchell, I Guessous, JM Gaspoz, N Bouatia-Naji, P Froguel, A Metspalu, T Esko, P Mitchell, I Guessous, JM Gaspoz, N Bouatia Naji, P Froguel, A Metspalu, T Esko BA Oostra, CM van Duijn, V Emilsson, H Brenner, I Borecki, CS Fox, Q Yang, BA Oostra, CM van Duijn, V Emilsson, H Brenner, I Borecki, CS Fox, Q Yang, BK Kra¨mer, PS Wild, BI Freedman, J Ding, Y Liu, AB Zonderman, MK Evans, A Adeyemo, CN Rotimi, D Cusi, P Gasparini, M Ciullo, D Toniolo, C Gieger, 21. Dupuis, J. et al. New genetic loci implicated in fasting glucose homeostasis and their impact on type 2 diabetes risk. Nat. Genet. 42, 105–116 (2010). C Meisinger, CA Bo¨ger, HE Wichmann, T Illig, I Rudan, W Ma¨rz, PP Pramstaller, 22. Morris, A. P. Transethnic meta-analysis of genomewide association studies. Genet. Epidemiol. 35, 809–822 (2011). EP Bottinger, BW Penninx, H Snieder, U Gyllensten, AF Wright, H Campbell, JF Wilson, SH Wild, GJ Navis, BM Buckley, I Ford, JW Jukema, B Paulweber, L Kedenko, p 23. Pers, T. H. et al. Biological interpretation of genome-wide association studies using predicted gene functions. Nat. Commun. 6, 5890 (2015). F Kronenberg, K Endlich, R Rettig, R Biffar, H Vo¨lzke, JK Fernandes, MM Sale, M Pruijm, GB Ehret, A To¨njes, M Stumvoll, JC Denny, RJ Carroll, N Hastie, O Polasek, PM Ridker, J Viikari, M Ka¨ho¨nen, O Raitakari, T Lehtima¨ki. 24. Trynka, G. et al. Chromatin marks identify critical cell types for ﬁne mapping complex trait variants. Nat. Genet. 45, 124–130 (2013). References k d 1. Eckardt, K. U. et al. Evolving importance of kidney disease: from subspecialty to global health burden. Lancet 382, 158–169 (2013). 2. El Nahas, M. The global challenge of chronic kidney disease. Kidney Int. 68, 2918–2929 (2005). 3. Baumeister, S. E. et al. Effect of chronic kidney disease and comorbid conditions on health care costs: A 10-year observational study in a general population. Am. J. Nephrol. 31, 222–229 (2010). 4. Fox, C. S. et al. Associations of kidney disease measures with mortality and end-stage renal disease in individuals with and without diabetes: a meta- analysis. Lancet 380, 1662–1673 (2012). y 5. Mahmoodi, B. K. et al. Associations of kidney disease measures with mortality and end-stage renal disease in individuals with and without hypertension: a meta-analysis. Lancet 380, 1649–1661 (2012). 6. Himmelfarb, J. & Tuttle, K. R. New therapies for diabetic kidney disease. N. Engl. J. Med. 369, 2549–2550 (2013). 7. Pattaro, C. et al. Genome-wide association and functional follow-up reveals new loci for kidney function. PLoS Genet. 8, e1002584 (2012). 7. Pattaro, C. et al. Genome-wide association and functional follow-up reveals new loci for kidney function. PLoS Genet. 8, e1002584 (2012). 8. Kottgen, A. et al. New loci associated with kidney function and chronic kidney disease. Nat. Genet. 42, 376–384 (2010). 8. Kottgen, A. et al. New loci associated with kidney function and chronic kidney disease. Nat. Genet. 42, 376–384 (2010). 9. Kottgen, A. et al. Multiple loci associated with indices of renal function and chronic kidney disease. Nat. Genet. 41, 712–717 (2009). 9. Kottgen, A. et al. Multiple loci associated with indices of renal function and chronic kidney disease. Nat. Genet. 41, 712–717 (2009). 10. Chambers, J. C. et al. Genetic loci inﬂuencing kidney function and chronic kidney disease. Nat. Genet. 42, 373–375 (2010). 9 Author contributions Ackermann, M Pruijm, GB Ehret, A To¨njes, P Kovacs, JC Denny, RJ 31. Stevens, L. A. et al. Estimating GFR using serum cystatin C alone and in combination with serum creatinine: a pooled analysis of 3,418 individuals with CKD. Am. J. Kidney Dis. 51, 395–406 (2008). y 32. Fuchsberger, C., Taliun, D., Pramstaller, P. P. & Pattaro, C. CKDGen consortium. GWAtoolbox: an R package for fast quality control and handling of GWAS meta-analysis data. Bioinformatics 28, 444–445 (2012). Author contributions 20. Smith, N. L. et al. Association of genome-wide variation with the risk of incident heart failure in adults of European and African ancestry: a prospective meta-analysis from the cohorts for heart and aging research in genomic epidemiology (CHARGE) consortium. Circ. Cardiovasc. Genet. 3, 256–266 (2010). Study design: C Helmer, B Stengel, J Chalmers, M Woodward, P Hamet, G Eiriksdottir, LJ Launer, TB Harris, V Gudnason, JR O’Connell, A Ko¨ttgen, E Boerwinkle, WHL Kao, P Mitchell, I Guessous, JM Gaspoz, N Bouatia-Naji, P Froguel, A Metspalu, T Esko, BA Oostra, CM van Duijn, V Emilsson, H Brenner, I Borecki, CS Fox, Q Yang, BK Kra¨mer, PS Wild, BI Freedman, J Ding, Y Liu, AB Zonderman, MK Evans, A Adeyemo, CN Rotimi, D Cusi, P Gasparini, M Ciullo, D Toniolo, C Gieger, C Meisinger, CA Bo¨ger, HE Wichmann, T Illig, I Rudan, W Ma¨rz, PP Pramstaller, EP Bottinger, BW Penninx, H Snieder, U Gyllensten, AF Wright, H Campbell, JF Wilson, SH Wild, GJ Navis, BM Buckley, I Ford, JW Jukema, B Paulweber, L Kedenko, F Kronenberg, K Endlich, R Rettig, R Biffar, H Vo¨lzke, JK Fernandes, MM Sale, M Pruijm, GB Ehret, A To¨njes, M Stumvoll, JC Denny, RJ Carroll, N Hastie, O Polasek, PM Ridker, J Viikari, M Ka¨ho¨nen, O Raitakari, T Lehtima¨ki. Interpretation of results: C Helmer, JC Lambert, M Metzger, B Stengel, V Chouraki, J Tremblay, J Chalmers, M Woodward, P Hamet, AV Smith, A Parsa, JR O’Connell, A Tin, A Ko¨ttgen, M Li, WHL Kao, Y Li, EG Holliday, J Attia, I Guessous, CA Peralta, AC Morrison, JF Felix, C Pattaro, G Li, IH de Boer, O Devuyst, H Lin, A Isaacs, V Emilsson, AD Johnson, CS Fox, M Olden, Q Yang, EJ Atkinson, M de Andrade, ST Turner, T Zeller, J Ding, Y Liu, M Nalls, A Adeyemo, D Shriner, D Cusi, E Salvi, V Mijatovic, D Ruggiero, R Sorice, AW Dreisbach, AY Chu, DI Chasman, CA Bo¨ger, IM Heid, M Gorski, B Tayo, C Fuchsberger, H Snieder, IM Nolte, W Igl, K Susztak, NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications Author contributions Study management: C Helmer, M Metzger, J Tremblay, J Chalmers, M Woodward, P Hamet, G Eiriksdottir, TB Harris, T Aspelund, V Gudnason, A Parsa, AR Shuldiner, BD Mitchell, E Boerwinkle, J Coresh, WHL Kao, R Schmidt, L Ferrucci, E Rochtchina, JJ Wang, J Attia, P Mitchell, I Guessous, JM Gaspoz, M Bochud, DS Siscovick, O Devuyst, P Froguel, T Esko, BA Oostra, CM van Duijn, V Emilsson, AK Dieffenbach, H Brenner, I Borecki, CS Fox, M Rheinberger, ST Turner, S Kloiber, PS Wild, J Ding, Y Liu, SLR Kardia, AB Zonderman, MK Evans, MC Cornelis, A Adeyemo, CN Rotimi, D Cusi, E Salvi, PB Munroe, P Gasparini, M Ciullo, R Sorice, C Sala, D Toniolo, AW Dreisbach, DI Chasman, C Gieger, C Meisinger, M Waldenberger, HE Wichmann, T Illig, W Koenig, I Rudan, I Kolcic, M Boban, T Zemunik, W Ma¨rz, H Kramer, PP Pramstaller, EP Bottinger, O Gottesman, BW Penninx, H Snieder, JH Smit, AF Wright, H Campbell, JF Wilson, SH Wild, W Lieb, GJ Navis, BM Buckley, I Ford, JW Jukema, A Hofman, OH Franco, M Adam, M Imboden, N Probst-Hensch, B Paulweber, L Kedenko, F Kronenberg, S Coassin, M Haun, HK Kroemer, K Endlich, M Nauck, R Rettig, R Biffar, S Stracke, U Vo¨lker, H Wallaschofski, H Vo¨lzke, KL Keene, MM Sale, B Ponte, D Ackermann, M Pruijm, GB Ehret, A To¨njes, P Kovacs, JC Denny, RJ Carroll, C Hayward, O Polasek, V Vitart, PM Ridker, J Viikari, M Ka¨ho¨nen, O Raitakari, T Lehtima¨ki. 25. Raychaudhuri, S. et al. Identifying relationships among genomic disease regions: predicting genes at pathogenic SNP associations and rare deletions. PLoS Genet. 5, e1000534 (2009). 26. Panagiotou, O. A., Willer, C. J., Hirschhorn, J. N. & Ioannidis, J. P. The power of meta-analysis in genome-wide association studies. Annu. Rev. Genomics Hum. Genet. 14, 441–465 (2013). 27. Barker, D. J. The fetal and infant origins of adult disease. BMJ 301, 1111 (1990). 28. Boubred, F. et al. Developmental origins of chronic renal disease: an integrative hypothesis. Int. J. Nephrol. 2013, 346067 (2013). yp p 29. Fox, C. S. et al. Predictors of new-onset kidney disease in a community-based population. JAMA 291, 844–850 (2004). population. JAMA 291, 844–850 (2004). 30. Coresh, J. et al. Calibration and random variation of the serum creatinine assay as critical elements of using equations to estimate glomerular ﬁltration rate. Am. J. Kidney Dis. 39, 920–929 (2002). NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 Genotyping: JC Lambert, J Tremblay, P Hamet, E Boerwinkle, WHL Kao, H Schmidt, GW Montgomery, L Ferrucci, M Bochud, BA Oostra, CM van Duijn, K Butterbach, I Borecki, M de Andrade, T Zeller, Y Liu, RJ Scott, SLR Kardia, M Nalls, FB Hu, GC Curhan, A Adeyemo, D Shriner, D Cusi, N Soranzo, P d’Adamo, D Ruggiero, M Ciullo, R Sorice, DI Chasman, H Grallert, T Zemunik, ME Kleber, EP Bottinger, O Gottesman, RJF Loos, AF Wright, JF Wilson, A Franke, D Ellinghaus, JW Jukema, S Trompet, AG Uitterlinden, F Rivadeneira, F Kronenberg, S Coassin, M Haun, F Ernst, G Homuth, HK Kroemer, M Nauck, U Vo¨lker, H Wallaschofski, MM Sale, GB Ehret, A To¨njes, M Stumvoll, P Kovacs, CM Shaffer, JC Denny, PM Ridker, T Lehtima¨ki. N Verweij, S Trompet, A Dehghan, B Kollerits, F Kronenberg, A Teumer, J Divers, KL Keene, MM Sale, WM Chen, GB Ehret, I Prokopenko, R Ma¨gi, JC Denny, RJ Carroll. notyping: JC Lambert, J Tremblay, P Hamet, E Boerwinkle, WHL Kao Design, performance and interpretation of zebraﬁsh experiments: M Garnaas, W Goessling. E Boerwinkle, J Coresh, M Li, WHL Kao, Y Li, H Schmidt, M Cavalieri, R Schmidt, JB Whitﬁeld, EG Holliday, JJ Wang, J Attia, P Mitchell, I Guessous, JM Gaspoz, M Bochud, CA Peralta, AC Morrison, JF Felix, C Pattaro, DS Siscovick, IH de Boer, M Bochud, CA Peralta, AC Morrison, JF Felix, C Pattaro, DS Siscovick, IH de Boer, M Rao, R Katz, O Devuyst, TH Pers, A Isaacs, H Brenner, M Garnaas, W Goessling, M Rao, R Katz, O Devuyst, TH Pers, A Isaacs, H Brenner, M Garnaas, W Goessling, BK Kra¨mer, M Rheinberger, ST Turner, D Czamara, S Kloiber, T Zeller, BI Freedman JM Stafford, J Ding, Y Liu, MA McEvoy, RJ Scott, SJ Hancock, JA Smith, JD Faul, SLR Kardia, AB Zonderman, M Nalls, MK Evans, FB Hu, GC Curhan, MC Cornelis, A Lupo, G Gambaro, G Malerba, M Ciullo, R Sorice, AW Dreisbach, AY Chu, DI Chasman, C Gieger, H Grallert, C Meisinger, M Waldenberger, CA Bo¨ger, HE Wichmann, IM Heid, M Gorski, T Illig, W Koenig, I Kolcic, M Boban, T Zemunik W Ma¨rz, B Tayo, H Kramer, SE Rosas, C Fuchsberger, D Ruderfer, EP Bottinger, O Gottesman, RJF Loos, Y Lu, H Snieder, H Campbell, A Franke, W Lieb, IM Leach, BM Buckley, I Ford, JW Jukema, S Trompet, A Dehghan, S Sedaghat, GA Thun, M Adam, M Imboden, N Probst-Hensch, B Kollerits, B Paulweber, L Kedenko, b A T dl h l k S l Ch O Gottesman, RJF Loos, Y Lu, H Snieder, H Campbell, A Franke, W Lieb, IM Leach, BM Buckley, I Ford, JW Jukema, S Trompet, A Dehghan, S Sedaghat, GA Thun, Supplementary Information accompanies this paper at http://www.nature.com/ naturecommunications Supplementary Information accompanies this paper at http://www.nature.com/ naturecommunications M Adam, M Imboden, N Probst-Hensch, B Kollerits, B Paulweber, L Kedenko, F Kronenberg, A Teumer, K Endlich, H Vo¨lzke, KL Keene, MM Sale, WM Chen, F Kronenberg, A Teumer, K Endlich, H Vo¨lzke, KL Keene, MM Sale, WM Chen, B Ponte, D Ackermann, M Pruijm, GB Ehret, A To¨njes, I Prokopenko, M Stumvoll, Competing ﬁnancial interests: J.T. and P.H. are consultants for Servier. J.C. received research grants and honoraria from Servier. K.S. obtained research support from Boehringer Ingelheim. The remaining authors declared no competing ﬁnancial interests. P Kovacs, R Ma¨gi, JC Denny, O Polasek, J Viikari, LP Lyytika¨inen, M Ka¨ho¨nen, O Raitakari, T Lehtima¨ki. Statistical methods and analysis: C Helmer, JC Lambert, M Metzger, V Chouraki, J Tremblay, P Hamet, AV Smith, T Aspelund, A Parsa, JR O’Connell, A Tin, A Ko¨ttgen, M Li, M Foster, WHL Kao, Y Li, H Schmidt, M Struchalin, NG Martin, RPS Middelberg, T Tanaka, E Rochtchina, EG Holliday, I Guessous, M Bochud, JF Felix, C Pattaro, G Li, R Katz, JN Hirschhorn, J Karjalainen, L Franke, TH Pers, L Yengo, N Bouatia-Naji, H Lin, T Nikopensius, T Esko, A Isaacs, A Demirkan, MF Feitosa, M Olden, MH Chen, Q Yang, SJ Hwang, M Garnaas, W Goessling, EJ Atkinson, M de Andrade, D Czamara, S Kloiber, C Mu¨ller, JM Stafford, J Ding, K Lohman, Y Liu, JA Smith, JD Faul, M Nalls, MC Cornelis, A Adeyemo, D Shriner, E Salvi, V Mijatovic, A Robino, S Ulivi, R Sorice, G Pistis, M Cocca, AY Chu, DI Chasman, LM Rose, CA Bo¨ger, IM Heid, M Gorski, ME Kleber, B Tayo, C Fuchsberger, A Saint-Pierre, D Taliun, D Ruderfer, Y Lu, IM Nolte, PJ van der Most, S Enroth, W Igl, F Murgia, L Portas, K Susztak, YA Ko, N Verweij, S Trompet, A Dehghan, S Sedaghat, GA Thun, M Adam, M Imboden, N Probst-Hensch, B Kollerits, A Teumer, J Divers, WM Chen, GB Ehret, I Prokopenko, R Ma¨gi, CM Shaffer, RJ Carroll, C Hayward, V Vitart, LP Lyytika¨inen, V Aalto. Author contributions Sample size requirements for matched case-control studies of gene-environment interaction. Stat. Med. 21, 35–50 (2002). 39. Rosner, B. in Fundamentals of Biostatistics (Thomson-Brooks/Cole, Duxbury, 2006). 40. Welter, D. et al. The NHGRI GWAS Catalog, a curated resource of SNP-trait associations. Nucleic Acids Res. 42, D1001–D1006 (2014). 10 NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunicat Design, performance and interpretation of zebraﬁsh experiments: M Garnaas, W Goessling. Drafting manuscript: J Chalmers, A Tin, A Ko¨ttgen, WHL Kao, M Bochud, CA Peralta, Drafting manuscript: J Chalmers, A Tin, A Ko¨ttgen, WHL Kao, M Bochud, CA Peralta, C Pattaro, IH de Boer, CS Fox, M Garnaas, W Goessling, N Soranzo, CA Bo¨ger, IM Heid, g p g C Pattaro, IH de Boer, CS Fox, M Garnaas, W Goessling, N Soranzo, CA Bo¨ger, IM Heid M Gorski, S Trompet, A Dehghan, A Teumer, KL Keene, MM Sale. M Gorski, S Trompet, A Dehghan, A Teumer, KL Keene, MM Sale. M Gorski, S Trompet, A Dehghan, A Teumer, KL Keene, MM Sale. Critical review of the manuscript: J Tremblay, J Chalmers, M Woodward, P Hamet, TB Harris, V Gudnason, A Parsa, AR Shuldiner, BD Mitchell, A Tin, A Ko¨ttgen, Bioinformatics: JC Lambert, V Chouraki, J Tremblay, P Hamet, AV Smith, T Aspelund, JR O’Connell, E Boerwinkle, Y Li, M Struchalin, GW Montgomery, RPS Middelberg, T Tanaka, C Pattaro, G Li, JN Hirschhorn, J Karjalainen, L Franke, TH Pers, L Yengo, T Esko, AD Johnson, M Olden, M Garnaas, W Goessling, D Czamara, C Mu¨ller, JA Smith, SLR Kardia, M Nalls, E Salvi, G Malerba, V Mijatovic, P d’Adamo, S Ulivi, R Sorice, C Sala, G Pistis, M Cocca, DI Chasman, H Grallert, M Waldenberger, CA Bo¨ger, IM Heid, M Gorski, ME Kleber, D Taliun, O Gottesman, S Enroth, K Susztak, YA Ko, D Ellinghaus, N Verweij, I Ford, S Trompet, F Rivadeneira, WM Chen, GB Ehret, R Ma¨gi, CM Shaffer, JC Denny, RJ Carroll, C Hayward, LP Lyytika¨inen, V Aalto. Additional information Supplementary Information accompanies this paper at http://www.nature.com/ naturecommunications Genotyping: JC Lambert, J Tremblay, P Hamet, E Boerwinkle, WHL Kao, H Schmidt, GW Montgomery, L Ferrucci, M Bochud, BA Oostra, CM van Duijn, K Butterbach, I Borecki, M de Andrade, T Zeller, Y Liu, RJ Scott, SLR Kardia, M Nalls, FB Hu, GC Curhan, A Adeyemo, D Shriner, D Cusi, N Soranzo, P d’Adamo, D Ruggiero, M Ciullo, R Sorice, DI Chasman, H Grallert, T Zemunik, ME Kleber, EP Bottinger, O Gottesman, RJF Loos, AF Wright, JF Wilson, A Franke, D Ellinghaus, JW Jukema, S Trompet, AG Uitterlinden, F Rivadeneira, F Kronenberg, S Coassin, M Haun, F Ernst, G Homuth, HK Kroemer, M Nauck, U Vo¨lker, H Wallaschofski, MM Sale, GB Ehret, A To¨njes, M Stumvoll, P Kovacs, CM Shaffer, JC Denny, PM Ridker, T Lehtima¨ki. 1 Center for Biomedicine, European Academy of Bozen/Bolzano (EURAC), afﬁliated to the University of Lu¨beck, Via Galvani 31, Bolzano 39100, Italy. 2 Interfaculty Institute for Genetics and Functional Genomics, Ernst-Moritz-Arndt-University Greifswald, Friedrich-Loefﬂer-Strae 15a, Greifswald 17487, Germany. 3 Institute for Community Medicine, University of Greifswald, Walther-Rathenau-Strasse 48, Greifswald 17487, Germany. 4 Department of Genetic Epidemiology, Institute of Epidemiology and Preventive Medicine, University of Regensburg, Franz-Josef-Strau-Allee 11, Regensburg 93053, Germany. 5 Department of Nephrology, University Hospital Regensburg, Franz-Josef-Strau-Allee 11, Regensburg 93053, Germany. 6 Preventive Medicine, Brigham and Women’s Hospital, 900 Commonwealth Avenue East, Boston, Massachusetts 02215, USA. 7 Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, 615 North Wolfe Street, Baltimore, Maryland 21205, USA. 8 Department of Life and Reproduction Sciences, University of Verona, Strada Le Grazie 8, Verona 37134, Italy. 9 Division of Genetics, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, New Research Building77 Avenue Louis Pasteur, Room 458, Boston, Massachusetts 02115, USA. 10 Institute of Genetics and Biophysics "Adriano Buzzati-Traverso"—CNR, Via P. Castellino 111, Napoli 80131, Italy. 11 Department of Internal Medicine IV, University Hospital Freiburg, Berliner Allee 29, Freiburg 79110, Germany. 12 Division of Nephrology/Tufts Evidence Practice Center, Tufts University School of Medicine, Tufts Medical Center, Boston, Massachusetts 02111, USA. 13 Department of Biostatistics, Boston University School of Public Health, 715 Albany Street, Boston, Massachusetts 02118, USA. 14 Department of Neurology, Boston University School of Medicine, 72 East Concord ST B603, Boston, Massachusetts 02118, USA. 15 Division of Endocrinology and Center for Basic and Translational Obesity Research, Boston Children’s Hospital, 300 Longwood Avenue, Boston, Massachusetts 02115, USA. 16 Medical and Population Genetics Program, Broad Institute of MIT and Harvard, Cambridge, Massachusetts 2142, USA. 17 NHLBI’s Framingham Heart Study and the Center for Population Reprints and permission information is available online at http://npg.nature.com/ reprintsandpermissions/ How to cite this article: Pattaro, C. et al. Genetic associations at 53 loci highlight cell types and biological pathways relevant for kidney function. Nat. Commun. 7:10023 doi: 10.1038/ncomms10023 (2016). This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ Cristian Pattaro1,, Alexander Teumer2,3,, Mathias Gorski4,5,, Audrey Y. Chu6,, Man Li7,, Vladan Mijatovic8,, Maija Garnaas9, Adrienne Tin7, Rossella Sorice10, Yong Li11, Daniel Taliun1, Matthias Olden4,5, Meredith Foster12, Qiong Yang13, Ming-Huei Chen13,14, Tune H. Pers15,16, Andrew D. Johnson17, Yi-An Ko18, Christian Fuchsberger1, Bamidele Tayo19, Michael Nalls20, Mary F. Feitosa21, Aaron Isaacs22,23, Abbas Dehghan24, Pio d’Adamo25, Adebowale Adeyemo26, Aida Karina Dieffenbach27,28, Alan B. Zonderman29, Ilja M. Nolte30, Peter J. van der Most30, Alan F. Wright31, Alan R. Shuldiner32,33, Alanna C. Morrison34, Albert Hofman24, Albert V. Smith35,36, Albert W. Dreisbach37, Andre Franke38, Andre G. Uitterlinden39, Andres Metspalu40,41, Anke Tonjes42, Antonio Lupo43, Antonietta Robino25, Åsa Johansson44, Ayse Demirkan22, Barbara Kollerits45, Barry I. Freedman46, Belen Ponte47, Ben A. Oostra48, Bernhard Paulweber49, Bernhard K. Kra¨mer50, Braxton D. Mitchell32,33, Brendan M. Buckley51, Carmen A. Peralta52, Caroline Hayward31, Catherine Helmer53,54, Charles N. Rotimi26, Christian M. Shaffer55, Christian Mu¨ller56,57, Cinzia Sala58, Cornelia M. van Duijn22, Aude Saint-Pierre1,59, Daniel Ackermann47, Daniel Shriner26, Daniela Ruggiero10, Daniela Toniolo58,60, Yingchang Lu61, Daniele Cusi62, Darina Czamara63, David Ellinghaus38, David S. Siscovick64, Douglas Ruderfer65, Christian Gieger66, Harald Grallert67,68,69, Elena Rochtchina70, Elizabeth J. Atkinson71, Elizabeth G. Holliday72,73, Eric Boerwinkle34, Erika Salvi62, Erwin P. Bottinger61, Federico Murgia74, Fernando Rivadeneira39, Florian Ernst2, Florian Kronenberg45, Frank B. Hu75, Gerjan J. Navis76, Gary C. Curhan77, George B. Ehret78, Georg Homuth2, Stefan Coassin45, Gian-Andri Thun79,80, Giorgio Pistis58, NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications 11 NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 ARTICLE Giovanni Gambaro81, Giovanni Malerba8, Grant W. Montgomery82, Gudny Eiriksdottir35, Gunnar Jacobs83, Guo Li64, H-Erich Wichmann84,85,86, Harry Campbell87, Helena Schmidt88, Henri Wallaschofski89,90, Henry Vo¨lzke3,90, Hermann Brenner27,28, Heyo K. Kroemer91, Holly Kramer19, Honghuang Lin92, I. Mateo Leach93, Ian Ford94, Idris Guessous95,96,97, Igor Rudan87, Inga Prokopenko98, Ingrid Borecki21, Iris M. Heid4,66, Ivana Kolcic99, Ivana Persico74, J. Wouter Jukema100,101,102,103, James F. Wilson87, Janine F. Felix24, Jasmin Divers104, Jean-Charles Lambert105, Jeanette M. Stafford104, Jean-Michel Gaspoz95, Jennifer A. Smith106, Jessica D. Faul107, Jie Jin Wang108, Jingzhong Ding109, Joel N. Hirschhorn15,16,110, John Attia71,72, John B. Whitﬁeld82, John Chalmers111, Jorma Viikari112, Josef Coresh7,113, Joshua C. Denny114, Juha Karjalainen115, Jyotika K. Fernandes116, Karlhans Endlich117, Katja Butterbach27, Keith L. Keene118, Kurt Lohman46, Laura Portas74, Lenore J. Launer119, Leo-Pekka Lyytika¨inen120, Loic Yengo121,122,123, Lude Franke115, Luigi Ferrucci124, Lynda M. Rose6, Lyudmyla Kedenko49, Madhumathi Rao12, Maksim Struchalin125,126, Marcus E. Kleber127, Margherita Cavalieri128, Margot Haun45, Marilyn C. Cornelis75, Marina Ciullo10, Mario Pirastu74, Mariza de Andrade71, Mark A. McEvoy129, Mark Woodward7,111,112,130, Martin Adam79,80, Massimiliano Cocca58, Matthias Nauck89,90, Medea Imboden79,80, Melanie Waldenberger67, Menno Pruijm131, Marie Metzger132, Michael Stumvoll42, Michele K. Evans133, Michele M. Sale134, Mika Ka¨ho¨nen135, Mladen Boban99, Murielle Bochud136, Myriam Rheinberger5, Niek Verweij93, Nabila Bouatia-Naji137,138, Nicholas G. Martin82,139, Nick Hastie31, Nicole Probst-Hensch79,80, Nicole Soranzo140, Olivier Devuyst141, Olli Raitakari142, Omri Gottesman61, Oscar H. Franco24, Ozren Polasek99, Paolo Gasparini25, Patricia B. Munroe143,144, Paul M. Ridker145, Paul Mitchell108, Paul Muntner146,147, Christa Meisinger68, Johannes H. Smit148, ICBP Consortiumw, AGEN Consortiumw, CARDIOGRAMw, CHARGe-Heart Failure Groupw, ECHOGen Consortiumw, Peter Kovacs149, Philipp S. Wild150, Philippe Froguel121,122,123, Rainer Rettig151, Reedik Ma¨gi40, Reiner Biffar152, Reinhold Schmidt128, Rita P.S. Middelberg82, Robert J. Carroll114, Brenda W. Penninx148, Rodney J. Scott153, Ronit Katz154, Sanaz Sedaghat24, Sarah H. Wild87, Sharon L.R. Kardia106, Sheila Ulivi155, Shih-Jen Hwang17, Stefan Enroth44, Stefan Kloiber63, Stella Trompet100, Benedicte Stengel132, Stephen J. Hancock72,73, Stephen T. Turner156, Sylvia E. Rosas18, Sylvia Stracke105,157, Tamara B. Harris119, Tanja Zeller56,57, Tatijana Zemunik99, Terho Lehtima¨ki120, Thomas Illig68, Thor Aspelund35,36, Tiit Nikopensius40,41, Tonu Esko15,40,41, Toshiko Tanaka124, Ulf Gyllensten44, Uwe Vo¨lker2,90, Valur Emilsson35,158, Veronique Vitart31, Ville Aalto159, Vilmundur Gudnason35,36, Vincent Chouraki105, Wei-Min Chen134, Wilmar Igl44, Winfried Ma¨rz160, Wolfgang Koenig161, Wolfgang Lieb83, Ruth J.F. Loos61,162, Yongmei Liu46, Harold Snieder30, Peter P. Pramstaller1,163,164, Afshin Parsa165, Jeffrey R. O’Connell32, Katalin Susztak18, Pavel Hamet166, Johanne Tremblay166, Ian H. de Boer154, Carsten A. Bo¨ger5,, Wolfram Goessling9,, Daniel I. Chasman6,145,, Anna Ko¨ttgen7,11,, W.H. Linda Kao7,113,,z & Caroline S. Fox17,167, Giovanni Gambaro81, Giovanni Malerba8, Grant W. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 Montgomery82, Gudny Eiriksdottir35, Gunnar Jacobs83, Guo Li64, H-Erich Wichmann84,85,86, Harry Campbell87, Helena Schmidt88, Henri Wallaschofski89,90, Henry Vo¨lzke3,90, Hermann Brenner27,28, Heyo K. Kroemer91, Holly Kramer19, Honghuang Lin92, I. Mateo Leach93, Ian Ford94, Idris Guessous95,96,97, Igor Rudan87, Inga Prokopenko98, Ingrid Borecki21, Iris M. Heid4,66, Ivana Kolcic99, Ivana Persico74, J. Wouter Jukema100,101,102,103, James F. Wilson87, Janine F. Felix24, Jasmin Divers104, Jean-Charles Lambert105, Jeanette M. Stafford104, Jean-Michel Gaspoz95, Jennifer A. Smith106, Jessica D. Faul107, Jie Jin Wang108, Jingzhong Ding109, Joel N. Hirschhorn15,16,110, John Attia71,72, John B. Whitﬁeld82, John Chalmers111, Jorma Viikari112, Josef Coresh7,113, Joshua C. Denny114, Juha Karjalainen115, Jyotika K. Fernandes116, Karlhans Endlich117, Katja Butterbach27, Keith L. Keene118, Kurt Lohman46, Laura Portas74, Lenore J. Launer119, Leo-Pekka Lyytika¨inen120, Loic Yengo121,122,123, Lude Franke115, Luigi Ferrucci124, Lynda M. Rose6, Lyudmyla Kedenko49, Madhumathi Rao12, Maksim Struchalin125,126, Marcus E. Kleber127, Margherita Cavalieri128, Margot Haun45, Marilyn C. Cornelis75, Marina Ciullo10, Mario Pirastu74, Mariza de Andrade71, Mark A. McEvoy129, Mark Woodward7,111,112,130, Martin Adam79,80, Massimiliano Cocca58, Matthias Nauck89,90, Medea Imboden79,80, Melanie Waldenberger67 Menno Pruijm131, Marie Metzger132, Michael Stumvoll42, Michele K. Evans133, Michele M. Sale134, Mika Ka¨ho¨nen135, Mladen Boban99, Murielle Bochud136, Myriam Rheinberger5, Niek Verweij93, Nabila Bouatia-Naji137,138, Nicholas G. Martin82,139, Nick Hastie31, Nicole Probst-Hensch79,80, Nicole Soranzo140 Olivier Devuyst141, Olli Raitakari142, Omri Gottesman61, Oscar H. Franco24, Ozren Polasek99, Paolo Gasparini25 Patricia B. Munroe143,144, Paul M. Ridker145, Paul Mitchell108, Paul Muntner146,147, Christa Meisinger68, Johannes H. Smit148, ICBP Consortiumw, AGEN Consortiumw, CARDIOGRAMw, CHARGe-Heart Failure Groupw ECHOGen Consortiumw, Peter Kovacs149, Philipp S. Wild150, Philippe Froguel121,122,123, Rainer Rettig151, Reedik Ma¨gi40, Reiner Biffar152, Reinhold Schmidt128, Rita P.S. Middelberg82, Robert J. Carroll114, Brenda W. Penninx148, Rodney J. Scott153, Ronit Katz154, Sanaz Sedaghat24, Sarah H. Wild87, Sharon L.R. Kardia106, Sheila Ulivi155, Shih-Jen Hwang17, Stefan Enroth44, Stefan Kloiber63, Stella Trompet100, Benedicte Stengel132, Stephen J. Hancock72,73, Stephen T. Turner156, Sylvia E. Rosas18, Sylvia Stracke105,157, Tamara B. Harris119, Tanja Zeller56,57, Tatijana Zemunik99, Terho Lehtima¨ki120, Thomas Illig68, Thor Aspelund35,36, Tiit Nikopensius40,41, Tonu Esko15,40,41, Toshiko Tanaka124, Ulf Gyllensten44, Uwe Vo¨lker2,90 Valur Emilsson35,158, Veronique Vitart31, Ville Aalto159, Vilmundur Gudnason35,36, Vincent Chouraki105, Wei-Min Chen134, Wilmar Igl44, Winfried Ma¨rz160, Wolfgang Koenig161, Wolfgang Lieb83, Ruth J.F. Loos61,162, Yongmei Liu46, Harold Snieder30, Peter P. Pramstaller1,163,164, Afshin Parsa165, Jeffrey R. O’Connell32, Katalin Susztak18, Pavel Hamet166, Johanne Tremblay166, Ian H. de Boer154, Carsten A. Bo¨ger5,, Wolfram Goessling9,, Daniel I. Chasman6,145,, Anna Ko¨ttgen7,11,, W.H. Linda Kao7,113,,z & Caroline S Fox17,167, Giovanni Gambaro81, Giovanni Malerba8, Grant W. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 Montgomery82, Gudny Eiriksdottir35, Gunnar Jacobs83, Guo Li64, H-Erich Wichmann84,85,86, Harry Campbell87, Helena Schmidt88, Henri Wallaschofski89,90, Henry Olivier Devuyst141, Olli Raitakari142, Omri Gottesman61, Oscar H. Franco24, Ozren Polasek99, Paolo Gasparini25, Patricia B. Munroe143,144, Paul M. Ridker145, Paul Mitchell108, Paul Muntner146,147, Christa Meisinger68, NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications 12 NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 Studies, 73 Mt Wayte Avenue, Suite 2, Framingham, Massachusetts 01702, USA. 18 Renal Electrolyte and Hypertension Division, Perelman School of Medicine, University of Pennsylvania, 415 Curie Boulevard, 405B Clinical Research Building, Philadelphia, Pennsylvania 19104-4539, USA. 19 Department of Public Health Sciences, Loyola Medical Center, 2160 S First Avenue, Maywood, Illinois 60153, USA. 20 Laboratory of Neurogenetics, Building 35—Porter Building, 1A1015, National Institute on Aging/NIH, Bethesda, Maryland 20892, USA. 21 Division of Statistical Genomics, Washington University School of Medicine, 4444 Forest Park Boulevard, Box 8506, St Louis, Missouri 63108, USA. 22 Genetic Epidemiology Unit, Department of Epidemiology, Erasmus University Medical Center, Dr Molewaterplein 50, PO Box 2040, Rotterdam 3000 CA, The Netherlands. 23 Centre for Medical Systems Biology Leiden, Dr Molewaterplein 50, PO Box 2040, Rotterdam 3000 CA, The Netherlands. 24 Department of Epidemiology, Erasmus University Medical Center, PO Box 2040, Rotterdam 3000 CA, The Netherlands. 25 Institute for Maternal and Child Health—IRCCS "Burlo Garofolo" and University of Trieste, via dell’Istria 65/1, Trieste 34137, Italy. 26 Center for Research on Genomics and Global Health, National Human Genome Research Institute, Building 12A, Room 4047, 12 South Dr, MSC 5635, Bethesda, Maryland 20892-5635, USA. 27 Division of Clinical Epidemiology and Aging Research, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 581, Heidelberg 69120, Germany. 28 German Cancer Consortium (DKTK), Im Neuenheimer Feld 581, Heidelberg 69120, Germany. 29 Laboratory of Personality and Cognition, National Institute on Aging, National Institutes of Health, NIH Biomedical Center, 251 Bayview Boulevard, Suite 100, Baltimore, Maryland 21224, USA. 30 Unit of Genetic Epidemiology and Bioinformatics, Department of Epidemiology, University Medical Center Groningen, PO Box 30001, Groningen 9700 RB, The Netherlands. 31 MRC Human Genetics Unit, Institute of Genetics and Molecular Medicine, University of Edinburgh, Crewe Road, Edinburgh EH4 2XU, UK. 32 Department of Medicine, University of Maryland School of Medicine, 685 West Baltimore Street, Baltimore, Maryland 21201, USA. 33 Geriatric Research and Education Clinical Center, Veterans Administration Medical Center, 10 North Greene Street, Baltimore, Maryland 21201, USA. 34 Human Genetics Center, University of Texas Health Science Center at Houston, 1200 Pressler St Suite 453E, Houston, Texas 77030, USA. 35 Icelandic Heart Association, Research Institute, Holtasmari 1, Kopavogur 201, Iceland. 36 University of Iceland, Sæmundargo¨tu 2, Reykjavik 101, Iceland. 37 Division of Nephrology, University of Mississippi, 2500 North State Street, Jackson, Mississippi 39216, USA. ARTICLE 38 Institute of Clinical Molecular Biology, Christian-Albrechts University of Kiel, Schittenhelmstrae 12, Kiel 24105, Germany. 39 Department of Internal Medicine, Erasmus University Medical Center, PO Box 1738, Rotterdam 3000 DR, The Netherlands. 40 Estonian Genome Center of University of Tartu (EGCUT), Riia 23B, Tartu 51010, Estonia. 41 Institute of Molecular and Cell Biology, University of Tartu and Estonian Biocenter, Riia 23, Tartu 51010, Estonia. 42 Department of Medicine, University of Leipzig, Liebigstrae 18, Leipzig 04103, Germany. 43 Division of Nephrology, Department of Medicine, University of Verona, Piazzale Aristide Stefani 1, Verona 37126, Italy. 44 Uppsala University, Department of Immunology, Genetics and Pathology, Biomedical Center, SciLifeLab, Uppsala University, Uppsala SE- 75108, Sweden. 45 Innsbruck Medical University, Division of Genetic Epidemiology, Schoepfstrae 41, Innsbruck 6020, Austria. 46 Internal Medicine Department, Wake Forest School of Medicine, Medical Center Boulevard, Winston-Salem, North Carolina 27157-1053, USA. 47 Nephrology Division, Department of Specialties of Internal Medicine, Geneva University Hospital, 4 rue Gabrielle-Perret-Gentil, Geneve 1211, Switzerland. 48 Department of Clinical Genetics, Erasmus University Medical Center, Dr Molewaterplein 50, PO Box 2040, Rotterdam 3000 CA, The Netherlands. 49 First Department of Internal Medicine, Paracelsus Medical University/Salzburger Landeskliniken, Mu¨llner Hauptstrae 48, Salzburg 5020, Austria. 50 University Medical Centre Mannheim, 5th Department of Medicine, University of Heidelberg, Theodor Kutzer Ufer 1-3, Mannheim 68167, Germany. 51 Department of Pharmacology and Therapeutics, University College Cork, Clinical Investigations Building, Western Rd, Cork, Ireland. 52 Division of Nephrology, University of California, San Francisco Medical School and San Francisco VA Medical Center, 4150 Clement Street, San Francisco, California 94121, USA. 53 INSERM, ISPED, Centre INSERM U897—Epidemiologie-Biostatistique, Bordeaux F-33000, France. 54 Universite´ Bordeaux, ISPED, Centre INSERM U897-Epidemiologie- Biostatistique, Bordeaux F-33000, France. 55 Vanderbilt University School of Medicine, 2215-B Garland Avenue 1224—MRB4 (Light Hall)Nashville, Tennessee 37232, USA. 56 University Heart Center Hamburg, Clinic for general and interventional cardiology, Martinistrae 52, Hamburg 20246, Germany. 57 German Center for Cardiovascular Research (DZHK), Partner Site Hamburg/Lu¨beck/Kiel, Martinistrae 52, Hamburg 20246, Germany. 58 Division of Genetics and Cell Biology, San Raffaele Scientiﬁc Institute, Via Olgettina 58, Milano 20132, Italy. 59 INSERM U1078, Etablissement Franc¸ais du Sang, 46 rue Fe´lix Le Dantec, CS 51819, Brest Cedex 2 29218, France. 60 Institute of Molecular Genetics-CNR, Via Abbiategrasso 207, Pavia 27100, Italy. 61 The Charles Bronfman Institute for Personalized Medicine, Icahn School of Medicine at Mount Sinai, New York, New York 10029, USA. ARTICLE 62 Department of Health Sciences, University of Milano, Via Antonio di Rudinı` 8, Milano 20142, Italy. 63 Max Planck Institute of Psychiatry, Kraepelinstrae 2–10, Munich 80804, Germany. 64 Cardiovascular Health Research Unit, Departments of Epidemiology and Medicine, University of Washington, 1730 Minor Ave, Suite 1360, Seattle, Washington 98101, USA. 65 Division of Psychiatric Genomics, Department of Psychiatry, Icahn School of Medicine at Mount Sinai, New York, New York 10029, USA. 66 Institute of Genetic Epidemiology, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, Ingolstaedter Landstrae 1, 85764 Neuherberg, Germany. 67 Research Unit of Molecular Epidemiology, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, Ingolstaedter Landstrae 1, 85764 Neuherberg, Germany. 68 Institute of Epidemiology II, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, 85764 Neuherberg, Germany. 69 German Center for Diabetes Research (DZD), Ingolstaedter Landstrae 1, Neuherberg 85764, Germany. 70 Westmead Millennium Institute, Centre for Vision Research, University of Sydney, C24 Westmead Hospital, New South Wales 2145, Australia. 71 Division of Biomedical Statistics and Informatics, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905, USA. 72 Centre for Clinical Epidemiology and Biostatistics, School of Medicine and Public Health, University of Newcastle, HMRI Building1, Kookaburra Circuit, New Lambton New South Wales 2305, Australia. 73 Clinical Research Design, Information Technology and Statistical Support, Hunter Medical Research Institute, Newcastle, 1 Kookaburra Circuit, New Lambton Heights, New South Wales 2305, Australia. 74 Institute of Population Genetics—CNR, Traversa La Crucca 3, 07040 Reg. Baldinca, Li Punti, Sassari, Italy. 75 Department of Nutrition, Harvard School of Public Health, 665 Huntington Avenue, Building 2, Boston, Massachusetts 02115, USA. 76 Department of Internal Medicine, University Medical Center Groningen, University of Groningen, Hanzeplein 1, Groningen 9713 GZ, The Netherlands. 77 Brigham and Women’s Hospital and Channing Laboratory, Harvard Medical School, 181 Longwood Avenue, Boston, Massachusetts 02115, USA. 78 Cardiology, Department of Specialties of Internal Medicine, Geneva University Hospitals, Rue Gabrielle-Perret-Gentil 4, Geneva 1205, Switzerland. 79 Swiss Tropical and Public Health Institute, PO Box 4002, Basel, Switzerland. 80 University of Basel, Petersplatz 1, Basel 4003, Switzerland. 81 Division of Nephrology, Department of Internal Medicine and Medical Specialties, Columbus-Gemelli University Hospital, Catholic University, Via Moscati 31, Rome 00168, Italy. 82 Genetic Epidemiology, Queensland Institute of Medical Research, QIMR, PO Royal Brisbane Hospital, Queensland 4029, Australia. 83 Institute of Epidemiology and Biobank popgen, Christian-Albrechts University, Niemannsweg 11, Kiel 24105, Germany. ARTICLE 84 Institute of Epidemiology I, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, Ingolsta¨dter Landstrae 1, Neuherberg 85764, Germany. 85 Institute of Medical Informatics, Biometry and Epidemiology, Ludwig-Maximilians-Universita¨t, Ingolsta¨dter Landstrasse 1, 85764 Neuherberg, Germany. 86 Klinikum Grosshadern, Ingolsta¨dter Landstrae 1, Neuherberg 85764, Germany. 87 Centre for Population Health Sciences, University of Edinburgh, Teviot Place, Edinburgh, EH8 9AG Scotland, UK. 88 Austrian Stroke Prevention Study, Institute of Molecular Biology and Biochemistry, Department of Neurology, Medical University Graz Harrachgasse 21 Graz 8010 Austria 89 Institute of Clinical Chemistry and Laboratory Medicine University Medicine Greifswald Ferdinand Studies, 73 Mt Wayte Avenue, Suite 2, Framingham, Massachusetts 01702, USA. 18 Renal Electrolyte and Hypertension Division, Perelman School of Medicine, University of Pennsylvania, 415 Curie Boulevard, 405B Clinical Research Building, Philadelphia, Pennsylvania 19104-4539, USA. 19 Department of Public Health Sciences, Loyola Medical Center, 2160 S First Avenue, Maywood, Illinois 60153, USA. 20 Laboratory of Neurogenetics, Building 35—Porter Building, 1A1015, National Institute on Aging/NIH, Bethesda, Maryland 20892, USA. 21 Division of Statistical Genomics, Washington University School of Medicine, 4444 Forest Park Boulevard, Box 8506, St Louis, Missouri 63108, USA. 22 Genetic Epidemiology Unit, Department of Epidemiology, Erasmus University Medical Center, Dr Molewaterplein 50, PO Box 2040, Rotterdam 3000 CA, The Netherlands. 23 Centre for Medical Systems Biology Leiden, Dr Molewaterplein 50, PO Box 2040, Rotterdam 3000 CA, The Netherlands. 24 Department of Epidemiology, Erasmus University Medical Center, PO Box 2040, Rotterdam 3000 CA, The Netherlands. 25 Institute for Maternal and Child Health—IRCCS "Burlo Garofolo" and University of Trieste, via dell’Istria 65/1, Trieste 34137, Italy. 26 Center for Research on Genomics and Global Health, National Human Genome Research Institute, Building 12A, Room 4047, 12 South Dr, MSC 5635, Bethesda, Maryland 20892-5635, USA. 27 Division of Clinical Epidemiology and Aging Research, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 581, Heidelberg 69120, Germany. 28 German Cancer Consortium (DKTK), Im Neuenheimer Feld 581, Heidelberg 69120, Germany. 29 Laboratory of Personality and Cognition, National Institute on Aging, National Institutes of Health, NIH Biomedical Center, 251 Bayview Boulevard, Suite 100, Baltimore, Maryland 21224, USA. 30 Unit of Genetic Epidemiology and Bioinformatics, Department of Epidemiology, University Medical Center Groningen, PO Box 30001, Groningen 9700 RB, The Netherlands. 31 MRC Human Genetics Unit, Institute of Genetics and Molecular Medicine, University of Edinburgh, Crewe Road, Edinburgh EH4 2XU, UK. ARTICLE 32 Department of Medicine, University of Maryland School of Medicine, 685 West Baltimore Street, Baltimore, Maryland 21201, USA. 33 Geriatric Research and Education Clinical Center, Veterans Administration Medical Center, 10 North Greene Street, Baltimore, Maryland 21201, USA. 34 Human Genetics Center, University of Texas Health Science Center at Houston, 1200 Pressler St Suite 453E, Houston, Texas 77030, USA. 35 Icelandic Heart Association, Research Institute, Holtasmari 1, Kopavogur 201, Iceland. 36 University of Iceland, Sæmundargo¨tu 2, Reykjavik 101, Iceland. 37 Division of Nephrology, University of Mississippi, 2500 North State Street, Jackson, Mississippi 39216, USA. 38 Institute of Clinical Molecular Biology, Christian-Albrechts University of Kiel, Schittenhelmstrae 12, Kiel 24105, Germany. 39 Department of Internal Medicine, Erasmus University Medical Center, PO Box 1738, Rotterdam 3000 DR, The Netherlands. 40 Estonian Genome Center of University of Tartu (EGCUT), Riia 23B, Tartu 51010, Estonia. 41 Institute of Molecular and Cell Biology, University of Tartu and Estonian Biocenter, Riia 23, Tartu 51010, Estonia. 42 Department of Medicine, University of Leipzig, Liebigstrae 18, Leipzig 04103, Germany. 43 Division of Nephrology, Department of Medicine, University of Verona, Piazzale Aristide Stefani 1, Verona 37126, Italy. 44 Uppsala University, Department of Immunology, Genetics and Pathology, Biomedical Center, SciLifeLab, Uppsala University, Uppsala SE- 75108, Sweden. 45 Innsbruck Medical University, Division of Genetic Epidemiology, Schoepfstrae 41, Innsbruck 6020, Austria. 46 Internal Medicine Department, Wake Forest School of Medicine, Medical Center Boulevard, Winston-Salem, North Carolina 27157-1053, USA. 47 Nephrology Division, Department of Specialties of Internal Medicine, Geneva University Hospital, 4 rue Gabrielle-Perret-Gentil, Geneve 1211, Switzerland. 48 Department of Clinical Genetics, Erasmus University Medical Center, Dr Molewaterplein 50, PO Box 2040, Rotterdam 3000 CA, The Netherlands. 49 First Department of Internal Medicine, Paracelsus Medical University/Salzburger Landeskliniken, Mu¨llner Hauptstrae 48, Salzburg 5020, Austria. 50 University Medical Centre Mannheim, 5th Department of Medicine, University of Heidelberg, Theodor Kutzer Ufer 1-3, Mannheim 68167, Germany. 51 Department of Pharmacology and Therapeutics, University College Cork, Clinical Investigations Building, Western Rd, Cork, Ireland. 52 Division of Nephrology, University of California, San Francisco Medical School and San Francisco VA Medical Center, 4150 Clement Street, San Francisco, California 94121, USA. 53 INSERM, ISPED, Centre INSERM U897—Epidemiologie-Biostatistique, Bordeaux F-33000, France. 54 Universite´ Bordeaux, ISPED, Centre INSERM U897-Epidemiologie- Biostatistique, Bordeaux F-33000, France. 55 Vanderbilt University School of Medicine, 2215-B Garland Avenue 1224—MRB4 (Light Hall)Nashville, Tennessee 37232, USA. 56 University Heart Center Hamburg, Clinic for general and interventional cardiology, Martinistrae 52, Hamburg 20246, Germany. ARTICLE 57 German Center for Cardiovascular Research (DZHK), Partner Site Hamburg/Lu¨beck/Kiel, Martinistrae 52, Hamburg 20246, Germany. 58 Division of Genetics and Cell Biology, San Raffaele Scientiﬁc Institute, Via Olgettina 58, Milano 20132, Italy. 59 INSERM U1078, Etablissement Franc¸ais du Sang, 46 rue Fe´lix Le Dantec, CS 51819, Brest Cedex 2 29218, France. 60 Institute of Molecular Genetics-CNR, Via Abbiategrasso 207, Pavia 27100, Italy. 61 The Charles Bronfman Institute for Personalized Medicine, Icahn School of Medicine at Mount Sinai, New York, New York 10029, USA. 62 Department of Health Sciences, University of Milano, Via Antonio di Rudinı` 8, Milano 20142, Italy. 63 Max Planck Institute of Psychiatry, Kraepelinstrae 2–10, Munich 80804, Germany. 64 Cardiovascular Health Research Unit, Departments of Epidemiology and Medicine, University of Washington, 1730 Minor Ave, Suite 1360, Seattle, Washington 98101, USA. 65 Division of Psychiatric Genomics, Department of Psychiatry, Icahn School of Medicine at Mount Sinai, New York, New York 10029, USA. 66 Institute of Genetic Epidemiology, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, Ingolstaedter Landstrae 1, 85764 Neuherberg, Germany. 67 Research Unit of Molecular Epidemiology, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, Ingolstaedter Landstrae 1, 85764 Neuherberg, Germany. 68 Institute of Epidemiology II, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, 85764 Neuherberg, Germany. 69 German Center for Diabetes Research (DZD), Ingolstaedter Landstrae 1, Neuherberg 85764, Germany. 70 Westmead Millennium Institute, Centre for Vision Research, University of Sydney, C24 Westmead Hospital, New South Wales 2145, Australia. 71 Division of Biomedical Statistics and Informatics, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905, USA. 72 Centre for Clinical Epidemiology and Biostatistics, School of Medicine and Public Health, University of Newcastle, HMRI Building1, Kookaburra Circuit, New Lambton New South Wales 2305, Australia. 73 Clinical Research Design, Information Technology and Statistical Support, Hunter Medical Research Institute, Newcastle, 1 Kookaburra Circuit, New Lambton Heights, New South Wales 2305, Australia. 74 Institute of Population Genetics—CNR, Traversa La Crucca 3, 07040 Reg. Baldinca, Li Punti, Sassari, Italy. 75 Department of Nutrition, Harvard School of Public Health, 665 Huntington Avenue, Building 2, Boston, Massachusetts 02115, USA. 76 Department of Internal Medicine, University Medical Center Groningen, University of Groningen, Hanzeplein 1, Groningen 9713 GZ, The Netherlands. 77 Brigham and Women’s Hospital and Channing Laboratory, Harvard Medical School, 181 Longwood Avenue, Boston, Massachusetts 02115, USA. 78 Cardiology, Department of Specialties of Internal Medicine, Geneva University Hospitals, Rue Gabrielle-Perret-Gentil 4, Geneva 1205, Switzerland. NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications ARTICLE 79 Swiss Tropical and Public Health Institute, PO Box 4002, Basel, Switzerland. 80 University of Basel, Petersplatz 1, Basel 4003, Switzerland. 81 Division of Nephrology, Department of Internal Medicine and Medical Specialties, Columbus-Gemelli University Hospital, Catholic University, Via Moscati 31, Rome 00168, Italy. 82 Genetic Epidemiology, Queensland Institute of Medical Research, QIMR, PO Royal Brisbane Hospital, Queensland 4029, Australia. 83 Institute of Epidemiology and Biobank popgen, Christian-Albrechts University, Niemannsweg 11, Kiel 24105, Germany. 84 Institute of Epidemiology I, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, Ingolsta¨dter Landstrae 1, Neuherberg 85764, Germany. 85 Institute of 13 ARTICLE Medicine, Primary Care and Emergency Medicine, Geneva University Hospitals, Faculty of Medicine, University of Geneva, Geneva 1211, Switzerland. 96 Community Prevention Unit, University Institute of Social and Preventive Medicine, Lausanne University Hospital, Route de la Corniche 10, Lausanne 1010, Switzerland. 97 Department of Epidemiology, Rollins School of Public Health, Emory University, 1518 Clifton Road, NE, Atlanta, Georgia 30322, USA. 98 Department of Genomics of Common Disease, School of Public Health, Imperial College London, London W12 0NN, UK. 99 Croatian Centre for Global Health, University of Split Medical School, Sˇoltanska 2, Split 21000, Croatia. 100 Department of Cardiology, Leiden University Medical Center, PO Box 9600, Leiden 2300 RC, The Netherlands. 101 Interuniversity Cardiology Institute of the Netherlands (ICIN), Moreelsepark 1, Utrecht 3511 EP, The Netherlands. 102 Einthoven Laboratory for Experimental Vascular Medicine, Albinusdreef 2, Leiden 2333 ZA, The Netherlands. 103 Durrer Center for Cardiogenetic Research, Meibergdreef 9, Amsterdam 1105 AZ, The Netherlands. 104 Division of Public Health Sciences, Department of Biostatistical Sciences, Wake Forest University Health Sciences, 2326 Medical Center Boulevard, Winston-Salem, North Carolina 27157-1063, USA. 105 INSERM U744, Institut Pasteur de Lille, 1 rue du Pr, Calmette, Lille Ce´dex 59019, France. 106 Department of Epidemiology, School of Public Health, University of Michigan, 1415 Washington Heights, Ann Arbor, Michigan 48109-2029, USA. 107 Survey Research Center, Institute for Social Research, University of Michigan, 426 Thompson Street, #3456, Ann Arbor, Michigan 48104, USA. 108 Centre for Vision Research, Westmead Millennium Institute, University of Sydney, C24 Westmead Hospital, New South Wales 2145, Australia. 109 Wake Forest School of Medicine, Department of Internal Medicine/Geriatrics, Medical center Boulevard, Winston-Salem, North Carolina 27157, USA. 110 Department of Genetics, Harvard Medical School, 77 Avenue Louis Pasteur, NRB 0330, Boston, Massachusetts 02115, USA. 111 University of Sydney, The George Institute for Global Health, Level 10, King George V Building, 83-117 Missenden Road, Camperdown, New South Wales 2050, Australia. 112 Department of Medicine, University of Turku, Turku University Hospital, PO Box 52, Turku 20521, Finland. 113 Welch Center for Prevention, Epidemiology and Clinical Research, 2024 East Monument St, Suite 2-600, Baltimore, Maryland 21287, USA. 114 Vanderbilt University School of Medicine, 448 Eskind Biomedical Library, 2209 Garland Avenue, Nashville, Tennessee 37212, USA. 115 Department of Genetics, University of Groningen, University Medical Centre Groningen, PO Box 72, Groningen 9700 AB, The Netherlands. 116 Division of Endocrinology, Medical University of South Carolina, 171 Ashley Avenue, Charleston, South Carolina 29425, USA. 117 Institute of Anatomy and Cell Biology, University of Greifswald, Friedrich-Loefﬂer-Strae 23c, Greifswald 17487, Germany. 118 Center for Health Disparities, Department of Biology, East Carolina University, 1001 East 10th Street, N209 Howell Science Complex Mailstop 551, Greenville, North Carolina 27858, USA. 119 Intramural Research Program, Laboratory of Epidemiology, Demography, and Biometry, National Institute on Aging, Gateway Building, 3C309, 7201 Winsconsin Avenue, Bethesda, Maryland 20892-9205, USA. 120 Department of Clinical Chemistry, Fimlab Laboratories, University of Tampere, School of Medicine, Tampere 33520, Finland. 121 CNRS UMR 8199, 1 Rue du Professeur Calmette, Lille 59000, France. 122 Lille Pasteur Institute, 1 Rue du Professeur Calmette, Lille 59000, France. 123 Lille II University, 42 Rue paul Duez, Lille 59000, France. 124 Clinical Research Branch, National Institute on Aging, 251 Bayview Blvd, Baltimore, Maryland 21250, USA. 125 Department of Epidemiology and Biostatistics, Erasmus University Medical Center, Dr Molewaterplein, Rotterdam 50-603015 GE, The Netherlands. 126 Department of Forensic Molecular Biology, Erasmus University Medical Center, Dr Molewaterplein, Rotterdam 50-603015 GE, The Netherlands. 127 Medical Clinic V, Medical Faculty Mannheim, Heidelberg University, Theodor-Kutzer-Ufer 1-3, Mannheim 68167, Germany. 128 Austrian Stroke Prevention Study, Department of Neurology, Division of Special Neurology, Medical University Graz, Auenbruggerplatz 22, Graz 8036, Austria. 129 Centre for Clinical Epidemiology and Biostatistics, University of Newcastle, Hunter Medical Research Institute, John Hunter Hospital, Locked Bag 1, HRMC, New South Wales 2310, Australia. 130 The George Institute for Global Health, Nufﬁeld Department of Population Health, University of Oxford, Old Road Campus, Roosevelt Drive, Oxford OX3 7LF, UK. 131 Service of Nephrology, Lausanne University Hospital, Rue du Bugnon 17, Lausanne 1005, Switzerland. 132 Inserm UMRS 1018, CESP Team 10, Universite´ Paris-Sud, 16 avenue Paul Vaillant Couturier, Villejuif 94807, France. 133 Health Disparities Research Section, Clinical Research Branch, National Institute on Aging, National Institutes of Health, NIH Biomedical Center, 251 Bayview Boulevard, Suite 100, Baltimore, Maryland 21224, USA. 134 Center for Public Health Genomics, Department of Medicine (Cardiovascular Medicine), University of Virginia, PO Box 800717, Charlottesville, Virginia 22908, USA. 135 Department of Clinical Physiology, Tampere University Hospital, University of Tampere, School of Medicine, Tampere 33521, Finland. 136 University Institute of Social and Preventive Medicine, Centre Hospitalier Universitaire Vaudois, University of Lausanne, Route de la Corniche 2, Epalinges CH-1066, Switzerland. 137 INSERM UMR970, Paris Cardiovascular Research Center (PARCC), 56 rue Leblanc, Paris F-75015, France. 138 Paris Descartes University, Faculty of medicine, Paris Cite´ Sorbonne, 12 Rue de l’e´cole de Me´decine, Paris F-75006, France. 139 Brigham and Women’s Hospital, Harvard Medical School, 77 Avenue Louis Pasteur, Boston, Massachusetts 02115, USA. 140 Wellcome Trust Sanger Institute, Hinxton CB10 1HH, UK. 141 University of Zurich, Institute of Physiology, Mechanisms of Inherited Kidney Disorders Group, Winterthurerstrasse 190, Zu¨rich 8057, Switzerland. 142 Research Centre of Applied and Preventive Cardiovascular Medicine, University of Turku, Turku University Hospital, Department of Clinical Physiology, PO Box 52, Turku 20521, Finland. 143 Department Clinical Pharmacology, William Harvey Research Institute, Queen Mary University of London, London EC1M 6BQ, UK. 144 NIHR Barts Cardiovascular Biomedical Research Unit, Queen Mary University of London, London EC1M 6BQ, UK. 145 Harvard Medical School, 900 Commonwealth Avenue East, Boston, Massachusetts 02115, USA. 146 University of Alabama at Birmingham, Department of Medicine, 1530 3rd Avenue, South Birmingham, Alabama 35294-0022, USA. 147 University of Alabama at Birmingham, Department of Epidemiology, 1530 3rd Avenue, South Birmingham, Alabama 35294-0022, USA. 148 Department of Psychiatry and EMGO þ Institute, VU University Medical Center, A.J. Ernststraat 1187, Amsterdam 1081 HL, The Netherlands. 149 IFB AdiposityDiseases, University of Leipzig, Liebigstrae 21, Leipzig 04103, Germany. 150 Medical University Center Mainz, Langenbeckstrae 1, Mainz 55131, Germany. 151 Institute of Physiology, University of Greifswald, Greifswald 17487, Germany. 152 Clinic for Prosthodontic Dentistry, Gerostomatology and Material Science, University of Greifswald, Rotgerberstrae 8, Greifswald 17475, Germany. 153 School of Biomedical Sciences and Pharmacy, University of Newcastle, Hunter Medical Research Institute, John Hunter Hospital, Locked Bag 1, HRMC, New South Wales 2310, Australia. 154 Kidney Research Institute, University of Washington, Box 359606, 325 9th Avenue, Seattle, Washington 98104, USA. 155 Institute for Maternal and Child Health—IRCCS "Burlo Garofolo", Via dell’Istria 65, Trieste 34137, Italy. 156 Department of Internal Medicine, Division of Nephrology and Hypertension, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905, USA. 157 Clinic for Internal Medicine A, University of Greifswald, Friedrich-Loefﬂer-Strae 23a, Greifswald 17475, Germany. 158 Faculty of Pharmaceutical Sciences, University of Iceland, Sæmundargata 2, Reykjavik 101, Iceland. 159 Department of Clinical Physiology, Turku University Hospital, Research Centre of Applied and Preventive Cardiovascular Medicine, University of Turku, PO Box 52, Turku 20521, Finland. 160 Synlab Academy, Synlab Services GmbH, Oberer Eselsberg 45, Ulm 89081, Germany. 161 Department of Internal Medicine II—Cardiology, University of Ulm Medical Centre, Albert-Einstein-Allee 23, Ulm 89081, Germany. 162 The Mindich Child Health and Development Institute, Ichan School of Medicine at Mount Sinai, New York, New York 10029, USA. 163 Department of Neurology, General Central Hospital, Via Lorenz Bohler 5, Bolzano 39100, Italy. 164 Department of Neurology, University of Lu¨beck, Ratzeburger Allee 160, Lu¨beck 23538, Germany. 165 University of Maryland Medical School, Division of Nephrology, 685 W. Baltimore Street, MSTF 314, Baltimore, Maryland 21201, USA. 166 CRCHUM, University of Montreal, CHUM Research Center, Technopoˆle Angus, 900 Saint-Denis, Montreal, Que´bec, Canada H2X 0A9 167 Division of Endocrinology Brigham and Women s Hospital Harvard Medical School 221 Longwood Avenue Boston Massachusetts ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 Medicine, Primary Care and Emergency Medicine, Geneva University Hospitals, Faculty of Medicine, University of Geneva, Geneva 1211, Switzerland. 96 Community Prevention Unit, University Institute of Social and Preventive Medicine, Lausanne University Hospital, Route de la Corniche 10, Lausanne 1010, Switzerland. 97 Department of Epidemiology, Rollins School of Public Health, Emory University, 1518 Clifton Road, NE, Atlanta, Georgia 30322, USA. 98 Department of Genomics of Common Disease, School of Public Health, Imperial College London, London W12 0NN, UK. 99 Croatian Centre for Global Health, University of Split Medical School, Sˇoltanska 2, Split 21000, Croatia. 100 Department of Cardiology, Leiden University Medical Center, PO Box 9600, Leiden 2300 RC, The Netherlands. 101 Interuniversity Cardiology Institute of the Netherlands (ICIN), Moreelsepark 1, Utrecht 3511 EP, The Netherlands. 102 Einthoven Laboratory for Experimental Vascular Medicine, Albinusdreef 2, Leiden 2333 ZA, The Netherlands. 103 Durrer Center for Cardiogenetic Research, Meibergdreef 9, Amsterdam 1105 AZ, The Netherlands. 104 Division of Public Health Sciences, Department of Biostatistical Sciences, Wake Forest University Health Sciences, 2326 Medical Center Boulevard, Winston-Salem, North Carolina 27157-1063, USA. 105 INSERM U744, Institut Pasteur de Lille, 1 rue du Pr, Calmette, Lille Ce´dex 59019, France. 106 Department of Epidemiology, School of Public Health, University of Michigan, 1415 Washington Heights, Ann Arbor, Michigan 48109-2029, USA. 107 Survey Research Center, Institute for Social Research, University of Michigan, 426 Thompson Street, #3456, Ann Arbor, Michigan 48104, USA. 108 Centre for Vision Research, Westmead Millennium Institute, University of Sydney, C24 Westmead Hospital, New South Wales 2145, Australia. 109 Wake Forest School of Medicine, Department of Internal Medicine/Geriatrics, Medical center Boulevard, Winston-Salem, North Carolina 27157, USA. 110 Department of Genetics, Harvard Medical School, 77 Avenue Louis Pasteur, NRB 0330, Boston, Massachusetts 02115, USA. 111 University of Sydney, The George Institute for Global Health, Level 10, King George V Building, 83-117 Missenden Road, Camperdown, New South Wales 2050, Australia. 112 Department of Medicine, University of Turku, Turku University Hospital, PO Box 52, Turku 20521, Finland. 113 Welch Center for Prevention, Epidemiology and Clinical Research, 2024 East Monument St, Suite 2-600, Baltimore, Maryland 21287, USA. 114 Vanderbilt University School of Medicine, 448 Eskind Biomedical Library, 2209 Garland Avenue, Nashville, Tennessee 37212, USA. 115 Department of Genetics, University of Groningen, University Medical Centre Groningen, PO Box 72, Groningen 9700 AB, The Netherlands. ARTICLE 116 Division of Endocrinology, Medical University of South Carolina, 171 Ashley Avenue, Charleston, South Carolina 29425, USA. 117 Institute of Anatomy and Cell Biology, University of Greifswald, Friedrich-Loefﬂer-Strae 23c, Greifswald 17487, Germany. 118 Center for Health Disparities, Department of Biology, East Carolina University, 1001 East 10th Street, N209 Howell Science Complex Mailstop 551, Greenville, North Carolina 27858, USA. 119 Intramural Research Program, Laboratory of Epidemiology, Demography, and Biometry, National Institute on Aging, Gateway Building, 3C309, 7201 Winsconsin Avenue, Bethesda, Maryland 20892-9205, USA. 120 Department of Clinical Chemistry, Fimlab Laboratories, University of Tampere, School of Medicine, Tampere 33520, Finland. 121 CNRS UMR 8199, 1 Rue du Professeur Calmette, Lille 59000, France. 122 Lille Pasteur Institute, 1 Rue du Professeur Calmette, Lille 59000, France. 123 Lille II University, 42 Rue paul Duez, Lille 59000, France. 124 Clinical Research Branch, National Institute on Aging, 251 Bayview Blvd, Baltimore, Maryland 21250, USA. 125 Department of Epidemiology and Biostatistics, Erasmus University Medical Center, Dr Molewaterplein, Rotterdam 50-603015 GE, The Netherlands. 126 Department of Forensic Molecular Biology, Erasmus University Medical Center, Dr Molewaterplein, Rotterdam 50-603015 GE, The Netherlands. 127 Medical Clinic V, Medical Faculty Mannheim, Heidelberg University, Theodor-Kutzer-Ufer 1-3, Mannheim 68167, Germany. 128 Austrian Stroke Prevention Study, Department of Neurology, Division of Special Neurology, Medical University Graz, Auenbruggerplatz 22, Graz 8036, Austria. 129 Centre for Clinical Epidemiology and Biostatistics, University of Newcastle, Hunter Medical Research Institute, John Hunter Hospital, Locked Bag 1, HRMC, New South Wales 2310, Australia. 130 The George Institute for Global Health, Nufﬁeld Department of Population Health, University of Oxford, Old Road Campus, Roosevelt Drive, Oxford OX3 7LF, UK. 131 Service of Nephrology, Lausanne University Hospital, Rue du Bugnon 17, Lausanne 1005, Switzerland. 132 Inserm UMRS 1018, CESP Team 10, Universite´ Paris-Sud, 16 avenue Paul Vaillant Couturier, Villejuif 94807, France. 133 Health Disparities Research Section, Clinical Research Branch, National Institute on Aging, National Institutes of Health, NIH Biomedical Center, 251 Bayview Boulevard, Suite 100, Baltimore, Maryland 21224, USA. 134 Center for Public Health Genomics, Department of Medicine (Cardiovascular Medicine), University of Virginia, PO Box 800717, Charlottesville, Virginia 22908, USA. 135 Department of Clinical Physiology, Tampere University Hospital, University of Tampere, School of Medicine, Tampere 33521, Finland. 136 University Institute of Social and Preventive Medicine, Centre Hospitalier Universitaire Vaudois, University of Lausanne, Route de la Corniche 2, Epalinges CH-1066, Switzerland. 137 INSERM UMR970, Paris Cardiovascular Research Center (PARCC), 56 rue Leblanc, Paris F-75015, France. ARTICLE 138 Paris Descartes University, Faculty of medicine, Paris Cite´ Sorbonne, 12 Rue de l’e´cole de Me´decine, Paris F-75006, France. 139 Brigham and Women’s Hospital, Harvard Medical School, 77 Avenue Louis Pasteur, Boston, Massachusetts 02115, USA. 140 Wellcome Trust Sanger Institute, Hinxton CB10 1HH, UK. 141 University of Zurich, Institute of Physiology, Mechanisms of Inherited Kidney Disorders Group, Winterthurerstrasse 190, Zu¨rich 8057, Switzerland. 142 Research Centre of Applied and Preventive Cardiovascular Medicine, University of Turku, Turku University Hospital, Department of Clinical Physiology, PO Box 52, Turku 20521, Finland. 143 Department Clinical Pharmacology, William Harvey Research Institute, Queen Mary University of London, London EC1M 6BQ, UK. 144 NIHR Barts Cardiovascular Biomedical Research Unit, Queen Mary University of London, London EC1M 6BQ, UK. 145 Harvard Medical School, 900 Commonwealth Avenue East, Boston, Massachusetts 02115, USA. 146 University of Alabama at Birmingham, Department of Medicine, 1530 3rd Avenue, South Birmingham, Alabama 35294-0022, USA. 147 University of Alabama at Birmingham, Department of Epidemiology, 1530 3rd Avenue, South Birmingham, Alabama 35294-0022, USA. 148 Department of Psychiatry and EMGO þ Institute, VU University Medical Center, A.J. Ernststraat 1187, Amsterdam 1081 HL, The Netherlands. 149 IFB AdiposityDiseases, University of Leipzig, Liebigstrae 21, Leipzig 04103, Germany. 150 Medical University Center Mainz, Langenbeckstrae 1, Mainz 55131, Germany. 151 Institute of Physiology, University of Greifswald, Greifswald 17487, Germany. 152 Clinic for Prosthodontic Dentistry, Gerostomatology and Material Science, University of Greifswald, Rotgerberstrae 8, Greifswald 17475, Germany. 153 School of Biomedical Sciences and Pharmacy, University of Newcastle, Hunter Medical Research Institute, John Hunter Hospital, Locked Bag 1, HRMC, New South Wales 2310, Australia. 154 Kidney Research Institute, University of Washington, Box 359606, 325 9th Avenue, Seattle, Washington 98104, USA. 155 Institute for Maternal and Child Health—IRCCS "Burlo Garofolo", Via dell’Istria 65, Trieste 34137, Italy. 156 Department of Internal Medicine, Division of Nephrology and Hypertension, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905, USA. 157 Clinic for Internal Medicine A, University of Greifswald, Friedrich-Loefﬂer-Strae 23a, Greifswald 17475, Germany. 158 Faculty of Pharmaceutical Sciences, University of Iceland, Sæmundargata 2, Reykjavik 101, Iceland. 159 Department of Clinical Physiology, Turku University Hospital, Research Centre of Applied and Preventive Cardiovascular Medicine, University of Turku, PO Box 52, Turku 20521, Finland. 160 Synlab Academy, Synlab Services GmbH, Oberer Eselsberg 45, Ulm 89081, Germany. 161 Department of Internal Medicine II—Cardiology, University of Ulm Medical Centre, Albert-Einstein-Allee 23, Ulm 89081, Germany. ARTICLE 162 The Mindich Child Health and Development Institute, Ichan School of Medicine at Mount Sinai, New York, New York 10029, USA. 163 Department of Neurology, General Central Hospital, Via Lorenz Bohler 5, Bolzano 39100, Italy. 164 Department of Neurology, University of Lu¨beck, Ratzeburger Allee 160, Lu¨beck 23538, Germany. 165 University of Maryland Medical School, Division of Nephrology, 685 W. Baltimore Street, MSTF 314, Baltimore, Maryland 21201, USA. 166 CRCHUM, University of Montreal, CHUM Research Center, Technopoˆle Angus, 900 Saint-Denis, Montreal, Que´bec, Canada H2X 0A9. 167 Division of Endocrinology, Brigham and Women s Hospital, Harvard Medical School, 221 Longwood Avenue, Boston, Massachusetts Medicine, Primary Care and Emergency Medicine, Geneva University Hospitals, Faculty of Medicine, University of Geneva, Geneva 1211, Switzerland. 96 Community Prevention Unit, University Institute of Social and Preventive Medicine, Lausanne University Hospital, Route de la Corniche 10, Lausanne 1010, Switzerland. 97 Department of Epidemiology, Rollins School of Public Health, Emory University, 1518 Clifton Road, NE, Atlanta, Georgia 30322, USA. 98 Department of Genomics of Common Disease, School of Public Health, Imperial College London, London W12 0NN, UK. 99 Croatian Centre for Global Health, University of Split Medical School, Sˇoltanska 2, Split 21000, Croatia. 100 Department of Cardiology, Leiden University Medical Center, PO Box 9600, Leiden 2300 RC, The Netherlands. 101 Interuniversity Cardiology Institute of the Netherlands (ICIN), Moreelsepark 1, Utrecht 3511 EP, The Netherlands. 102 Einthoven Laboratory for Experimental Vascular Medicine Albinusdreef 2 Leiden 2333 ZA The Netherlands 103 Durrer Center for Cardiogenetic NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications 14 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 ICBP Consortium Goncalo R. Abecasis168, Linda S. Adair169, Myriam Alexander170, David Altshuler171,172, Najaf Amin24, Dan E. Arking173, Pankaj Arora174, Yurii Aulchenko24, Stephan J.L. Bakker76, Stefania Bandinelli175, Ines Barroso140, Jacques S. Beckmann176, John P. Beilby177, Richard N. Bergman178, Sven Bergmann176, Joshua C. Bis179, Michael Boehnke168, Lori L. Bonnycastle180, Stefan R. Bornstein181, Michiel L. Bots182, Jennifer L. Bragg-Gresham168, Stefan-Martin Brand183, Eva Brand184, Peter S. Braund185, Morris J. Brown186, Paul R. Burton187, Juan P. Casas188, Mark J. Caulﬁeld189, Aravinda Chakravarti173, John C. Chambers190, Giriraj R. Chandak191, Yen-Pei C. Chang192, Fadi J. Charchar193, Nish Chaturvedi194, Yoon Shin Cho195, Robert Clarke196, Francis S. Collins180, Rory Collins196, John M. Connell197, Jackie A. Cooper198, Matthew N. Cooper199, Richard S. Cooper200, Anna Maria Corsi201, Marcus Do¨rr202, Santosh Dahgam203, John Danesh170, George Davey Smith204, Ian N.M. Day204, Panos Deloukas140, Matthew Denniff185, Anna F. ARTICLE Dominiczak205, Yanbin Dong206, Ayo Doumatey26, Paul Elliott190, Roberto Elosua207, Jeanette Erdmann208, Susana Eyheramendy209, Martin Farrall210, Cristiano Fava211, Terrence Forrester212, F. Gerald R. Fowkes87, Ervin R. Fox213, Timothy M. Frayling214, Pilar Galan215, Santhi K. Ganesh216, Melissa Garcia217, Tom R. Gaunt204, Nicole L. Glazer179, Min Jin Go195, Anuj Goel210, Ju¨rgen Gra¨ssler181, Diederick E. Grobbee182, Leif Groop218, Simonetta Guarrera219, Xiuqing Guo220, David Hadley221, Anders Hamsten222, Bok-Ghee Han195, Rebecca Hardy223, Anna-Liisa Hartikainen224, Simon Heath225, Susan R. Heckbert226, Bo Hedblad211, Serge Hercberg215, Dena Hernandez20, Andrew A. Hicks1, Gina Hilton173, Aroon D. Hingorani227, Judith A. Hoffman Bolton7, Jemma C. Hopewell196, Philip Howard228, Steve E. Humphries198, Steven C. Hunt229, Kristian Hveem230, M. Arfan Ikram24, Muhammad Islam231,232, Naoharu Iwai233,234, Marjo-Riitta Jarvelin190, Anne U. Jackson168, Tazeen H. Jafar231,232, Charles S. Janipalli191, Toby Johnson189, Sekar Kathiresan235, Kay-Tee Khaw170, Hyung-Lae Kim195, Sanjay Kinra236, Yoshikuni Kita237, Mika Kivimaki227, Jaspal S. Kooner238, M.J. Kranthi Kumar191, Diana Kuh223, Smita R. Kulkarni239, Meena Kumari240, Johanna Kuusisto241, Tatiana Kuznetsova242, Markku Laakso241, Maris Laan243, Jaana Laitinen244, Edward G. Lakatta245, Carl D. Langefeld246, Martin G. Larson247, Mark Lathrop225, Debbie A. Lawlor204, Robert W. Lawrence199, Jong-Young Lee195, Nanette R. Lee248, Daniel Levy247, Yali Li249, Will T. Longstreth250, Jian’an Luan251, Gavin Lucas207, Barbara Ludwig181, Massimo Mangino252, K. Radha Mani191, Michael G. Marmot227, Francesco U.S. Mattace-Raso24, Giuseppe Matullo253, Wendy L. McArdle254, Colin A. McKenzie212, Thomas Meitinger255, Olle Melander211, Pierre Meneton256, James F. Meschia257, Tetsuro Miki258,259, Yuri Milaneschi124, Karen L. Mohlke260, Vincent Mooser261, Mario A. Morken180, Richard W. Morris262, Thomas H. Mosley263, Samer Najjar264, Narisu Narisu180, Christopher Newton-Cheh174, Khanh-Dung Hoang Nguyen173, Peter Nilsson211, Fredrik Nyberg203, Christopher J. O’Donnell247, Toshio Ogihara265, Takayoshi Ohkubo266, Tomonori Okamura233,234, Rick Twee-Hee Ong267, Halit Ongen210, N. Charlotte Onland-Moret182, Paul F. O’Reilly190, Elin Org243, Marco Orru268, Walter Palmas269, Jutta Palmen198, Lyle J. Palmer199, Nicholette D. Palmer246, Alex N. Parker270, John F. Peden210, Leena Peltonen140, Markus Perola271, Vasyl Pihur173, Carl G.P. Platou230, Andrew Plump272, Dorairajan Prabhakaran273, Bruce M. Psaty179, Leslie J. Raffel220, Dabeeru C. Rao274, Asif Rasheed275, Fulvio Ricceri253, Kenneth M. Rice276, Annika Rosengren277, Jerome I. Rotter220, Megan E. Rudock278, Siim So˜ber243, Tunde Salako279, Danish Saleheen275, Veikko Salomaa271, Nilesh J. Samani185, Steven M. Schwartz226, Peter E.H. Schwarz280, Laura J. Scott168, James Scott238, Angelo Scuteri245, Joban S. Sehmi238, Mark Seielstad281, Sudha Seshadri14, Pankaj Sharma282, Sue Shaw-Hawkins189, Gang Shi274, Nick R.G. Shrine187, Eric J.G. Sijbrands24, Xueling Sim283, Andrew Singleton20, Marketa Sjo¨gren211, Center for Statistical Genetics, Department of Biostatistics, University of Michigan, School of Public Health, Ann Arbor, Michigan 48103, USA. 169Department of Nutrition, University of North Carolina, Chapel Hill, North Carolina 27599, USA. 170Department of Public Health and Primary Care, University of Cambridge, Cambridge, CB1 8RN, UK. 171Department of Medicine, Harvard Medical School, Boston, Massachusetts 02115, USA. 172Department of Genetics, Harvard Medical School, Boston, Massachusetts 02115, USA. 173Center for Complex Disease Genomics, McKusick-Nathans Institute of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, USA. 174Center for Human Genetic Research, Cardiovascular Research Center, Massachusetts General Hospital, Boston, Massachusetts 02114, USA. 175Geriatric Rehabilitation Unit, Azienda Sanitaria Firenze (ASF), 50125 Florence, Italy. 176De´partement de Ge´ne´tique Me´dicale, Universite´ de Lausanne, Lausanne 1015, Switzerland. 177Pathology and Laboratory Medicine, University of Western Australia, 6009 Crawley, Western Australia, Australia. 178Department of Physiology and Biophysics, Keck School of Medicine, University of Southern California, Los Angeles, California 90033, USA. 179Cardiovascular Health Research Unit, Department of Medicine, University of Washington, Seattle 98195, Washington, USA. 180National Human Genome Research Institute, National Institutes of Health, Bethesda, Maryland 20892, USA. 181Department of Medicine III, Medical Faculty Carl Gustav Carus at the Technical University of Dresden, Dresden 01307, Germany. 182Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Heidelberglaan 100, Utrecht 3508 GA, The Netherlands. 183Leibniz-Institute for Arteriosclerosis Research, Department of Molecular Genetics of Cardiovascular Disease, University of Mu¨nster, Mu¨nster, Germany. 184University Hospital Mu¨nster, Internal Medicine D, Mu¨nster, Germany. 185Department of Cardiovascular Sciences, University of Leicester, Glenﬁeld Hospital, Leicester LE3 9QP, UK. 186Clinical Pharmacology Unit, University of Cambridge, Addenbrookes Hospital, Hills Road, Cambridge CB2 2QQ, UK. 187Department of Health Sciences, University of Leicester, University Rd, Leicester LE1 7RH, UK. 188Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London WC1E 7HT, UK. 189Clinical Pharmacology and The Genome Centre, William Harvey Research Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London EC1M 6BQ, UK. 190Department of Epidemiology and Biostatistics, School of Public Health, Imperial College London, Norfolk Place, London W2 1PG, UK. 191Centre for Cellular and Molecular Biology (CCMB), Council of Scientiﬁc and Industrial Research (CSIR), Uppal Road, Hyderabad 500 007, India. 192University of Maryland, School of Medicine, Baltimore, Maryland 21201, USA. 193School of Science and Engineering, University of Ballarat, Ballarat 3353, Australia. 194International Centre for Circulatory Health, National Heart & Lung Institute, Imperial College, London, UK. 195Center for Genome Science, National Institute of Health, Seoul, Korea. 196Clinical Trial Service Unit and Epidemiological Studies Unit, University of Oxford, Oxford OX3 7LF, UK. 197University of Dundee, Ninewells Hospital & Medical School, Dundee DD1 9SY, UK. 198Centre for Cardiovascular Genetics, University College London, London WC1E 6JF, UK. 199Centre for Genetic Epidemiology and Biostatistics, University of Western Australia, Crawley, Western Australia, Australia. 200Department of Preventive Medicine and Epidemiology, Loyola University Medical School, Maywood, Illinois, USA. 201Tuscany Regional Health Agency, Florence, Italy. 202Department of Internal Medicine B, Ernst-Moritz-Arndt-University Greifswald, Greifswald 17487, Germany. 203Occupational and Environmental Medicine, Department of Public Health and Community Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg 40530, Sweden. 204MRC Centre for Causal Analyses in Translational Epidemiology, School of Social & Community Medicine, University of Bristol, Bristol BS8 2BN, UK. 205BHF Glasgow Cardiovascular Research Centre, University of Glasgow, 126 University Place, Glasgow G12 8TA, UK. 206Georgia Prevention Institute, Department of Pediatrics, Medical College of Georgia, 30912 Augusta, Georgia, USA. 207Cardiovascular Epidemiology and Genetics, Institut Municipal d’Investigacio Medica, Barcelona Biomedical Research Park, 88 Doctor Aiguader, Barcelona 08003, Spain. 208Medizinische Klinik II, Universita¨t zu Lu¨beck, 23562 Lu¨beck, Germany. 209Department of Statistics, Pontiﬁcia Universidad Catolica de Chile, Vicun˜a Mackena, Santiago 4860, Chile. 210Department of Cardiovascular Medicine, The Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford OX3 7BN, UK. 211Department of Clinical Sciences, Lund University, SE-205 02 Malmo¨, Sweden. 212Tropical Medicine Research Institute, University of the West Indies, Mona, Kingston, Jamaica. 213Department of Medicine, University of Mississippi Medical Center, 2500 North State St, Jackson, Mississippi 39216, USA. 214Genetics of Complex Traits, Peninsula Medical School, University of Exeter, EX1 2LU Exeter, UK. 215U557 Institut National de la Sante´ et de la Recherche Me´dicale, U1125 Institut National de la Recherche Agronomique, Universite´ Paris 13, F-93017 Bobigny, France. 216Department of Internal Medicine, Division of Cardiovascular Medicine, University of Michigan Medical Center, Ann Arbor 48108, Michigan, USA. 217Laboratory of Epidemiology, Demography, Biometry, National Institute on Aging, National Institutes of Health, Bethesda, Maryland 20892, USA. 218Department of Clinical Sciences, Diabetes and Endocrinology Research Unit, University Hospital, SE-205 02 Malmo¨, Sweden. 219Human Genetics Foundation (HUGEF), Via Lagrange 35, Torino 10123, Italy. 220Medical Genetics Institute, Cedars-Sinai Medical Center, Los Angeles, 90048 California, USA. 221Division of Community Health Sciences, St George’s University of London, London SW17 0RE, UK. 222Atherosclerosis Research Unit, Department of Medicine, Karolinska Institute, SE-171 77 Stockholm, Sweden. 223MRC Unit for Lifelong Health & Ageing, London WC1B 5JU, UK. 224Institute of Clinical Medicine/Obstetrics and Gynecology, University of Oulu, 90014 Oulu, Finland. 225Centre National de Ge´notypage, Commissariat de L’Energie Atomique, Institut de Ge´nomique, 91000 Evry, France. 226Department of Epidemiology, University of Washington, Seattle, Washington 98195, USA. 227Epidemiology Public Health, UCL, London WC1E 6BT, UK. 228William Harvey Research Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London EC1M 6BQ, UK. 229Cardiovascular Genetics, University of Utah, School of Medicine, Salt Lake City, 84132 Utah, USA. 230HUNT Research Centre, Department of Public Health and General Practice, Norwegian University of Science and Technology, Levanger 7600, Norway. 231Department of Community ICBP Consortium NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications 15 NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications Nicholas L. Smith226, Maria Soler Artigas187, Tim D. Spector252, Jan A. Staessen244, Alena Stancakova241, Nanette I. Steinle192, David P. Strachan221, Heather M. Stringham168, Yan V. Sun106, Amy J. Swift180, Yasuharu Tabara258,259, E-Shyong Tai284, Philippa J. Talmud198, Andrew Taylor240, Janos Terzic285, Dag S. Thelle286, Martin D. Tobin189, Maciej Tomaszewski185, Vikal Tripathy273, Jaakko Tuomilehto287, Ioanna Tzoulaki190, Manuela Uda268, Hirotsugu Ueshima288, Cuno S.P.M. Uiterwaal184, Satoshi Umemura289, Pim van der Harst93, Yvonne T. van der Schouw182, Wiek H. van Gilst93, Erkki Vartiainen271, Ramachandran S. Vasan247, Gudrun Veldre243, Germaine C. Verwoert24, Margus Viigimaa290, D.G. Vinay191, Paolo Vineis291, Benjamin F. Voight235, Peter Vollenweider292, Lynne E. Wagenknecht246, Louise V. Wain187, Xiaoling Wang206, Thomas J. Wang247, Nicholas J. Wareham251, Hugh Watkins210, Alan B. Weder216, Peter H. Whincup221, Kerri L. Wiggins179, Jacqueline C.M. Witteman24, Andrew Wong223, Ying Wu259, Chittaranjan S. Yajnik239, Jie Yao220, J.H. Young293, Diana Zelenika225, Guangju Zhai252, Weihua Zhang190, Feng Zhang252, Jing Hua Zhao251, Haidong Zhu206, Xiaofeng Zhu249, Paavo Zitting294, Ewa Zukowska-Szczechowska295 netics, Department of Biostatistics, University of Michigan, School of Public Health, Ann Arbor, Michigan 48103, USA. 170 168Center for Statistical Genetics, Department of Biostatistics, University of Michigan, School of Public Health, Ann Ar 169D f N i i U i i f N h C li Ch l Hill N h C li 27599 USA 170D f P 296Laboratory for Statistical Analysis, RIKEN Center for Integrative Medical Sciences, Yokohama, Kanagawa 230-0045, Japan. 297Department of Human Genetics and Disease Diversity, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, Tokyo 113-8510, Japan. 298Institute of Biomedical Sciences, Academia Sinica, Taipei 115, Taiwan. 299School of Chinese Medicine, China Medical University, Taichung 404, Taiwan. 300Cardiovascular Institute and Fu Wai Hospital, Chinese Academy of Medical Sciences, Peking Union Medical College, Beijing 100037, China. 301Department of Gene Diagnostics and Therapeutics, Research Institute, National Center for Global Health and Medicine, Tokyo 162-8655, Japan. 302Laboratory for Endocrinology and Metabolism, RIKEN Center for Integrative Medical Sciences, Yokohama, Kanagawa 230-0045, Japan. 303Laboratory for Medical Informatics, RIKEN ARTICLE 290Tallinn University of Technology, Institute of Biomedical Engineering, Ehitajate tee 5, Tallinn 19086, Estonia. 291Department of Epidemiology and Public Health, Imperial College, Norfolk Place, London W2 1PG, UK. 292Department of Internal Medicine, Centre Hospitalier Universitaire Vaudois, Lausanne 1011, Switzerland. 293Department of Medicine, Johns Hopkins University, Baltimore, 21205 Maryland, USA. 294Lapland Central Hospital, Department of Physiatrics, Box 8041, Rovaniemi 96101, Finland. 295Department of Internal Medicine, Diabetology, and Nephrology, Medical University of Silesia, Zabrze 41-800, Poland. NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications Health Sciences, Aga Khan University, 74800 Karachi, Pakistan. 232Department of Medicine, Aga Khan University, 74800 Karachi, Pakistan. 233Department of Genomic Medicine, National Cerebral and Cardiovascular Research Center, Suita 565-8565, Japan. 234Department of Preventive Cardiology, National Cerebral and Cardiovascular Research Center, Suita 565-8565, Japan. 235Medical Population Genetics, Broad Institute of Harvard and MIT, 5 Cambridge Center, Cambridge, Massachusetts 02142, USA. 236Division of Non-communicable disease Epidemiology, The London School of Hygiene and Tropical Medicine London, Keppel Street, London WC1E 7HT, UK. 237Department of Health Science, Shiga University of Medical Science, Otsu 520-2192, Japan. 238National Heart and Lung Institute, Imperial College London, London W12 0HS, UK. 239Diabetes Unit, KEM Hospital and Research Centre, Rasta Peth, Pune, 411011 Maharashtra, India. 240Genetic Epidemiology Group, Epidemiology and Public Health, UCL, London WC1E 6BT, UK. 241Department of Medicine, University of Kuopio, Kuopio University Hospital, Kuopio 70210, Finland. 242Studies Coordinating Centre, Division of Hypertension and Cardiac Rehabilitation, Department of Cardiovascular Diseases, University of Leuven, Campus Sint Rafae¨l, Kapucijnenvoer 35, Block D, Box 7001, Leuven 3000, Belgium. 243Institute of Molecular and Cell Biology, University of Tartu, Riia 23, Tartu 51010, Estonia. 244Finnish Institute of Occupational Health, Finnish Institute of Occupational Health, Aapistie 1, Oulu 90220, Finland. 245Gerontology Research Center, National Institute on Aging, Baltimore, Maryland 21224, USA. 246Wake Forest University Health Sciences, Winston-Salem, North Carolina 27157, USA. 247National Heart, Lung and Blood Institute Framingham Heart Study, Framingham, 01702-5827 Massachusetts, USA. 248Ofﬁce of Population Studies Foundation, University of San Carlos, Talamban, Cebu 6000, Philippines. 249Department of Epidemiology and Biostatistics, Case Western Reserve University, 2103 Cornell Road, Cleveland, Ohio 44106, USA. 250Department of Medicine and Neurology, University of Washington, Seattle, Washington 98195, USA. 251MRC Epidemiology Unit, Institute of Metabolic Science, Cambridge CB2 0QQ, UK. 252Department of Twin Research and Genetic Epidemiology, King’s College London, SE1 7EH London, UK. 253Department of Genetics, Biology and Biochemistry, University of Torino, Via Santena 19, Torino 10126, Italy. 254ALSPAC Laboratory, University of Bristol, Bristol BS8 2BN, UK. 255Institute of Human Genetics, Helmholtz Zentrum Munich, German Research Centre for Environmental Health, Neuherberg 85764, Germany. 256U872 Institut National de la Sante´ et de la Recherche Me´dicale, Centre de Recherche des Cordeliers, 75006 Paris, France. 257Mayo Clinic, Jacksonville, 32224 Florida, USA. 258Department of Basic Medical Research and Education, Ehime University Graduate School of Medicine, Toon 791-0295, Japan. 259Department of Geriatric Medicine, Ehime University Graduate School of Medicine, Toon 791-0295, Japan. 260Department of Genetics, University of North Carolina, Chapel Hill, North Carolina 27599, USA. 261Division of Genetics, GlaxoSmithKline, Philadelphia, Pennsylvania 19101, USA. 262Department of Primary Care and Population Health, UCL, London NW3 2PF, UK. 263Department of Medicine (Geriatrics), University of Mississippi Medical Center, Jackson, 39216 Mississippi, USA. 264Laboratory of Cardiovascular Science, Intramural Research Program, National Institute on Aging, NIH, Baltimore, 21224-6825 Maryland, USA. 265Department of Geriatric Medicine, Osaka University Graduate School of Medicine, Suita 565-0871, Japan. 266Tohoku University, Graduate School of Pharmaceutical Sciences and Medicine, Sendai 980-8578, Japan. 267Genome Institute of Singapore, Agency for Science, Technology and Research, Singapore 138672, Singapore. 268Istituto di Neurogenetica e Neurofarmacologia, Consiglio Nazionale delle Ricerche, Cittadella Universitaria di Monserrato, 09042 Monserrato, Cagliari, Italy. 269Columbia University, New York, 10027 New York, USA. 270Amgen, 1 Kendall Square, Building 100, Cambridge, Massachusetts 02139, USA. 271National Institute for Health and Welfare, Helsinki 00271, Finland. 272Merck Research Laboratory, 126 East Lincoln Avenue, Rahway, New Jersey 07065, USA. 273South Asia Network for Chronic Disease, Public Health Foundation of India, C-1/52, SDA, New Delhi 100016, India. 274Division of Biostatistics, Washington University School of Medicine, Saint Louis, Missouri 63110, USA. 275Center for Non-Communicable Diseases, 74800 Karachi, Pakistan. 276Department of Biostatistics, University of Washington, Seattle, 98105 Washington, USA. 277Department of Emergency and Cardiovascular Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg 41685, Sweden. 278Epidemiology & Prevention, Division of Public Health Sciences, Wake Forest University, School of Medicine, Winston-Salem, North Carolina 27157, USA. 279University of Ibadan, PMB 5017 Ibadan, Nigeria. 280Prevention and Care of Diabetes, Department of Medicine III, Medical Faculty Carl Gustav Carus at the Technical University of Dresden, Dresden 01307, Germany. 281Department of Laboratory Medicine, Institute of Human Genetics, University of California, San Francisco, 513 Parnassus Avenue, San Francisco, California 94143, USA. 282Imperial College Cerebrovascular Unit (ICCRU), Imperial College, London W6 8RF, UK. 283Centre for Molecular Epidemiology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 117597, Singapore. 284Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 119074, Singapore. 285Faculty of Medicine, University of Split, 21000 Split, Croatia. 286Department of Biostatistics, Institute of Basic Medical Sciences, University of Oslo, Oslo 0317, Norway. 287Diabetes Prevention Unit, National Institute for Health and Welfare, Helsinki 00271, Finland. 288Lifestyle-related Disease Prevention Center, Shiga University of Medical Science, Otsu 520-2192, Japan. 289Department of Medical Science and Cardiorenal Medicine, Yokohama City University School of Medicine, Yokohama 236-0004, Japan. 290Tallinn University of Technology, Institute of Biomedical Engineering, Ehitajate tee 5, Tallinn 19086, Estonia. 291Department of Epidemiology and Public Health, Imperial College, Norfolk Place, London W2 1PG, UK. 292Department of Internal Medicine, Centre Hospitalier Universitaire Vaudois, Lausanne 1011, Switzerland. 293Department of Medicine, Johns Hopkins University, Baltimore, 21205 Maryland, USA. 294Lapland Central Hospital, Department of Physiatrics, Box 8041, Rovaniemi 96101, Finland. 295Department of Internal Medicine, Diabetology, and Nephrology, Medical University of Silesia, Zabrze 41-800, Poland. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 ARTICLE Nicholas L. Smith226, Maria Soler Artigas187, Tim D. Spector252, Jan A. Staessen244, Alena Stancakova241, Nanette I. Steinle192, David P. Strachan221, Heather M. Stringham168, Yan V. Sun106, Amy J. Swift180, Yasuharu Tabara258,259, E-Shyong Tai284, Philippa J. Talmud198, Andrew Taylor240, Janos Terzic285, Dag S. Thelle286, Martin D. Tobin189, Maciej Tomaszewski185, Vikal Tripathy273, Jaakko Tuomilehto287, Ioanna Tzoulaki190, Manuela Uda268, Hirotsugu Ueshima288, Cuno S.P.M. Uiterwaal184, Satoshi Umemura289, Pim van der Harst93, Yvonne T. van der Schouw182, Wiek H. van Gilst93, Erkki Vartiainen271, Ramachandran S. Vasan247, Gudrun Veldre243, Germaine C. Verwoert24, Margus Viigimaa290, D.G. Vinay191, Paolo Vineis291, Benjamin F. Voight235, Peter Vollenweider292, Lynne E. Wagenknecht246, Louise V. Wain187, Xiaoling Wang206, Thomas J. Wang247, Nicholas J. Wareham251, Hugh Watkins210, Alan B. Weder216, Peter H. Whincup221, Kerri L. Wiggins179, Jacqueline C.M. Witteman24, Andrew Wong223, Ying Wu259, Chittaranjan S. Yajnik239, Jie Yao220, J.H. Young293, Diana Zelenika225, Guangju Zhai252, Weihua Zhang190, Feng Zhang252, Jing Hua Zhao251, Haidong Zhu206, Xiaofeng Zhu249, Paavo Zitting294, Ewa Zukowska-Szczechowska295 Nicholas L. Smith226, Maria Soler Artigas187, Tim D. Spector252, Jan A. Staessen244, Alena Stancakova241, Nanette I. Steinle192, David P. Strachan221, Heather M. Stringham168, Yan V. Sun106, Amy J. Swift180, Yasuharu Tabara258,259, E-Shyong Tai284, Philippa J. Talmud198, Andrew Taylor240, Janos Terzic285, Dag S. Thelle286, Martin D. Tobin189, Maciej Tomaszewski185, Vikal Tripathy273, Jaakko Tuomilehto287, Ioanna Tzoulaki190, Manuela Uda268, Hirotsugu Ueshima288, Cuno S.P.M. Uiterwaal184, Satoshi Umemura289, Pim van der Harst93, Yvonne T. van der Schouw182, Wiek H. van Gilst93, Erkki Vartiainen271, Ramachandran S. Vasan247, Gudrun Veldre243, Germaine C. Verwoert24, Margus Viigimaa290, D.G. Vinay191, Paolo Vineis291, Benjamin F. Voight235, Peter Vollenweider292, Lynne E. Wagenknecht246, Louise V. Wain187, Xiaoling Wang206, Thomas J. Wang247, Nicholas J. Wareham251, Hugh Watkins210, Alan B. Weder216, Peter H. Whincup221, Kerri L. Wiggins179, Jacqueline C.M. Witteman24, Andrew Wong223, Ying Wu259, Chittaranjan S. Yajnik239, Jie Yao220, J.H. Young293, Diana Zelenika225, Guangju Zhai252, Weihua Zhang190, Feng Zhang252, Jing Hua Zhao251, Haidong Zhu206, Xiaofeng Zhu249, Paavo Zitting294, Ewa Zukowska-Szczechowska295 16 ARTICLE ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 Health Sciences, Aga Khan University, 74800 Karachi, Pakistan. 232Department of Medicine, Aga Khan University, 74800 Karachi, Pakistan. 233Department of Genomic Medicine, National Cerebral and Cardiovascular Research Center, Suita 565-8565, Japan. 234Department of Preventive Cardiology, National Cerebral and Cardiovascular Research Center, Suita 565-8565, Japan. 235Medical Population Genetics, Broad Institute of Harvard and MIT, 5 Cambridge Center, Cambridge, Massachusetts 02142, USA. 236Division of Non-communicable disease Epidemiology, The London School of Hygiene and Tropical Medicine London, Keppel Street, London WC1E 7HT, UK. 237Department of Health Science, Shiga University of Medical Science, Otsu 520-2192, Japan. 238National Heart and Lung Institute, Imperial College London, London W12 0HS, UK. 239Diabetes Unit, KEM Hospital and Research Centre, Rasta Peth, Pune, 411011 Maharashtra, India. 240Genetic Epidemiology Group, Epidemiology and Public Health, UCL, London WC1E 6BT, UK. 241Department of Medicine, University of Kuopio, Kuopio University Hospital, Kuopio 70210, Finland. 242Studies Coordinating Centre, Division of Hypertension and Cardiac Rehabilitation, Department of Cardiovascular Diseases, University of Leuven, Campus Sint Rafae¨l, Kapucijnenvoer 35, Block D, Box 7001, Leuven 3000, Belgium. 243Institute of Molecular and Cell Biology, University of Tartu, Riia 23, Tartu 51010, Estonia. 244Finnish Institute of Occupational Health, Finnish Institute of Occupational Health, Aapistie 1, Oulu 90220, Finland. 245Gerontology Research Center, National Institute on Aging, Baltimore, Maryland 21224, USA. 246Wake Forest University Health Sciences, Winston-Salem, North Carolina 27157, USA. 247National Heart, Lung and Blood Institute Framingham Heart Study, Framingham, 01702-5827 Massachusetts, USA. 248Ofﬁce of Population Studies Foundation, University of San Carlos, Talamban, Cebu 6000, Philippines. 249Department of Epidemiology and Biostatistics, Case Western Reserve University, 2103 Cornell Road, Cleveland, Ohio 44106, USA. 250Department of Medicine and Neurology, University of Washington, Seattle, Washington 98195, USA. 251MRC Epidemiology Unit, Institute of Metabolic Science, Cambridge CB2 0QQ, UK. 252Department of Twin Research and Genetic Epidemiology, King’s College London, SE1 7EH London, UK. 253Department of Genetics, Biology and Biochemistry, University of Torino, Via Santena 19, Torino 10126, Italy. 254ALSPAC Laboratory, University of Bristol, Bristol BS8 2BN, UK. 255Institute of Human Genetics, Helmholtz Zentrum Munich, German Research Centre for Environmental Health, Neuherberg 85764, Germany. 256U872 Institut National de la Sante´ et de la Recherche Me´dicale, Centre de Recherche des Cordeliers, 75006 Paris, France. 257Mayo Clinic, Jacksonville, 32224 Florida, USA. 258Department of Basic Medical Research and Education, Ehime University Graduate School of Medicine, Toon 791-0295, Japan. 259Department of Geriatric Medicine, Ehime University Graduate School of Medicine, Toon 791-0295, Japan. ARTICLE 260Department of Genetics, University of North Carolina, Chapel Hill, North Carolina 27599, USA. 261Division of Genetics, GlaxoSmithKline, Philadelphia, Pennsylvania 19101, USA. 262Department of Primary Care and Population Health, UCL, London NW3 2PF, UK. 263Department of Medicine (Geriatrics), University of Mississippi Medical Center, Jackson, 39216 Mississippi, USA. 264Laboratory of Cardiovascular Science, Intramural Research Program, National Institute on Aging, NIH, Baltimore, 21224-6825 Maryland, USA. 265Department of Geriatric Medicine, Osaka University Graduate School of Medicine, Suita 565-0871, Japan. 266Tohoku University, Graduate School of Pharmaceutical Sciences and Medicine, Sendai 980-8578, Japan. 267Genome Institute of Singapore, Agency for Science, Technology and Research, Singapore 138672, Singapore. 268Istituto di Neurogenetica e Neurofarmacologia, Consiglio Nazionale delle Ricerche, Cittadella Universitaria di Monserrato, 09042 Monserrato, Cagliari, Italy. 269Columbia University, New York, 10027 New York, USA. 270Amgen, 1 Kendall Square, Building 100, Cambridge, Massachusetts 02139, USA. 271National Institute for Health and Welfare, Helsinki 00271, Finland. 272Merck Research Laboratory, 126 East Lincoln Avenue, Rahway, New Jersey 07065, USA. 273South Asia Network for Chronic Disease, Public Health Foundation of India, C-1/52, SDA, New Delhi 100016, India. 274Division of Biostatistics, Washington University School of Medicine, Saint Louis, Missouri 63110, USA. 275Center for Non-Communicable Diseases, 74800 Karachi, Pakistan. 276Department of Biostatistics, University of Washington, Seattle, 98105 Washington, USA. 277Department of Emergency and Cardiovascular Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg 41685, Sweden. 278Epidemiology & Prevention, Division of Public Health Sciences, Wake Forest University, School of Medicine, Winston-Salem, North Carolina 27157, USA. 279University of Ibadan, PMB 5017 Ibadan, Nigeria. 280Prevention and Care of Diabetes, Department of Medicine III, Medical Faculty Carl Gustav Carus at the Technical University of Dresden, Dresden 01307, Germany. 281Department of Laboratory Medicine, Institute of Human Genetics, University of California, San Francisco, 513 Parnassus Avenue, San Francisco, California 94143, USA. 282Imperial College Cerebrovascular Unit (ICCRU), Imperial College, London W6 8RF, UK. 283Centre for Molecular Epidemiology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 117597, Singapore. 284Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 119074, Singapore. 285Faculty of Medicine, University of Split, 21000 Split, Croatia. 286Department of Biostatistics, Institute of Basic Medical Sciences, University of Oslo, Oslo 0317, Norway. 287Diabetes Prevention Unit, National Institute for Health and Welfare, Helsinki 00271, Finland. 288Lifestyle-related Disease Prevention Center, Shiga University of Medical Science, Otsu 520-2192, Japan. 289Department of Medical Science and Cardiorenal Medicine, Yokohama City University School of Medicine, Yokohama 236-0004, Japan. 323The Cardiovascular Institute, University of Pennsylvania, Philadelphia, Pennsylvania 19104, USA. 324Institut fu¨r integrative und experimentelle Genomik, Universita¨t zu Lu¨beck, Lu¨beck 23562, Germany. 325Deutsches Herzzentrum Mu¨nchen, Technische Universita¨t Mu¨nchen, Mu¨nchen 80636, Germany. 326Department of Medicine, Stanford University School of Medicine, Stanford, 94305-5101 California, USA. 327Division of Cardiovascular and Neuronal Remodelling, Multidisciplinary Cardiovascular Research Centre, Leeds Institute of Genetics, Health and Therapeutics, University of Leeds, Leeds LS2 9JT, UK. 328Klinik und Poliklinik fu¨r Innere Medizin II, Universita¨t Regensburg, 93053 Regensburg, Germany. 329Institut fu¨r Medizinische Biometrie und Statistik, Universita¨t zu Lu¨beck, 23562 Lu¨beck, Germany. 330Science Center, Tampere University Hospital, Tampere 33521, Finland. 331The John & Jennifer Ruddy Canadian Cardiovascular Genetics Centre, University of Ottawa Heart Institute, Ottawa, Ontario, Canada K1Y 4W7. 332deCODE Genetics, Reykjavik 101, Iceland. 333Faculty of Medicine, University of Iceland, Reykjavik 101, Iceland. 334Hudson Alpha Institute, Huntsville, 35806 Alabama, USA. 335Cardiovascular Research Methods Centre, University of Ottawa Heart Institute, 40 Ruskin Street, Ottawa, Ontario, Canada K1Y 4W7. 336The Blavatnik School of Computer Science, Department of Molecular Microbiology and Biotechnology, Tel-Aviv University, Tel-Aviv 6997801, Israel. 337The International Computer Science Institute, Berkeley, 94704 California, USA. 338Biostatistics and Epidemiology, University of Pennsylvania, Philadelphia, 19104 Pennsylvania, USA. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 Center for Integrative Medical Sciences, Yokohama, Kanagawa 230-0045, Japan. 304Saw Swee Hock School of Public Health, National University of Singapore, Singapore 119077, Singapore. 305Department of Medicine, Khoo Teck Puat Hospital, Singapore 768828, Singapore. 306Department of Epidemiology and Public Health, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 117597, Singapore. 307Singapore Eye Research Institute, Singapore National Eye Centre, Singapore 168751, Singapore. 308Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore 117597, Singapore. 309Centre for Eye Research Australia, University of Melbourne, East Melbourne, 3002 Victoria, Australia. 310Center for Human Genetics, Duke University Medical Center, Durham, 27710 North Carolina, USA. 311Department of Statistics and Applied Probability, National University of Singapore, Singapore 117546, Singapore. 312NUS Graduate School for Integrative Science and Engineering, National University of Singapore, Singapore 119077, Singapore. 313Center for Genome Science, National Institute of Health, Osong Health Technology Administration Complex, 187 Chungcheongbuk-do, Korea. 314Department of Preventive Medicine, Seoul National University College of Medicine, Seoul 08826, Korea. 315Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, 70112 Louisiana, USA. 316Department of Clinical Gene Therapy, Osaka University Graduate School of Medicine, Suita, Osaka 565-0871, Japan. 317Department of Geriatric Medicine and Nephrology, Osaka University Graduate School of Medicine, Suita, Osaka 565-0871, Japan. 318Department of Allergy and Rheumatology, Graduate School of Medicine, University of Tokyo, Tokyo 113-0033, Japan. 319Laboratory for Genotyping Development, RIKEN Center for Integrative Medical Sciences, Yokohama, Kanagawa 230-0045, Japan. 320Laboratory of Molecular Medicine, Human Genome Center, Institute of Medical Science, University of Tokyo, Tokyo 108- 8639, Japan. 321Laboratory for International Alliance, RIKEN Center for Integrative Medical Sciences, Yokohama, Kanagawa 230-0045, Japan. 322Laboratory for Cardiovascular Diseases, RIKEN Center for Integrative Medical Sciences, Yokohama, Kanagawa 230-0045, Japan. NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications AGEN Consortium Yukinori Okada296,297, Jer-Yuarn Wu298,299, Dongfeng Gu300, Fumihiko Takeuchi301, Atsushi Takahashi296, Shiro Maeda302, Tatsuhiko Tsunoda303, Peng Chen304, Su-Chi Lim305,306, Tien-Yin Wong307,308,309, Jianjun Liu267, Terri L. Young310, Tin Aung307,308, Yik-Ying Teo267,283,306,311,312, Young Jin Kim313, Daehee Kang314, Chien-Hsiun Chen298,299, Fuu-Jen Tsai299, Li-Ching Chang298, S.-J. Cathy Fann298, Hao Mei315, James E. Hixson34, Shufeng Chen300, Tomohiro Katsuya316,317, Masato Isono301, Eva Albrecht66, Kazuhiko Yamamoto318, Michiaki Kubo319, Yusuke Nakamura320, Naoyuki Kamatani321, Norihiro Kato301, Jiang He315, Yuan-Tsong Chen298, Toshihiro Tanaka297,322 17 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 ARTICLE 339Cardiovascular Research Center and Cardiology Division, Massachusetts General Hospital, Boston, 02114 Massachusetts, USA. 340Center for Human Genetic Research, Massachusetts General Hospital, Boston, 02114 Massachusetts, USA. 341Division of Research, Kaiser Permanente, Oakland, 94611 California, USA. 342Institute for Human Genetics, University of California, San Francisco, San Francisco, 94143 California, USA. 343Department of Cardiovascular Medicine, Cleveland Clinic 7255 Old Oak Blvd, Cleveland, Ohio 44130, USA. 344Medizinische Klinik und Poliklinik, Johannes-Gutenberg Universita¨t Mainz, Universita¨tsmedizin, 55122 Mainz, Germany. 345Institut fu¨r Klinische Chemie und Laboratoriumsmediizin, Johannes-Gutenberg Universita¨t Mainz, Universita¨tsmedizin, 55122 Mainz, Germany. 346INSERM UMRS 937, Pierre and Marie Curie University (UPMC, Paris 6) and Medical School, 75005 Paris, France. 347Boston University, School of Public Health, Boston, 02118 Massachusetts, USA. 348University of Minnesota School of Public Health, Division of Epidemiology and Community Health, School of Public Health (A.R.F.), Minneapolis, 55454 Minnesota, USA. 349University of Washington, Department of Internal Medicine, Seattle, 98195-6420 Washington, USA. 350University of Texas, School of Public Health, Houston, 77030 Texas, USA. 351Department of Medicine, Landspitali University Hospital, Reykjavik 101, Iceland. 352Institute of Epidemiology, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, 85764 Neuherberg, Germany. 353Division of Endocrinology and Diabetes, Graduate School of Molecular Endocrinology and Diabetes, University of Ulm, 89069 Ulm, Germany. 354Division of Endocrinology, Department of Medicine, Medical University of Graz, 8010 Graz, Austria. 355Synlab Center of Laboratory Diagnostics Heidelberg, 69037 Heidelberg, Germany. 356Division of Clinical Chemistry, Department of Medicine, Albert Ludwigs University, 79085 Freiburg, Germany. 357Clinical Institute of Medical and Chemical Laboratory Diagnostics, Medical University Graz, 8010 Graz, Austria. 358Cardiology Group Frankfurt-Sachsenhausen, 60594 Frankfurt, Germany. 359The Center for Applied Genomics, Children’s Hospital of Philadelphia, 19104 Philadelphia, Pennsylvania, USA. 360Cardiovascular Research Institute, Medstar Health Research Institute, Washington Hospital Center, Washington, DC 20010, USA. 361Genetics Division and Drug Discovery, GlaxoSmithKline, King of Prussia, Pennsylvania 19406, USA. 362The Institute for Translational Medicine and Therapeutics, School of Medicine, University of Pennsylvania, Philadelphia, 19104-5158 Pennsylvania, USA. 363Department of Cardiovascular Surgery, University of Leicester, Leicester LE1 7RH, UK. 364Division of Cardiovascular and Diabetes Research, Multidisciplinary Cardiovascular Research Centre, Leeds Institute of Genetics, Health and Therapeutics, University of Leeds, Leeds LS2 9JT, UK. 365LIGHT Research Institute, Faculty of Medicine and Health, University of Leeds, Leeds LS2 9JT, UK. 372Department of Internal Medicine, Division of Cardiology, Medical University Graz, Graz 8036, Austria. 373Department of Medical Sciences, Uppsala University, Uppsala 75185, Sweden. 374Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota 55905, USA. 375Institute of Medical Epidemiology, Biostatistics and Informatics, Martin Luther University of Halle-Wittenberg, Halle-Wittenberg, Halle (Saale) 06097, Germany. 376Department of Cardiology, Erasmus University Medical Center, Rotterdam 3000 CA, The Netherlands. 377Medical Clinic 2, University of Lu¨beck, Lu¨beck 23538, Germany. 378Department of Medical Epidemiology and Biostatistics, Karolinska Institute, Stockholm 17177, Sweden. 379Martin Luther University, Halle-Wittenberg, Halle (Saale) 06097, Germany. 380Department of Epidemiology, University of Alabama at Birmingham, Birmingham, Alabama 35294-0022, USA. 381Division of Cardiology, University of Maryland Hospital, Baltimore, Maryland 21201, USA. 382Institute of Medical Biometry, Epidemiology, and Informatics, Johannes Gutenberg University, Mainz 55101, Germany. 383Institute for Community Medicine, Ernst-Moritz-Arndt-Universita¨t, Greifswald 17475, Germany. CHARGe-Heart Failure Group CHARGe-Heart Failure Group Laura R. Loehr366,367,368, Wayne D. Rosamond367, Emelia Benjamin247, Talin Haritunians220, David Couper369, Joanne Murabito247, Ying A. Wang13, Bruno H. Stricker24, Patricia P. Chang366, James T. Willerson370,371 366Department of Medicine, University of North Carolina at Chapel Hill, North Carolina 27516, USA. 367Department of Epidemiology, University of North Carolina at Chapel Hill, North Carolina 27599-7435, USA. 368National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina 27709, USA. 369Department of Biostatistics, University of North Carolina at Chapel Hill, North Carolina 27514, USA. 370University of Texas, Houston Health Science Center, Houston, Texas 77030, USA. 371Texas Heart Institute, Houston, Texas 77225-0345, USA. CHARGe Heart Failure Group Laura R. Loehr366,367,368, Wayne D. Rosamond367, Emelia Benjamin247, Talin Haritunians220, David Couper369, Joanne Murabito247, Ying A. Wang13, Bruno H. Stricker24, Patricia P. Chang366, James T. Willerson370,371 366Department of Medicine, University of North Carolina at Chapel Hill, North Carolina 27516, USA. 367Department of Epidemiology, University of North Carolina at Chapel Hill North Carolina 27599 7435 USA 368National Institute of Environmental Health Sciences Research Triangle Park North Carolina 366Department of Medicine, University of North Carolina at Chapel Hill, North Carolina 27516, USA. 367Department of Epidemiology, University of North Carolina at Chapel Hill, North Carolina 27599-7435, USA. 368National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina 27709, USA. 369Department of Biostatistics, University of North Carolina at Chapel Hill, North Carolina 27514, USA. 370University of Texas, Houston Health Science Center, Houston, Texas 77030, USA. 371Texas Heart Institute, Houston, Texas 77225-0345, USA. NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 NATURE COMMUNICATIONS \| DOI: 10.1038/ncomms10023 CARDIOGRAM Muredach P. Reilly323, Heribert Schunkert57,324,325, Themistocles L. Assimes326, Alistair Hall327, Christian Hengstenberg328, Inke R. Ko¨nig329, Reijo Laaksonen330, Ruth McPherson331, John R. Thompson187, Unnur Thorsteinsdottir332,333, Andreas Ziegler329, Devin Absher334, Li Chen335, L. Adrienne Cupples13,247, Eran Halperin336,337, Mingyao Li338, Kiran Musunuru140,339,340, Michael Preuss324,329, Arne Schillert329, Gudmar Thorleifsson332, George A. Wells335, Hilma Holm332, Robert Roberts331, Alexandre F.R. Stewart331, Stephen Fortmann326, Alan Go341, Mark Hlatky326, Carlos Iribarren341, Joshua Knowles326, Richard Myers334, Thomas Quertermous326, Steven Sidney341, Neil Risch342, Hua Tang343, Stefan Blankenberg344, Renate Schnabel344, Christoph Sinning344, Karl J. Lackner345, Laurence Tiret346, Viviane Nicaud346, Francois Cambien346, Christoph Bickel344, Hans J. Rupprecht344, Claire Perret346, Carole Proust346, Thomas F. Mu¨nzel344, Maja Barbalic34, Ida Yii-Der Chen220, Serkalem Demissie-Banjaw246,347, Aaron Folsom348, Thomas Lumley275, Kristin Marciante349, Kent D. Taylor220, Kelly Volcik350, Solveig Gretarsdottir332, Jeffrey R. Gulcher332, Augustine Kong332, Kari Stefansson332,333, Gudmundur Thorgeirsson333,351, Karl Andersen333,351, Marcus Fischer328, Anika Grosshennig324,329, Patrick Linsel-Nitschke324, Klaus Stark328, Stefan Schreiber38, Zouhair Aherrahrou57,324, Petra Bruse57,324, Angela Doering352, Norman Klopp352, Patrick Diemert324, Christina Loley324,329, Anja Medack57,324, Janja Nahrstedt324,329, Annette Peters68, Arnika K. Wagner324, Christina Willenborg57,324, Bernhard O. Bo¨hm353, Harald Dobnig354, Tanja B. Grammer355, Michael M. Hoffmann356, Andreas Meinitzer357, Bernhard R. Winkelmann358, Stefan Pilz354, Wilfried Renner357, Hubert Scharnagl357, Tatjana Stojakovic357, Andreas Tomaschitz354, Karl Winkler356, Candace Guiducci16, Noel Burtt16, Stacey B. Gabriel16, Sonny Dandona331, Olga Jarinova331, Liming Qu338, Robert Wilensky323, William Matthai323, Hakon H. Hakonarson359, Joe Devaney360, Mary Susan Burnett360, Augusto D. Pichard360, Kenneth M. Kent360, Lowell Satler360, Joseph M. Lindsay360, Ron Waksman360, Christopher W. Knouff361, Dawn M. Waterworth361, Max C. Walker361, Stephen E. Epstein360, Daniel J. Rader323,362, Christopher P. Nelson185, Benjamin J. Wright363, Anthony J. Balmforth364, Stephen G. Ball365 NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications 18 ECHOGen Consortium ECHOGen Consortium Stephan B. Felix202, Norbert Watzinger372, Jayashri Aragam247, Robert Zweiker372, Lars Lind373, Richard J. Rodeheffer374, Karin Halina Greiser375, Jaap W. Deckers376, Jan Stritzke377, Erik Ingelsson378, Iftikhar Kullo374, Johannes Haerting375, Thorsten Reffelmann202, Margaret M. Redﬁeld374, Karl Werdan379, Gary F. Mitchell247, Donna K. Arnett380, John S. Gottdiener381, Maria Blettner382, Nele Friedrich383 19 19 NATURE COMMUNICATIONS \| 7:10023 \| DOI: 10.1038/ncomms10023 \| www.nature.com/naturecommunications
https://openalex.org/W3109150160	https://seer.ufrgs.br/gearte/article/download/103750/59194	Spanish; Castilian	null	Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural	Revista Gearte	2,020	cc-by	12,401	ISSN 2357-9854\|e-ISSN 2596-3198 (online) ISSN 2357-9854\|e-ISSN 2596-3198 (online) VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. http://dx.doi.org/10.22456/2357-9854.103750 PALABRAS CLAVE Trabajo artesanal. Artesanía. Artesanos. Capitalismo. Creatividad. Trabajo artesanal. Artesanía. Artesanos. Capitalismo. Creatividad. RESUMO — Trabalho artesanal, artesanato e indústrias criativas: reflexões em relação às transformações da atividade sociocultural — O objetivo do artigo é desenvolver uma crítica da compreensão do artesanato e do trabalho artesanal desde o desenvolvimento histórico capitalista. Nesse sentido se realiza uma primeira reflexão entre a indústria e o trabalho artesanal como parte da organização técnica e política dos artesãos no capitalismo industrial. Depois se apresenta uma reflexão sobre a indústria cultural e o artesanato, entendendo as relações sociais, políticas e simbólicas que oferecem reconhecimento do aparecimento do artesanato e do artesão como expressão cultural. Logo, se complementa a reflexão com a análise das indústrias criativas e a produção artesanal na atual classificação de ocupações. No final do artigo realizam-se reflexões conceituais e políticas que definem trajetórias do estudo e intervenção. PALAVRAS CHAVE p PALAVRAS-CHAVE Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural Daniel Roberto Vega Torres (Fundación Universitaria Juan de Castellanos, Tunja/Boyacá, Colombia) Daniel Roberto Vega Torres (Fundación Universitaria Juan de Castellanos, Tunja/Boyacá, Colombia) RESUMEN — Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural — El objetivo del artículo es desarrollar una crítica de la comprensión de la artesanía y del trabajo artesanal desde el desarrollo histórico capitalista. Para ello se realiza una primera reflexión entre la industria y el trabajo artesanal como parte de la organización técnica y política de los artesanos en el capitalismo industrial. Luego se desarrolla una reflexión sobre la Industria cultural y la artesanía, entendiendo las implicaciones sociales políticas y simbólicas que permiten reconocer la emergencia de la artesanía y del artesano. Después se complementa la reflexión mediante el análisis de las industrias creativas y la producción artesanal en la clasificación de ocupaciones actual. Al final se realizan reflexiones conceptuales y políticas que definen trayectorias de estudio e intervención. PALABRAS CLAVE PALABRAS CLAVE PALABRAS CLAVE Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. PALAVRAS-CHAVE Trabalho artesanal. Artesanato. Artesãos. Capitalismo. Criatividade. ABSTRACT — Craft work, crafts and creative industries: reflections on the transformations of socio-cultural activity. — The objective of the article is to develop a critique of the understanding of crafts and craft work from the capitalist historical development. For this, a first reflection is made between industry and artisan work as part of the technical and political organization of artisans in industrial capitalism. Then a reflection is developed on the cultural industry and crafts, understanding the political and symbolic social implications that allow us to recognize the emergence of crafts and artisans. Then the reflection is complemented by analyzing the creative industries and artisan production in the current occupation classification. At the end, conceptual and political reflections are made that define trajectories of study and intervention. KEYWORDS Craft work. Crafts. Artisans. Capitalism. Creativity. 461 VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. Disponível em: http://seer.ufrgs.br/gearte a GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. , j transformaciones de la actividad sociocultural. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. g p Disponível em: http://seer.ufrgs.br/gearte Disponível em: http://seer.ufrgs.br/gearte vista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Introducción ¿Cuál es la compresión contemporánea de la artesanía o de lo artesanal?, ¿cuándo se habla de artesanal se podría hablar de un campo constituido o acaso puede pensarse como una práctica cultural natural de los seres humanos? y de forma actual, ¿qué relación tiene con las actividades creativas y/o artísticas? Estos cuestionamientos trazan trayectorias sobre la comprensión de la actividad artesanal y de la artesanía como una cuestión que requiere de análisis no solamente técnico sino también sociocultural, pues de allí puede tenerse un juicio claro de las posibilidades de desarrollo e intervención política sobre dicho fenómeno social. Actualmente el desarrollo de trabajos académicos sobre la actividad artesanal ha tenido un fuerte aumento en relación a diferentes condiciones, si trazáramos grosso modo la tendencia en las publicaciones sobre la artesanía (Figura 1), podría evaluarse un crecimiento tanto por la actividad propiamente dicha, como por la palabra como uso descriptivo y hasta metafórico. Así, esto expresa una consolidación de intereses políticos culturales que van desde las condiciones propiamente técnicas, empresariales, culturales, arqueológicas, artísticas, entre otras. Figura 1 – Producción de textos sobre “artesanía” por año, tomado de la base de datos Google Académico en el periodo 1988 a 2018 Fuente: Google Académico, Elaboración propia. (2019). 0 2000 4000 6000 8000 10000 12000 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Figura 1 – Producción de textos sobre “artesanía” por año, tomado de la base de datos Google Académico en el periodo 1988 a 2018 De esta manera, el peso de la producción confirma el desarrollo de un área de interés académico que puede indicar de forma directa o indirecta el 462 Disponível em: http://seer.ufrgs.br/gearte reconocimiento que la artesanía ha tenido como estudio en Iberoamérica. La complejidad de dicho proceso no deja de ser problemático si pensamos que la artesanía como estudio se compone de diferentes aristas que se pueden extrapolar a diferentes áreas del conocimiento, o que pueden considerarse divergentes entre sí, sin observar una base común de organización que cobije las diferentes expresiones del trabajo artesanal o de artesanía. Introducción Un segundo momento del desarrollo de la industria cultural y la artesanía, como eje de la producción patrimonial cultural que permite un proceso de subsunción indirecta del saber hacer local a la acumulación capitalista; luego, la relación entre las industrias creativas y la actividad artística-artesanal, como parte de la subordinación directa del trabajo artesanal al capital, entendiendo su desarrollo en el mercado mundial. Por último, se presentará una relación de categorías que pueden ser utilizadas para el estudio e intervención de lo artesanal. reconocimiento de identidades laborales y políticas. Un segundo momento del desarrollo de la industria cultural y la artesanía, como eje de la producción patrimonial cultural que permite un proceso de subsunción indirecta del saber hacer local a la acumulación capitalista; luego, la relación entre las industrias creativas y la actividad artística-artesanal, como parte de la subordinación directa del trabajo artesanal al capital, entendiendo su desarrollo en el mercado mundial. Por último, se presentará una relación de categorías que pueden ser utilizadas para el estudio e intervención de lo artesanal. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las Disponível em: http://seer.ufrgs.br/gearte g p Disponível em: http://seer.ufrgs.br/gearte , j transformaciones de la actividad sociocultural. Disponível em: http://seer.ufrgs.br/gearte Introducción En este sentido, tratar de responder la cuestión sobre la comprensión contemporánea de la artesanía puede que tenga que ir más allá de la apariencia naturalizada del objeto de estudio, y por el contrario definir una crítica sobre su conformación sociocultural que corresponda a sus posibilidades concretas. reconocimiento que la artesanía ha tenido como estudio en Iberoamérica. La complejidad de dicho proceso no deja de ser problemático si pensamos que la artesanía como estudio se compone de diferentes aristas que se pueden extrapolar a diferentes áreas del conocimiento, o que pueden considerarse divergentes entre sí, sin observar una base común de organización que cobije las diferentes expresiones del trabajo artesanal o de artesanía. En este sentido, tratar de responder la cuestión sobre la comprensión contemporánea de la artesanía puede que tenga que ir más allá de la apariencia naturalizada del objeto de estudio, y por el contrario definir una crítica sobre su conformación sociocultural que corresponda a sus posibilidades concretas. En ese sentido, este estudio parte de una reflexión sobre lo artesanal como eje de problematización de la actividad humana, que corresponda con la historia concreta del desarrollo de producción dominante capitalista, tomando como hipótesis de estudio que solo mediante la comprensión del trabajo en un modo dominante de producción histórica se puede definir las condiciones de desarrollo y las transformaciones sobre las cuales se puede ejercer una comprensión pertinente para su correspondiente intervención y dirección. En ese sentido, la reflexión sobre lo artesanal no debe partir de un concepto “ideal”, como usualmente se realiza a la hora de definir las características de dicha actividad, y donde los debates sobre los límites de lo que debe ser o no debe ser lo artesanal extrapolan la coherencia del estudio científico, derivando de esta manera en una imposición normativa del ejercicio administrativo. Por el contrario, pensar lo artesanal debe fundamentarse en las formas de desarrollo de las condiciones productivas generales de la sociedad. De esta manera, mediante el desarrollo de la producción capitalista se podrá obtener un mecanismo de comprensión claro del fenómeno social. Así, se dispondrán de tres ejes de desarrollo e institucionalización de la actividad artesanal para la exposición: el primero desde una comprensión propia de la relación trabajo industrial y trabajo artesanal que permite la disyuntiva productiva y el 463 reconocimiento de identidades laborales y políticas. Industria y Trabajo artesanal Uno de los aspectos principales trabajados en la historia de la formación del capitalismo industrial en el siglo XIX ha permitido profundizar en los cambios sociales de la producción, de donde puede permitirse evaluar cambios significativos entre lo que era propiamente pre-industrial como aquello que constituiría el desarrollo de una sociedad industrial. Ese periodo de división de la sociedad moderna sobre sus bases materiales de producción permitió que tanto el desarrollo de la cooperación de trabajadores, la construcción de las grandes fábricas y la maquinaria como eje de la producción permitiera un cambio en la consideración de la evaluación de la producción manual con herramientas simples, que estaba considerada como parte de los oficios de las ciudades en los diferentes territorios de la sociedad occidental. Tanto en la historia de los imperios como en la transformación de sus colonias, este cambio produjo una fuerte reacción en la organización de los trabajadores de oficios y gremios artesanales, que vieron vulnerados su forma de organización social y política, pues la dependencia de los gremios de artesanos a la vida cotidiana de las poblaciones tanto de las grandes urbes como de los pueblos de las colonias terminarían por evidenciar de manera violenta la proletarización de la población trabajadora. Este cambio de las condiciones generales de la producción condujo a la formación de la clase proletaria de una manera mucho más rápida en Inglaterra, 464 pero también se produjo un cambio en las condiciones de los pequeños propietarios de la clase media, que reaccionaban de forma organizada a una política de protección de los oficios artesanales en gremios con fuerte influencia en la organización política local y nacional, como aconteció en los territorios de la actual Alemania y Francia. En ese sentido, la actividad gremial de artesanos pudo verse contrarrestada por mecanismos sociales de lucha revolucionaria y contrarrevolucionaria de los gremios de artesanos (ENGELS, 1980). Para el caso, la proletarización de la población rural en migración a las ciudades y el dominio de la clase burguesa en la consolidación de un modo de producción capitalista dominante definieron la identidad de los artesanos. La identificación de los artesanos como clase media corresponde a un proceso de organización política en las transformaciones materiales de producción industrial dominante, esto ocurre tanto en las sociedades de temprana industrialización, como en las colonias y excolonias. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. Disponível em: http://seer.ufrgs.br/gearte g p Disponível em: http://seer.ufrgs.br/gearte a GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Industria y Trabajo artesanal Cuando se consolida el proceso de identificación de lo artesanal en el capitalismo industrial decimonónico se produce un cambio más que técnico propiamente político, en donde los artesanos, en el sentido de hombres que socialmente pueden organizarse por sus oficios reconocidos social y culturalmente pueden integrar una clase social en el contexto histórico de dominio del trabajo productivo masculino. frente a los cambios económicos de producción y consumo en la sociedad. En Latinoamérica se encuentra este proceso político de sociabilidad de los artesanos desde una identificación positiva como clase social frente a los comerciantes, lo que deriva en una lucha indirecta de la transformación de la producción industrial en Europa, y que en este caso genera un movimiento internacional en la región que lleva a consolidar sociabilidades políticas de artesanos (LOAIZA, 2011). Cuando se consolida el proceso de identificación de lo artesanal en el capitalismo industrial decimonónico se produce un cambio más que técnico propiamente político, en donde los artesanos, en el sentido de hombres que socialmente pueden organizarse por sus oficios reconocidos social y culturalmente pueden integrar una clase social en el contexto histórico de dominio del trabajo productivo masculino. Son los artesanos organizados en valores tradicionales de producción, pero no solo de valores técnicos sino también culturales, se aferran a producir discursos reaccionarios en su mayoría tanto a la clase burguesa de empresarios en creciente dominio del mercado, como de los trabajadores proletarios que no constituían una clase que tuviese las capacidades simbólicas que los oficios y la propiedad de los productores artesanales tenían en su haber, esos recursos de distinción familiar, religiosa, ociosa y cultural que frecuentemente constituía un escenario más cercano de identificación a las condiciones de la pequeña burguesía. Este notable ejercicio de desdoblamiento de la clase media que comentaba Marx (1980), fue parte del desarrollo general en cada uno de los países, pues solo en algunos casos de fuerte poder político como en Francia, Italia y Alemania pudieron permanecer los artesanos agremiados en su mercado interno (ZARCA, 1993), lo que actualmente da su valor diferencial en organización frente a otros sistemas de organización laboral de los artesanos. Pero esta situación no corresponde con las colonias y excolonias que debieron supeditarse a la integración del mercado de producción industrializada en la dependencia de mercados. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. vista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. sponível em: http://seer.ufrgs.br/gearte Disponível em: http://seer.ufrgs.br/gearte Industria y Trabajo artesanal La experiencia de la organización laboral de los gremios artesanales en las colonias ya había tenido una reacción temprana en la liberalización de políticas comerciales que afectaban los oficios locales, quienes habían apoyado movilizaciones iniciales frente a las políticas imperiales de comercio y producción (LOAIZA, 2011). En un medio de producción local la diferenciación por oficios es más común para la identificación de los trabajadores, en ese sentido, la importancia del reconocimiento de los gremios no se presentaba como una identidad general de “artesanos”, sino que se identificaban por oficios, zapateros, joyeros, pintores, escultores, etc (MORA DE JARAMILLO, 1974). En ese sentido, la diferenciación y distinción dentro de los gremios tenía que ver más con el material de trabajo y con la habilidad técnica que permitía un consumo de producción suntuosa, frente a labores de trabajo manual con materiales rústicos que representaban la vida pre-moderna. Solo mediante la irrupción de la industrialización y la grande fábrica puede entenderse que exista una necesidad de identificación de los pequeños propietarios y trabajadores manuales que permitieran identificar su posición política 465 Disponível em: http://seer.ufrgs.br/gearte frente a los cambios económicos de producción y consumo en la sociedad. En Latinoamérica se encuentra este proceso político de sociabilidad de los artesanos desde una identificación positiva como clase social frente a los comerciantes, lo que deriva en una lucha indirecta de la transformación de la producción industrial en Europa, y que en este caso genera un movimiento internacional en la región que lleva a consolidar sociabilidades políticas de artesanos (LOAIZA, 2011). Cuando se consolida el proceso de identificación de lo artesanal en el capitalismo industrial decimonónico se produce un cambio más que técnico propiamente político, en donde los artesanos, en el sentido de hombres que socialmente pueden organizarse por sus oficios reconocidos social y culturalmente pueden integrar una clase social en el contexto histórico de dominio del trabajo productivo masculino. frente a los cambios económicos de producción y consumo en la sociedad. En Latinoamérica se encuentra este proceso político de sociabilidad de los artesanos desde una identificación positiva como clase social frente a los comerciantes, lo que deriva en una lucha indirecta de la transformación de la producción industrial en Europa, y que en este caso genera un movimiento internacional en la región que lleva a consolidar sociabilidades políticas de artesanos (LOAIZA, 2011). g p Disponível em: http://seer.ufrgs.br/gearte Industria y Trabajo artesanal La situación de los artesanos en su lucha política pierde vigencia frente al crecimiento de las ciudades y los cambios en la forma de consumo e integración laboral mediante la formación de la clase obrera industrial que corresponde a un aspecto general de transformación de la vida familiar y cultural urbanizada. Lo artesanal como forma de producción se mantiene poco a poco en lo local, en la formación de lo rural como forma de vida contrapuesta a la modernidad que se vive en las ciudades. El trabajador autónomo en su oficio, propietario de su saber y de sus herramientas corresponde cada vez más a lo tradicional, a lo pre-moderno, y su identidad como artesano ya no va a competir con el sistema de reconocimiento y producción capitalista por su baja productividad, sino al sistema de reconocimiento de lo cultural, como símbolo del saber y el hacer. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. , j , transformaciones de la actividad sociocultural. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las Industria y Trabajo artesanal Con la llegada del siglo XX, la consolidación del capitalismo industrial y su proceso imperialista definió la transformación de las condiciones laborales en los 466 diferentes países tanto centrales como periféricos a la acumulación del capital. La situación de los artesanos en su lucha política pierde vigencia frente al crecimiento de las ciudades y los cambios en la forma de consumo e integración laboral mediante la formación de la clase obrera industrial que corresponde a un aspecto general de transformación de la vida familiar y cultural urbanizada. Lo artesanal como forma de producción se mantiene poco a poco en lo local, en la formación de lo rural como forma de vida contrapuesta a la modernidad que se vive en las ciudades. El trabajador autónomo en su oficio, propietario de su saber y de sus herramientas corresponde cada vez más a lo tradicional, a lo pre-moderno, y su identidad como artesano ya no va a competir con el sistema de reconocimiento y producción capitalista por su baja productividad, sino al sistema de reconocimiento de lo cultural, como símbolo del saber y el hacer. diferentes países tanto centrales como periféricos a la acumulación del capital. La situación de los artesanos en su lucha política pierde vigencia frente al crecimiento de las ciudades y los cambios en la forma de consumo e integración laboral mediante la formación de la clase obrera industrial que corresponde a un aspecto general de transformación de la vida familiar y cultural urbanizada. Lo artesanal como forma de producción se mantiene poco a poco en lo local, en la formación de lo rural como forma de vida contrapuesta a la modernidad que se vive en las ciudades. El trabajador autónomo en su oficio, propietario de su saber y de sus herramientas corresponde cada vez más a lo tradicional, a lo pre-moderno, y su identidad como artesano ya no va a competir con el sistema de reconocimiento y producción capitalista por su baja productividad, sino al sistema de reconocimiento de lo cultural, como símbolo del saber y el hacer. diferentes países tanto centrales como periféricos a la acumulación del capital. , j transformaciones de la actividad sociocultural. Disponível em: http://seer.ufrgs.br/gearte g p Disponível em: http://seer.ufrgs.br/gearte g p Disponível em: http://seer.ufrgs.br/gearte TE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. La industria cultural y la Artesanía Cabe anotar que cuando se habla de forma dominante de producción capitalista no tiene que corresponder a que exista únicamente una producción industrial con una proletarización general de la población trabajadora, sino que por el contrario, es el poder que ejerce la transformación de una forma de producir lo que termina afectando directa o indirectamente la organización social del trabajo de manera amplia. En este sentido, aunque en el siglo XX comienza el crecimiento de las ciudades y la forma de vida moderna corresponde con un consumo de mercancías industrializadas en un mercado mundial, es necesario identificar que el desarrollo capitalista también permitió la constitución de un proceso diferencial de la vida comunitaria local o rural. Si no se integraba la población a la vida urbana, es posible que la manera de vivir, pensar y producir tuviese que definirse de manera contrapuesta, pero en muchos casos dependiente del sistema dominante. Este proceso de identificación de la producción artesanal tiene que ver más con un proceso de comprensión y estudio de lo cultural como eje del reconocimiento, es decir, en un primer momento los estudios sobre las poblaciones 467 Disponível em: http://seer.ufrgs.br/gearte rurales, derivadas de la etnografía, la antropología y la sociología definieron características de la población en los discursos que definían formas de producir y vivir. Es así que en este mecanismo de caracterización de la población se ejerce un reconocimiento externo, de los académicos hacía la población, que terminan por definir la actividad social local pre-moderna como una actividad cultural o folclórica, es decir que hace parte del estudio eurocéntrico colonial de la mirada de la producción y forma de vida que encajaba en la comprensión de la “naturaleza” de dichas poblaciones, especialmente de indios, negros y asiáticos. La formación de la artesanía como estudio corresponde a un ejercicio de poder de definición de las formas de vida y las formas de producción rural, pre- industrial y pre-moderna que fueron clasificados desde la noción colonizadora de la diversidad en las formas de producción local, dependientes de los territorios tanto por sus condiciones físicas (recursos materiales, recursos ambientales) como de sus condiciones sociales (cultura, memoria, tradición, aprendizaje). Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las sponível em: http://seer.ufrgs.br/gearte , j transformaciones de la actividad sociocultural. g p Disponível em: http://seer.ufrgs.br/gearte TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las rmaciones de la actividad sociocultural. La industria cultural y la Artesanía En ese sentido, el objetivo del conocimiento no se encontraba en los sujetos propiamente tales, sino en los objetos, en la “artesanía”, es el objeto la razón de estudio, saber y clasificación, una clasificación que ponía de manifiesto el folclor y la cultural de lo local en un ejercicio homogéneo de identificación a lo largo de los territorios colonizados o periféricos a la vida moderna. Mediante la “racialización” de las poblaciones periféricas (CAMPOS, 2012), rurales o minoritarias, adquieren un valor naturalizado que exige del discurso académico crear la correspondencia entre el ser social como ser natural, en esto consiste la clave de los discursos sobre la producción de artesanías a lo largo del siglo XX, pues se concentra como un debate sobre la aprehensión de lo distinto a la vida occidental, urbana e industrial. De esta manera se reproduce ideológicamente el mito de la artesanía como parte de la naturaleza humana. La artesanía no solo encubre un estudio etnográfico y colonizador de la no- modernidad, de la otredad, sino que además fortalece esa distinción en la descripción del objeto sobre el sujeto, pues en este caso no se puede ser artesano 468 sin antes presentar de forma clara cuál es la artesanía que produce. Este juego de dependencia del artesano y la artesanía en donde se institucionaliza la actividad artesanal puede ser el paso clave de su desarrollo como objeto de estudio cultural. sin antes presentar de forma clara cuál es la artesanía que produce. Este juego de dependencia del artesano y la artesanía en donde se institucionaliza la actividad artesanal puede ser el paso clave de su desarrollo como objeto de estudio cultural. sin antes presentar de forma clara cuál es la artesanía que produce. Este juego de dependencia del artesano y la artesanía en donde se institucionaliza la actividad artesanal puede ser el paso clave de su desarrollo como objeto de estudio cultural. sin antes presentar de forma clara cuál es la artesanía que produce. Este juego de dependencia del artesano y la artesanía en donde se institucionaliza la actividad artesanal puede ser el paso clave de su desarrollo como objeto de estudio cultural. La genealogía del término artesanía, que puede ser más extenso en su desarrollo práctico y científico (MEJÍA, 2004), pero corresponde en este caso a un uso específico de identificación cultural de las poblaciones rurales o periféricas. vista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. 4 VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las sponível em: http://seer.ufrgs.br/gearte , j transformaciones de la actividad sociocultural. Disponível em: http://seer.ufrgs.br/gearte La industria cultural y la Artesanía Periféricas en la producción, porque los medios continúan siendo la fuerza física o la herramienta simple, a diferencia del desarrollo técnico y tecnológico industrial. Periféricas en la sociedad, pues son las poblaciones no occidentales o pre- modernas las que constituyen el objeto de análisis y clasificación. Periféricas en los recursos, pues se trabaja sobre materiales como la madera, el carbón, el barro, el algodón, o las fibras naturales en general, frente a los metales y polímeros industrializados en la gran industria. Periféricas en la división sexual del trabajo, pues las mujeres van a tener una mayor participación en dicha producción, las actividades domésticas de reproducción social de la vida comunitaria van a integrar una forma de vida cultural y étnica definida en la producción de artesanía. Las mujeres van a sostener el proceso de organización del trabajo artesanal, pues son principalmente ellas las que van a producir un proceso de identificación y organización comunitaria como artesanas, artesanas que producen artesanías (VEGA, 2013), y es allí donde las unidades productivas que en el silo XIX tenían locales en las zonas urbanas, van a trasladarse a los hogares de las familias rurales donde la división sexual del trabajo, entre las labores de agricultura se entrelazan en obligaciones dominantes sobre la producción doméstica de las mujeres en su saber hacer propiamente dado para los útiles domésticos. En ese sentido, el mobiliario y utensilios domésticos que fueron organizando los mercados en los centros urbanos y en las plazas de las ciudades mediante la migración interna de la población rural, va a ser el paisaje principal de reconocimiento de la actividad artesanal a lo largo del siglo XX. Que se haya identificado la artesanía 469 con la producción doméstica femenina es un ejercicio de naturalización de la dominación masculina de la reproducción de la vida social. con la producción doméstica femenina es un ejercicio de naturalización de la dominación masculina de la reproducción de la vida social. con la producción doméstica femenina es un ejercicio de naturalización de la dominación masculina de la reproducción de la vida social. vista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. , j , transformaciones de la actividad sociocultural. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. sponível em: http://seer.ufrgs.br/gearte , j transformaciones de la actividad sociocultural. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. La industria cultural y la Artesanía Sin embargo, este proceso que inicia en la primera mitad del siglo XX como un mecanismo de reconocimiento, naturalización y clasificación, tiene su continuidad en el proceso de integración a la producción capitalista en la medida que se subordina este tipo de producción y se territorializa el mercado local como parte de la actividad del consumo turístico y cultural de la clase media y la clase trabajadora urbana. Parte del desarrollo de la industria cultural de la artesanía se integra indirectamente mediante el consumo en la sociedad capitalista y no en la transformación de los medios de producción. Este ejercicio de integración es lo que Néstor García Canclini consideraba como la “hibridación” de la cultura, lo que en sus bases materiales consistía en lo que Marx llamaba como la subordinación indirecta de la producción al capital. Este mecanismo de dominación de la cultura no corresponde directamente a una racionalidad instrumental de los productores, como mencionaba Adorno y Horkheimer, sino que es en el desarrollo desigual y combinado del capital en el mercado mundial donde se identifican las formas no occidentales de vida en comunidad, que terminan por ser parte del circuito total de acumulación y concentración del capital. Así, el modo de desarrollo fordista de la gran industria que constituye la forma de organización política, económica y social del trabajo a lo largo del siglo XX, presenta una integración indirecta de las formas dominadas de producción cultural rural mediante la posibilidad de consumo de la clase trabajadora o de los obreros urbanos. Sin este proceso de consumo, la producción artesanal como mercado local no tendría como reproducirse, lo que podría haber llevado a permanecer a la población de las comunidades rurales en una producción de subsistencia. Es precisamente ese paso de la producción de subsistencia a la producción de un mercado local dependiente de la sociedad capitalista industrial lo que define el espacio social donde se estructura el campo social de la actividad artesanal en el siglo XX. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. vista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. sponível em: http://seer.ufrgs.br/gearte EGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las Disponível em: http://seer.ufrgs.br/gearte transformaciones de la actividad sociocultural. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. La industria cultural y la Artesanía 470 Más reconocido el término, y más socializado desde la academia, es posible que la artesanía haya tenido ese crecimiento en los estudios que aparece en la figura 1, como mecanismo de reconocimiento y relación entre lo rural y lo urbano, tanto desde un proceso de intervención económica para el mejoramiento de las condiciones productivas de las artesanas, integradas en su contexto rural y su aplicación en la organización del mercado en la comercialización hacia las ciudades; como desde la crítica socioeconómica de la producción desde las desigualdades, la pobreza y la marginalidad en donde se desarrolla la actividad artesanal, especialmente en los países del denominado “tercer mundo”. Allí la formación cultural de las poblaciones como mecanismo etnológico de reconocimiento y auto-reconocimiento van a permitir que las formas de lucha de las comunidades tomen como recurso los símbolos, tradiciones y expresiones locales frente a la globalización. La institucionalización de la artesanía se constituye como un correlato al desarrollo capitalista, colonialista y patriarcal que constituye el fordismo. El Estado como forma de promoción y protección de las políticas públicas termina validando en un principio las condiciones folclóricas de la vida en lo local al ser patrimonio cultural, tanto como por incentivar procesos de reconocimiento formal de las actividades, empezando por la formación de leyes correspondientes a la planificación de la actividad en cada país. Este proceso que se consolida en la segunda mitad del siglo XX permite tanto naturalizar la actividad de la artesanía, como delimitar la participación de las categorías de artesanos en la población. Proceso que debido a la informalidad de la actividad, inscrita en la cotidianidad del trabajo flexible pago y no pago de las mujeres, sufrió siempre problemas en la consolidación de la identificación de las trabajadoras. No obstante, como ley, se ha consolidado en los países la necesidad de reglamentar la actividad, siendo la conceptualización el principal problema para la definición de quién es artesano y qué es artesanía. 471 Disponível em: http://seer.ufrgs.br/gearte Por otro lado, la identificación de las artesanas en las zonas rurales presenta problemas comunes de acceso y apropiación de los recursos que les permitan mayor justicia en el mercado. La estabilidad económica, la competencia con comerciantes, los bajos precios de venta y la baja calificación técnica, entre otras, son una constante en la población artesana de objetos utilitarios, folclóricos y culturales. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. transformaciones de la actividad sociocultural. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las sponível em: http://seer.ufrgs.br/gearte VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Disponível em: http://seer.ufrgs.br/gearte Disponível em: http://seer.ufrgs.br/gearte j transformaciones de la actividad sociocultural. La industria cultural y la Artesanía Esta manera limita la identidad de la población trabajadora, o que depende del trabajo a fragmentarse en asociaciones locales y regionales, pero que por la misma reproducción del objeto de su actividad cultural no consiguen superar su dependencia a la clase trabajadora formal de ingresos medio y alto. Puesto que además de luchar por la producción industrializada de baja calidad y de menor precio, de productores de otros países y regiones, deben competir por la diferenciación de productos entre los mismos productores para que exista una estabilidad financiera. Pese a estas situaciones, las propuestas de alternativas al desarrollo capitalista utilizan las representaciones de la artesanía como parte de la contra- globalización, en donde lo valores comunitarios y de cuidado son exaltados en las poblaciones marginadas y excluidas de la acumulación y concentración capitalista. La economía social y solidaria adhiere en sus principios una forma de producción que considere el trabajo artesanal como respuesta hacia el daño ecológico social provocado por la producción dominante capitalista (KELLER, 2011), y que contribuye a crear símbolos locales de resistencia mediante prácticas que consideran valores propios y valores de uso como el eje principal de la transformación productiva hacía otra sociedad. Sin embargo, este proceso basado en la artesanía puede tener problemas prácticos en la crítica al sistema capitalista, cuando la misma actividad artesanal puede encubrir valores excluyentes, discriminatorios y políticamente fragmentados. 472 Industrias creativas y actividad artística Si por un lado el fordismo representaba el modo de desarrollo de la sociedad industrial, por otro lado la crisis de los setenta y la reestructuración productiva que trajo consigo disminuyó la concentración de trabajadores en la grande industria, y por el contrario, existió un crecimiento del sector servicios, donde actividades de tipo administrativo y de gestión permitieron modificar la forma en que se organizaba el trabajo, lo que de manera correspondiente trajo consigo cambios en la manera como se observaba la producción artesanal. El pós-fordismo, como expresión de los cambios productivos en que se fundamentaba el capital, principalmente de un proceso de acumulación del capitalismo industrial a una sociedad de dominio del capital financiero permitía la descentralización y desterritorialización de la producción en los países industrializados (CUNNINGHAM, 2002), obteniendo mayor ganancia en la explotación laboral en diferentes países subdesarrollados. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. vista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las g p Disponível em: http://seer.ufrgs.br/gearte Disponível em: http://seer.ufrgs.br/gearte j transformaciones de la actividad sociocultural. , j transformaciones de la actividad sociocultural. sponível em: http://seer.ufrgs.br/gearte La industria cultural y la Artesanía Sin el desarrollo del diseño industrial aplicado a la producción artesanal no podría comprenderse la situación actual de la artesanía, pues es en la intervención de profesionales técnicos calificados como poco a poco se desarrolla un mercado de artesanías cada vez más especializado (QUIÑONES; BARRERA, 2006), donde la correspondencia entre las pequeñas empresas y emprendimientos urbanos y rurales van a obtener mayor representatividad en el mercado nacional e internacional. producción propiamente dicha, ya no podría dependerse del consumo de objetos tradicionales y folclóricos, sino que debería llevarse a cabo un desarrollo específico técnico y tecnológico de la producción que ofreciera un mejoramiento en los productos para competir de manera más rentable en el mercado. En este sentido aparece el diseño, y el diseño artesanal, como eje principal de configuración de un nuevo proceso de la actividad artesanal. Sin el desarrollo del diseño industrial aplicado a la producción artesanal no podría comprenderse la situación actual de la artesanía, pues es en la intervención de profesionales técnicos calificados como poco a poco se desarrolla un mercado de artesanías cada vez más especializado (QUIÑONES; BARRERA, 2006), donde la correspondencia entre las pequeñas empresas y emprendimientos urbanos y rurales van a obtener mayor representatividad en el mercado nacional e internacional. producción propiamente dicha, ya no podría dependerse del consumo de objetos tradicionales y folclóricos, sino que debería llevarse a cabo un desarrollo específico técnico y tecnológico de la producción que ofreciera un mejoramiento en los productos para competir de manera más rentable en el mercado. En este sentido aparece el diseño, y el diseño artesanal, como eje principal de configuración de un nuevo proceso de la actividad artesanal. Sin el desarrollo del diseño industrial aplicado a la producción artesanal no podría comprenderse la situación actual de la artesanía, pues es en la intervención de profesionales técnicos calificados como poco a poco se desarrolla un mercado de artesanías cada vez más especializado (QUIÑONES; BARRERA, 2006), donde la correspondencia entre las pequeñas empresas y emprendimientos urbanos y rurales van a obtener mayor representatividad en el mercado nacional e internacional. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. Disponível em: http://seer.ufrgs.br/gearte vista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. g p Disponível em: http://seer.ufrgs.br/gearte La industria cultural y la Artesanía Industrias creativas y actividad artística Industrias creativas y actividad artística Si por un lado el fordismo representaba el modo de desarrollo de la sociedad industrial, por otro lado la crisis de los setenta y la reestructuración productiva que trajo consigo disminuyó la concentración de trabajadores en la grande industria, y por el contrario, existió un crecimiento del sector servicios, donde actividades de tipo administrativo y de gestión permitieron modificar la forma en que se organizaba el trabajo, lo que de manera correspondiente trajo consigo cambios en la manera como se observaba la producción artesanal. El pós-fordismo, como expresión de los cambios productivos en que se fundamentaba el capital, principalmente de un proceso de acumulación del capitalismo industrial a una sociedad de dominio del capital financiero permitía la descentralización y desterritorialización de la producción en los países industrializados (CUNNINGHAM, 2002), obteniendo mayor ganancia en la explotación laboral en diferentes países subdesarrollados. Esto trajo consigo que la disminución de un trabajo formal en la sociedad industrial, y un crecimiento en la población laboral informal que tenía consigo el peso de una transformación productiva sin acceso directo a sus beneficios productivos. Con el desarrollo de políticas neoliberales desde los años setenta y ochenta, se concibe un mecanismo de individualización de la clase obrera, fragmentando las condiciones de reconocimiento en la promoción de competencia para el mercado, creando un discurso de éxito para los mejores, como esfuerzo voluntarista del empresario. Cada uno sería un empresario de sí, pues el recurso principal de explotación que es el trabajo sería ofertado como capital, capital humano, para obtener beneficios productivos en el mercado. El micro empresario transforma al pequeño propietario de la clase media tradicional en un emprendedor capitalista a pequeña escala. La artesanía como actividad cultural no corresponderá a esa protección como patrimonio de lo local desde el Estado, sino que por el contrario, en un proyecto neoliberal, tendrá que afrontar las transformaciones productivas desde la 473 producción propiamente dicha, ya no podría dependerse del consumo de objetos tradicionales y folclóricos, sino que debería llevarse a cabo un desarrollo específico técnico y tecnológico de la producción que ofreciera un mejoramiento en los productos para competir de manera más rentable en el mercado. En este sentido aparece el diseño, y el diseño artesanal, como eje principal de configuración de un nuevo proceso de la actividad artesanal. La industria cultural y la Artesanía Así, la participación de profesionales del diseño y del arte van a crear trayectorias de mercado de productos cada vez más diferenciados para una proyección de lujo y arte (AVELINO, 2007), donde las correspondencias entre los productos ya no van a desarrollarse hacía lo tradicional, hacia la memoria como un recurso patrimonial que debe protegerse y mantenerse, en esa ilusión mítica y estática de la artesanía que tanto impacto a principios del siglo XX, sino que debía individualizarse la producción, enseñar a producir con técnicas y valores estéticos y artísticos cada vez más autónomos, donde el artesano no se debía a su oficio a la manera tradicional, sino que los artesanos deberían producir distinciones como marca, es decir su nombre, su local, o su unidad productiva como signo distintivo que corresponde a la técnica, a la obra y al productor artesano-artista, en un régimen de identificación y representación que puede llegar a validarse en la legitimidad del campo artístico. En este espacio de transición del fordismo al pós-fordismo, las mujeres, los jóvenes y los adultos mayores como población excluida directamente del mercado laboral formal, atienden a la participación de sus actividades auxiliares en micro- emprendimientos de artesanía, como trabajadoras autónomas que pueden llegar 474 a trasladarse en ocupaciones transitorias entre la profesión, las artesanía o microempresa artesanal y el trabajo de cuidado (FIGUEIREDO, 2015). Estas manifestaciones promedian los ejemplos de la población que por algún motivo recurre a actividades de trabajo auxiliar o complementario, lo que incentiva un mercado de bienes de producción cada vez más especializado en las ciudades, que consiste en materia prima de industria nacionales e internacionales, lo que de alguna manera termina por subordinar directa y formalmente el trabajo de la artesanía al desarrollo capitalista. Por ello, el crecimiento de la producción académica en el área de administración y diseño ha sido constante conforme se requiere intervenir en el trabajo artesanal para una mayor productividad. a trasladarse en ocupaciones transitorias entre la profesión, las artesanía o microempresa artesanal y el trabajo de cuidado (FIGUEIREDO, 2015). Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. g p Disponível em: http://seer.ufrgs.br/gearte Disponível em: http://seer.ufrgs.br/gearte La industria cultural y la Artesanía Estas manifestaciones promedian los ejemplos de la población que por algún motivo recurre a actividades de trabajo auxiliar o complementario, lo que incentiva un mercado de bienes de producción cada vez más especializado en las ciudades, que consiste en materia prima de industria nacionales e internacionales, lo que de alguna manera termina por subordinar directa y formalmente el trabajo de la artesanía al desarrollo capitalista. Por ello, el crecimiento de la producción académica en el área de administración y diseño ha sido constante conforme se requiere intervenir en el trabajo artesanal para una mayor productividad. De manera correspondiente, si el modo de desarrollo fordista que considera condiciones de producción de la gran industria, en donde las industrias culturales eran subordinadas por mecanismos de una sociedad de consumo en las condiciones que el capitalismo industrial dominante permitía en la estructuración de un mercado nacional. El correlato del proceso pós-fordista de la reestructuración productiva va a ser la clasificación y organización de las industrias creativas, pues no va a ser lo “cultural”, entendido esto como valores simbólicos destinados al mercado de masas, sino la creatividad, entendida como el valor del productor en sí que se ofrece al mercado desde sus habilidades y condiciones simbólicas y corporales, lo que termina por dar acento al proceso de transformación neoliberal que va más allá de la simple reflexión de la palabra cultural y de la palabra creatividad. En este sentido, el paso de las industrias culturales a las industrias creativas como modo de desarrollo no es una etiqueta y sí un mecanismo de dominación de la producción y del productor en sí. Las industrias creativas que se estructuran como política de organización laboral en Inglaterra en los años noventa, y que toman fuerza en los países desarrollados en la primera década del siglo XXI (UNDP, 2013), hasta convertirse en política de clasificación mundial de la producción cultural, consideran el trabajo artesanal como parte de ese gran conglomerado de actividades creativas que 475 permiten evaluar un desarrollo de las habilidades de productores artísticos y culturales, que van desde las actividades de mayor desarrollo tecnológico y científico, como la producción cinematográfica y el desarrollo de tecnologías 4.0 como aplicaciones y formatos virtuales de comunicación, exposición y comercialización, hasta actividades de carácter más técnico como la artesanía, con avances de diseño en aplicación de materiales y transformación productiva material. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. g p Disponível em: http://seer.ufrgs.br/gearte vista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Disponível em: http://seer.ufrgs.br/gearte La industria cultural y la Artesanía En ese sentido la manera de proteger la actividad cultural es con la protección de derechos de producción desde los territorios, como formas de mantener una autonomía en la producción que salvaguarde la competencia en el mercado mundial. Sin embargo este tipo de desarrollo no escapa de la lógica de la acumulación del capital, puesto que la productividad no se mide propiamente en la transformación técnica y tecnológica de las unidades productivas, sino en la renta monopolista que busca ganancias indirectas en el valor agregado que obtienen los productos en el mercado internacional. denominaciones de origen (JURADO; AGUILAR; JUAJIBIOY, 2014). En ese sentido la manera de proteger la actividad cultural es con la protección de derechos de producción desde los territorios, como formas de mantener una autonomía en la producción que salvaguarde la competencia en el mercado mundial. Sin embargo este tipo de desarrollo no escapa de la lógica de la acumulación del capital, puesto que la productividad no se mide propiamente en la transformación técnica y tecnológica de las unidades productivas, sino en la renta monopolista que busca ganancias indirectas en el valor agregado que obtienen los productos en el mercado internacional. Así, tanto en la subordinación directa en la transformación técnica de los productos para un consumo de lujo, como de la subordinación indirecta en la territorialización de la producción mediante la renta monopolística, la actividad artesanal constituye hoy un escenario de debate entre una gran cantidad de personas que transitan entre las posibilidades de transformación en el mercado mundial, donde la precarización y la exclusión de los trabajos formales obliga a una participación cada vez menos valorizada en lo general, pues aún en el caso que exista un proceso de integración a la valorización estética como forma de legitimidad simbólica y material de las obras en el campo artístico, que se observa en el desarrollo cada vez más autónomo y de auto-reconocimiento de los artesanos como artistas, esto solo corresponde a un pequeño grupo de casos exitosos en la creciente población de trabajadores artesanales, artesanos y artesanas que aun trabajando en las condiciones desiguales de su local continúan con la actividad. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. Disponível em: http://seer.ufrgs.br/gearte La industria cultural y la Artesanía Este gran abanico de actividades laborales que constituyen las industrias creativas, da paso a un debate sobre el trabajo creativo dentro del proceso neoliberal de la globalización. El desarrollo de actividades culturales y creativas constituyen un objeto de estudio contemporáneo, en donde se busca poner en contexto la producción cultural en función de aquellos artesanos artistas y diseñadores que transformen lo tradicional en un medio de desarrollo estético y artístico que participe de exposiciones y mercados especializados en el consumo de lujo (FREITAG, 2014). En ese sentido, existe un proceso de homogeneización global del trabajo artesanal que corresponde con una transformación empresarial y artística de trabajadores autónomos que participan de un mercado global de exposición y reconocimiento. Los Estados participan de la promoción de la actividad artesanal en la forma de desarrollo artístico y competitivo mediante concursos y premios a los mejores productores, pero que en relación a políticas de fondo para el total de la población de artesanas y artesanos ha sido más bien débil, por el costo de asumir constantemente una actividad que transita entre la informalidad, la marginalidad y el trabajo ocasional (SCRASE, 2010). Por último, es posible identificar mecanismos de protección del trabajo artesanal en lo cultural desde un ejercicio de territorialización de la marca, es decir, mientras que en las zonas urbanas se obliga a los productores a obtener un reconocimiento individual como trabajadores autónomos mediantes las marcas o signos distintivos, en las zonas rurales donde la producción mantiene los saberes comunitarios se territorializa la marca con las indicaciones geográficas o las 476 denominaciones de origen (JURADO; AGUILAR; JUAJIBIOY, 2014). En ese sentido la manera de proteger la actividad cultural es con la protección de derechos de producción desde los territorios, como formas de mantener una autonomía en la producción que salvaguarde la competencia en el mercado mundial. Sin embargo este tipo de desarrollo no escapa de la lógica de la acumulación del capital, puesto que la productividad no se mide propiamente en la transformación técnica y tecnológica de las unidades productivas, sino en la renta monopolista que busca ganancias indirectas en el valor agregado que obtienen los productos en el mercado internacional. denominaciones de origen (JURADO; AGUILAR; JUAJIBIOY, 2014). Disponível em: http://seer.ufrgs.br/gearte Reflexiones sobre las categorías en lo artesanal Al tener en claro que solo desde las condiciones de la totalidad concreta del modo de producción capitalista se pueden observar las diferentes manifestaciones de lo artesanal, es posible llegar a algunas reflexiones sobre cómo crear categorías correspondientes a la actividad artesanal. Esto se realiza como crítica a las definiciones ligeras que se han realizado desde la conceptualización formal e ideal 477 de lo que debería ser la actividad artesanal, empezando por la reflexión transhistórica de Richard Sennet (2008), que contempla a lo artesanal como un trabajo “bien hecho”, esta vaguedad de la definición recuerda la crítica de Hobsbawm a Foucault quien decía de él que “rebuscaba en la historia para pescar ilustraciones al servicio de una narrativa preconstruida” (HOBSBAWM, 2016, p. 50), esto no implica que el ensayo sobre el artesano pierda importancia desde su filosofía pragmática en correspondencia a su “sistema” filosófico sobre el saber hacer, saber ser y saber-saber. Sin embargo, para él la artesanía solo sería un ejemplo más de ese hacer bien hecho, lo que elimina cualquier proceso real de comprensión e intervención política. de lo que debería ser la actividad artesanal, empezando por la reflexión transhistórica de Richard Sennet (2008), que contempla a lo artesanal como un trabajo “bien hecho”, esta vaguedad de la definición recuerda la crítica de Hobsbawm a Foucault quien decía de él que “rebuscaba en la historia para pescar ilustraciones al servicio de una narrativa preconstruida” (HOBSBAWM, 2016, p. 50), esto no implica que el ensayo sobre el artesano pierda importancia desde su filosofía pragmática en correspondencia a su “sistema” filosófico sobre el saber hacer, saber ser y saber-saber. Sin embargo, para él la artesanía solo sería un ejemplo más de ese hacer bien hecho, lo que elimina cualquier proceso real de comprensión e intervención política. Por otro lado, también se hace una crítica a las reflexiones latinoamericanas que busca poner límites desde los conceptos formales de lo que es o no es artesanía. Aún comprendiendo la necesidad de implementación de políticas públicas de la actividad artesanal, este esfuerzo ha perdido el horizonte promoviendo debates inacabados sobre la mejor forma de conceptualizar la actividad artesanal, partiendo siempre de una serie de generalizaciones que no constituyen un horizonte confiable de clasificación, pues en su mayoría solo reproduce lugares comunes de la artesanía como cultura naturalizada de la población indígena, afroamericana o “contemporánea”. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. Disponível em: http://seer.ufrgs.br/gearte VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las t f i d l ti id d i lt l vista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. , j transformaciones de la actividad sociocultural. Disponível em: http://seer.ufrgs.br/gearte Reflexiones sobre las categorías en lo artesanal La artesanía para estos documentos entonces es indígena, artesanía tradicional, artesanía contemporánea, neo-artesanía, industrianato, etc. Todas estas clasificaciones parten de esquemas conceptuales abstractos y ofrecen un resultado didáctico pero carente de crítica histórica. Teniendo en cuenta esto, la propuesta de categorías parte de una reflexión del trabajo concreto que constituye el mecanismo de interpretación general de la actividad humana y que permite entender el proceso de subordinación frente a un sistema histórico dominante de producción material. Entender que solo mediante el desarrollo de una división el trabajo desde el modo dominante de producción se 478 puede definir de manera más dinámica lo artesanal. A continuación se explica la actividad artesanal en relación a una división social, técnica y simbólica del trabajo. puede definir de manera más dinámica lo artesanal. A continuación se explica la actividad artesanal en relación a una división social, técnica y simbólica del trabajo. puede definir de manera más dinámica lo artesanal. A continuación se explica la actividad artesanal en relación a una división social, técnica y simbólica del trabajo. El trabajo artesanal se entiende como una categoría histórica y geográfica que representa una dimensión de la división del trabajo en una sociedad específica, y que constituye un medio de reconocimiento técnico, simbólico y social de organización de la producción, que como tal define su condición diferencial en relación a los desarrollos técnicos y tecnológicos en otras áreas de producción industrial y de servicios. En ese sentido representa formas de trabajo de la producción técnica básica o primaria, dada principalmente por la determinación negativa que se deriva de los avances tecnológicos de una sociedad. Por ello, hablar de trabajo artesanal puede ser relativo al estado de la producción técnica y tecnológica en un periodo histórico, y no puede pensarse como una categoría natural del saber-hacer humano. La actividad puede pensarse desde la división técnica del trabajo artesanal de una forma histórica de producción donde el trabajador y su producto comparten una serie de características técnicas simples o básicas, en donde se considera la habilidad individual y la herramienta -o maquinaria simple- como el apoyo principal en la producción, este tipo de organización depende del reconocimiento del oficio y de su tamaño, pues el ejercicio de la actividad debe mantener las proporciones ajustadas a un número que permita que el trabajador pueda encargarse de la mayor parte del proceso de producción. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. Disponível em: http://seer.ufrgs.br/gearte sponível em: http://seer.ufrgs.br/gearte g p Disponível em: http://seer.ufrgs.br/gearte , j transformaciones de la actividad sociocultural. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. Reflexiones sobre las categorías en lo artesanal En esto consiste el reconocimiento de actividades destinadas a la construcción, los alimentos, o la fabricación de productos utilitarios a base de recursos naturales como el cuero, la madera, el metal, entre otros. Esta definición del trabajo artesanal tiene mayor relación con la clasificación del a Organización Internacional del Trabajo, en dónde se presenta la técnica, la habilidad y la profesionalización como indicadores de reconocimiento. La división simbólica del trabajo artesanal constituye un proceso de reconocimiento cultural, en donde se encuentra en relación símbolos, textos, 479 contextos y significados en la producción, allí aparece la división principal de subjetividades entre la artesanía y el artesano como agente cultural. La artesanía no constituye sólo la división técnica, que presupone el trabajo artesanal, sino que además debe estar constituido por una serie de contenidos institucionalizados en la memoria y las tradiciones que correspondan con la identificación de valores históricos y culturales, lo que define a cabalidad la importancia del objeto artesanal como producto de la cultura, como parte del patrimonio de un lugar o de una nación. Ser artesano o artesana depende de la producción de dicho objeto cultural, donde se construye la narración del valor mediante su legitimidad en el campo social que lo constituye. Por eso, el artesano no depende únicamente de la técnica. Así, un trabajador artesanal como el albañil no va a ser fácilmente reconocido como artesano a menos que constituya junto con él una artesanía que corresponda simbólicamente con la legitimidad del campo como marco de representación válido. contextos y significados en la producción, allí aparece la división principal de subjetividades entre la artesanía y el artesano como agente cultural. La artesanía no constituye sólo la división técnica, que presupone el trabajo artesanal, sino que además debe estar constituido por una serie de contenidos institucionalizados en la memoria y las tradiciones que correspondan con la identificación de valores históricos y culturales, lo que define a cabalidad la importancia del objeto artesanal como producto de la cultura, como parte del patrimonio de un lugar o de una nación. Ser artesano o artesana depende de la producción de dicho objeto cultural, donde se construye la narración del valor mediante su legitimidad en el campo social que lo constituye. Por eso, el artesano no depende únicamente de la técnica. g p Disponível em: http://seer.ufrgs.br/gearte Disponível em: http://seer.ufrgs.br/gearte Reflexiones sobre las categorías en lo artesanal Así, un trabajador artesanal como el albañil no va a ser fácilmente reconocido como artesano a menos que constituya junto con él una artesanía que corresponda simbólicamente con la legitimidad del campo como marco de representación válido. Por último, la división social del trabajo artesanal exige la diferenciación de las posiciones de clase entre los productores, en ese sentido no todos los artesanos, trabajadores artesanales o productores en general pueden corresponder a posiciones de clase homogéneas Para ello, los recursos de distinción sobre el modo de producción dominante dan claridad a los procesos de distinción de clase. Por un lado se presentan los pequeños propietarios, posicionados en la clase media tradicional de productores que se reconocen en la consolidación de micro y pequeños emprendimientos, y que pueden constituir tanto los empresarios de oficios artesanales a pequeña escala, como productores artesanales de la artesanía que compiten en mercados para el posicionamiento de la marca y de su producto. Por otro lado se encuentran los trabajadores asalariados, que como trabajadores artesanales pueden depender de pequeñas empresas, siendo contratados en subordinación a los pequeños propietarios o empresarios, esto corresponde con la división de los trabajadores formalmente integrados a unidades productivas mediante la venta de su fuerza de trabajo (MAZAUD, 2009). Estos 480 casos de asalariamiento dependen principalmente de unidades productivas urbanas, y donde la clase de pequeños propietarios es más fuerte y reconocida en el mercado nacional. Por último las trabajadoras y trabajadores por cuenta propia, como agentes de producción que no se reconocen dentro de la clase media, sino que son trabajadores que transitan desde la producción precarizada hasta la formalizada en las zonas rurales y periurbanas. En otras palabras, es la clase-que- vive-del-trabajo y que corresponde con la población en situación más vulnerable para encontrar una continuidad y sostenibilidad en la producción artesanal. casos de asalariamiento dependen principalmente de unidades productivas urbanas, y donde la clase de pequeños propietarios es más fuerte y reconocida en el mercado nacional. Por último las trabajadoras y trabajadores por cuenta propia, como agentes de producción que no se reconocen dentro de la clase media, sino que son trabajadores que transitan desde la producción precarizada hasta la formalizada en las zonas rurales y periurbanas. En otras palabras, es la clase-que- vive-del-trabajo y que corresponde con la población en situación más vulnerable para encontrar una continuidad y sostenibilidad en la producción artesanal. Reflexiones sobre las categorías en lo artesanal Desde la división del trabajo sobre el desarrollo del capital, como forma dominante de producción, se puede obtener un ejercicio claro de las diferentes formas en que aparecen los trabajadores artesanales, en especial, entendiendo que existen diversas formas de clasificación que pueden coincidir en algunos casos, mientras que en otros son totalmente contradictorios. Esta dificultad no se resuelve únicamente con la interdisciplinariedad del análisis, sino con el desarrollo de una comprensión histórica concreta del proceso, que sea crítica en la reflexión de las subjetividades, de los discursos y del contenido ideológico que corresponde con la reproducción del sistema capitalista. De esta forma se pueden reconocer los aspectos de movilización que sean adecuados para la transformación política y económica de los trabajadores artesanales. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. vista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. Disponível em: http://seer.ufrgs.br/gearte EGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las Disponível em: http://seer.ufrgs.br/gearte VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. 481 Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. , j transformaciones de la actividad sociocultural. Reflexiones políticas La organización internacional del trabajo debe constituir un escenario de justicia y libertad para el desarrollo de las habilidades creativas y humanas, siendo esto lo que Marx podría entender como trabajo creativo fuera de las formas de dominación del capital (MARX, 2011). En ese sentido, cuando se expone la relación entre las habilidades autónomas de producción, no es posible identificar el peso culturalista que ha cargado consigo la clase trabajadora de artesanos, y que por el contrario ha sido una camisa de fuerza para organizar políticamente su protección y salvaguarda de la vida en comunidad. Salir de la subordinación directa 481 o indirecta del capital exige un reconocimiento crítico de la institucionalización de o indirecta del capital exige un reconocimiento crítico de la institucionalización de la artesanía como campo u objeto de conocimiento. No es posible reincidir en la fragmentación del trabajo si se consideran únicamente los valores culturales locales como eje de la organización política de los individuos. o indirecta del capital exige un reconocimiento crítico de la institucionalización de la artesanía como campo u objeto de conocimiento. No es posible reincidir en la fragmentación del trabajo si se consideran únicamente los valores culturales locales como eje de la organización política de los individuos. o indirecta del capital exige un reconocimiento crítico de la institucionalización de la artesanía como campo u objeto de conocimiento. No es posible reincidir en la fragmentación del trabajo si se consideran únicamente los valores culturales locales como eje de la organización política de los individuos. En ese sentido debe existir una reflexión amplia de las posibilidades de organización de los trabajadores desde sus condiciones de resistencia y colectividad en respuestas amplias e intersectoriales de organización para alcanzar un control político de sus condiciones de producción tanto desde el Estado como desde las bases sociales. Uno de los problemas principales es no reconocer este proceso amplio de participación de los trabajadores en definir políticas públicas sectoriales que se mantienen al margen de cambios estructurales significativos en la población, y que por el contrario, en regímenes de democracia subordinada al mercado, terminan por validar prácticas de protección y promoción de las mercancías por encima de la población vulnerable. vista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. j nsformaciones de la actividad sociocultural. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las g p sponível em: http://seer.ufrgs.br/gearte , j transformaciones de la actividad sociocultural. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Disponível em: http://seer.ufrgs.br/gearte Conclusiones La comprensión actual de la artesanía tiene que buscar superar la hegemonía cultural que pretende subordinar la expresión del saber hacer autónomo y comunitario a la acumulación del capital en escala internacional. La industrialización de la creatividad como mecanismo de transformación y clasificación de la producción actual que busca homogeneizar la diversidad local de los productores culturales debe tener un contrapeso en la reflexión crítica sobre la forma de intervenir políticamente para generar un bienestar a los individuos más allá de su dependencia a la producción de mercancías con valor cultural local. En ese sentido, comprender el desarrollo y génesis de la actividad artesanal requiere de su contextualización en el desarrollo desigual del capital. Por otro lado, aunque la producción académica sobre lo artesanal tenga un aumento significativo en el presente siglo, es necesario reconocer que las herramientas conceptuales sobre el trabajo artesanal tienen límites específicos en la reproducción ideológica dominante del capital de forma directa o indirecta, como tal toda intervención es una intervención política, y es desde allí como debe optarse por ser críticos con el desarrollo del trabajo artesanal. Reflexiones políticas De esta manera existe una doble relación entre el reconocimiento y fortalecimiento de la producción social alternativa al capitalismo, lo que se ha constituido desde los procesos de organización social y solidaria de los artesanos y artesanas, teniendo su auge principal en la organización sindicalizada y cooperativista de los trabajadores artesanales, llegando a reivindicar sus derechos como productores culturales. Sin embargo otra relación para generar esa integración a los procesos alternativos de producción debe consistir en el reconocimiento de la actividad artesanal como parte de una división social, técnica y simbólica del trabajo de forma internacional, lo que corresponde a un ejercicio de organización e identificación más allá de los límites de la profesión o del oficio, y más allá de las fronteras nacionales, pues es en la autonomía que se deriva de la afirmación de los artesanos y artesanas como trabajadores como podría conseguirse un proceso de integración de mayor fuerza en la protección y promoción de su bienestar. 482 Conclusiones Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. CAMPOS, A. Racialización, racialismo y racismo un discernimiento necesario. Universidad de La Habana, 273, p. 184-199. 2012. Disponible en: https://www.academia.edu/6283861/Raci alizaci%C3%B3n_Racialismo_y_Racismo._Un_discernimiento_necesario. Acceso en: 12 fev. 2020. Disponível em: http://seer.ufrgs.br/gearte AVELINO, C. O design, a arte e o artesanato deslocando o centro. Cadernos EBAPE, Rio de Janeiro, v. 5, n. 4, 2007. j transformaciones de la actividad sociocultural. CUNNINGHAM, S. From Cultural to Creative Industries: Theory, Industry and Policy Implications. Media International Australia, v. 102, n. 1, p. 54-65, 2002. sponível em: http://seer.ufrgs.br/gearte ELINO, C. O design, a arte e o artesanato deslocando o centro. Cadernos EBAPE, Rio de eiro, v. 5, n. 4, 2007. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las t f i d l ti id d i lt l Disponível em: http://seer.ufrgs.br/gearte Referências México: FCE, 1980. MAZAUD, C. Entre le métier et l’entreprise Renouvellement et transformations de l’artisanat français. 2009. Tese (Doutorado em Sociologia) - Université de Nantes, Nates, 2009. MEJÍA, D. La artesanía de México: historia, mutación y adaptación de un concepto. México: El Colegio de Michoacan. 2004. MORA DE JARAMILLO, Y. Clasificación y notas sobre técnicas y el desarrollo histórico de las artesanías colombianas. Revista Colombiana de Antropología, v. 16, p. 283-354, 1974. QUIÑONES, A. y BARRERA, G. Conspirando con los artesanos: Crítica y propuesta del Diseño en la artesanía. Bogotá: Editorial Pontifica Universidad Javeriana, 2006. SCRASE, T. Precarious production: Globalisation and artisan labour in the Third World. Third World Quarterly, v. 24, n. 3, p. 449-461, 2010. SENETT, R. The Craftman. London: Yale University Press, 2008. SENETT, R. The Craftman. London: Yale University Press, 2008. UNDP, United Nations Development Programme. Creative economy report. Special Edition. UNESCO, 2013. Disponible en: http://www.unesco.org/culture/pdf/creative-economy-report- 2013.pdf. Acceso en: 10 fev. 2020. VEGA, D. El campo artesanal: Aporte teórico social y pedagógico. Tunja: Fundación Universitaria Juan de Castellanos, 2013. VEGA, D. El campo artesanal: Aporte teórico social y pedagógico. Tunja: Fundación Universitaria Juan de Castellanos, 2013. ZARCA, B. L'artisanat. La plus populaire des classes moyennes? Vingtième Siècle, revue d'histoire, v. 37, p. 55-68, 1993. ZARCA, B. L'artisanat. La plus populaire des classes moyennes? Vingtième Siècle, revue d'histoire, v. 37, p. 55-68, 1993. Referências AVELINO, C. O design, a arte e o artesanato deslocando o centro. Cadernos EBAPE, Rio de Janeiro, v. 5, n. 4, 2007. AVELINO, C. O design, a arte e o artesanato deslocando o centro. Cadernos EBAPE, Rio de Janeiro, v. 5, n. 4, 2007. CAMPOS, A. Racialización, racialismo y racismo un discernimiento necesario. Universidad de La Habana, 273, p. 184-199. 2012. Disponible en: https://www.academia.edu/6283861/Raci alizaci%C3%B3n_Racialismo_y_Racismo._Un_discernimiento_necesario. Acceso en: 12 fev. 2020. CAMPOS, A. Racialización, racialismo y racismo un discernimiento necesario. Universidad de La Habana, 273, p. 184-199. 2012. Disponible en: https://www.academia.edu/6283861/Raci alizaci%C3%B3n_Racialismo_y_Racismo._Un_discernimiento_necesario. Acceso en: 12 fev. 2020. CUNNINGHAM, S. From Cultural to Creative Industries: Theory, Industry and Policy Implications. Media International Australia, v. 102, n. 1, p. 54-65, 2002. CUNNINGHAM, S. From Cultural to Creative Industries: Theory, Industry and Policy Implicatio Media International Australia, v. 102, n. 1, p. 54-65, 2002. ENGELS, F. Revolución y contrarrevolución en Alemania. In: MARX, K. e ENEGELS, F. Obras Escogidas Tomo I. Editorial Progreso, Moscou. 1980. Disponible en: https://webs.ucm.es/info/bas/es/marx-eng/52rca/index.htm. Acceso en: 19 enero 2020. FIGUEIREDO, M., et al. Empreendedorismo feminino no artesanato: Uma análise crítica do caso das rendeiras dos Morros da Mariana. RECADM, v. 14, n. 2, p. 110-123. 2015. FREITAG, V. Entre arte y artesanía: elementos para pensar el oficio artesanal en la actualidad. El Artista, v. 11, p. 129-143. 2014. 483 HOBSBAWM, E. Pierre Bourdieu, sociología crítica e historia social, New Left Review, v. 101, p. 41-52. 2016. Disponible en: https://newleftreview.es/issues/101/articles/eric-hobsbawm-pierre- HOBSBAWM, E. Pierre Bourdieu, sociología crítica e historia social, New Left Review, v. 101, p. 41-52. 2016. Disponible en: https://newleftreview.es/issues/101/articles/eric-hobsbawm-pierre- bourdieu.pdf. Acceso en: 4 enero 2020. HOBSBAWM, E. Pierre Bourdieu, sociología crítica e historia social, New Left Review, v. 101, p. 41-52. 2016. Disponible en: https://newleftreview.es/issues/101/articles/eric-hobsbawm-pierre- bourdieu.pdf. Acceso en: 4 enero 2020. JURADO, G. S. B., AGUILAR, A. C. Q., & JUAJIBIOY, J. C. J. Riesgos y tensiones de las marcas colectivas y denominaciones de origen de las creaciones colectivas artesanales indígenas. Apuntes. Revista de estudios sobre patrimonio cultural, v. 27, n. 1, 2014. KELLER, P. Trabalho artesanal e cooperado: realidades, mudanças e desafios. Sociedade e Cultura, Goiânia, v. 14, n. 1, p. 29-40, 2011. LOAIZA, G. Sociabilidad, religión y política en la definición de la nación: Colombia 1820-1886. Bogotá: Colección Bicentenario Universidad externado de Colombia, 2011. MARX, K. Grundrisse: Manuscritos econômicos de 1857-1858. São Paulo: Boitempo, 2011. MARX, K. Teorías sobre la plusvalía I, Tomo IV de El Capital. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Recebido em 31 de maio de 2019 Aceito em 28 de julho de 2020 VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Disponível em: http://seer.ufrgs.br/gearte VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. Daniel Roberto Vega Torres Possui graduação em Sociología — Universidad Nacional de Colombia — Bogotá (2010) e mestrado em Maestría en Historia — Universidad Pedagógica y Tecnológica de Colombia (2014). Doutor em Ciências Sociais Aplicadas da Universidade Estadual de Ponta Grossa e docente investigador — Fundación Universitaria Juan de Castellanos. Tem experiência na área de Sociologia da cultura, pesquisa em Educação. VEGA TORRES, Daniel Roberto. Trabajo artesanal, artesanía e industrias creativas: reflexiones en torno a las transformaciones de la actividad sociocultural. 484 Revista GEARTE, Porto Alegre, v. 7, n. 3, p. 461-484, set./dez. 2020. Disponível em: http://seer.ufrgs.br/gearte ORCID: https://orcid.org/0000-0001-9951-157X E-mail: dvega@jdc.edu.co Currículo: http://lattes.cnpq.br/8158197997704522 Recebido em 31 de maio de 2019 Aceito em 28 de julho de 2020 484
https://openalex.org/W2951125499	https://ejtcm.gumed.edu.pl/articles/22.pdf	English	null	A novel approach to visualization of the right ventricular outflow tract	European Journal of Translational and Clinical Medicine	2,019	cc-by-sa	3,535	Ewa Nowicka RESEARCH ARTICLE A novel approach to visualization of the right ventricular outflow tract EUROPEAN JOURNAL OF TRANSLATIONAL AND CLINICAL MEDICINE 2018;1(2):36-40 Agata Ewa Kaczyńska , Adam Kosiński , Rafał Kamiński , Miłosz Andrzej Zajączkowski , Marta Gleinert-Rożek , Department of Clinical Anatomy, Medical University of Gdańsk, Poland S R I E T Y V I O N F U G L D A A C I N D S E K M Ewa Nowicka RESEARCH ARTICLE A novel approach to visualization of the right ventricular outflow tract EUROPEAN JOURNAL OF TRANSLATIONAL AND CLINICAL MEDICINE 2018;1(2):36-40 Agata Ewa Kaczyńska , Adam Kosiński , Rafał Kamiński , Miłosz Andrzej Zajączkowski , Marta Gleinert-Rożek , Department of Clinical Anatomy, Medical University of Gdańsk, Poland S R I E T Y V I O N F U G L D A A C I N D S E K M S R I E T Y V I O N F U G L D A A C I N D S E K M EUROPEAN JOURNAL OF TRANSLATIONAL AND CLINICAL MEDICINE 2018;1(2):36-40 RESEARCH ARTICLE Citation Kaczyńska AE, Kosiński A, Kamiński R, Zajączkowski MA, Nowicka E, Gleinert-Rożek M. A novel approach to visu- alization of the right ventricular outflow tract. Eur J Transl Clin Med. 2018;1(2):36-40. DOI: 10.31373/ejtcm/100603 Kaczyńska AE, Kosiński A, Kamiński R, Zajączkowski MA, Nowicka E, Gleinert-Rożek M. A novel approach to visu- alization of the right ventricular outflow tract. Eur J Transl Clin Med. 2018;1(2):36-40. DOI: 10.31373/ejtcm/100603 Abstract Introduction: Nowadays, heart is one of the most demanding organs for imaging procedures. This is related to its irregular shape and complex internal structure. Increased demand for imaging complex cardiac structures has resulted in the development of novel 3D modeling techniques. Not only did the methods of imaging the organs of the living patients developed in recent years, but also new methods of post-mortem analysis. Acquired 3D models have a number of applications, both clinical and educational.f Detailed knowledge of the morphology of right ventricular outflow tract (RVOT) is extremely important in terms of cardio-invasive therapeutic procedures. Its significance was noticed during the exploration of the optimal pa- cing sites in the area of right ventricle. What is more, accurate analysis of the RVOT morphology and spatial struc- ture is also the basis for the treatment of ventricular arrhythmias which foci are located within the outflow tract. The aim of this study was to elaborate the most accurate technique of preparing interior models of the right ven- tricle and digitizing them to the 3D form. For this purpose we used a silicone molding of the heart cavities with digital photogrammetry.f ords: 3D models / cardiac imaging / silicone molding / photogrammetry / right ventricular outflow tract A novel approach to visualization of the right ventricular outflow tract Ewa Nowicka Agata Ewa Kaczyńska , Adam Kosiński , Rafał Kamiński , Miłosz Andrzej Zajączkowski , Marta Gleinert-Rożek , Introduction the increased interest in these techniques, in particu- lar in connection with cardiac visualization. Its imaging poses a great challenge, e.g. due to the unique func- tioning of this organ, the diversity of dimensions and The progress of technologies that enable the ima- ging of complex anatomical structures has resulted in Corresponding author: Agata Ewa Kaczyńska, Department of Clinical Anatomy, Medical University of Gdansk, Poland e-mail: agata.kaczynska@gumed.edu.pl No external funds. Available online: www.ejtcm.gumed.edu.pl Copyright ® Medical University of Gdańsk This is Open Access article distributed under the terms of the Creative Commons Attribution-ShareAlike 4.0 International. Material and methods shapes depending on the phase of the heart cycle. The visualization of the inside of the heart is extremely de- manding due to the complex internal structure of the atria and the chambers [1-3]. 30 hearts fixed in a formalin solution were used for the study. Hearts, without macroscopically visible pa- thological changes, belonged to adults of both sexes.i The first stage of the research was to construct ri- ght ventricle models using a silicone molding compo- und (Xiameter 4250 S Green). The applied technique has already been used in the case of modeling the heart's appendages [10]. Before modeling, the hearts were thoroughly rinsed with cold water to remove blood clots from the inside. The silicone was poured into the heart cavity from the apex of the right ventric- le, because filling RVOT with silicone was more advan- tageous when it was deposited inside the heart in a gravitational manner, for instance, with the apex po- inting upwards. For this purpose, the apex of the right and left ventricles was cut to pour the plastic mass (Fi- gure 1A). Next, the venous outlets of the right atrium were sutured and a catheter balloon was inserted into the pulmonary trunk above the level of the valve. The catheter balloon was filled with air so as not to distort the surface of the free wall of the RVOT and the pul- monary trunk (Figure 1B, Figure 1C). Thus prepared and dried hearts were placed in a glass vessel with the atria directed towards its bottom and with the apex to- wards the top (Figure 1D). After preparing the molding mass by adding an activator, the syringe was filled with silicone (Janet 100 ml), and then the mass was poured into the heart. The tip of the syringe was placed in the incised part of the right ventricle and the silicone was poured inside. Occasionally, the wall of the right ventricle and atrium was slightly pressed to remove air bubbles from among the trabeculations (Figure 1E). When the right part of the heart was completely filled with silicone, the heart was covered tightly with foil. The silicone-filled heart was left to harden for 24 hours. Material and methods After this time, the ventricular wall was incised along its edge, starting with the incision of the apex of the ri- ght ventricle to the base of the right appendage of the heart, including the wall of the right atrium (Figure 1F). After making the incision, the model of the interior of the heart was carefully removed and it was left to dry for 24 hours (Figure 2A, Figure 2B). Despite obtaining a large amount of the informa- tion about the internal structure of the heart, thanks to the use of commonly used imaging techniques, inc- luding magnetic resonance, computed tomography and echocardiography [4-6], there is still little atten- tion paid to the structure that is a part of the right ven- tricle - the outflow tract.f The right ventricular outflow tract (RVOT) is the space located between the supraventricular crest and the pulmonary valve. Exploration of the exact mor- phological architecture of this structure is extremely important in terms of the cardio-invasive therapeutic procedures. During the search for the optimal stimula- tion sites from the right ventricle area, it was proved that the most appropriate location for an electrode placement is the RVOT region. It was shown that the stimulation of this area is characterized by a fewer complications compared to the often used stimulation of the apex of the right ventricle [7-8]. Analysis of RVOT morphology is also extremely im- portant in the context of treatment of the ventricular arrhythmias, whose source is located within this area. Accurate assessment of the arrhythmia and the preci- se location of its foci allows to increase the efficiency of ablation and shorten its duration [9]. Knowledge of the RVOT morphology and its clo- sest neighborhood allows proper optimization of both the implantation of the stimulation system and the ablation of arrhythmogenic foci in this area, as well as limiting the risk of complications. For this purpo- se, it is extremely important to examine the internal structure of the right ventricle, taking into account the thickness of its muscular wall, the location of vessels and elements of the conductive system, as well as the position of heart valves and cavities in close proximity. Acquiring such information may appear crucial, due to the lack of precise literature data in the subject of the morphological structure of RVOT and its surroundin- gs. Corresponding author: 37 A novel approach to visualization... Material and methods Proposing an adequate anatomical nomenclature of this area and determining the exact boundaries of RVOT is a very important aspect of the research, which would significantly improve the work of clinicians.i The obtained 30 silicone models of the interior of the right heart, including the right atrium and the right ventricle, were subjected to the virtual visualization procedure to form the 3D model. A photogrammetry technique was used which uses serial images of the object in order to transform the physical model into virtual form. A scale was applied to every model, which allows for precise spatial measurements. The aim of this study was to develop an optimal method of obtaining a model of the right ventricle cavity and to transform its physical form into a virtual form that would give the possibility to perform signifi- cant measurements of the RVOT area. With the use of silicone modeling and photogram- metry techniques, 30 RVOT models were obtained. 38 Eur J Transl Clin Med 2018;1(2):36-40 Exemplary models and their modifications, as well as exemplary measurements that can be made on the models, are shown in Figure 3A ‒ Figure 3E. a number of applications in many areas of medicine. Computer simulations give the opportunity to explore the exact anatomy of the heart and help to understand the structural defects of this organ. What is more, the virtual heart model makes it possible to analyze some physiological aspects without carrying out complica- ted in vivo tests [11–13]. Discussion Intensive development of heart imaging tech- niques, including spatial modeling, allows its more accurate virtual reproduction. Such a model can be subjected to various computer analyses, e.g. specific measurements and simulations. In addition, the 3D model that has been subjected to the appropriate virtual processing can be successfully duplicated to the physical form thanks to the use of the 3D printing technology. Such created 3D heart models, both vir- tual and those obtained using silicone molding, have According to some reports 3D models of the hearts of patients with complex cardiac disease were con- structed in order to plan the optimal surgical approach [14]. Commonly used methods of cardiac imaging were used for this purpose: computed tomography and ma- gnetic resonance. Next, a 3D model was printed which was based on the virtual heart model that accurately reflected the patient's heart and defect. Such a model helped during the search for the most advantageous method of the resection of the aneurysm of the ven- tricle and the tumor in the right ventricle, as well as the reconstruction of the area after excision of the lesion. What is more, the surgeon was also able to es- timate the risk of complications of the planned surgery and to choose the technique with the least risk of failure. Figure 1 A-F. Stages of RVOT modeling with silicone mass. A – incision of the ventricles apex, B,C – placement of a cathe- ter in the pulmonary trunk, D – placement of the heart in the glass vessel, E – pouring the molding mass, F – incision of the ventricle and atrium in order to take out done model Figure 2 A, B. Finished model of the interior of right atrium and right ventricle A B Figure 2 A, B. Finished model of the interior of right atrium and right ventricle Figure 2 A, B. Finished model of the interior of right atrium and right ventricle A B A B C D E F A B C D E F A B C D E F A B A B A A A D C C D B E F E F B F F E 39 A novel approach to visualization... Discussion Figure 3 A-E Figure 3 A-E gu e 3 A – 3D model of the right ventricle and right atrium, B – view of the right ventricle and RVOT, C – separated view of RVOT and pulmonary valve, D – section through RVOT and exemplary measurement, E – exemplary measurement of RVOT A B C D E Measure Select two points Distance: 2.982 cm Measure Select two points Distance: 2.468 cm A Measure Select tw B D C B A E D E A – 3D model of the right ventricle and right atrium, B – view of the right ventricle and RVOT, C – separated view of RVOT and pulmonary valve, D – section through RVOT and exemplary measurement, E – exemplary measurement of RVOT and those with defects. Grabherr et al. compared the post-mortem imaging techniques, taking into account their advantages and disadvantages [18]. The photo- grammetry method was mentioned as one that does not require a large financial investment, in particular if other methods of post-mortem imaging of structures are not available, for example, scanners that use a beam of light. The authors also emphasized that in addition to its low cost, the most important advanta- ges of the photogrammetry are the simplicity of this method and the availability of various computer so- ftware that support this technique. In addition, a mo- del made with the use of the photogrammetry retains its natural color, which is extremely important in foren- sic medicine. Modeling the heart with the use of 3D technology is also widely used in the case of valve defects. In Vu- kicevic et al., a mitral valve replacement was perfor- med - specific for a given patient's case - using images obtained by transesophageal 3D echocardiography and computed tomography [15]. In Maragiannis et al., also on the basis of the images from computed tomo- graphy, a copy of the patient’s severely stenotic aor- tic valve was constructed [16]. It was also proven that such models are not only a source of information abo- ut the structure of the valve defect, but also retain the functional properties of a given defect. Such models are an ideal material for planning valve reconstruc- tion and for designing the most beneficial tools for in this type of surgery. Discussion This was also proven in the work of Schievano et al., where on the basis of the images obtained from magnetic resonance imaging, the RVOT model and the pulmonary trunk were made to analy- ze the possibility of percutaneous implantation of the pulmonary valve [17].i The procedure of the photogrammetry and the characteristics of its accuracy were described by Bob- kowska et al. [19]. That study showed that the silico- ne model of the right ventricular interior, which was virtualized using photogrammetry and then scaled, is suitable for analyzing the shape and size of the struc- ture. The use of imaging and visualization techniques has also become more important in these areas of me- dical sciences which deal with biometric analysis in the context of face modeling and its correct identification The wide range of applications of in vivo heart mo- deling with the use of imaging methods is an extremely important element of cardiac diagnostics. However, possibility of mapping post-mortem structures should be also mentioned, both in the case of normal hearts 40 Eur J Transl Clin Med 2018;1(2):36-40 The models of the interior of the heart provide an ideal source of information on the morphologi- cal structure and spatial architecture of the interior of the ventricles and atria. The combination of both techniques, molding the heart with silicone and pho- togrammetry are the original approach to obtaining in- formation about the internal structure of the right ven- tricle, in particular the RVOT, which still attracted too little attention. Due to existing inconsistencies in the morphology of this area, boundaries and spatial struc- ture, RVOT is a very interesting cardiac component for anatomical research. The results of such studies may be useful for many clinicians in their daily work. The use of 3D RVOT modeling will contribute to the acquisition of additional information about the mor- phology of this area and may be the basis for further research related to, for example, determining the loca- tion of other structures in the close vicinity of the RVOT. [20]. An accurate reflection of the dimensions and spa- tial structure of a given object is undoubtedly a great advantage of modern virtualization methods, thanks to which they are widely used, particularly in medical sciences. Conclusions Post-mortem heart modeling with the use of the molding mass and then transforming a given model into a virtual form combines methods that are prima- rily characterized by low cost and ease of performing a 3D heart model. The obtained virtual models are widely used because they can be used for measure- ment analysis and to modify a given spatial model. In addition, the acquired heart model can be reproduced using 3D printing. References 1. Kim MS, Hansgen AR, Wink O, Quaife RA, Carroll JD. Rapid prototyping: A new tool in understanding and treating structural heart dis- ease. Circulation. 2008;117(18):2388-94. 2. Greil GF, Wolf I, Kuettner A, Fenchel M, Miller S, Martirosian P, et al. Stereolithographic reproduction of complex cardiac m based on high spatial resolution imaging. Clin Res Cardiol. 2007;96(3):176-85. 3. Lopez-Perez A, Sebastian R, Ferrero JM. Three-dimensional cardiac computational modelling: methods, features and appl omed Eng Online. 2015;14(1):1-31. 4. Machaj I, Janczewska E, Truszewski Z, Trzebicki J, Gaciong Z. Nieinwazyjne badania obrazujące w kardiologii. Med Ogólna i Nauk o Zdrowiu. 2015;21(4):362-8. 4. Machaj I, Janczewska E, Truszewski Z, Trzebicki J, Gaciong Z. Nieinwazyjne badania obrazujące w kardiologii. Med Ogólna i Nauk o Zdrowiu. 2015;21(4):362-8. 5. Greil GF, Beerbaum P, Razavi R, Miller O. Imaging the right ventricle. Heart. 2008;94(6):803-8. 5. Greil GF, Beerbaum P, Razavi R, Miller O. Imaging the right ventricle. Heart. 2008;94(6):803-8. 6. Sonecki P. Nowe techniki w kardiologii: echokardiografia 3D. Kardiol po Dyplomie. 2010;9(11):36-45. 7. Ho SY, Nihoyannopoulos P. Anatomy, echocardiography, and normal right ventricular dimensions. Heart. 2006;92(SUPPL. 1):2-14. 7. Ho SY, Nihoyannopoulos P. Anatomy, echocardiography, and normal right ventricular dimensions. Heart. 2006;92(SUPPL. 1):2-14. ii 8. Lieberman R, Grenz D, Mond HG, Gammage MD. Selective site pacing: defining and reaching the selected site. Pacing Clin Electrophysiol. 2004;27(6 Pt 2):883-6. 8. Lieberman R, Grenz D, Mond HG, Gammage MD. Selective site pacing: defining and reaching the selected site. Pacing Clin Electrophysiol. 2004;27(6 Pt 2):883-6. 9. Pytkowski M, Maciąg A, Sterliński M, Jankowska A. Lokalizacja ogniska arytmii u chorych z zaburzeniami rytmu serca pochodzącymi z drogi odpływu prawej komory. Folia Cardiol Excerpta. 2006;1(4):211-20. t 10. Kamiński R, Kosiński A, Brala M, Piwko G, Lewicka E, Dąbrowska-Kugacka A, et al. Variability of the left atrial appendage in hum PLoS One. 2015;10(11):1-9. 11. Trunk P, Mocnik J, Trobec R, Gersak B. 3D heart model for computer simulations in cardiac surgery. Comput Biol Med. 2007;37(10):1398-403. ii 11. Conclusions Trunk P, Mocnik J, Trobec R, Gersak B. 3D heart model for computer simulations in cardiac surgery. Comput Biol Med. 2007;37(10):1398-403. 12. Dankowski R, Baszko A, Sutherland M, Firek L, Kałmucki P, Wróblewska K, et al. 3D heart model printing for preparation of percutaneous structural interventions: description of the technology and case report. Kardiol Pol. 2014;72(6):546-51. 12. Dankowski R, Baszko A, Sutherland M, Firek L, Kałmucki P, Wróblewska K, et al. 3D heart model printing for preparation of percutaneous structural interventions: description of the technology and case report. Kardiol Pol. 2014;72(6):546-51. 13. Rao AS, Menon PG. Presurgical planning using image-based in silico anatomical and functional characterization of Tetralogy of Fallot with associated anomalies. Interact Cardiovasc Thorac Surg. 2015;20(2):149-56. 13. Rao AS, Menon PG. Presurgical planning using image-based in silico anatomical and functional characterization of Tetralogy of Fallot with associated anomalies. Interact Cardiovasc Thorac Surg. 2015;20(2):149-56. 14. Jacobs S, Grunert R, Mohr FW, Falk V. 3D-Imaging of cardiac structures using 3D heart models for planning in heart surgery: a prelimi- nary study. Interact Cardiovasc Thorac Surg. 2008;7(1):6-9. 15. Vukicevic M, Mosadegh B, Min JK, Little SH. Cardiac 3D Printing and its Future Directions. J Am Coll Cardiol Img. 2017;10(2): tii 16. Maragiannis D, Jackson MS, Igo SR, Schutt RC, Connell P, Grande-Allen J, et al. Replicating Patient-Specific Severe Aortic Valve Stenosis with Functional 3D Modeling. Circ Cardiovasc Imaging. 2015;8(10):e003626. 17. Schievano S, Migliavacca F, Coats L, Khambadkone S, Carminati M, Wilson N, et al. Percutaneous pulmonary valve implantation based on rapid prototyping of right ventricular outflow tract and pulmonary trunk from MR data. Radiology. 2007;242(2):490-7. 17. Schievano S, Migliavacca F, Coats L, Khambadkone S, Carminati M, Wilson N, et al. Percutaneous pulmonary valve implantation based on rapid prototyping of right ventricular outflow tract and pulmonary trunk from MR data. Radiology. 2007;242(2):490-7. 18. Grabherr S, Egger C, Vilarino R, Campana L, Jotterand M, Dedouit F. Modern post-mortem imaging: an update on recent developments. Forensic Sci Res. 2017;2(2):52-64. 18. Grabherr S, Egger C, Vilarino R, Campana L, Jotterand M, Dedouit F. Modern post-mortem imaging: an update on recent developments. Forensic Sci Res. 2017;2(2):52-64. 19. Bobkowska K, Przyborski M, Kaczyńska A, Kosiński A. Digital Photogrammetry in the Analysis of the Ventricles’ Shape and Size. In: Pro- ceedings - 2017 Baltic Geodetic Congress (Geomatics), BGC Geomatics 2017. 2017. p. 169-73. 19. Bobkowska K, Przyborski M, Kaczyńska A, Kosiński A. Conclusions Digital Photogrammetry in the Analysis of the Ventricles’ Shape and Size. In: Pro- ceedings - 2017 Baltic Geodetic Congress (Geomatics), BGC Geomatics 2017. 2017. p. 169-73.
https://openalex.org/W2937715855	https://www.epj-conferences.org/10.1051/epjconf/201920503014/pdf	English	null	Spatially resolved coherent 2D fluorescence spectroscopy within a high-NA microscope	EPJ web of conferences	2,019	cc-by	1,801	Donghai Li1,, Matthias Nuss1 , Sebastian Goetz1 , Verena Kolb2 , Jens Pflaum2,3 , Chiara Trovatello4 , Giulio Cerullo4,5 , and Tobias Brixner1,6 Donghai Li1,, Matthias Nuss1 , Sebastian Goetz1 , Verena Kolb2 , Jens Pflaum2,3 , Chiara Trovatello4 , Giulio Cerullo4,5 , and Tobias Brixner1,6 1Institut für Physikalische und Theoretische Chemie, Universität Würzburg, Am Hubland, 97074 Würzburg, Germany 1Institut für Physikalische und Theoretische Chemie, Universität Würzburg, Am Hubland, 97074 Würzburg, Germany 2Experimental Physics VI, Universität Würzburg, Am Hubland, 97074 Würzburg, Germany 3Bavarian Center for Applied Energy Research, Magdalene-Schoch-Str. 3, 97074 Würzburg, Germany 2Experimental Physics VI, Universität Würzburg, Am Hubland, 97074 Würzburg, Germany 3Bavarian Center for Applied Energy Research, Magdalene-Schoch-Str. 3, 97074 Würzburg, Germany 4Dipartimento di Fisica, Politecnico di Milano, Piazza Leonardo da Vinci 32, 20133 Milano, Italy 5Istituto di Fotonica e Nanotecnologie, CNR, Piazza Leonardo da Vinci 32, 20133 Milano, Italy 6Center for Nanosystems Chemistry, Theodor-Boveri-Weg, 97074 Würzburg, Germany 4Dipartimento di Fisica, Politecnico di Milano, Piazza Leonardo da Vinci 32, 20133 Milano, Italy 5Istituto di Fotonica e Nanotecnologie, CNR, Piazza Leonardo da Vinci 32, 20133 Milano, Italy 6Center for Nanosystems Chemistry, Theodor-Boveri-Weg, 97074 Würzburg, Germany Abstract. We have developed coherent two-dimensional (2D) fluorescence micro-spectroscopy which probes the nonlinear optical response at surfaces via fluorescence detection with sub-micron spatial resolution. This enables the investigation of microscopic variations in laterally heterogeneous film samples which are of interests for sub- wavelength opto-electronic devices. 1 Introduction Coherent two-dimensional (2D) spectroscopy reveals the third-order nonlinear response of the system under investigation. Most geometries detect a coherently emitted four-wave- mixing signal following three-pulse excitation. However, when a fully collinear geometry is chosen [1, 2], incoherent population-based observables can also be utilized as has been demonstrated for fluorescence [1-3], mass spectroscopy [4], or electron currents [5]. Collinear geometry offers a possibility to couple the laser pulses into a high-NA fluorescence microscope. With such a method, a spatial resolution below half of the wavelength of the exciting electromagnetic field can be achieved according to the diffraction limit. In this work, we establish spatially resolved coherent 2D fluorescence microscopy and present two examples to demonstrate its applications. * Corresponding author: donghai.li@phys-chemie.uni-wuerzburg.de © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). , (201 E Web of Conferences PJ 9) 205 0 UP 2018 3014 , (201 E Web of Conferences PJ 9) 205 0 UP 2018 3014 https://doi.org/10.1051/e onf /201920 pjc 503014 2 Experimental methods The experimental setup is described in detail in a previous publication [6]. Basically, a femtosecond oscillator is used to provide broadband fs laser pulses with a spectrum ranging from 670 nm to 810 nm. A pulse shaper is used to generate pulse sequences with © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). , (201 E Web of Conferences PJ 9) 205 0 UP 2018 3014 https://doi.org/10.1051/e onf /201920 pjc 503014 controllable inter-pulse time delays and phases for the linear and 2D spectral measurements. The laser focus in the microscope is mapped by a piezo scanning stage. Sample fluorescence is detected by an avalanche photodiode (APD). Linear spectra as well as nonlinear 2D spectra can be obtained at different positions on the sample surface with spatial resolution of 260 nm. For linear spectra, a two-pulse delay scan is conducted, where fluorescence intensity is recorded as a function of time delay between the two pulses. Subsequent Fourier transformation results in a spectrum which is the product of the laser spectrum and the fluorescence excitation spectrum. So, by dividing by the laser spectrum, the spatially resolved excitation spectrum can be obtained. While in the 2D spectral measurements, relative time delays between four pulses are scanned, namely coherence time τ, the population time T, and signal time t. A 27-step phase-cycling scheme is used, allowing the extraction of the absorptive spectrum from the rephasing and non-rephasing nonlinear contributions. 3 Results and discussions As the first sample, a regularly structured sample surface made of fluorinated zinc phthalocyanine (F16ZnPc) is fabricated, the process is detailed in reference [6]. The absorption spectrum of a planar F16ZnPc film is fitted by multiple peaks as shown in Fig. 1(a) and indicates the absorption features that result from different coexisting molecular conformations upon preparation. By measuring linear spectra from an area of interest, a gradient spectral change across the molecular “hills” can be seen in Fig. 1(b). Typical linear spectra are shown in Figs. 1(c) and 1(d) for two exemplary locations. The two excitation spectra show differences in the ratio between intensities of high- and low-frequency components. 2D absorptive spectra are measured at the same positions and are shown below the corresponding linear spectra [Figs. 1(e) and 1(f)]. Along the diagonal, the observed difference in the ratio between the two components is similar to the linear spectra. Broadening in the anti-diagonal direction in the two 2D spectra indicates strong coupling between the high- and low-frequency components. between the high and low frequency components. Fig. 1. (a) Absorption spectrum (black solid line) of a planar F16ZnPc film, fitting curve (yellow dotted line) with multiple component peaks plotted in different colors. (b) Map of ratio between intensities at 700 and 770 nm in linear spectra. (c, d) Spatially resolved linear product spectra (red) and excitation spectra (black) from two exemplary positions of the sample. (e, f) Corresponding spatially resolved 2D absorptive spectra from the two exemplary positions. A d l ll f l i M S f ( b 10 g q y p y p Fig. 1. (a) Absorption spectrum (black solid line) of a planar F16ZnPc film, fitting curve (yellow Fig. 1. (a) Absorption spectrum (black solid line) of a planar F16ZnPc film, fitting curve (yellow dotted line) with multiple component peaks plotted in different colors. (b) Map of ratio between intensities at 700 and 770 nm in linear spectra. (c, d) Spatially resolved linear product spectra (red) and excitation spectra (black) from two exemplary positions of the sample. (e, f) Corresponding spatially resolved 2D absorptive spectra from the two exemplary positions. Fig. 1. (a) Absorption spectrum (black solid line) of a planar F16ZnPc film, fitting curve (yellow dotted line) with multiple component peaks plotted in different colors. (b) Map of ratio between intensities at 700 and 770 nm in linear spectra. 3 Results and discussions (c, d) Spatially resolved linear product spectra (red) and excitation spectra (black) from two exemplary positions of the sample. (e, f) Corresponding spatially resolved 2D absorptive spectra from the two exemplary positions. As a second sample, a small area of monolayer region on a MoSe2 surface (about 10 µm in diameter) is studied. The absorption spectrum of the MoSe2 monolayer has two peaks 2 2 , (201 E Web of Conferences PJ 9) 205 0 UP 2018 3014 https://doi.org/10.1051/e onf /201920 pjc 503014 corresponding to transition bands labeled A and B [7]. 2D fluorescence micro-spectra at different waiting times were measured as shown in Fig. 2(a). A strong diagonal peak of transition band A and two cross peaks corresponding to energy transfer from A to B and from B to A can be clearly seen. The cross-peak dynamics are extracted as the curves in Fig. 2(b). As waiting time increases, the B-to-A transfer peak rises. The fine structure of peak A changes with waiting time dramatically. Plotting the evolution of different regions within diagonal peak A as marked by green and purple dashed lines in Fig. 2(a), long-lived out-of- phase oscillations lasting up to 2 ps can be observed as shown in Fig. 2(c). corresponding to transition bands labeled A and B [7]. 2D fluorescence micro-spectra at different waiting times were measured as shown in Fig. 2(a). A strong diagonal peak of transition band A and two cross peaks corresponding to energy transfer from A to B and from B to A can be clearly seen. The cross-peak dynamics are extracted as the curves in Fig. 2(b). As waiting time increases, the B-to-A transfer peak rises. The fine structure of peak A changes with waiting time dramatically. Plotting the evolution of different regions within diagonal peak A as marked by green and purple dashed lines in Fig. 2(a), long-lived out-of- phase oscillations lasting up to 2 ps can be observed as shown in Fig. 2(c). p g p p g ( ) Fig. 2. (a) 2D absorptive spectra at different waiting times. In the map for T = 500fs, four regions of interest are marked with different colors. (b) Averaged intensity in the marked cross-peak areas as a function of waiting time (red: transfer from A to B, blue: transfer from B to A). 4 Conclusions In conclusion, we established the method of spatially resolved 2D fluorescence microscopy, where high temporal and spatial resolutions are achieved at the same time. Two samples were studied. The first one shows that 2D spectra can be measured from distinct small domains, and localized structural information can be extracted. The second study shows that ultrafast energy transfer dynamics within a microscopic target can be well detected by this method. 3 Results and discussions (c) Averaged intensity in the two marked regions (green and purple) within peak A as a function of waiting time. Fig. 2. (a) 2D absorptive spectra at different waiting times. In the map for T = 500fs, four regions of interest are marked with different colors. (b) Averaged intensity in the marked cross-peak areas as a function of waiting time (red: transfer from A to B, blue: transfer from B to A). (c) Averaged intensity in the two marked regions (green and purple) within peak A as a function of waiting time. References 1. P. Tian, D. Keusters, Y. Suzaki, and W. S. Warren, Science 300, 1553 (2003). 2. P. F. Tekavec, G. A. Lott, and A. H. Marcus, J. Chem. Phys. 127, 214307 (2007). 3. S. Draeger, S. Roeding, and T. Brixner, Opt. Express 25, 3259 (2017). 1. P. Tian, D. Keusters, Y. Suzaki, and W. S. Warren, Science 300, 1553 (2003). 2. P. F. Tekavec, G. A. Lott, and A. H. Marcus, J. Chem. Phys. 127, 214307 (2007) 3. S. Draeger, S. Roeding, and T. Brixner, Opt. Express 25, 3259 (2017). 4. S. Roeding and T. Brixner, Nat. Commun. 9, 259 (2018). 5. M. Aeschlimann, T. Brixner, A. Fischer, C. Kramer, P. Melchior, W. Pfeiffer, C. Schneider, C. Strüber, P. Tuchscherer, and D. V. Voronine, Science 333, 1723 (2011). 6. S. Goetz, D. Li, V. Kolb, J. Pflaum, and T. Brixner, Opt. Express 26, 3915 (2018). 7. Z. Nie, C. Trovatello, E. A. A. Pogna, S. Dal Conte, P. B. Miranda, E. Kelleher, C. Zhu, I. C. E. Turcu, Y. Xu, K. Liu, G. Cerullo, and F. Wang, Appl. Phys. Lett. 112, 031108 (2018) 3
https://openalex.org/W2026164108	https://research-repository.st-andrews.ac.uk/bitstream/10023/3004/1/Randorn2008IntJnlPhoto426872Synthesis.pdf	English	null	Synthesis of Visible-Light-Activated Yellow Amorphous<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mrow><mml:mtext>TiO</mml:mtext></mml:mrow><mml:mtext>2</mml:mtext></mml:msub></mml:mrow></mml:math>Photocatalyst	International journal of photoenergy	2,008	cc-by	5,460	Hindawi Publishing Corporation International Journal of Photoenergy Volume 2008, Article ID 426872, 6 pages doi:10.1155/2008/426872 Hindawi Publishing Corporation International Journal of Photoenergy Volume 2008, Article ID 426872, 6 pages doi:10.1155/2008/426872 Chamnan Randorn,1 John T. S. Irvine,1 and Peter Robertson2 Chamnan Randorn,1 John T. S. Irvine,1 and Peter Robertson2 1School of Chemistry, University of St. Andrews, St. Andrews, Fife KY16 9ST, UK 2Centre for Research in Energy and the Environment, School of Engineering, The Robert Gordon University, Schoolhill, Aberdeen AB10 1FR, UK Chamnan Randorn,1 John T. S. Irvine,1 and Peter Robertson2 1School of Chemistry, University of St. Andrews, St. Andrews, Fife KY16 9ST, UK 2Centre for Research in Energy and the Environment, School of Engineering, The Robert Gordon University, Schoolhill, Aberdeen AB10 1FR, UK 1School of Chemistry, University of St. Andrews, St. Andrews, Fife KY16 9ST, UK 2Centre for Research in Energy and the Environment, School of Engineering, The Robert Gordon University, Schoolhill, Aberdeen AB10 1FR, UK 1School of Chemistry, University of St. Andrews, St. Andrews, Fife KY16 9ST, UK 2Centre for Research in Energy and the Environment, School of Engineering, The Robert Gordon University, Schoolhill, Aberdeen AB10 1FR, UK Correspondence should be addressed to John T. S. Irvine, jtsi@st-and.ac.uk Received 10 March 2008; Accepted 22 April 2008 Received 10 March 2008; Accepted 22 April 2008 Recommended by Russell. Howe Recommended by Russell. Howe Visible-light-activated yellow amorphous TiO2 (yam-TiO2) was synthesised by a simple and organic-free precipitation method. TiN, an alternative precursor for TiO2 preparation, was dissolved in hydrogen peroxide under acidic condition (pH∼1) adjusted by nitric acid. The yellow precipitate was obtained after adjusting pH of the resultant red brown solution to 2 with NH4OH. The BET surface area of this sample was 261 m2/g. The visible light photoactivity was evaluated on the basis of the photobleaching of methylene blue (MB) in an aqueous solution by using a 250 W metal halide bulb equipped with UV cutoﬀﬁlter (λ > 420 nm) under aerobic conditions. Yam-TiO2 exhibits an interesting property of being both surface adsorbent and photoactive under visible light. It was assigned to the η2-peroxide, an active intermediate form of the addition of H2O2 into crystallined TiO2 photocatalyst. It can be concluded that an active intermediate form of titanium peroxo species in photocatalytic process can be synthesised and used as a visible-light-driven photocatalyst. Copyright © 2008 Chamnan Randorn et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 1. INTRODUCTION simple synthesis with elimination of the calcination step would mean lower costs for chemicals and energy consump- tion. There have been only a few works that studied or mentioned amorphous TiO2. Among them, it was found that commercial amorphous TiO2 had negligible photoactivity because it contains high concentrations of defects which cause rapid e−-h+ recombination [13]. However, synthetic amorphous TiO2 has been recently shown to exhibit sig- niﬁcantly photoactivity rates in aqueous methanol solutions under full solar spectrum (200 to >1000 nm) using a 400 W Xe arc lamp [14]. Titaniumdioxide(TiO2)has been the prevailing material in the ﬁelds of photocatalysis and solar energy conversion due to being chemically and biologically inert and environmen- tally friendly. Although TiO2 is now used in various practical applications, only a small UV fraction of solar light can be utilised because of its large band gap of 3.2 eV. The development of a titania-related photocatalyst that shows a high level of activity under visible light would therefore be a major advance. Yellow TiO2 prepared by nitrogen doping has been regarded as a visible-light-sensitive photocatalyst [1–4] as well as Cr or V doped TiO2 [5, 6]. Additionally, metal ion doped SrTiO3 [7–9], some metal sulﬁdes such as AgGaS2 and AgInZn7S9 [10], (Ga1−yZny)(N1−xOx) [11], and Sm2Ti2O7 [12] have been reported as visible-light-driven photocatalysts. There may be some disadvantages of doped materials, for instance, low surface area because of using high temperature, the requirement of the expensive chemicals and synthetic methods, and an increase in carrier recombination. Hydrogen peroxide (H2O2) has been reported as an important factor in the photocatalytic process because it has occurred as an intermediate in the mechanism of the photocatalytic process using TiO2 as photocatalysts, and it has been reported that the reactivity can be accelerated by the addition of H2O2 [15–17]. Moreover, titanium- peroxo species were found to be the active species in the titanium silicalite-1 (TS-1)/H2O2/H2O system for partial hydrocarbon oxidation [18, 19]. In general, titanium-peroxo species or aqueous peroxotitanate are yellow or yellow-green depending on pH. There have been some works that used Amorphous TiO2 is of interest for photocatalysis because of its high surface area and high adsorption. 2. EXPERIMENTAL Figure 1: XRD patterns of as-prepared yam-TiO2 (a) and after calcined at temperatures of (b) 400◦C, (c) 800◦C, (d) 850◦C, (e) 900◦C, (f) 1100◦C (∗: Vaseline speciﬁc peaks, A: anatase, R: rutile). 2.1. Materials Yam-TiO2 wassynthesised by a peroxide-based route using TiN as precursor. 50 mL of 30% H2O2 (Fisher Scientiﬁc, Leicestershire, UK) was added to 1.0 g of TiN (Alfa Aesar, Lancashire, UK) under acidic condition (pH ≤1) adjusted by HNO3 acid solution. After aging at room temperature- without any stirring for 24 hours, a clear red-brown solution was obtained. The obtained precursor solution was stable for several days under ambient atmosphere. Yellow precipitate is formed from this solution after adjusting the pH of the solution to 2 by slowly adding ammonia solution with the constant stirring. The precipitate was ﬁltered and washed with distilled water several times. It was then dried at room temperature. The dried powder was also calcined at a constant heating rate of 5◦C/min at various temperatures and held at these temperatures for 2 hours. diation, 2 mL of the mixture was collected and centrifuged at the irradiation time intervals (hourly). The photoactivity was examinedby monitoring the reduction of the absorbance at 665 nm. The spectrophotometric measurements were carried out using Perkin Elmer Lambda35 UV/Vis spectrometer. A ten-fold dilution was used for the high absorbance solutions. 1. INTRODUCTION Moreover, International Journal of Photoenergy 2 40 80 120 160 200 240 Relative intensity (%) 20 30 40 50 60 70 2Θ a b c d e f A R ∗∗ Figure 1: XRD patterns of as-prepared yam-TiO2 (a) and after calcined at temperatures of (b) 400◦C, (c) 800◦C, (d) 850◦C, (e) 900◦C, (f) 1100◦C (∗: Vaseline speciﬁc peaks, A: anatase, R: rutile). 40 80 120 160 200 240 Relative intensity (%) 20 30 40 50 60 70 2Θ a b c d e f A R ∗∗ titanium peroxo complex as a starting material for TiO2 preparation [20–22]. As mentioned in the above discussion, H2O2 is important in photocatalysis at TiO2, amorphous TiO2 oﬀers some enhancement of activity and noting the yellow colour of the catalytically active titanium-peroxo species in the TS- 1/H2O2/H2O system, it would be interesting to see if yellow amorphous titanium-peroxo species can act as a photocatalyst under visible light. In this work, therefore, we investigate the visible-light-driven photoactivity of yellow amorphous TiO2 (hereinafter referred to as “yam-TiO2”) prepared by a peroxide-based route, using titanium nitride (TiN) as a precursor. 3.1. Characterisation The XRD pattern of the as-prepared powder conﬁrmed the amorphous structure of yam-TiO2 as no diﬀraction peaks can be observed, Figure 1. Yam-TiO2 was quite stable under air atmosphere (the sample used throughout this work was kept for several months in the aerobic vial). However, colour and phase transformation of yellow amorphous to white anatase/rutile occurred on heat treatment. Anatase was presented after calcination at 400◦C and remained as a major phase until 900◦C. This indicated that a high temperature stable anatase phase can be produced by this method. Phase transformation from anatase to rutile can be observed at about 800–850◦C which was higher than TiO2 prepared by conventional methods. XRD data were collected with a Stoe Stadi-P Transmis- sion X-ray diﬀractometer, using Cu Kα1 radiation in the range 2θ = 20–80◦. The BET surface area measurement and pore analysis were carried out by nitrogen adsorption with use of Micromeritics ASAP 2020 V3.00 H surface area analyser. The measurement was carried out at liquid nitrogen temperature after degassing the powder sample at 120◦C. TEM imaging was obtained using a JEOL-JEM 2011 electron microscope. Diﬀuse reﬂectance spectroscopy was carried out with a Perkin Elmer Lambda35 UV/Vis spectrometer using BaSO4 as a reference. Furthermore, it was found that the particle sizes of TiO2 after calcination at 600 and 900◦C calculated by Scherrer equation were 31 and 51 nm, respectively. The obtained phase transformation temperature and particle size were consistent with the previous work that synthesised TiO2 by peroxide-based route and using H2TiO3 as precursor [20]. However, the stability of anatase phase in this work was higher than TiO2 synthesised by the peroxide-based route using TiCl3 as precursor [21]. The peroxide-based route, therefore, is interesting for TiO2 preparation because of the high temperature anatase-rutile phase transformation; however, it depends on precursor and synthesis condition. TiN is interesting to be used as an alternative precursor in a peroxide-based route because of its air and moisture stability, organic and chloride ion-free route, simplicity, and low cost. 2.2. Evaluation of photocatalytic activity Figure 4: The emitting wavelength of metal halide lamp (a) and UV-Vis absorption spectrum of MB (b). Figure 2: TEM image of as-prepared yam-TiO2. 0 10 20 30 40 50 60 70 80 90 100 R (%) 390 450 500 550 600 650 700 750 800 (nm) a b cerns about the bleaching of MB due to the photoreduction to colourless form and photoabsorption of MB [23–25]. There are two main forms of MB, the blue colour of oxidised form (MB) and the colourless reduced form (leuco form, LMB). MB can be photoreduced to LMB by TiO2 under UV light under anaerobic condition and in the presence of a sacriﬁcial electron acceptor (SED). It was found that MB itself can act as an SED. However, this process is reversed on the addition of O2 to the anaerobic system. Although LMB is readily oxidised back to MB by oxygen, this reaction depends on pH. The rate increases with pH, LMB is moderately stable under acidic condition but reacts rapidly with air under basic condition [23]: Figure 3: The UV-Vis diﬀuse reﬂectance of (a) yam-TiO2, (b) commercial rutile TiO2. MB + SED TiO2+UV −−−−−→LMB + SED2+; 2LMB + O2 −→2MB + 2H2O. (1) Accordingto our focus on evaluation of the visible light photoactivity of yellow amorphous TiO2, TEM, BET and diﬀuse reﬂectance were only examined with the as-prepared powder. Figure 2 shows the TEM image of the as-prepared powder. The ultraﬁne particles with amorphous structure can be seen. As a result, a high BET surface area of 261 m2/g was obtained. The diﬀuse reﬂectance spectrum shows the red shift of the absorption edge into the visible region (Figure 3(a)) compared with commercial rutile TiO2 (TIOXIDE) (Figure 3(b)). (1) As continuous air bubbling of a neutral solution was utilised in this work and no reducing agent was added to the system, it is likely that the dominant bleaching process is the photomineralisation of MB rather than the reductive formation of LMB. This system, MB/TiO2/air-saturated water system, has widely been utilised for the demonstration of semiconductor photocatalysis, as reviewed by Mills and Wang [23]. It is possible or even probable that partial rather than complete mineralisation into CO2, NH4+, NO3 −, and SO42 is occurring as the main observable in bleaching MB. 2.2. Evaluation of photocatalytic activity Visible light activity was evaluated on the basis of the decomposition of methylene blue (MB) in an aqueous solution. The sample powder was suspended in 200 mL of an 1 × 10−4 M MB solution by air bubbling. This mixture was ﬁrst suspended in the dark for 2 hours to reach the adsorption equilibrium before irradiation with a high intensity discharge 250 W iron doped metal halide UV bulb (UV Light Technology Ltd., Bermingham, UK) equipped with UV cutoﬀﬁlter, λ ≥420 nm (Borosilicate Coated Glass HM07, UQG(optic)Ltd., Cambridge UK). The mixture temperature was controlled at about 25◦C using a water bath for infrared radiation and lamp heating removal. After irra- 3 3 Chamnan Randorn et al. 20 nm Figure 2: TEM image of as-prepared yam-TiO2. 0 10 20 30 40 50 60 70 80 90 100 R (%) 390 450 500 550 600 650 700 750 800 (nm) a b Figure 3: The UV-Vis diﬀuse reﬂectance of (a) yam-TiO2, (b) commercial rutile TiO2. 300 400 500 600 700 (nm) a b Figure 4: The emitting wavelength of metal halide lamp (a) and UV-Vis absorption spectrum of MB (b). 20 nm Figure 2: TEM image of as-prepared yam-TiO2. 0 10 20 30 40 50 60 70 80 90 100 R (%) 390 450 500 550 600 650 700 750 800 (nm) a b Figure 3: The UV-Vis diﬀuse reﬂectance of (a) yam-TiO2, (b) commercial rutile TiO2. 300 400 500 600 700 (nm) a b Figure 4: The emitting wavelength of metal halide lamp (a) and UV-Vis absorption spectrum of MB (b). cerns about the bleaching of MB due to the photoreduction to colourless form and photoabsorption of MB [23–25] There are two main forms of MB, the blue colour of oxidised form (MB) and the colourless reduced form (leuco form LMB). MB can be photoreduced to LMB by TiO2 under UV light under anaerobic condition and in the presence of a sacriﬁcial electron acceptor (SED). It was found that MB itself can act as an SED. However, this process is reversed on the addition of O2 to the anaerobic system. Although LMB is readily oxidised back to MB by oxygen, this reaction depends on pH. The rate increases with pH, LMB is moderately stable under acidic condition but reacts rapidly with air under basic condition [23]: 20 nm Figure 2: TEM image of as-prepared yam-TiO2. 2.2. Evaluation of photocatalytic activity g Yam-TiO2 is attributed to the η2-peroxide (η2-TiOOH) because of the similar red shift and that yam-TiO2 trans- formed to white powder on heating similar to the η2-titani- um peroxospecies in the TS-1/H2O2/H2O system reported by Bonino et al. [18]. Lin and Frei [19] found that the η2- peroxide species obtained upon loading H2O2 into TS-1 was photodissociated eﬃciently under irradiation with visible or near UV light. Moreover, η2-peroxide was assigned to the adsorption structure formed by the addition of H2O2 on the surface of rutile TiO2. This structure was preferable to pro- duce hydroxyl radical that accelerates the photoactivity [17]. The photoabsorption of MB can be an issue for visible- light-driven photocatalyst investigation because MB can absorb visible light, especially, in the range 600–700 nm. Yan et al. [25] concluded that MB was not an appropriate substrate for a visible-light photocatalytic activity test par- ticularly in the range of 540–680 nm irradiation. However, the most important and intense emitting wavelength of the metal halide lamp used in this work is in the range of about 200–580 nm [26]. After equipping with a UV cutoﬀﬁlter, the emitting wavelength is in the range of 420–580, which has little overlap with the absorption of MB as shown in Figure 4. 3.2. Visible light photocatalytic activity The photobleaching of MB has been widely studied in heterogeneous photocatalysis. However, there are some con- International Journal of Photoenergy 4 0 10 20 30 40 50 60 70 80 90 100 Decolourisation (%) a b c d Light 5 h Dark 2 h Figure 5: % decolourisation of MB under the condition of dark 2 hours and light 5 hours of (a) MB, (b) MB + 0.5 g/L byam-TiO2, (c) MB + 0.5 g/L yam-TiO2, (d) MB + 0.2 g/L yam-TiO2. 0 10 20 30 40 50 60 70 80 90 100 Decolourisation (%) a b c d As a result, the photoabsorption of MB should not signif- icantly eﬀect the determination of visible light photocatalytic activity test in this work. y Figure 5 shows the photocatalytic degradation of MB comparing between amorphous TiO2 prepared in acidic condition (yam-TiO2) and in basic condition (labeled byam- TiO2), including blanktest (2 × 10−5 M MB). It was found that 10% bleaching of MB was shown after visible light irra- diation of only MB for 5 hours (Figure 5(a)), whereas slightly higher, 15% photobleaching of MB, was obtained by using 0.5 g/L byam-TiO2 as photocatalyst (Figure 5(b)). However, the latter amount was obtained after 45% of 1 × 10−4 M MB was adsorbed, that indicated that the initial concentration of MB before photobleaching was about 5.5 × 10−5 M. Com- parision with using 0.5 g/L yam-TiO2, about 97% adsorption and 2% photocatalytic activity was measured (Figure 5(c)). It can be concluded that the adsorption capacity of yam-TiO2 was signiﬁcantly higher than byam-TiO2; however, the extent of adsorption must hinder the hetereogeneous photoactivity. Therefore, a decreased amount of yam-TiO2 to 0.2 g/L was performed for photoactivity comparison (Figure 5(d)). It can be seen that about 46% photobleaching of MB after adsorption was obtained, which was considerably higher than using byam-TiO2 in spite of using less amount and higher concentration remained after adsorption. Moreover, only 9% photobleaching was observed for 0.5 g/L of Degussa P25 in 1 × 10−5 MB under visible light for 2 hours (compared to 93% under UV for 1 hour) showing that the yam-TiO2 was much more eﬀective, ∼40% under visible light for 2 hours. An important point to note is that the adsorbed MB seemed much less intense in colour and indeed the colour had changed to pale purple/violet. 3.2. Visible light photocatalytic activity This could indicate an oxidative adsorption process in the dark with the oxidised form, MB•+ being formed and adsorbed onto the yam-TiO2; however, conﬁrmation requires further study. 0 10 20 30 40 50 60 70 80 Decolourisation (%) a b c d Light 5 h Dark 2 h Figure 5: % decolourisation of MB under the condition of dark 2 hours and light 5 hours of (a) MB, (b) MB + 0.5 g/L byam-TiO2, (c) MB + 0.5 g/L yam-TiO2, (d) MB + 0.2 g/L yam-TiO2. 0.001 0.05 0.1 0.15 0.2 0.25 0.3 0.347 A 200 300 400 500 600 700 800 (nm) Figure 6: UV-Vis absorption spectra of MB. 0 10 20 30 40 50 60 70 80 90 100 R (%) 390 450 500 550 600 650 700 750 800 (nm) a b Figure 7: The UV-Vis diﬀuse reﬂectance of (a) yam-TiO2, (b) byam-TiO2. Figure 5: % decolourisation of MB under the condition of dark 2 hours and light 5 hours of (a) MB, (b) MB + 0.5 g/L byam-TiO2, (c) MB + 0.5 g/L yam-TiO2, (d) MB + 0.2 g/L yam-TiO2. 0.001 0.05 0.1 0.15 0.2 0.25 0.3 0.347 A 200 300 400 500 600 700 800 (nm) Figure 6: UV-Vis absorption spectra of MB. 0 10 20 30 40 50 60 70 80 90 100 R (%) 390 450 500 550 600 650 700 750 800 (nm) a b Figure 7: The UV-Vis diﬀuse reﬂectance of (a) yam-TiO2, (b) byam-TiO2. 0.001 0.05 0.1 0.15 0.2 0.25 0.3 0.347 A 200 300 400 500 600 700 800 (nm) Figure 6: UV-Vis absorption spectra of MB. 0.001 0.05 0.1 0.15 0.2 0.25 0.3 0.347 A 200 300 400 500 600 700 800 (nm) Figure 6: UV-Vis absorption spectra of MB. The irreversibility of the photobleaching process on continuous air bubbling for 24 hours in the dark after complete decolourisation of MB and the disappearance of the bands associated with MB (294 nm and 665 nm) with no appearance of band associated with LMB (256 nm) [24] as shown in Figure 6. were observed, further conﬁrming that photoreduction of MB to LMB was not the dominant process. 0 10 20 30 40 50 60 70 80 90 100 R (%) 390 450 500 550 600 650 700 750 800 (nm) a b Figure 7: The UV-Vis diﬀuse reﬂectance of (a) yam-TiO2, (b) byam-TiO2. 3.2. Visible light photocatalytic activity The ﬁrst way, to completely remove adsorbed MB on the surface by using photoactivity itself until the original colour of yam-TiO2 was recovered. Due to a large amount of MB being adsorbed on the surface, 48 hours UV irradiation was used for adsorbed MB removal, after colourless solution was obtained by visible light photoac- tivity test, in order to reduce the duration of adsorbed MB removal and to evaluate the stability of yam-TiO2 under UV light simultaneously. Figure 8 shows the recyclability by using UV irradiation for adsorbed MB removal. The 0.5 g/L ratio of catalyst to volume of MB was chosen because of the incomplete MB adsorption, hence, the visible-light photoac- tivity can be monitored, and a quantity of catalyst will be suﬃcient for removal of the high amount of adsorbed MB in 48 hours. About 97% adsorption and 2% photoactivity in 5 hours were obtained in the preliminary photoactivity test (Figure 8(a)). After the mixture was irradiated with UV light for 48 hours, the violet powder of visible light irradiated/adsorbed MB was changed to yellow powder as the original colour with a small amount of violet specks. Some violet specks indicated incomplete removal of adsorbed MB. Thereafter, the colourless aqueous solution was removed from the mixture by rinsing and the remaining powder left to dry at room temperature. 0 Dark 2 h Light 1 h Light 2 h Light 3 h Light 4 h Light 5 h Figure 8: Recyclability with removal of adsorbed MB (a) MB + 0.5 g/L yam-TiO2, (b) 1st reuse, (c) 2nd reuse, (d) only photoactivity of 1st reuse, (e) only photoactivity of 2nd reuse. Figure 8: Recyclability with removal of adsorbed MB (a) MB + 0.5 g/L yam-TiO2, (b) 1st reuse, (c) 2nd reuse, (d) only photoactivity of 1st reuse, (e) only photoactivity of 2nd reuse. y The ﬁrst reuse was performed by readding 200 mL of 1 × 10−4 M MB. A slight decrease of adsorption and photoactivity was obtained for the ﬁrst and second recycles (Figure 8(b), (c)). A small decrease of adsorption capac- ity and photoactivity presumably caused by incomplete adsorbed MB removal and decrease of eﬃciency or stability of amorphous structure after prolonged UV irradiation, which the colour changing from yellow to pale yellow can be observed by the naked eye (XRD data showed that it remained amorphous structure). 3.2. Visible light photocatalytic activity However, if considering only photoactivity by using MB concentration after adsorp- tion, about 2 × 10−5 M estimated from the percentage of bleaching, as an initial concentration, about 80% bleaching of MB under visible light irradiation can be reached in 5 hours both in the ﬁrst and second recycles (Figure 8(d), (e)). 0 10 20 30 40 50 60 70 80 90 100 Decolourisation (%) Dark 2 h Light 1 h Light 2 h Light 3 h Light 4 h Light 24 h a b c d e Figure 9: Recyclability without removal of adsorbed MB (a) MB + 0.5 g/L yam-TiO2, (b) 1st reuse, (c) 2nd reuse, (d) only photoactivity of 1st reuse, (e) only photoactivity of 2nd reuse. 0 10 20 30 40 50 60 70 80 90 100 Decolourisation (%) Dark 2 h Light 1 h Light 2 h Light 3 h Light 4 h Light 24 h Decolourisation (%) The second way of recyclability testing was performed by readding 200 mL 1 × 10−4 M MB without removal of adsorbed MB. The results are shown in Figure 9. The adsorp- tion capacity signiﬁcantly decreased in the ﬁrst and second reuses as expected, 75% and 22%, respectively (Figure 9(b), (c)), due to MB was adsorbed 97% of 1 × 10−4 M MB in the preliminary photoactivity test before recycling (Figure 9(a)). The photobleaching of MB in the ﬁrst reuse, including adsorption and photoactivity, was 85% in 4 hours and 99% in 24 hours (Figure 9(b)), while 51% and 95%, respectively, in the second reuse (Figure 9(c)). However, considering only the photoactivity, the percentage of bleaching of MB under visible light irradiation can be about 40% in 5 hours and more than 90% in 24 hours both in the ﬁrst and second recycles (Figure 9(d), (e)). These results show the reusability of yam-TiO2. 0 Dark 2 h Light 1 h Light 2 h Light 3 h Light 4 h Light 24 h Figure 9: Recyclability without removal of adsorbed MB (a) MB + 0.5 g/L yam-TiO2, (b) 1st reuse, (c) 2nd reuse, (d) only photoactivity of 1st reuse, (e) only photoactivity of 2nd reuse. light-driven photocatalyst even on prolonged irradiation, although the reaction rate decreased. The color change probably due to some η2-peroxide being converted to a hydroxide form upon photo irradiation [17, 19]. 3.2. Visible light photocatalytic activity 0 10 20 30 40 50 60 70 80 90 100 R (%) 390 450 500 550 600 650 700 750 800 (nm) a b These results show that yam-TiO2 can act as a visible- light-driven photocatalyst better than byam-TiO2, in parallel with a red shift of the absorption edge into the visible region as shown in Figure 7. However, an increase in visible absorption edge does not guarantee consistency in visible light photoactivity, the recombination and surface area are also important factors. Figure 7: The UV-Vis diﬀuse reﬂectance of (a) yam-TiO2, (b) byam-TiO2. Recyclability and stability of photocatalyst are important factors for any practical applications. Therefore, we have attempted to investigate cyclability and stability. In general, the photoactivity of crystalline TiO2 is observed by monitoring the degradation of MB in an aqueous solution, and adsorbed MB is not an issue probably due to small amount of MB that was adsorbed on crystallined TiO2. The recyclability of crystallined TiO2 normally performed by elimination of adsorbed MB on TiO2 by oxidation at temperatures higher than about 200◦C, which is the decomposition temperature of MB. However, removal of adsorbed MB by calcination cannot be used in the case of amorphous form because crystallisation will be induced. As a result, the cyclability of yam-TiO2 was measured by two diﬀerent methods, with and without removal of at temperatures higher than about 200◦C, which is the decomposition temperature of MB. However, removal of adsorbed MB by calcination cannot be used in the case of amorphous form because crystallisation will be induced. As a result, the cyclability of yam-TiO2 was measured by two diﬀerent methods, with and without removal of Chamnan Randorn et al. 5 0 10 20 30 40 50 60 70 80 90 100 Decolourisation (%) Dark 2 h Light 1 h Light 2 h Light 3 h Light 4 h Light 5 h a b c d e Figure 8: Recyclability with removal of adsorbed MB (a) MB + 0.5 g/L yam-TiO2, (b) 1st reuse, (c) 2nd reuse, (d) only photoactivity of 1st reuse, (e) only photoactivity of 2nd reuse. 0 10 20 30 40 50 60 70 80 90 100 Decolourisation (%) Dark 2 h Light 1 h Light 2 h Light 3 h Light 4 h Light 5 h adsorbed MB. REFERENCES [17] T. Hirakawa, K. Yawata, and Y. Nosaka, “Photocatalytic reactivity for O·− 2 and OH· radical formation in anatase and rutile TiO2 suspension as the eﬀect of H2O2 addition,” Applied Catalysis A, vol. 325, no. 1, pp. 105–111, 2007. [1] T. Sano, N. Negishi, K. Koike, K. Takeuchi, and S. Matsuzawa, “Preparation of a visible light-responsive photocatalyst from a complex of Ti4+ with a nitrogen-containing ligand,” Journal of Materials Chemistry, vol. 14, no. 3, pp. 380–384, 2004. [18] F. Bonino, A. Damin, G. Ricchiardi, et al., “Ti-peroxo species in the TS-1/H2O2/H2O system,” Journal of Physical Chemistry B, vol. 108, no. 11, pp. 3573–3583, 2004. [2] T. Ihara, M. Miyoshi, Y. Iriyama, O. Matsumoto, and S. Sugihara, “Visible-light-active titanium oxide photocatalyst realized by an oxygen-deﬁcient structure and by nitrogen doping,” Applied Catalysis B, vol. 42, no. 4, pp. 403–409, 2003. [19] W. Lin and H. Frei, “Photochemical and FT-IR probing of the active site of hydrogen peroxide in Ti silicalite sieve,” Journal of the American Chemical Society, vol. 124, no. 31, pp. 9292– 9298, 2002. [3] C. Shifu, L. Xuqiang, L. Yunzhang, and C. Gengyu, “The preparation of nitrogen-doped TiO2−xNx photocatalyst coated on hollow glass microbeads,” Applied Surface Science, vol. 253, no. 6, pp. 3077–3082, 2007. [20] Y. Gao, Y. Masuda, W.-S. Seo, H. Ohta, and K. Koumoto, “TiO2 nanoparticles prepared using an aqueous peroxotitanate solu- tion,” Ceramics International, vol. 30, no. 7, pp. 1365–1368, 2004. [4] T. Matsumoto, N. Iyi, Y. Kaneko, et al., “High visible-light photocatalytic activity of nitrogen-doped titania prepared from layered titania/isostearate nanocomposite,” Catalysis Today, vol. 120, no. 2, pp. 226–232, 2007. [21] K. Melghit, S. S. Al-Rabaniah, and I. Al-Amri, “Low tem- perature preparation and characterization of nanospherical anatase TiO2 and its photocatalytic activity on Congo red degradation under sunlight,” Ceramics International, vol. 34, no. 3, pp. 479–483, 2008. [5] J. C.-S. Wu and C.-H. Chen, “A visible-light response vanadium-doped titania nanocatalyst by sol-gel method,” Journal of Photochemistry and Photobiology A, vol. 163, no. 3, pp. 509–515, 2004. [22] E. Kanezaki, T. Sakamoto, A. Ookubo, and K. Ooi, “Solid-state chemistry of thermally induced yellow colouring in synthetic hydrous titanium oxide from TiCl3,” Journal of the Chemical Society, Faraday Transactions, vol. 88, no. 24, pp. 3583–3586, 1992. [6] H. Kato and A. Kudo, “Visible-light-response and photocat- alytic activities of TiO2 and SrTiO3 photocatalysts codoped with antimony and chromium,” Journal of Physical Chemistry B, vol. 106, no. 3.2. Visible light photocatalytic activity The intense yellow can be recovered after further addition of H2O2 as well known for crystalline TiO2. For this reason, yam-TiO2 can be reused several times with addition of H2O2. As mentioned above about the gradual colour changing of yam-TiO2 from yellow to pale yellow after irradiation, it was found that it remained pale yellow and can be a visible- International Journal of Photoenergy 6 4. CONCLUSION [13] B. Ohtani, Y. Ogawa, and S. Nishimoto, “Photocatalytic activity of amorphous-anatase mixture of titanium(IV) oxide particles suspended in aqueous solutions,” Journal of Physical Chemistry B, vol. 101, no. 19, pp. 3746–3752, 1997. An active intermediate form of an addition of H2O2 into crystallined TiO2 photocatalyst can be synthesised by peroxide-based route. It can be used as a visible-light-driven photocatalyst itself with the high surface area of amorphous form. [14] Z. Zhang and P. A. Maggard, “Investigation of photocatalytically-active hydrated forms of amorphous titania, TiO2·nH2O,” Journal of Photochemistry and Photobiology A, vol. 186, no. 1, pp. 8–13, 2007. [15] M. R. Hoﬀmann, S. T. Martin, W. Choi, and D. W. Bahnemann, “Environmental applications of semiconductor photocatalysis,” Chemical Reviews, vol. 95, no. 1, pp. 69–96, 1995. ACKNOWLEDGMENT The authors would like to thank the Royal Thai Government for Chamnan Randorn’s scholarship and EPSRC. The authors would like to thank the Royal Thai Government for Chamnan Randorn’s scholarship and EPSRC. [16] A. Houas, H. Lachheb, M. Ksibi, E. Elaloui, C. Guillard, and J.-M. Herrmann, “Photocatalytic degradation pathway of methylene blue in water,” Applied Catalysis B, vol. 31, no. 2, pp. 145–157, 2001. REFERENCES 19, pp. 5029–5034, 2002. [23] A. Mills and J. Wang, “Photobleaching of methylene blue sensitised by TiO2: an ambiguous system?” Journal of Photo- chemistry and Photobiology A, vol. 127, no. 1–3, pp. 123–134, 1999. [7] R. Konta, T. Ishii, H. Kato, and A. Kudo, “Photocatalytic activities of noble metal ion doped SrTiO3 under visible light irradiation,” Journal of Physical Chemistry B, vol. 108, no. 26, pp. 8992–8995, 2004. [24] S.-K. Lee and A. Mills, “Novel photochemistry of leuco- Methylene blue,” Chemical Communications, no. 18, pp. 2366– 2367, 2003. [8] S. Otsuka-Yao-Matsuo and M. Ueda, “Visible light-induced photobleaching of methylene blue aqueous solution using (Sr1−xLax)TiO3+δ−TiO2 composite powder,” Journal of Photo- chemistry and Photobiology A, vol. 168, no. 1-2, pp. 1–6, 2004. [25] X. Yan, T. Ohno, K. Nishijima, R. Abe, and B. Ohtani, “Is methylene blue an appropriate substrate for a photocatalytic activity test? A study with visible-light responsive titania,” Chemical Physics Letters, vol. 429, no. 4–6, pp. 606–610, 2006. [9] S. Otsuka-Yao-Matsuo, T. Omata, and M. Yoshimura, “Pho- tocatalytic behavior of cerium titanates, CeTiO4 and CeTi2O6 and their composite powders with SrTiO3,” Journal of Alloys and Compounds, vol. 376, no. 1-2, pp. 262–267, 2004. [26] M. Cristea and B. Laﬁtte, “On the emitted spectrum chang- ing in metal-halide lamps,” Journal of Optoelectronics and Advanced Materials, vol. 6, no. 1, pp. 87–94, 2004. [10] A. Kudo, “Development of photocatalyst materials for water splitting,” International Journal of Hydrogen Energy, vol. 31, no. 2, pp. 197–202, 2006. [11] M. Yashima, K. Maeda, K. Teramura, T. Takata, and K. Domen, “Crystal structure and optical properties of (Ga1−xZnx)(N1−xOx) oxynitride photocatalyst (x = 0.13),” Chemical Physics Letters, vol. 416, no. 4–6, pp. 225–228, 2005. [12] M. Uno, A. Kosuga, M. Okui, K. Horisaka, and S. Yamanaka, “Photoelectrochemical study of lanthanide titanium oxides, Ln2Ti2O7 (Ln = La, Sm, and Gd),” Journal of Alloys and Compounds, vol. 400, no. 1-2, pp. 270–275, 2005.
W2136278133.txt	https://healtheconomicsreview.biomedcentral.com/counter/pdf/10.1186/s13561-014-0007-5	en		Risk adjustment in aging societies	Health economics review	2,014	cc-by	11,060	von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 RESEARCH Open Access Risk adjustment in aging societies Viktor von Wyl1,2* and Konstantin Beck1,3 Abstract Background: In Switzerland, age is the predominant driver of solidarity transfers in risk adjustment (RA). Concerns have been voiced regarding growing imbalances in cost sharing between young and old insured due to demographic changes (larger fraction of elderly >65 years and rise in average age). Particularly young adults aged 19–25 with limited incomes have to shoulder increasing solidarity burdens. Between 1996 and 2011, monthly intergenerational solidarity payments for young adults have doubled from CHF 87 to CHF 182, which corresponds to the highest absolute transfer increase of all age groups. Results: By constructing models for age-specific RA growth and for calculating the lifetime sum of RA transfers we investigated the causes and consequences of demographic changes on RA payments. The models suggest that the main driver for RA increases in the past was below average health care expenditure (HCE) growth in young adults, which was only half as high (average 2% per year) compared with older adults (average 4% per year). Shifts in age group distributions were only accountable for 2% of the CHF 95 rise in RA payments. Despite rising risk adjustment debts for young insured the balance of lifetime transfers remains positive as long as HCE growth rates are greater than the discount rate used in this model (3%). Moreover, the life-cycle model predicts that the lifetime rate of return on RA payments may even be further increased by demographic changes. Nevertheless, continued growth of RA contributions may overwhelm vulnerable age groups such as young adults. We therefore propose methods to limit the burden of social health insurance for specific age groups (e.g. young adults in Switzerland) by capping solidarity payments. Conclusions: Taken together, our mathematical modelling framework helps to gain a better understanding of how demographic changes interact with risk adjustment and how redistribution of funds between age groups can be controlled without inducing further selection incentives. Those methods can help to construct more equitable systems of health financing in light of population aging. Keywords: Risk adjustment; Demography; Health insurance; Intergenerational solidarity JEL codes: I13; J11 Background Societies in highly industrialized countries in Western Europe, North America or Japan have undergone profound demographic changes over the past decades. Lifeexpectancy has increased substantially, owing to reductions of mortality, better life-styles and greater medical possibilities [1,2]. It is estimated that the average lifespan in OECD countries rose by more than 6 years between 1970 and 2000 [3]. Although this growing life expectancy is * Correspondence: viktor.vonwyl@css-institut.ch 1 CSS-Institute for Empirical Health Economics, Tribschenstrasse 21, 6002 Luzern, Switzerland 2 Institute for Social and Preventive Medicine, University of Bern, Finkenhubelweg 11, 3012 Bern, Switzerland Full list of author information is available at the end of the article commonly perceived as positive, it puts strains on the welfare systems of industrialized societies. For example, the ratio of retired individuals to active workers is increasingly shifting towards the elderly, and fewer active workers have to support more retired persons [4]. Apart from pension systems, health insurance systems are also affected by those demographic trends. The impact of aging societies – increasing average age and a rising share of elderly (>65 years) – on health care expenditures (HCE) has long been recognized. For example, recent projections of Swiss health care costs adjusted for expected demographic changes predict substantial overall increases in HCE owing to a higher proportion of elderly in the population and higher cost growth for older insured [5]. As for the latter, the higher © 2014 von Wyl and Beck; licensee Springer. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 cost increase for older age groups is well documented by long-term observations. For example, Mendelson and Schwarz analyzed Health Care Financing Administration Data from 1977 through 1987 and noticed a disproportionally high cost growth among the elderly aged 65 and more [6], later termed “steepening” by Buchner & Wasem [7]. Such longitudinal analyses of age-stratified costprofiles were also performed for Switzerland and reached similar results [8,9]. However, the reasons for the accelerated cost growth among the elderly are still debated. In particular, the effects of age, medical progress, and interactions thereof are not fully understood, and their disentanglement in statistical models is very challenging. One prominent explanation termed “red herring hypothesis” was put forth by Zweifel et al. [10,11], which states that the cost increases are not linked to age per se but rather to proximity of death (which is more likely at older age), although opposing studies found significant age effects on HCE increases [12-14]. Despite this unresolved debate, there is unanimous agreement that health care expenditures will rise further in the future, which in turn will have implications for health insurance and premium financing. In settings with competing health insurers and community-rated premiums (e.g. Belgium, The Netherlands, and Switzerland), many of the effects of population aging on health financing are mediated through risk adjustment (RA). Risk adjustment (or risk equalization) is a necessary means to prevent risk selection because individual health care expenditures can vary greatly and in part even predictably, whereas health insurance premiums do not [15]. Therefore, premiums systematically do not match costs for certain age groups (e.g. elderly), which leads to incentives for “cream skimming” and discrimination of insured. Risk selection is also harmful from a societal perspective, because it can create losses in welfare and efficiency [15]. Risk adjustment reduces incentives for risk selection. It operates by estimating the difference between groupspecific average health care expenditures and the overall average. The difference between these two amounts is then taxed from groups with below average costs and passed on as a subsidy to groups with greater than average costs. Thus, risk adjustment should equal out risk differences in portfolios of insurers to eliminate “cream skimming”. In general, risk adjustment leads to re-distribution of money from younger, healthier individuals to older, sicker insured, thereby establishing an intergenerational solidarity because age is one of the main drivers of risk adjustment transfers (especially in the context of the Swiss risk adjustment formula). Hence, if the share of elderly in a population grows over time (and thus average health care expenditures increase), this means that Page 2 of 14 younger, healthier individuals have to contribute more to risk adjustment in order to achieve risk equalization. Solidarity across age groups is also established by other transfer schemes in mandatory health insurance. In particular, young adults also benefit from tax-financed premium subsidies. But quantitatively risk adjustment is by far the largest solidarity component in mandatory health insurance: In 2012 young adults contributed CHF 1.37 billion to risk adjustment, but received only CHF 0.52 billion in premium subsidies [16]. As a consequence risk adjustment should play a key role in any attempt to re-distribute the burden of rising HCE in aging societies. Nevertheless, the knowledge on interactions between population aging and risk adjustment is still partial. What is more, frameworks for the implementation of fair (as defined normatively by society) and stable intergenerational solidarity transfers within risk adjustment are, to our knowledge, still lacking. In this paper, we aim to address two questions. First, we aim to investigate how the Swiss risk adjustment scheme (or any scheme) responds to population aging. We will analyze possible effects both from a cross-sectional (i.e. different age groups at single time-points) and from a lifetime perspective (i.e. following an age-cohort of insured over time). Second, we aim to seek ways how solidarity enforced by risk adjustment can be maintained in long term without financially overwhelming especially vulnerable age groups. The remainder of this paper is structured as follows. First follows a brief explanation of the Swiss health care setting. Second, to gain a better understanding of the processes leading to an increasing premium burden for young insured we develop a simple model of risk adjustment payments over time from the perspective of young adults, in which we include variables for health care growth and demographic changes. Although the model is generic and can accommodate any age splits we will center these calculations around young adults aged 19 to 25 years for reasons that will be explained in the methods section. The third section outlines the construction of a mathematical model to assess the balance of risk adjustment over a lifecycle. By use of those models from the second and the third section we assess the importance of demographic change for the increase of solidarity transfers from young to old using data from the Swiss risk adjustment statistics. Moreover, we will sketch out ideas on how to reduce and stabilize the levels of risk adjustment payments for specific age groups, again using the young adults as an example. The results section describes applications of the mathematical models within stochastic simulations and tests different reform suggestions for how they reduce the premium burden for 19 to 25 year old insured. The paper concludes with a discussion of the findings. von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 Methods Setting The Swiss system of social health insurance is influenced by Enthoven’s concept of managed competition [17]. A large number of health insurers (124 in 1996; 62 in 2011) compete for customers and are obliged to accept any person willing to enroll, independent of age or health status. Mandatory health insurance is organized on a pay-as-you-go basis. Benefit packages are strictly defined and comprehensive. Insurance is not linked to employment, and each insured has to pay the premiums directly to the insurer, although state-funded premium subsidies are granted to individuals in need on the basis of taxable income. Currently 30% of all insured receive such financial assistance [16]. Premiums for mandatory insurance are charged as age-independent community rates with two exceptions. Children between 0 to 18 years of age are granted riskrated premiums with regard to age on a mandatory basis. Moreover, the health insurance law states that insurance companies can grant premium reductions to young adults aged between 19 to 25 years. An internal risk adjustment scheme (i.e. without supplemental funds from the government), defined by retrospective redistribution of premiums across sex and 15 age groups, was introduced in 1993 and left unchanged until 2011 [18]. Starting in 2012, the Swiss risk adjustment was reformed to be based on prospective payments and to additionally include prior hospitalization as a crude morbidity indicator [19]. Further reform steps are currently discussed in the Swiss parliament and will likely include the introduction of pharmaceutical cost groups. Variable notations for modeling analyses In the following, we will develop a simple model with only two groups of insured, which we term young adults and adults. We are using the following notation: C = Average health care expenditures x = Number of individuals p = Proportion in the general population of insured older than 18 years a = Risk adjustment payment Y = Indicator for young adults (e.g. 19–25 year olds) A = Indicator for adults (e.g. >25 year olds) i = Indicator variable for the 30 risk groups (ordered by age) that were included in the Swiss risk adjustment scheme until 2011 (15 age groups, male/female). The indicator i = 1,..,k corresponds to risk groups for young adults with the cut-off denoted by k, and the remaining k + 1 to 30 risk groups represent adults. Definition of risk adjustment equations We define average health care expenditures for young adults Y, adults A and overall as Page 3 of 14 Xk Y ¼ Xi¼1 C k C i xi x i¼1 i X30 A ¼ Xi¼kþ1 C 30 C i xi x i¼kþ1 i X30 C i xi and C ¼ Xi¼1 30 x i¼1 i ð1Þ The proportion of young adults and of adults is defined, Xk X30 x = x , and pA = 1 − pY. respectively, by pY ¼ i i¼1 i¼1 i Thus, we can also define average health care expenditures as C A ¼ pY C Y þ 1−pY C C ð2Þ Risk adjustment transfers for young adults and adults can then be written as H− C aH ¼ C ð3Þ for H = Y or H = A. We will use equation (3) as starting point for the development of a model of changes in the amount of risk adjustment transfers for young adults. Cross-sectional analysis of impact of demographic changes on risk adjustment transfers For this analysis, we consider the intergenerational part of risk adjustment transfers defined in equation (3). Given equation (2), we can write the intergenerational transfer per person for young adults at time point t = 0 as A Y− C ¼C Y − pY C Y − 1−pY C aY0 ¼ C − ð4Þ − Note that because C Y < C risk adjustment transfers for young adults are also negative (aY0 < 0), meaning that they have to make payments into the fund. For time points t > 0 the intergenerational part of risk adjustment transfers becomes Y ð1 þ ΔY Þt Y ð1 þ ΔY Þt − pY ð1 þ d Y Þt C aYt ¼ C A ð1 þ ΔA Þt ; − 1 − pY ð1 þ d Y Þt C ð5Þ whereby (1 + ΔY)t and (1 + ΔA)t stand for average health care expenditure (HCE) growth for young adults and adults with average growth rates of ΔY and ΔA, respectively. The expression (1 + dY)t denotes average changes in the fraction of young adults in the population of all inY, C A and sured older than 18 years (decrease if dY <0). C pY denote starting point values at time 0 (time index is left away for the sake of simplicity). We now combine equations (4) and (5) into a difference equation. In addition, we separate the terms into von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 Page 4 of 14 those which are independent of the demographic change dY (first row) and those that are dependent on dY (second row). aYt − aY0 ¼ h i h i Y ð1 þ Δ Y Þt − C A ð1 þ Δ A Þt − ð1 − p Y Þ C Y−C A C h i Y ð1 þ Δ Y Þt A ð1 þ Δ A Þ t − C þ pY ð1 þ d Y Þt C ð6Þ If t = 1, then equation (6) simplifies to Y A ΔA ΔY − C aY1 − aY0 ¼ 1 − pY C Y ð1 þ ΔY Þ A ð1 þ ΔA Þ − C þ d Y pY C ð7Þ This equation (7) can be transformed into a change rate ΔaY through division by aY0 . ΔaY ¼ aY1 − aY0 =aY0 ð8Þ which is important in the context of lifetime redistribution described in the next section. Equation (7) allows for comparative static analyses of the impact of different growth parameter combinations on changes of risk adjustment transfers over time. For example, if there was only a change in the composition of the population (i.e. dY ≠ 0 and ΔY, ΔA = 0), then expression (7) simplifies to Y . The sign of this simplified expression A −C ð d Y pY Þ C depends on dY only (change in proportion of insured aged 19 to 25 years), which for Switzerland turns out to be negative (<0, c.f. Results). For young adults with negative transfers (aY < 0) this means that the amount to be paid increases, because equation (7) becomes even more negative. Alternatively, if the only change present was high cost growth among adults (i.e. ΔA > 0 and Δ Y, dY = 0), then expression (7) becomes ðpY −1Þ A C ΔA , which is negative because (pY − 1) < 0. Finally, high cost growth for adults and an increase in their share of the population (i.e. dY <0, ΔA > 0 and ΔY = 0) leads to elevated risk adjustment contributions adults by for young A Δ A þ ð d Y pY Þ the amount of ðd Y pY þ pY −1Þ C A −C Y , because this expression is negative. A more C formal analysis of partial derivatives of equation (7) with respect to different growth parameters is given in Appendix A.1. To summarize, from our two generation model we can derive the following conclusions with respect to population aging. All other things equal, the risk adjustment debt for young adults increases if HCE growth rates are higher for adults than for young adults and/or if the proportion of young adults in the population is shrinking. Basic model for lifetime transfers in risk adjustment (cohort analysis) Given that risk adjustment contributions made by specific age groups can change (as demonstrated in the previous section), what are possible implications for the lifetime balance of risk adjustment transfers? In particular, will the current young generations pay more into risk adjustment than they will ever receive back when they grow old? In the following we will address those questions by developing a discrete-time overlapping generations model, similar in spirit to those developed for the analysis of pension systems (e.g. [20]). Several features of the Swiss risk adjustment make such an analogy quite fitting. First, Swiss social health insurance, of which risk adjustment is an integral part, is a pay-as-you-go system. Because Switzerland only has an internal risk adjustment system (without any tax-financed contributions as for example in Germany) this means that in any given year the contributions made into the fund must equal the benefits paid (henceforth: “symmetry property”). In addition, risk adjustment transfers follow an age-gradient similar to pension schemes, with younger age groups contributing and older generations profiting from risk adjustment (although the addition of further morbidity-criteria may weaken the age dependency). The overall idea for the model is as follows. The curve of risk adjustment transfers ordered by amount and weighted by group size roughly resembles two triangles: One below the zero line (net payers aged 19 to 60 years) and one above (net beneficiaries, aged 61 years and older, Figure 1a). As mentioned above, we will make use of the fact that for internal risk adjustment schemes the two areas defined by the zero-line and the riskadjustment curve are of equal size (symmetry property). Initially, we assume that the shape of those triangles in a given year resembles the pattern of risk adjustment transfers for a single person over a lifetime. In other words, the x-axis interpretation in Figure 1a changes from “age” to “time”. In the simplest model we further assume that a person only lives for two generations: one in which contributions are made and the second in which payments are received. Subsequently we will expand the model to more generations and by modeling growth of risk adjustment payments (Figure 1a and b). The model notation is the same as for the previous section. Two overlapping generations, no population change, undiscounted payments As a convention the superscripts N and P denote negative transfers (contributions into fund) and positive transfers (payments from the fund), respectively. von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 Page 5 of 14 Figure 1 Schematic drawing outlining the concept for modelling lifetime risk adjustment transfers. a shows monthly risk adjustment payments in 2011, ordered by amount of payment (which corresponds to increasing age). Because the volume of payments into and out of the fund are symmetric in Switzerland, the area under the curves below and above the zero-line are identical. For the model of lifetime payments the volume of transfers into the fund (and hence out of the fund because of the symmetry) are approximated by a triangle confined by the most negative transfer to the point where the curve of transfers crosses the zero-line on the x-axis. b generalizes the model of lifetime payments for 3 time periods (two as a net-payer into and one as a beneficiary of risk adjustment). The area of the dotted triangle below the 0 line in Figure 1a (AN 0) can be approximated by the 0 at baseline Y area of a triangle 0; pN ; a 0 0 ) in Figure 1b. Y N P AN 0 ¼ a0 p0 =2 ¼ A0 ð9Þ The notation is similar as in the previous crosssectional model, with aY0 denoting the risk adjustment of the youngest age group (young adults) and pN 0 representing the proportion of net-payers (aged 19 to 60 years) at baseline (time 0). Because of the symmetry property P AN 0 ¼ A0 must hold. We now turn to the case of two time periods with a cost increase of (1 + ΔaY) between the two periods. The variable ΔaY stands for the increase of risk adjustment contributions to be made by young adults in the second time period (cf. Equation 8). Again, there is one time period with payments made into and one with payments received out of the fund. Turning to Figure 1b the young age period with negative transfers is defined by the Y triangle 0; pN and the old age period by triangle 0 ; a0 N p0 ; 1; amax . Because of the symmetry property of in3 ternal risk adjustment we can define both triangles in terms of AN 0. P Y N N AN 1 ¼ A1 ¼ a0 ð1 þ ΔaY Þp0 =2 ¼ ð1 þ ΔaY ÞA0 ð10Þ If ΔaY ≥ 0 then it is straightforward to show that the N inequality AN 0 ≤A1 holds. This means that the payments made into the fund at time 0 are surpassed by the payments received out of the fund at time point 1. More than two net-payer generations, no population change, present-value perspective In order to enhance realism of the model we split the life phase of payments to the fund into several separate time periods and allow cost growth between periods. This is shown in Figure 1b for three phases. As a person ages she transits two periods with decreasing contributions into the fund, i.e. the blue area including the points N N Y 0; p0 =2; aY0 =2; aY0 and the red triangle pN 0 =2; p0 ; a1 =2 , as well as a third life phase with payments out of the max . fund as shown by the green triangle pN 0 ; 1; a3 Y The term a0 =2 follows from the second intercept theN orem: If we divide the line (0, pN 0 ) of triangle ( 0; p0 ; Y a0 ) into halves, then the length of the vertical down ward line 0; aY0 also reduces to 0; aY0 =2 at point N pN 0 =2. This implies that the area of triangle (0; p0 =2; Y N Y a0 =2) is one fourth of the larger triangle (0; p0 ; a0 ), a property used in expression (11). The steepening dotted black lines symbolize cost growth between different time periods (also note that the area of the red triangle is larger than what the corresponding blue segment for the same period would be). Moreover, we now discount all payments at a rate of r. Let’s first focus on the blue area 0; pN =2; aY0 =2; aY0 0 N Y and the red area pN 0 =2; p0 ; a1 =2 in Figure 1b. We denote the sum of those negative transfers over n (here two) time periods by TnN . TnN ¼ AN 0 n−1 X ð1 þ ΔaY Þi i¼0 1 ð1 þ r Þi " n−i n 2 # n−ði þ 1Þ 2 − n ð11Þ von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 Page 6 of 14 The mechanism of this equation is easy to demonstrate for the case of two periods (n = 2) with net contributions into and a third period (n + 1) with payments out of the fund. When i = 0 then the first term in the square Y bracket defines the full triangle at time 0 0; pN 0 ; a0 . In N order to obtain the blue segment 0; p0 =2; aY0 =2; aY0 we N Y have to subtract the smaller triangle pN 0 =2; p0 ; a0 =2 . The N Y area of the triangle in the second period pN 0 =2; p0 ; a0 ð1 þ ΔaY Þ=2Þ takes cost growth into account. Because of the symmetry property for internal risk adjustment, the max area of the green triangle pN representing 0 ; 1; a3 insured with positive payments out of the fund at time point 3 (n + 1) can be defined according to the following equation. n P Tnþ1 ¼ APnþ1 ¼ AN 0 ð1 þ ΔaY Þ 1 ð1 þ r Þn ð12Þ In order to show that over a lifetime the present value of transfers received from risk adjustment are equal or greater than the amounts paid in we have to verify the following inequality. " # n−1 X 1 n−i 2 n−ði þ 1Þ 2 i ð1 þ ΔaY Þ − ð13Þ n n ð1 þ r Þi i¼0 ≤ð1 þ ΔaY Þn 1 ð1 þ r Þn It is commonplace in the literature to discount payments at rates between 2% and 3%. In contrast, we estimated the increase of ΔaY at 4.5% p.a. between 1996 and 2011 using Swiss risk adjustment statistics ([21], not shown). Under those circumstances the expression i 1þΔaY as a function of i is monotonically increasing, 1þr which is an important result on the way to proof inequality (13). In order to demonstrate that expression (13) is true Y we consider a triangle 0; pN 0 ; a0 ð1 þ ΔaY Þ . It is obvious from Figure 1b that the area of this triangle is an overestimate of the actual area of all colored segments, but the mathematical formulation is more tractable. If this triangle is still smaller in size than the one determined by the right hand side of the inequality (green triangle in Figure 1b), then inequality (13) must hold. Y The triangle 0; pN corresponds to a 0 ; a0 ð1 þ ΔaY Þ 1 situation where the index i in (1 + ΔaY)i and ð1þr of the Þi left hand side of equation (13) is held fixed at i = n − 1. This leads to (1 + ΔaY)n − 1 and ð1þr1Þn−1 , which are the lar- gest possible values in the iterations (provided that ΔaY. and r are both positive and the function is monotonically increasing). Equation (13) can then be rewritten as follows. " # n−1 X n−i 2 n − ði þ 1Þ 2 ð1 þ ΔaY Þ − n n ð1 þ r Þn−1 i¼0 1 ≤ð1 þ ΔaY Þn ð1 þ r Þn ð14Þ n−1 1 It is easy to show that the sum of the square brackets is just 1 (an expansion of the sum cancels out all terms except for the first and the last, which are 1 and 0, respectively), leading after some re-arrangements to the following equation. 1≤ ð1 þ ΔaY Þ ð1 þ r Þ ð15Þ Given the observed value for ΔaY and an assumed discount rate of 3% inequality (15) holds true. Further note that the right hand side of this inequality can be interpreted as a return rate for the payments made into the fund. Moreover, inequality equation (15) fully integrates with equations (7) and (8) defining ΔaY, and the discussion of effects of different change parameters on risk adjustment payments for young adults applies. In particular, the expected greater increase in costs for adults (compared with young adults) and the decreasing fraction of young adults in the population (as a result of demographic changes) both increase ΔaY and therefore will enhance the return rate. More than two net-payer generations, population change, present-value perspective It is straightforward to show that the above reasoning can be generalized for any n generations of net payers (the return rate will actually remain the same). But what happens if, in addition, the population as a whole gets older in average age? We model this change by introducing an additional growth rate dN, which stands for changes in the proportion of net contributors (aged 19 to 60 years) into risk adjustment. Turning again to the scenario with multiple time steps on the left hand side and with one time step on the right hand side, we obtain the following inequality. n−1 X ð1 þ ΔaY Þi ð1 þ d N Þi i¼0 " # n−i 2 n−ði þ 1Þ 2 − n n ð1 þ r Þi 1 ð16Þ ≤ð1 þ ΔaY Þn ð1 þ d N Þn 1 ð1 þ r Þn Again, studying the combination of growth rates 1 ð1 þ ΔaY Þi ð1 þ d N Þi ð1þr is a key step in the analysis of Þi inequality (16). Because dN was positive in the past (the fraction of net payers - corresponding to all age groups von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 Page 7 of 14 between 19 and 60 years - increased at 0.33% p.a., although not continuously [21]) it is reasonable to assume that all growth rates of net payer fractions combined as a function of i are also monotonically increasing. Proceeding like above we fix all growth rates on the left hand side at i = n − 1 and at i = n for the right hand side term of equation (16). After rearrangements and simplifications we obtain inequality (17). 1≤ ð1 þ ΔaY Þð1 þ d N Þ ð1 þ r Þ ð17Þ This inequality (17) holds true if ΔaY, dN, r ≥ 0 and ΔaY ≥ r or dN ≥ r. Because demographic changes are likely to increase ΔaY, the return rate on the right-hand side of the equal sign is also expected to become larger, all other things equal. In summary, the models developed in this section suggest that, all other things equal, insured will on average receive more (discounted) payments out of the risk adjustment fund than they will have contributed over a lifetime, and that the expected demographic changes are likely to have an increasing effect on the return rate. How can the solidarity burden be redistributed? The calculations from the basic models suggest that, if unchecked, the solidarity burden for young adults will continue to grow at high rate, which is mainly due to health care expenditure growth among older generations. While those younger generations may still be net beneficiaries of risk adjustment payments over lifetime, their owed risk adjustment debt may nonetheless overwhelm their financial means (cf. [22]). In Switzerland, one such vulnerable group are the young adults aged 19–25 years, which are entitled to premium reductions by law. This relaxation from the community-rate principle was introduced to provide relief for young adults who have limited disposable income. However, by 2011 those rebates have all but disappeared [22]. As already observed by Beck in 2004 [23], the reason for those diminishing rebates is that the Swiss risk adjustment scheme ignores the possibility for premium reductions to young adults and overcharges this age group. This problem has been recognized by federal authorities (e.g. [24]), but no convincing solutions have been presented so far. In particular, many proposed solutions neglect that any reduction of solidarity transfers (be it as a premium reduction or a reduction of risk adjustment contributions) must be compatible with risk adjustment so as not to induce selection incentives for insurers. McGuire et al. [25] and Beck et al. (Beck K, Buchner F, van Kleef R, von Wyl V: Theory of risk equalization: Are we on the wrong track? submitted) have provided methodologies for how to limit solidarity transfers within risk adjustment systems. In line with their suggestions we develop a method that can correct for the expected shifts in demography and keep solidarity contributions for specific age groups stable over time. To this end, we introduce two additional parameters in the model described by equation (7). γ = Growth factor determining an upper limit of risk adjustment payment growth. For example, this parameter can be used to decouple risk adjustment contribution growth for young adults from the growth rate of the remaining adults, which is the main driver for the observed increases in younger age groups (c.f. Results). ρ = Factor for reduction of risk adjustment payments (e.g. a premium rebate for young adults). In addition to the stabilization by γ a further reduction of nominal payments can be granted. For example, in Switzerland it is currently discussed to charge only 50% of the nominal risk adjustment contributions from 19 to 25 year old individuals (i.e. ρ = 0.5) [22]. Solution for two risk adjustment groups In general, the reductions are implemented by subtracting an amount ut from the nominal RA-payment aYt (defined in Equation 7) so that current RA-payments (at time t) for young adults after correction are equivalent to a baseline payment times a pre-specified growth rate γ. aY0 ð1 þ γ Þt ¼ aYt −ut ð18Þ Solving for ut yields equation (19). ut ¼ aYt −aY0 ð1 þ γ Þt ð19Þ Corrected payments as defined by (18) can be modified further by a second parameter ρ that defines the rebate (<1) on nominal RA payments for specific groups (e.g. 50% to 19 to 25 year olds). This reduction is applied to the stabilized risk adjustment contribution aY0 ð1 þ γ Þt for young adults defined in (18), and then equation (19) becomes ut ¼ aYt −ρaY0 ð1 þ γ Þt : ð20Þ Thus, the degree of fairness is determined by the parameters γ and ρ. The full equation for reduced intergenerational risk adjustment transfers in young adults reads as ãtY ¼ aYt − ut ¼ ρaY0 ð1 þ γ Þt ð21Þ For the transfers to sum to zero the payments benefitting older generations must also be shortened by a certain amount. We denote this deduction by vt. von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 ãtA ¼ aAt − vt Page 8 of 14 ð22Þ Moreover, we set the restriction that the sum of intergenerational transfers between young adults and adults must equal to zero according to equation (23) 1 − pYt aAt − vt þ pYt aYt − ut ¼ 0 ð23Þ A Solving for at −vt and plugging into (22) yields expression (24) ãtA ¼ − pYt ρaY0 ð1 þ γ Þt ; 1 − pYt ð24Þ which corresponds to the new intergenerational risk adjustment payment of young adults to adults. An extension of the reasoning in this section to several age groups is given in Appendix A.2. Additionally, in Section A.3 of the appendix we show that the lifetime balance of payments stays positive even after reductions of risk adjustment for specific young age groups. Results Retrospective analysis of intergenerational solidarity transfers in Switzerland Next, we illustrate the mathematical models defined by equations (7) and (16) from the methods sections by retrospectively analyzing solidarity transfers between generations over the period of 1996 to 2011 within Swiss risk adjustment. As an example, we center this and the following analysis on the group of young adults aged 19–25 years, and the rationale for that decision is detailed in the methods section and in a companion paper (von Wyl V, Beck K: Distribution of premium burden for mandatory health insurance in Switzerland, submitted). To inform the models we used data from the official Swiss risk adjustment statistics 1996 through 2011 [21], which are displayed in Table 1. The second column shows the proportion of young adults in the Swiss population of individuals older than 18 years. Over the 15 year observation period this proportion has decreased by 0.8% points, thus the share of young adults has shrunken slightly. Columns 3 and 4 show average costs for the two age groups, whereas columns 5 and 6 represent relative changes using 1996 as the base year (100%). Overall, health care expenditures for young adults have increased by 32% at an average growth rate of 1.88% and even by 78% for adults older than 25 years (average growth rate 3.94%). Over time, intergenerational risk adjustment payments for young adults have risen from CHF 87 per month in 1996 to CHF 182 in the year 2011 (Table 1, column 7). Yet the effect of those solidarity transfers on the adults’ side remained rather small. In 1996 each adult received CHF 11 per month in solidarity transfers from 19 to 25 year olds. Fifteen years later those payments have only risen by CHF 10 to a total of CHF 21 per month (Table 1, column 8). Overall the solidarity burden for young adults has experienced the highest absolute Table 1 Evolution of risk adjustment transfers between 1996 and 2011 Health care expenditures Cost increase (CHF per month) (base year 1996) Monthly risk adjustment payment into (<0) or from (>0) the fund Year Proportion of young adults Young adults Adults Young adults Adults Young adults Adults 1996 11.1% 61 159 100% 100% −87 11 1997 10.8% 61 166 101% 105% −94 11 1998 10.6% 61 175 101% 110% −101 12 1999 10.6% 62 181 101% 114% −107 13 2000 10.5% 65 192 106% 121% −114 13 2001 10.4% 68 203 111% 128% −121 14 2002 10.5% 69 210 113% 133% −127 15 2003 10.5% 71 220 117% 139% −133 16 2004 10.5% 73 234 121% 147% −143 17 2005 10.5% 75 246 123% 155% −153 18 2006 10.4% 72 247 118% 156% −157 18 2007 10.3% 73 257 120% 162% −165 19 2008 10.3% 76 267 125% 169% −172 20 2009 10.3% 79 275 129% 173% −176 20 2010 10.3% 80 280 131% 177% −180 21 2011 10.3% 80 283 132% 178% −182 21 von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 growth of any age groups between 1996 and 2011, but with little effect for adults due to the demographic constellation (i.e. much more adults than young adults). Next, we investigated the importance of the three model parameters (ΔY, ΔA , dY) for the increase of solidarity payments by young adults. Following the discussion of partial derivatives in Appendix A.1 (equations (A.1.1)(A.1.3)) we quantified the magnitude of change of RA payments for a one unit parameter increase. To this end we used health care expenditures in 2011 and averages of cost growth and demographic change rates over the 15 year observation period for calculations (Table 1), which yielded values of CHF 21.75, CHF −254.0 and CHF 72.20 for dy, ΔA and ΔY, respectively. In other words, solidarity payments owed by young adults are most sensitive to health care expenditure growth among adults, whereas shifts in population composition (i.e. fewer young adults) seem to play a less important role. This finding was confirmed when we performed correlation analyses of observed annual changes in dy, ΔA and ΔY with observed increases in RA contributions of young adults ΔaY. While spearman correlations of ΔaY with dy (rho = −0.45) and ΔY (rho = 0.11) did not reach statistical significance, the HCE growth rate of adults ΔA showed a strong, statistically significant correlation with ΔaY (rho = 0.81, p-value < 0.001). Along those lines we decomposed the impact of demography on the cumulative observed increase of risk adjustment payments between 1996 and 2011 using equation (7). Over this time period the monthly risk adjustment contribution rose by CHF 95 for young adults (Table 1, column 7). The demography-dependent term of equation (7) sums to CHF 2, whereas the independent terms amount to CHF 93. Thus, 98% of the increase was driven by health care expenditure growth and only 2% were attributable to demographic changes, meaning the decreasing proportion of young adults in the population. How will solidarity transfers develop in the future? Based on the algorithms for RA modifications developed in the methods section and on data presented in Table 1 we performed projections of solidarity transfers between young adults and adults 15 years into the future, starting in 2011 and ending in 2026. Moreover, we tested the impact of two different reform suggestions to relief the premium burden for younger generations. This was done by repeated stochastic simulations based on equation (7). In particular, we used the cost and demography information for the year 2011 from Table 1 as baseline and progressed stepwise in 1 year intervals into the future by randomly selecting, for each analysis year, HCE growth parameters for young adults and adults from the distribution of growth rates observed between 1996 and 2011 (assuming that they came from a Gaussian distribution). Page 9 of 14 The parameter values for the demographic change dy stemmed from predictions of demographic changes in Switzerland [26] (conservative scenario) and was held fixed at dy = −0.01234 per year. Subsequently, HCE and risk adjustment contributions were calculated. Each simulation run, consisting of the full 15 year observation period, was repeated 10′000 times. We explored three scenarios. The first scenario represents the status quo (before the 2012 risk adjustment reform) without any post-hoc modifications of nominal risk adjustment transfers. For the second scenario risk adjustment contributions by young adults were stabilized to increase only with their own HCE growth rate (1 + ΔY)t (note that alternate choices could consider income growth). Thus restriction (18) becomes aY0 ð1 þ ΔY Þt ¼ aYt −ut . In the third scenario the risk adjustment payments were further reduced by 50% after capping growth (i.e., γ = ΔY and ρ = 0.5), which yields the condition 0:5aY0 ð1 þ ΔY Þt ¼ 0:5 aYt −ut . The results of those simulations are displayed in Figure 2. The solid lines represent the status quo scenario and suggest that solidarity contributions by young adults (blue line) will grow steadily at a greater-than-linear rate from CHF −182 in 2011 to CHF −378 [95% simulation interval −322; −439] in 2026. In contrast, per-capita contributions received by older generations will increase only moderately from CHF 21 in 2011 to CHF 35 [30; 41] in 2026. The second scenario, which includes a stabilization of RA growth for young adults (dotted line) leads to risk adjustment payments of −244 [−205; −287] by young adults in 2026, whereas adults would receive CHF 25 [21; 30]. If additional reductions on risk adjustment (here: 50%) were granted to young adults this would lead to a solidarity burden of −122 [−103; −143] (starting from CHF −91 in 2011), and adults would receive monthly contributions of CHF 13 [11; 15] (up from CHF 10 in 2011). Along the same lines we modeled lifetime redistribution through risk adjustment. In particular, we simulated the life cycle of a 19 year old insured starting in 2011 and progressed in one year steps towards the age of 82 (average life expectancy in Switzerland for the year 2011 [27]). Using risk adjustment contributions of the year 2011 as a basis we modeled yearly cost increases for all 15 age groups (by drawing random yearly growth rates from a multivariate normal distribution that was informed with observed cost increases for each age group [21]). All age-specific risk adjustment transfers were summed up and discounted at 3%. In total, 10′000 such life cycles were simulated and repeated for the two different reform suggestions. Overall, those simulations suggest that for the baseline scenario discounted payments of CHF 83′164 are made into the fund and CHF 114′707 are received out of the von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 Page 10 of 14 Figure 2 Projections of risk adjustment transfers from young adults to adults. Monthly intergenerational risk adjustment transfers of young adults (blue lines) and adults older than 25 years (red lines). The solid lines project transfers without any modifications, and the dashed lines show hypothetical transfers after capping transfer growth for young adults by their health care expenditure growth rate. fund (difference CHF 30′761 [12′091, 55′467]). When capping risk adjustment payment growth at 2% for young adults the net surplus dropped to CHF 17′069 [2′ 336, 36′833]. If an additional rebate of 50% on stabilized risk adjustment payments is granted to young adults, then the lifetime balance of discounted payment increased to CHF 19′030 [4′297, 38′795] compared with the scenario without rebate in addition to stabilization. Discussion The view that Western societies will have to face rising healthcare expenditures because of population aging is largely uncontested. Increasing health costs are also observed in Switzerland, which has led to shifts in solidarity transfers over the past years as witnessed by the doubling of risk adjustment payments made by young adults since 1996. In this paper we have analyzed those intergenerational solidarity transfers between young and old via a mathematical model and estimated the lifetime balance of discounted risk adjustment transfers. We observed that population aging continuously tends to increase risk adjustment contributions from younger generations. In a model analysis informed by data from Swiss risk adjustment and focusing on young insured aged 19–25 years we found that disparate health care utilization patterns between young and old insured were at the base of the problem of rising solidarity transfers from young adults to older generations. While health care expenditures for young adults grew only moderately, cost growth was almost twice as high among all adults aged 26 years and more. This development has widened the gap of health care expenditure levels between the two age groups. Since risk adjustment operates at closing the gap in order to prevent risk selection, this mechanism has led to increasing intergenerational solidarity payments owed by young adults. But how problematic is the rising solidarity burden for younger insured? As for Switzerland, we believe that there is ample societal justification for lowering the burden of risk adjustment payments for young adults aged 19 to 25 years. First and foremost, the current legislation is inconsistent with regards to solidarity contributions by young adults. While the law explicitly allows premium reductions for 19 to 25 year olds, risk adjustment legislation demands that all insured aged 19 and more are treated equal, thus leaving no room for premium rebates. Second, as outlined in detail in [22], there is mounting evidence that young adults may not be able to carry their solidarity burden, which leads to inefficiencies and inequalities in health financing. Data presented in the companion paper suggest that young adults have to rely heavily on their parents for health insurance premium financing. Along the same lines, the proportion of young adults who are eligible for premium subsidies has risen (although not steadily) from 41% in 2000 (the first year with detailed statistics available) to 44% in 2011 [16]. Third, voting data also presented in the companion paper suggest that the current system with strictly communityrated premiums reflects preferences of net contributors, and young adults in particular, less well than those of older generations who benefit from intergenerational solidarity in the status quo situation (von Wyl V, Beck K: Distribution of premium burden for mandatory health insurance in Switzerland, submitted). It is furthermore remarkable that older generations of today contributed less to solidarity at younger age because of partially age-rated premiums and the possibility for opting-out prior to the reform in 1996. von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 Finally, it can be argued that the demographic changes and the observed steepening of cost growth across age gradients represent new developments that were not part of voters’ expectations at the time of the referendum for the new health insurance act implemented in 1996, which introduced community rating and a health insurance mandate. Taken together, those trends have the potential to endanger the generational contract, because they can lead to overwhelming solidarity transfers from young to old (as shown by our model) and thus deteriorate the acceptance of the community-rating principle among the young. Such tendencies are already observed in surveys on the Swiss pension system, which is also under pressure as a result of the demographic change [28]. Consequently, the subject of lowering the growing financial burden of social health insurance for young insured has gained political momentum and has even found its way on the regulator’s agenda. Several reform suggestions have been discussed in the news media, ranging from excluding young adults from risk adjustment, abolishment of premiums for children to relieve young families, or the introduction of new premium age groups for individuals above 50 years to levy more health insurance costs on older generations. What is rarely considered in those suggestions, however, is that in settings with internal risk adjustment (as in Switzerland) such rebates must either be implemented through risk adjustment itself or must at least be reflected, otherwise new selection incentives are created. The premium rebates for young adults are an example of an inconsistent regulation that collided with risk adjustment and led to unwanted results [22]. Within this study we have developed a framework for how such reforms of health premium financing could be implemented in a consistent manner, and we have investigated possible effects on the balance of lifetime risk adjustment payments. We argue that any premium reductions must be implemented directly at the level of risk adjustment payments, thereby giving insurers the possibility to pass on those risk adjustment reductions as premium rebates to young adults. In that regard our method resembles the work of McGuire et al. [25]. Our approach extends theirs by further introducing the concept of stabilization of payments at predefined levels over time. For example, as our analysis has shown risk adjustment payments by young adults increase at a rate that is approximately proportional to health care expenditure growth in older age groups. Without stabilization intergenerational risk adjustment contributions may otherwise soon again rise to levels that exceed disposable means of young adults. Despite an ever growing solidarity burden, over lifetime the younger generations of today can still expect a Page 11 of 14 net gain from risk adjustment, all other things equal and given that the current level of health care can still be financed in the future. Interestingly, the models suggest that the expected demographic changes may even increase their lifetime rate of return on risk adjustment contributions. As the share of older insured increases over time, former net-recipients of funds from risk adjustment may turn into net-contributors, and therefore the number of net-payers into the fund increases. For example, the group of 51 to 55 year old women still received a small payment from the fund in 1996, whereas 15 years later the same age group has become a net-payer. In addition, the observed greater average HCE growth for older age groups compared with young adults further increases the return rate of risk adjustment. Similar results are known from modelling studies of pay-as-you-go social security systems [20,29] or health insurance [30], which found the rate of return also to be dependent on cost growth among older generations (i.e. recipients of payments) and/or increases in the proportion of net payers. Interestingly, an application of generational accounting for health financing to Switzerland corroborates our observation that currently young generations will receive more payments from social health insurance than they will have to contribute over a lifetime [31]. But when these authors performed a more comprehensive analysis including all money streams for health financing the lifetime balance of payments turned negative for several younger age groups. Although developed for the Swiss setting, our models of demographic effects on risk adjustment and on the life-time balance of transfers also apply to other countries with risk adjustment, such as the Netherlands or Germany. But since both countries raise premiums for mandatory insurance in an income-dependent manner, the direct effects of demographic changes on risk adjustment will be less felt than in Switzerland (whereas health financing of those countries will be more affected by the expected decrease of the ratio of active to retired workers). Similar to our framework for controlling age-dependent solidarity transfers, the U.S. has implemented a rule that age-based premiums are only allowed to vary within a ratio of 3 (oldest age group) to 1 (youngest) [25]. In contrast, solidarity restrictions between age groups as a means to improve equity are usually not needed in settings with income-dependent premiums. However, our modelling findings hinge on assumptions regarding rates of risk adjustment growth, the proportion of net payers, and discount rates. Also, further modifications to risk adjustment such as the planned introduction of pharmaceutical cost groups in Switzerland will have a yet unclear impact on the balance. For example, the proportion of net payers pN may suddenly change if additional morbidity criteria are introduced into the risk von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 Page 12 of 14 adjustment model. Moreover, the impact of population aging on the parameters ΔaY and pN is not continuous and difficult to predict in the long run. An aging population drives up the population average of HCE from which risk adjustment contributions are measured, and age groups which have formerly received payments from risk adjustment may become net-contributors in subsequent years. Finally, our model focused only on risk adjustment payments and ignored any other components of health financing such as direct and indirect taxes, private insurance, and premium subsidies. Furthermore, it should be noted that our model does not address issues of sustainability and affordability of health finance. Generational accounting may offer suitable tools for such more comprehensive analyses [32]. Conclusions In summary, the framework developed in this paper may help to construct a more equitable system of health financing in light of population aging and to strengthen the acceptance of the intergenerational contract in social health insurance systems with community-rating. Appendix A.1 Analysis of change rates We now analyze equation (7) by calculating the partial derivatives of aY1 −aY0 with respect to the variables dY, ΔY, ΔA. Y ð1 þ Δ Y Þ A ð1 þ ΔA Þ−C ∂ aY1 −aY0 =∂d Y ¼ pY C ðA:1:1Þ This partial derivative (A.1.1) indicates that changes in risk adjustment contributions resulting from demographic trends are also dependent on the absolute difference of cost levels between young adults and adults at time t=1. Because HCE and growth rates of adults are A ð1 þ ΔA Þ > C Y larger than those of young adults, i.e. C ð1 þ ΔY Þ expression (A.1.1) is positive. Going back to equation (7) describing the change in risk adjustment payments for young adults it becomes apparent that an increase in the proportion of young adults by dY reduces their risk adjustment debt, because this will make expression (7) less negative. The partial derivative with respect to HCE growth for adults is as follows. A ∂ aY1 −aY0 =∂ΔA ¼ d Y pY þ pY −1 C ðA:1:2Þ As long as pY represents a minority this expression is negative since (dYpY + pY − 1) < (2pY − 1) < 0, which means that all other things equal a one unit increase in HCE growth for adults leads to an increase of the risk adjustment debt for young adults. Equation (A.1.3) shows the partial derivative with respect to HCE growth in young adults. Y ∂ aY1 −aY0 =∂ΔY ¼ − d Y pY þ pY −1 C ðA:1:3Þ Expression (A.1.3) has the inverse sign of (A.1.2). Thus, all other things equal, a one unit increase in HCE growth for young adults decreases their risk adjustment debt. A.2 Solidarity reductions for several risk adjustment groups The algorithm described in equations (18)-(24) can also accommodate reductions for several age groups i = 1,..,m. For the simple case where modifications of RA transfers only apply to net contributors to risk adjustment this can be written as ãi;t ¼ ai;t − ui;t ðA:2:1Þ for age groups i = 1,..,m. The solidarity cap is implemented according to equation (21), whereby ρi is set to 1 if no post-hoc modifications to risk adjustment payments are applied. Likewise, if no modifications to γi apply then it can simply be set to 0. t ui;t ¼ ai;t − ρi ai;0 1 þ γ i ðA:2:2Þ Next, we derive the solidarity transfers for older age groups m+1 to 30. We denote those groups of beneficiaries from intergenerational solidarity transfers by subscript r and treat them as one group (for sake of simplicity). As in the simple case with only two age groups discussed above the sum of intergenerational solidarity transfers must equal to 0 in internal risk adjustment systems, and therefore the new transfers from young insured to the (old) reference group can be written as ãr;t ¼ − Xm xi;t t ρi ai;0 1 þ γ i ; i¼1 x r;t ðA:2:3Þ with r symbolizing the reference population consisting of all m+1,…,30 remaining groups and the variables xi,t and xr,t being the number of individuals in group i and the reference group r, respectively. Modifications of transfers within recipients of RA payments (e.g. higher payments for specific age groups) are also possible, but they require an additional calculation step, in which the total volume of intergenerational RApayments from younger age groups are redistributed within older generations. This can be achieved, for each age group j = m+1 to 30, by defining weights wj,t according to equation von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 wj;t ¼ xj;t πj X30 xr;t π j¼mþ1 j ; Page 13 of 14 ðA:2:4Þ whereby πj denotes the group specific factor (analogous to ρi). A.3. Implication of restricted solidarity for lifetime redistribution How is the lifetime balance of payments into and from the risk adjustment fund affected by the modifications proposed in equations (18) to (24)? As defined by equation (18) rebates are defined as a reduction of nominal payments. By use of inequality equation (16) the lifetime balance in settings with premium reductions for specific age groups g (which also correspond to specific time periods as an individual transits through all age groups over lifetime) can therefore be defined by TnN − aY0 X ð1 þ ΔaY Þi ð1 þ d N Þi i∈g 1 ð1 þ r Þi " n−i n 2 # n − ð i þ 1Þ 2 − ρi n ðA:3:1Þ 1 ≤T Pnþ1 − aY0 ð1 þ ΔaY Þnþ1 ð1 þ d Y Þnþ1 nþ1 " # ð1 þ r Þ 2 2 X n−j n − ðj þ 1Þ − ρj n n j∈g Note that the indices i and j have somewhat different interpretations. The left-hand side of equation (A.3.1) sums up all payments made over n time periods, and hence i refers to time. The right hand side of equation (A.3.1) represents, according to our convention, the volume of all contributions made by net payers at time n+1. Hence the parameter j has the interpretation of age groups. This distinction does not alter the conclusions, however. We have already demonstrated that the inequality N T n ≤T Pnþ1 holds if the combination of growth rates yields a monotonically increasing function. Therefore we only have to proof that the reductions up to period n are smaller than the sum of all reductions at time point n+1. We will first demonstrate this for constant, timeindependent reductions ρ. Since all factors before the sum sign combined are monotonically increasing we can fix i at n for all growth rates on the left-hand side of equation (A.3.2), which makes them constant. aY0 ð1 " # X n − i2 n − ði þ 1Þ2 1 ρi þ ΔaY Þ ð1 þ d N Þ − ð1 þ r Þn i∈g n n n n ðA:3:2Þ 1 ð1 þ r Þnþ1 # ≤ aY0 ð1 þ ΔaY Þnþ1 ð1 þ d Y Þnþ1 " X n − j2 n − ðj þ 1Þ − n n j∈g 2 ρj Because all elements of the sums over i and j are timeindependent and the same on both sides of the equation they can be dropped. Thus, the remaining inequality must hold true given our assumption of monotonically increasing growth terms in combination. Using the same approach, it is simple to demonstrate that the inequality also holds when growth of aY0 is capped by a pre-defined, positive growth rate γi (by replacing the parameter ΔaY in equation (A.3.2), proof not shown). Competing interests VVW and KB are employees of CSS Insurance, Switzerland, but the employer had no role in study design, analysis, manuscript preparation or decision to publish. Authors’ contributions VVW and KB conceived the study. VVW developed the methods, conducted the analyses and wrote the manuscript. KB revised it critically for important intellectual content. Both authors read and approved the final manuscript. Acknowledgements The authors wish to thank Florian Buchner, Jürgen Wasem and two anonymous reviewers for helpful comments. Author details 1 CSS-Institute for Empirical Health Economics, Tribschenstrasse 21, 6002 Luzern, Switzerland. 2Institute for Social and Preventive Medicine, University of Bern, Finkenhubelweg 11, 3012 Bern, Switzerland. 3Department of Economics, University of Zurich, Blümlisalpstrasse 10, 8006 Zurich, Switzerland. Received: 17 July 2013 Accepted: 24 January 2014 References 1. Christensen K, Doblhammer G, Rau R, Vaupel JW: Ageing populations: the challenges ahead. Lancet 2009, 374:1196–1208. 2. Vaupel JW: Biodemography of human ageing. Nature 2010, 464:536–542. 3. Breyer F, Costa-Font J, Felder S: Ageing, health, and health care. Oxf Rev Econ Policy 2010, 26:674–690. 4. Meier V, Werding M: Ageing and the welfare state: securing sustainability. Oxf Rev Econ Policy 2010, 26:655–673. 5. Colombier C, Weber W: Projecting health-care expenditure for Switzerland: further evidence against the 'red-herring' hypothesis. Int J Health Plan M 2011, 26:246–263. 6. Mendelson DN, Schwartz WB: The effects of aging and population growth on health care costs. Health Aff (Millwood) 1993, 12:119–125. 7. Buchner F, Wasem J: “Steeping” of health expenditure profiles. Geneva Pap R I-Iss P 2006, 31:581–599. 8. Steinmann L, Telser H, Zweifel P: Aging and future healthcare expenditure: a consistent approach. Forum Health Econ Policy 2007, 10:1–30. 9. Felder S, Werblow A: Does the age profile of health care expenditure really steepen over time? New evidence from Swiss Cantons. Geneva Pap R I-Iss P 2008, 33:710–727. 10. Zweifel P, Felder S, Meiers M: Ageing of population and health care expenditure: a red herring? Health Econ 1999, 8:485–496. 11. Felder S, Werblow A, Zweifel P: Do red herrings swim in circles? Controlling for the endogeneity of time to death. J Health Econ 2010, 29:205–212. von Wyl and Beck Health Economics Review 2014, 4:7 http://www.healtheconomicsreview.com/content/4/1/7 Page 14 of 14 12. Seshamani M, Gray A: Ageing and health-care expenditure: the red herring argument revisited. Health Econ 2004, 13:303–314. 13. Seshamani M, Gray AM: A longitudinal study of the effects of age and time to death on hospital costs. J Health Econ 2004, 23:217–235. 14. Wong A, van Baal PH, Boshuizen HC, Polder JJ: Exploring the influence of proximity to death on disease-specific hospital expenditures: a carpaccio of red herrings. Health Econ 2011, 20:379–400. 15. Wynand PMM, RPE v d V: Risk adjustment in competitive health plan markets. Handbook Health Econ 2000, 1:755–845. 16. Federal Office of Public Health: Statistical Data on Health and Accident Insurance for 2012 (in German). 2012 [http://www.bag.admin.ch/themen/ krankenversicherung/01156/index.html?lang=de] 17. Enthoven AC: Consumer-choice health plan (second of two parts). A national-health-insurance proposal based on regulated competition in the private sector. New Engl J Med 1978, 298:709–720. 18. Beck K, Spycher S, Holly A, Gardiol L: Risk adjustment in Switzerland. Health Policy 2003, 65:63–74. 19. Beck K, Trottmann M, Zweifel P: Risk adjustment in health insurance and its long-term effectiveness. J Health Econ 2010, 29:489–498. 20. Feldstein M, Liebman JB: Social security. Handbook Public Econ 2002, 4:2245–2324. 21. Common Institution under the Federal Health Insurance Act: Risk Adjustment Statistics. 2013 [http://www.kvg.org/de/statistik-_content—1– 1052.html] 22. von Wyl V, Beck K: Sind junge Risiken auch gute Risiken? Eine kritische Betrachtung der Stellung der Jugendlichen in der obligatorischen Grundversicherung; [http://www.css-institut.ch/downloadpdf.html?fileRef=/ content/dam/css-institut/de/documents/kvg_und_junge_v3.pdf.pdf] 23. Beck K: Risiko Krankenversicherung. Haupt Verlag: Bern, Switzerland; 2004. 24. Federal Office of Internal Affairs: Moderate Prämienanpassung auf 2013 (in German). 2013 [http://www.news.admin.ch/message/index.html?lang= de&msg-id=46122]. 25. McGuire TG, Glazer J, Newhouse JP, Normand S-L, Shi J, Sinaiko AD, Zuvekas S: Integrating risk adjustment and enrollee premiums in health plan payment. J Health Econ 2013, 32:1263–1277. 26. Federal Office of Statistics: Szenarien zur Bevölkerungsentwicklung 2010–2060. Neuchâtel, Switzerland: Federal Office of Statistics; 2010. 27. Organization for Economic Co-operation and Development (OECD): Health Data 2013. 2013 [http://www.oecd.org/health/healthdata] 28. Rossini S, Fischer A: Mémoire sociale et pérennité des politiques de solidarité. Les jeunes et la solidarité: quelle compréhension? Lausanne, Switzerland: Haute école de travail social et de la santé EESP; 2012. 29. Samuelson PA: An exact consumption-loan model of interest with or without the social contrivance of money. J Polit Econ 1958, 66:467–482. 30. Jack W: Intergenerational risk sharing and health insurance financing. Econ Record 1998, 74:153–161. 31. Hagist C, Klusen N, Plate A, Raffelhüschen B: Social health insurance-the major driver of unsustainable fiscal policy? Freiburg, Germany: Diskussionsbeiträge des Forschungszentrums Generationenverträge, Albert-Ludwigs-Universität Freiburg; 2005. 32. Auerbach AJ, Gokhale J, Kotlikoff LJ: Generational accounting: a meaningful way to evaluate fiscal policy. J Econ Perspect 1994, 8:73–94. doi:10.1186/s13561-014-0007-5 Cite this article as: von Wyl and Beck: Risk adjustment in aging societies. Health Economics Review 2014 4:7. Submit your manuscript to a journal and beneﬁt from: 7 Convenient online submission 7 Rigorous peer review 7 Immediate publication on acceptance 7 Open access: articles freely available online 7 High visibility within the ﬁeld 7 Retaining the copyright to your article Submit your next manuscript at 7 springeropen.com
https://openalex.org/W4254706590	https://figshare.com/articles/journal_contribution/A_Review_on_Breathing_Behaviors_of_Metal-Organic-Frameworks_MOFs_for_Gas_Adsorption/14637477/1/files/28112847.pdf	English	null	A Review on Breathing Behaviors of Metal-Organic-Frameworks (MOFs) for Gas Adsorption	null	2,023	cc-by	21,685	Materials 2014, 7, 3198-3250; doi:10.3390/ma7043198 Materials 2014, 7, 3198-3250; doi:10.3390/ma7043198 materials ISSN 1996-1944 www.mdpi.com/journal/materials OPEN ACCESS Materials 2014, 7 Materials 2014, 7 3199 1. Introduction Metal-organic frameworks (MOFs) are a new class of crystalline microporous materials that exhibit tunable functionalities, large surface areas, framework flexibility, catalytic activities [1], and gas separation capability among others [2]. Their frameworks are formed by the covalent linkages between metals or metal oxides and organic moieties. Even though, compared with zeolites, MOFs suffer a major drawback of relatively lower thermal stability, the flexible structure of MOFs contrast to rigid zeolite frameworks, yielding the unique breathing phenomena or gate-opening effect. The pore diameter of MOFs enlarges or shrinks while external stimuli are applied or removed, such as gas molecules, solvents, or pressure [3]. The features of breathing phenomena and large surface areas enable MOFs suitable materials for applications in green and renewable energy as media of gas storages, sensing and separation [4–6]. Scientists and engineers strive to understand and substantiate the origin of the framework flexibility such that properties of MOFs can be designed at the synthesis stage. Given that a large variety of MOFs family and publications, this review focuses on summarizing the MOFs’ breathing effect, to which a hydrothermal synthesis route is applied. Readers can find more information in a critical review published recently by Férey’s group, in which the authors elucidated rationales of the breathing effects [7]. This review article will describe the breathing phenomena of MOFs in following sequences: (1) definition of MOFs and summary of synthesis conditions; (2) summary of MOFs exhibiting the breathing phenomena; (3) post-synthetic medication to alter breathing phenomena; (4) outlooks of MOFs’ breathing phenomena. Mays Alhamami, Huu Doan and Chil-Hung Cheng * Department of Chemical Engineering, Ryerson University, 350 Victoria Street, Toronto, ON M5B 2K3, Canada, E-Mails: mays.alhamami@ryerson.ca (M.A.); hdoan@ryerson.c Department of Chemical Engineering, Ryerson University, 350 Victoria Street, Toronto, ON M5B 2K3, Canada, E-Mails: mays.alhamami@ryerson.ca (M.A.); hdoan@ryerson.ca (H.D.) * Author to whom correspondence should be addressed; E-Mail: chilhung.cheng@ryerson.ca; Tel.: +1-416-979-5000 (ext. 2131); Fax: +1-416-979-5083. Received: 1 March 2014; in revised form: 31 March 2014 / Accepted: 11 April 2014 Published: 21 April 2014 Received: 1 March 2014; in revised form: 31 March 2014 / Accepted: 11 April 2014 Published: 21 April 2014 Abstract: Metal-organic frameworks (MOFs) are a new class of microporous materials that possess framework flexibility, large surface areas, “tailor-made” framework functionalities, and tunable pore sizes. These features empower MOFs superior performances and broader application spectra than those of zeolites and phosphine-based molecular sieves. In parallel with designing new structures and new chemistry of MOFs, the observation of unique breathing behaviors upon adsorption of gases or solvents stimulates their potential applications as host materials in gas storage for renewable energy. This has attracted intense research energy to understand the causes at the atomic level, using in situ X-ray diffraction, calorimetry, Fourier transform infrared spectroscopy, and molecular dynamics simulations. This article is developed in the following order: first to introduce the definition of MOFs and the observation of their framework flexibility. Second, synthesis routes of MOFs are summarized with the emphasis on the hydrothermal synthesis, owing to the environmental-benign and economically availability of water. Third, MOFs exhibiting breathing behaviors are summarized, followed by rationales from thermodynamic viewpoint. Subsequently, effects of various functionalities on breathing behaviors are appraised, including using post-synthetic modification routes. Finally, possible framework spatial requirements of MOFs for yielding breathing behaviors are highlighted as the design strategies for new syntheses. Keywords: metal-organic frameworks (MOFs); breathing; gas adsorption; gas storage; flexibility; post-synthetic modifications (PSM) 2. What Are MOFs? Metal-organic frameworks (MOFs), or called metal-ligand coordination polymers, are organic-inorganic hybrid extended networks that are constructed via covalent linkages between metal ions/metal clusters and organic ligands. The strategies of forming versatile frameworks utilize the principles of reticular synthesis [8]. The majority of metal ions/metal clusters are transition metals with various geometries, due to their versatile coordination numbers. These geometries include square–planer, tetrahedron, and octahedron among others [9]. The organic ligands contain halides, cyanides, neutral organic molecules (4,4′-bipyridine), and anionic organic molecules (benzenedicarboxylic acid). Each constituent is schematically drawn in Figure 1. Together, both organic and inorganic components can form n-dimensional (n = 1, 2, and 3) motifs which show flexibility upon interactions with guest molecules, schematically shown in Figure 2 [7,10–12]. Kitagawa’s group classifies three types of MOFs exhibiting flexibility [10]. For the first type, rigid 2D layers are covalently connected by flexible pillars. Upon the absorption of guest molecules, the elongation of pillars leads to the expansion of frameworks. Vice versa, frameworks shrink upon the removal of guest molecules (Figure 2a). Regarding the second type, the framework topology remains the same upon the adsorption of guest molecules. The rotation of organic moieties, caused by strong guest-host interactions, induces the volume change (Figure 2b). In the case of the third type, the interpenetrated 3D grids slide apart, causing the pores open or close upon the adsorption of guest molecules (Figure 2c). Materials 2014, 7 3200 Materials 2014, 7 eometrics of metal ions and organic ligands. Reprinted with permission from [9]. Figure 1. Geometrics of metal ions and organic ligands. Reprinted with permission from [9]. Copyright 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. Copyright 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. Figure 2. The illustration dynamic behaviors of the metal-organic frameworks’ (MOFs) structure upon interacting with guest molecules. (a) 2-dimension; (b) 1-dimension; (c) 3 dimension. Red spheres stand for the metals; lines stand for the organic linkers. Reprinted with permission from [11] Copyright 2013 American Chemical Society Figure 2. The illustration dynamic behaviors of the metal-organic frameworks’ (MOFs) structure upon interacting with guest molecules. (a) 2-dimension; (b) 1-dimension; (c) 3 dimension. Red spheres stand for the metals; lines stand for the organic linkers. Reprinted with permission from [11]. Copyright 2013 American Chemical Society. Figure 2. The illustration dynamic behaviors of the metal-organic frameworks’ (MOFs) structure upon interacting with guest molecules. (a) 2-dimension; (b) 1-dimension; (c) 3 dimension. 2. What Are MOFs? Red spheres stand for the metals; lines stand for the organic linkers. Reprinted with permission from [11]. Copyright 2013 American Chemical Society. 3201 Materials 2014, 7 Materials 2014, 7 The first coordination polymer, Ni(CN)2(NH3)·C6H6, was synthesized by Hofmann and Küspert in 1897 [13], called Hofmann Complex. The coordination polymer was formed by reacting slowly of C6H6 with Ni(CN)2 in a NH3 solution. The structure of coordination polymer was later refined using X-ray crystallography, as a two-dimensional layer structure formed by covalent linkages between Ni and CN groups [14]. The parallel layers are held up by extruding NH3 groups, forming interplanar cavities in which reside benzene molecules. The work inspired research towards designing various inorganic-organic hybrid polymers using diamines, such as 4,4′-bipyridine. The adoption of longer chain length of organic linkers in MOFs’ framework enlarged the interplanar space by 33%, from 8.28 Å to 10.9 Å. The framework with longer organic ligands creates larger cavities for molecule encapsulation. In addition to aforementioned bidentate linear organic molecules, other multidentate organic linkers, such as 4,4′,4′′,4′′′-tetracyanotetraphenylmethane was first applied by Robson’s group to yield 3-dimensional infinite framework with a large unit cell volume of 4200 Å [15,16], in which about two-thirds of the volume is occupied by removable solvents molecules. The refined crystallographic structure is shown in Figure 3. Robson’s work shed the light of controlling the cavity size of MOFs with predetermined size and shape of inorganic nodes, and organic linkers. Figure 3. Crystallographic structure of coordination polymer, CuI[C(C6H4·CN)4]n. An adamantine-like cavity is highlighted using bold lines. Reprinted with permission from [15]. Copyright 1989 American Chemical Society. Researches on MOFs have seen explosively increasing since 2000 [9,17]. In the open literature, there are several families of MOFs systematically synthesized and coined their nomenclatures using the abbreviation of institution’s name. Pioneer works done by Yaghi and O’Keefe has named their resulting solids as MOF–n (where n = 1, 2, 3, and etc.) [8]. Others include MILs (Matériaux de l’lnstitut Lavoisier), HKUSTs (Hong-Kong University of Science and Technology), POST (Pohang University of Science and Technology), etc. Researches on MOFs have seen explosively increasing since 2000 [9,17]. In the open literature, there are several families of MOFs systematically synthesized and coined their nomenclatures using the abbreviation of institution’s name. Pioneer works done by Yaghi and O’Keefe has named their resulting solids as MOF–n (where n = 1, 2, 3, and etc.) [8]. Others include MILs (Matériaux de l’lnstitut Lavoisier), HKUSTs (Hong-Kong University of Science and Technology), POST (Pohang University of Science and Technology), etc. Researches on MOFs have seen explosively increasing since 2000 [9,17]. 2.1. Single-Crystal-to-Single-Crystal (SCSC) Phenomena The application of MOFs on gas adsorption was first reported using a three-dimensional framework: M2(4,4′-bpy)3(NO3)4·xH2O with M = Co, Ni, Zn and 4,4′-bpy = 4,4′-bipyridine [18]. The MOFs were synthesized at room temperature for 7 days in an acetone/EtOH mixture. Detail X-ray crystallography analyses indicated that the Co center forms pentagonal bipyramid geometry, bridged by 4,4′-bpy organic linkers. The linkage forms infinite 1-dimensional chains with an interdigited, or tongue-and-groove, structure. The cavities have a window size of 3 Å × 6 Å and 3 Å × 3 Å along a and b axes respectively, which were suitable for gas adsorption. The authors proposed that the adsorption feature evolved from the affinity of organic linkers inside the framework cavities. The adsorption properties can be further fine-tuned via the guest-host interaction using versatile organic linkers [19,20]. Kitagawa’s group adopted various framework regulators, AF6 − anions (A = Si, Ge, and P), to manipulate selectively the framework connectivity and pore dimensions upon hydrolysis. The resulted compounds demonstrated a higher adsorption capacity towards methane. It is proposed that the guest-host interaction and the size exclusion effect between the pore size and kinetic diameters of adsorbates lead to the selectivity of gas adsorption. However, it is noted that no structure transformation of MOFs was observed, meaning that the framework exhibits the rigidity upon the adsorption of gas molecules. The guest-host interaction also accounts for the single-crystal-to-single-crystal (SCSC) transformation that was first successfully discovered in [Ni2(4,4′-bpy)3(NO3)4]·6EtOH upon the removal of solvent [21]. Their detail single crystal X-ray diffraction data indicated that the unit cell volume shrunk 2.3% after encapsulated ethanol was evaporated or removed from the channels while the framework crystallinity still remains integrated. At room temperature, the adsorptions of H2O, MeOH, and EtOH of the MOF were demonstrated to be reversible, showing the inhaling-exhaling cycle, despite the relative small difference of the unit cell volume. A similar behavior was noticed later using mixed organic ligands, 2,5-dicarboxypyridine (H2pydc) and 4,4′-bipyridine (4,4′-bpy). A MOF of [Fe(pydc)(4,4′-bpy)]·H2O was synthesized hydrothermally and was characterized as a non-interpenetrated square-grid structure with a cavity size of 11.5 Å × 8.8 Å [22]. The authors reported a larger magnitude of volume contraction upon the removal of solvent, 6.6%–8.2%, possibly owing to strong edge-to-face aromatic π–π stacking interactions between the aromatic groups of pydc and 4,4′-bpy in adjacent layers [22]. Materials 2014, 7 In the open literature, there are several families of MOFs systematically synthesized and coined their nomenclatures using the abbreviation of institution’s name. Pioneer works done by Yaghi and O’Keefe has named their resulting solids as MOF–n (where n = 1, 2, 3, and etc.) [8]. Others include MILs (Matériaux de l’lnstitut Lavoisier), HKUSTs (Hong-Kong University of Science and Technology), POST (Pohang University of Science and Technology), etc. Materials 2014, 7 Materials 2014, 7 3202 2.1. Single-Crystal-to-Single-Crystal (SCSC) Phenomena 2.1. Single-Crystal-to-Single-Crystal (SCSC) Phenomena The doubly-bridged layer structure exhibits its robustness and sliding of layers during the SCSC transformation. The unit cell volume change of MOF crystals caused by the SCSC transformation becomes noticeable in {(ZnI2)3(TPT)2·6C6H5NO2}n, where TPT stands for 2,4,6-tris(4-pyridyl)triazine. The geometry of zinc atoms is tetrahedral, linking to two iodine atoms and two TPT molecules individually. The network adopts a (10,3)-b configuration, in which contains 10 TPT molecules and 10 zinc atoms, forming an interpenetrated 3-D framework, shown in Figure 4. The randomly-packed nitrobenzene molecules occupy 60% of the unit cell volume. The Bravais lattice of the interpenetrated framework changes from monoclinic to triclinic upon the removal of solvent, yielding the unit cell volume contraction of 20%–23% [23]. The schematic diagram is illustrated in Figure 5. The shrinkage of unit cell volume is owing to shorter distances between Zn-Zn and ligand-ligand, changing from 6.9 Å to 5.2 Å and 8.2 Å to 5.1 Å respectively. It is noteworthy that the MOF exhibited a reversible swelling 3203 Materials 2014, 7 Materials 2014, 7 behavior as solvent re-entered the channels and showed a wide range of chemical stability upon exposing to various organic solvents, such as mesitylene, and cis-stilbene among others. This type of flexibility evolved from the crystal structure transformation is classified as the breathing behavior [24]. Figure 4. The crystal structure of {(ZnI2)3(TPT)2·6C6H5NO2}n. (Left) the net formed by ten molecules of TPT and ten Zn atoms. (Right) the interpenetrated framework viewed along the b-axis. C: gray, N: blue, Zn: magenta. Reprinted with permission from [23]. Copyright© 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. Figure 5. Schematic drawing of SCSC of {(ZnI2)3(TPT)2·6C6H5NO2}n upon the removal of nitrobenzene as the guest molecules. The contraction of unit cell volumes is viewed along the (100) direction. Reprinted with permission from [23]. Copyright© 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. Figure 5. Schematic drawing of SCSC of {(ZnI2)3(TPT)2·6C6H5NO2}n upon the removal of nitrobenzene as the guest molecules. The contraction of unit cell volumes is viewed along the (100) direction. Reprinted with permission from [23]. Copyright© 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. A similar solvent-induced reversible framework transition was reported for MOROF-1, a mesoporous MOF of which a polychlorinated triphenylmethyl radicals(PTM), functionalized with hree carboxylic groups (TC), was used as the organic ligand (PTMTC), illustrated in Figure 6a [25]. A similar solvent-induced reversible framework transition was reported for MOROF-1, a mesoporous MOF of which a polychlorinated triphenylmethyl radicals(PTM), functionalized with three carboxylic groups (TC), was used as the organic ligand (PTMTC), illustrated in Figure 6a [25]. A similar solvent-induced reversible framework transition was reported for MOROF-1, a mesoporous MOF of which a polychlorinated triphenylmethyl radicals(PTM), functionalized with three carboxylic groups (TC), was used as the organic ligand (PTMTC), illustrated in Figure 6a [25]. A similar solvent-induced reversible framework transition was reported for MOROF-1, a mesoporous MOF of which a polychlorinated triphenylmethyl radicals(PTM), functionalized with three carboxylic groups (TC), was used as the organic ligand (PTMTC), illustrated in Figure 6a [25]. 3204 Materials 2014, 7 The occurrence of SCSC transformation upon the solvent removal and re-adsorption clearly indicates the flexibility of metal-organic frameworks. The feature is unique and has not been observed in rigid porous materials, such as zeolites. Some authors called this phenomena “breathing”, which is described as a reversible structural transition of organic-inorganic hybrid materials whose unit cell dimensions experience “large variations (>5 Å)” upon the exposure of external stimuli [26,27]. Given their framework flexibility and selectivity of solvent re-adsorption, it is generally recognized that the flexible MOFs would find larger applications in gas separation, sequestration and purification if they exhibit similar breathing behaviors while gas molecules are external stimuli. Materials 2014, 7 The adoption of the multidentate bulky linkers was designated for the formation of open framework with even larger pore diameter. The copper metal centers adopt a geometry of square pyramidal polyhedrons, and link with ligands, forming a two-dimensional honeycomb layer structure (Figure 6b). Different layers are organized via π–π and van der Waals interactions, yielding an open framework (Figure 6c). The window sizes measured along (001) and (100) directions are about 2.8–3.1 nm and 0.5–0.7 nm, respectively. The MOF experienced a huge volume contraction, 25%–35%, upon the removal of solvent and became an amorphous closed-pore material. The MOF exhibited high selectivity towards methanol and ethanol than other several organic solvents upon re-adsorption. Figure 6. (a) The structure of copper(II) tricarboxylate; (b) the connection of building blocks to form a hexagonal pore structure; (c) the mesoporous structure formed by the stacking of hexagonal pore structure along (001) direction. Cu atom: orange; C atom: light blue and violet; O atom: red; Cl atom: green; and N atom: navy. Reprinted with permission from [25], Copyright 2003 Macmillan Publishers Ltd: Nature Materials. 3205 2.2. Hydrothermal Synthesis versus Solvothermal Synthesis Materials 2014, 7 3206 Materials 2014, 7 Figure 7. (a) Non-interpenetrated structure of [Cd(bipy)(bdc)] formed using low concentration of reactants; (b) Interpenetrated structure of [Cd(bipy)(bdc)] formed using high concentration of reactants. Reprinted with permission from [39]. Copyright 2009 American Chemical Society. The pH of starting precursor solutions plays significant roles at the structure of final products as well. From the system of Al3+/H4btec/H2O/NaOH, where H4btec represents pyromellitic acid, MIL-118 and MIL-121 were formed under low pH environment, while MIL-120 was the preferable phase at higher pH synthesis conditions [40]. It was proposed, based on detail crystallography analyses, that corner-sharing [AlO6] octahedra were formed under low pH solutions; however, edge-sharing alumina polyhedra were observed at high pH environment. The polarity of solvent affects profoundly the final structure formation perhaps via influencing The pH of starting precursor solutions plays significant roles at the structure of final products as well. From the system of Al3+/H4btec/H2O/NaOH, where H4btec represents pyromellitic acid, MIL-118 and MIL-121 were formed under low pH environment, while MIL-120 was the preferable phase at higher pH synthesis conditions [40]. It was proposed, based on detail crystallography analyses, that corner-sharing [AlO6] octahedra were formed under low pH solutions; however, edge-sharing alumina polyhedra were observed at high pH environment. The pH of starting precursor solutions plays significant roles at the structure of final products as well. From the system of Al3+/H4btec/H2O/NaOH, where H4btec represents pyromellitic acid, MIL-118 and MIL-121 were formed under low pH environment, while MIL-120 was the preferable phase at higher pH synthesis conditions [40]. It was proposed, based on detail crystallography analyses, that corner-sharing [AlO6] octahedra were formed under low pH solutions; however, edge-sharing alumina polyhedra were observed at high pH environment. The polarity of solvent affects profoundly the final structure formation, perhaps via influencing the solubility of organic linkers as well as its proteolysis properties [41]. For the system of Fe3+/H2BDC-NH2/solvent (H2BDC-NH2 = 2-amino-terephthalic acid), only Fe-MIL-88B-NH2 was formed when acetonitrile and methanol were used as solvents [42]. Only Fe-MIL-53-NH2 was obtained if water was the solvent. However, a mixture Fe-MIL-88B-NH2 and Fe-MIL-101-NH2 was isolated when N,N’-dimethylformamide (DMF) was used as the solvent. Similarly, pure Cr-MIL-100 was obtained from Cr3+/H3BTC/H2O system (H3BTC = 1,3,5-benzenetricarboxylic acid), and Cr-MIL-96 co-existed with Cr-MIL-100 when a small quantity of methanol was added into H2O [43]. Further increase of methanol content led to the formation of pure Cr-MIL-96 phase. 2.2. Hydrothermal Synthesis versus Solvothermal Synthesis Syntheses of Metal-organic frameworks can take place at room temperature, or using solvothermal synthesis, microwave-assisted synthesis, electrochemical synthesis, or sonochemical synthesis among others [28,29]. Being the most prevail approach to form novel solid state materials by materials scientists and chemists, the solvothermal synthesis can be carried out using H2O or organic solvents as the medium [30]. The majority of this article focuses on the formations of MOFs from aqueous solutions, in contrast to those from organic solvents, because the former case is easy to scale-up for mass production, cost-effective, and environmental benign in terms of solvent [31,32]. Solvothermal techniques are widely applied to synthesize novel metal oxides materials from a reaction mixture inside a closed reaction vessel, under the supercritical temperature of the solvent which can be either aqueous or organic liquids [33,34]. Some researchers reckon the processes as hydrothermal syntheses while H2O is used as the solvent [35,36]. The main purposes of solvothermal techniques are to widen the synthesis avenue for synthesizing new materials as well as to manipulate the shape of material crystals. Compared with other synthesis methods such as ceramic, sol-gel techniques, the main advantages of solvothermal techniques are (1) simple one-pot/one-step preparation/synthesis; (2) ability to control the morphology of material crystals; (3) efficient synthesis conditions, with milder synthesis temperature and shorter synthesis duration, possibly due to reduced viscosity and dielectric constants [37]. Syntheses of MOFs originally adopted the precipitation route with careful diffusion control or slow evaporation of organic ligands and solvent molecules [15]. Later, syntheses of MOFs take up the solvothermal synthesis pathway which can overcome the solubility and integrity concerns of a desired secondary building unit (SBU) generated in situ in the reaction mixture. However, the structure of yielded materials is sensitive to subtle changes in concentration, solvent polarity, pH and temperature. The concentration of starting materials determined the structure of final products. For MOF-5 system, the dilution of reactants leads to the formation of non-interpenetrated structure (IRMOF-10, -12, -14, and -16), subsequently possessing higher porosity compared with individual concentrated counterpart [38]. A similar concentration-dependent trend of MOF structure was observed in the primitive cubic type (pcu) MOF of [Cd(bipy)(bdc)]·3DMF·H2O, whereas bipy = 4,4′-bipyridine, and bdc = 1,4-benzenedicarboxylic acid, shown in Figure 7 [39]. It is speculated that the formation of a sub-lattice in the voids of non-interpenetrated structures can be further reduced while lowering the reactant concentrations. 2.2. Hydrothermal Synthesis versus Solvothermal Synthesis Regarding the Al3+/H2BDC-NH2/DMF system, a mixture of Al-MIL-53-NH2 and Al-MIL-101-NH2 was observed when DMF was used as the solvent. However, increasing the H2O content or using H2O as the sole solvent ensured the formation of pure Al-MIL-53-NH2 [44]. Using in situ X-ray scattering technique illustrated that the formation of an intermediate phase, MOF-235, stabilized by DMF, leads to the crystallization of Al-MIL-101-NH2. The presence of H2O causes the hydrolysis of MOF-235 phase, which subsequently facilitate the formation of Al-MIL-53-NH2. The schematic diagram of propose mechanism is illustrated in Figure 8. Materials 2014, 7 3207 Materials 2014, 7 Figure 8. Schematic diagram of the effect of solvent polarity on final structure of MOF. C: gray; H: white; N: blue; O: red; Al: yellow; Cl: green. Reprinted with permission from [44]. Copyright© 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. Figure 8. Schematic diagram of the effect of solvent polarity on final structure of MOF. C: gray; H: white; N: blue; O: red; Al: yellow; Cl: green. Reprinted with permission from [44]. Copyright© 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. MOF. C: gray; H: white; N: blue; O: red; Al: yellow; Cl: green. Reprinted with permission from [44]. Copyright© 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. As scientists and engineers envision the promising performance of MOFs at the gas separation and purification, syntheses of MOFs at a large scale arises as another challenge and attracts research ttention [2,45,46]. It is widely recognized that several synthesis concerns have to be circumvented prior to practical applications, such as synthesis temperature and duration, yield, availability and cost of starting materials, synthesis steps, and quantity of solvents [28]. The role of solvent becomes critical nd it is highly desired to minimize or avoid the use of organic solvents. Given this concern, this work mainly focuses on reviewing the breathing behaviors of MOFs synthesized via the hydrothermal route. It is noteworthy that major MOFs are synthesized using organic solvents, except some MIL series. As summarized in Table 1, it is clearly demonstrated that the applied synthesis temperature is the highest among MOFs, and the synthesis duration is usually longer than its counterparts synthesized rom organic solvents, which are summarized in Table 2. As scientists and engineers envision the promising performance of MOFs at the gas separation and purification, syntheses of MOFs at a large scale arises as another challenge and attracts research attention [2,45,46]. 2.2. Hydrothermal Synthesis versus Solvothermal Synthesis It is widely recognized that several synthesis concerns have to be circumvented prior to practical applications, such as synthesis temperature and duration, yield, availability and cost of starting materials, synthesis steps, and quantity of solvents [28]. The role of solvent becomes critical and it is highly desired to minimize or avoid the use of organic solvents. Given this concern, this work mainly focuses on reviewing the breathing behaviors of MOFs synthesized via the hydrothermal route. As scientists and engineers envision the promising performance of MOFs at the gas separation and purification, syntheses of MOFs at a large scale arises as another challenge and attracts research attention [2,45,46]. It is widely recognized that several synthesis concerns have to be circumvented prior to practical applications, such as synthesis temperature and duration, yield, availability and cost of starting materials, synthesis steps, and quantity of solvents [28]. The role of solvent becomes critical and it is highly desired to minimize or avoid the use of organic solvents. Given this concern, this work mainly focuses on reviewing the breathing behaviors of MOFs synthesized via the hydrothermal route. It is noteworthy that major MOFs are synthesized using organic solvents, except some MIL series. As summarized in Table 1, it is clearly demonstrated that the applied synthesis temperature is the highest among MOFs, and the synthesis duration is usually longer than its counterparts synthesized from organic solvents, which are summarized in Table 2. It is noteworthy that major MOFs are synthesized using organic solvents, except some MIL series. As summarized in Table 1, it is clearly demonstrated that the applied synthesis temperature is the highest among MOFs, and the synthesis duration is usually longer than its counterparts synthesized from organic solvents, which are summarized in Table 2. 3208 Materials 2014, 7 Table 1. MOFs Synthesized from Hydrothermal Conditions. Table 1. MOFs Synthesized from Hydrothermal Conditions. 2.2. Hydrothermal Synthesis versus Solvothermal Synthesis MOF Class Metal Center Organic Linker Mineralizing Agent Temperature (°C) Time (h) Pore Diameter (Å) Space Group Bravis Lattice Reference MIL-47 VCl3 BDC – 200 96 as: 7.9 × 12.0 ht: 10.5 × 11.0 as: Pnma ht: Pnma as: orthorhombic ht: orthorhombic [47] MIL-53 Al(NO3)3·9H2O BDC – 220 72 as: 7.3 × 7.7 ht: 8.5 × 8.5 lt: 2.6 × 13.6 as: Pnma ht: Imma lt: Cc as: orthorhombic ht: orthorhombic lt: monoclinic [26] MIL-69 NDC KOH 210 16 H: 2.7 × 19.4 DeH: 8.5 × 8.5 C2/c Monoclinic [48] MIL-96 BTC – 210 24 Cavity free: 8.8 Pore opening: 2.5–3.5 P63/mmc Hexagonal [49] MIL-100 Fe metal HF 160 8 Cage: 25. 29 Window: 5, 9 Fd m Cubic [50,51] HNO3 12 – – – [45] FeCl3 6H2O – Fe(NO3)3 9H2O Al(NO3)3·9H2O HNO3 210 3.5 5.2 × 8.8 Fd3m Cubic [52] Cr metal HF 220 96 Internal: 6.6; External: 25 [50] MIL-110 Al(NO3)3·9H2O HNO3 210 72 16 P6322 Hexagonal [53] MIL-118 Al(NO3)3·9H2O H4BTEC – 210 24 as: 7.3 × 7.7 ht: 8.5 × 8.5 lt: 2.6 × 13.6 as: C2/c ht: Pbam lt: Pnam as: monoclinic ht: orthorhombic lt: rthorhombic [54] MIL-120 Al(NO3)3·9H2O H4BTEC NaOH 210 24 5.4 × 4.7 C2/c monoclinic [55] Materials 2014, 7 3209 Table 1. Cont. MOF Class Metal Center Organic Linker Mineralizing Agent Temperature (°C) Time (h) Pore Diameter (Å) Space Group Bravis Lattice Reference MIL-121 Al(NO3)3·9H2O H4BTEC – 210 24 8.7 × 5.7 C2/c monoclinic [40] MIL-122 Al(NO3)3·9H2O H4NTC – 210 24 – P21/c monoclinic [56] Ga(NO3)3·xH2O * In(NO3)3·5H2O CuSO4∙5H2O L KBr 85 1.5 Hydrated: 18.99 × 15.14 Dehydrated: 18.56 × 12.81 P triclinic [11] Cu-MOF- SiF6 CuSiF6 BPED – rt 72 13.5 × 13.5 P4/ncc Tetragonal [57] ELM-12 Cu(OTf)2 4,4′-bpy MeOH rt >168 11 × 11.5 Pbna Orthorhombic [58] EtOH Pn21a 1-PrOH Pnna 1-BuOH * Cu(NO3)3·3H2O HDHBC 4,4′-bpy Diethyl ether rt a few days 3.6 × 4.2 P2/c Monoclinic [59] 4,4′-bpy = 4,4′-bipyridine; BDC = 1,4-benzene-dicarboxylic acid; BTC = 1,3,5-benzene-tricarboxylic acid; BPED = meso-1,2-bis(4-pyridyl)-1,2-ethanediol; H4BTEC = 1,2,4,5-benzene-tetracarboxylic acid; H4NTC = 1,4,5,8,-naphthalene-tetracarboxylic acid; HDHBC = 2,5-dihydroxybenzoic acid; L = 4,4′-(1,4-(trans-2- butene)diyl)bis(1,2,4-triazole); NDC = 2,6-naphthalene-dicarboxylic acid; OTf = trifluoromethanesulfonate; rt: room temperature; * haven’t been named. Materials 2014, 7 3210 Table 2. MOFs Synthesized from Solvothermal Conditions. 2.2. Hydrothermal Synthesis versus Solvothermal Synthesis MOF Class Metal Center Organic Linker Solvent Mineralizing Agent Temprature (°C) Time (h) Pore Diameter (Å) Space Group Bravis Lattice Reference MOF-5 Zn(NO3)2·6H2O BDC DMF C6H15N; C6H5Cl; H2O2 rt 0.5–4 8 Fm3m Cubic [60,61] Zn(OAc)2·2H2O BDC DMF C6H15N rt 2.5 7.56 Fm3m Cubic [62] MOF-74 Zn(OAc)2·2H2O H4DHTP DMF – rt 18 – – – [62] Zn(NO3)2∙4H2O 2-Propanol H2O 105 20 10.3 × 5.5 R Trigonal [63] MOF-177 Zn(NO3)2∙6H2O H3BTB DEF – 100 23 10.8 P 1c Trigonal [62,64] MOF-199 (HKUST-1) Cu(NO3)2∙2.5H2O BTC DMF C6H15N; EtOH; H2O 85 24 Cage: 13.2 × 11.1; Aperture: 6.9 Fm3m Cubic [64–66] IR-MOF-0 Zn(OAC)2·2H2O ADC DMF C6H15N rt over- night – Fm3m Cubic [62] MIL-88 FeCl3·6H2O BDC DMF – 150 0.2 3 × 14 P 2c Hexagonal [67,68] * Co(CF3SO3)2 H2BDP DEF – 150 144 – C2/c Monoclinic [69] CFA2 Cu(OAc)2·H2O H2PHBPZ DEF; MeOH – 110 72 12–14.5 I41/a Tetragonal [70] CFA3 Ag2O H2PHBPZ EtOH NH4OH rt 120 – P21/c Monoclinic [70] * Zn(NO3)2∙6H2O TCPOM 4,4′-bpy DMF – 100 24 – C2/c Monoclinic [71] * Zn(NO3)2∙6H2O BDC DABCO DMF – 120 48 7.5 × 7.5 I4/mcm Tetragonal [72,73] * Cd(NO3)2∙4H2O BPNDC 4,4′-bpy DMF – 120 24 Cavity: 12 × 7 × 4; Aperture: 2 × 6 C2/c Monoclinic [74] * La(NO3)2·6H2O H3BTB DMF – 95 48 – P6522 Hexagonal [75] * Zn(NO3)2∙6H2O 2,5-BME-BDC; DABCO DMF – 120 48 – C2/m Monoclinic [76] Table 2. MOFs Synthesized from Solvothermal Conditions. MOF Class Metal Center Organic Linker Solvent Mineralizing Agent Tem Materials 2014, 7 3211 Table 2. Cont. Table 2. Cont. Table 2. Cont. 2.2. Hydrothermal Synthesis versus Solvothermal Synthesis MOF Class Metal Center Organic Linker Solvent Mineralizing Agent Temprature (°C) Time (h) Pore Diameter (Å) Space Group Bravis Lattice Reference * Zn(NO3)2∙6H2O 2,3-BME- BDC; DABCO DMF – 120 48 – C2/m Monoclinic [76] * Zn(NO3)2∙6H2O DE-BDC; DABCO DMF – 120 48 – C2/m Monoclinic [76] * Zn(NO3)2∙6H2O DP-BDC; DABCO DMF – 120 48 – C2/m Monoclinic [76] * Zn(NO3)2∙6H2O DiP-BDC; DABCO DMF – 120 48 – C2/m Monoclinic [76] * Zn(NO3)2∙6H2O DB-BDC; DABCO DMF – 120 48 – P4/mmm Tetragonal [76] * Zn(NO3)2∙6H2O DPe-BDC; DABCO DMF – 120 48 – I4/mcm Tetragonal [76] * Zn(NO3)2∙6H2O BA-BDC; DABCO DMF – 120 48 – P4/mmm Tetragonal [76] * Zn(NO3)2∙6H2O BPy-BDC; DABCO DMF – 120 48 – P4/mmm Tetragonal [76] CID-23 Zn(NO3)2∙6H2O BPA; IP DMF – 120 48 5.6 × 8.6 P triclinic [77] FMOF-1 AgNO3 NaTz MeOH – rt n/a 12.2 × 7.3; 4.9 × 6.6 I d Tetragonal [78,79] * Zn(OAc)2·2H2O H3PBC; 2,2’-bpy DMF – 170 96 4.1 × 4.1; 2.7 × 2.7 C2/c Monoclinic [80] 3212 Materials 2014, 7 Table 2. Cont. Table 2. Cont. 2.2. Hydrothermal Synthesis versus Solvothermal Synthesis MOF Class Metal Center Organic Linker Solvent Mineralizing Agent Temprature (°C) Time (h) Pore Diameter (Å) Space Group Bravis Lattice Reference UCY-3 Cd(NO3)2 4H2O H3CIP DMF – 100 20 5–7 C2/c Monoclinic [81] FMOF-2 Zn(NO3)2∙6H2O H2hfipbb DMF; EtOH – 110 a few hours – P2/n Monoclinic [82] SUMOF-6 Ln(NO3)6 salt H2BPYDC DMF – 120 24 19.0 × 5.3 P2/n Monoclinic [83] * CuCl2·2H2O H2DMCAPZ DMF; MeOH – 90 24 6.7 × 7.2 9.7 × 9.6 I41md Tetragonal [84] * Zn(NO3)2∙6H2O BDC; BTZ; PYZ DMA; MeOH – 120 120 13.5 × 8.5 C2m Monoclinic [85] * Zn(NO3)2∙6H2O BDC; TZ; PYZ DMF – 120 120 14.5 × 5.0 P Triclinic [85] 2,2’-bpy = 2,2’-bipyridine; 2,5-BME-BDC = 2,5-bis(2-methoxyethoxy)-1,4-benzenedicarboxylic acid; 2,3-BME-BDC = 2,3-bis(2-methoxyethoxy)-1,4-benzenedicarboxylic acid; 4,4′-bpy = 4,4′-bipyridine; ADC = Acetylenedicarboxylic acid; BA-BDC = 2,5-Bis(allyloxy)-1,4-benzenedicarboxylic acid; BDC = 1,4-benzene-dicarboxylic acid; BPA = 1,4-bis(4-pyridyl)acetylene; BPNDC = Benzophenone-4,4′-dicarboxylate; BPy-BDC = 2,5-Bis(prop-2-ynyloxy)-1,4-benzenedicarboxylic acid; BTC = 1,3,5-benzene- tricarboxylic acid; BTZ = 1H-benzotriazole; DABCO = 1,4-diazabicyclo-[2.2.2]octane; DB-BDC = 2,5-di-butoxy-1,4-benzenedicarboxylic acid; DE-BDC = 2,5-di- ethoxy-1,4-benzenedicarboxylic acid; DEF = N,N-diethylformamide; DP-BDC = 2,5-di-propoxy-1,4-benzenedicarboxylic acid; DPe-BDC = 2,5-di-pentoxy-1,4- benzenedicarboxylic acid; DiP-BDC = 2,5-di-iso-propoxy-1,4-benzenedicarboxylic acid; DMA = N,N-dimethylacetamide; DMF = N,N-dimethylformamide; EtOH = Ethanol; H2BDP = 1,4-benzene-dipyrozolate; H2BPYDC = 2,2’-bipyridine-5,5’-dicarboxylic acid; H2DMCAPZ = 3,5-Dimethyl-4-Carboxypyrazole; H2hfipbb = 2,2’-bis(4- carboxyphenyl)hexafluoropropane; H2PHBPZ = 3,3’,5,5’-tetraphenyl-1H,1’H-4,4′-bipyrazole; H3CIP = 5-(4-carboxybenzylideneamino)isophthalic acid; H3PBC = 4- phosphono-benzoic acid; H3BTB = Benzenetribenoic acid or 1,3,5-tris(4-carboxyphenyl)benzene acid; H4DHTP = 2,5-dihydroxyterephtalic acid; IP = Isophthalic acid; MeOH = Methanol; NaTz = sodium perfluorinated ligand 3,5-bis(trifluoromethyl)-1,2,4-triazolate; PYZ = pyrazine; TCPOM = Tetrakis[4-(carboxyphenyl)oxamethyl]methane; TZ = 1,2,3-1H-triazole; * haven’t been named. Materials 2014, 7 Materials 2014, 7 Materials 2014, 7 3213 3. Characterizations of Breathing Behaviors Despite the variety of MOFs, only several classes exhibit the unique breathing phenomena. For example, MIL-53 series are formed by connections of corner-sharing MO4(OH)2 octahedra linked by 1,4-benzenedicarboxylic (BDC) acids. Since MIL-53 (Al, Cr) are synthesized hydrothermally, denoted as as-synthesized MIL-53 (MIL-53as), the channels of MOFs are filled with disordered BDC and H2O molecules, demonstrating the narrow-pore (np) form. Shown in Figure 9, this np structure is caused by the hydrogen-bond interactions between the hydrogen atoms of the water molecules and the oxygen atoms of the carboxylic group and the μ2-hydroxo group [86]. Upon the dehydration at elevated temperatures, the MIL-53s exhibit a porous structure which is referred as the large-pore (lp) form, due to the absence of the interactions. Figure 9. Illustration of the breathing behavior of MIL-53 using heat as the external stimulus. (a) dehydrated lp form; (b) hydrated np form. Reprinted with permission from [87]. Copyright 2013 American Chemical Society. An unprecedented adsorption behavior is observed as the hydrated MIL-53s are exposed to CO2 molecules. The hydrated MIL-53s remain the np form when CO2 pressure is below 5 bar. Further increasing the gas pressure turns the framework into the coexistence of np and lp forms. The framework turns into a pure lp form when the gas pressure reaches above 15 bar (about 7.2 mmol∙g−1), shown in Figure 10. The unit cell volume expands from 1012.8 Å3 (hydrated) to 1522.5 Å3 (hydrated + CO2). This behavior is different from their dehydrated counterparts that exhibit two-step structure transition upon CO2 adsorption (lp → np → lp), during which the first plateau occurs at around 3 mmol∙g−1 while the second plateau starts evolving around 8 bar (7.7 mmol∙g−1). The lag of the structure transition might be due to the steric hindrance in the presence of H2O molecules prior to the CO2 adsorption. Furthermore, the hydrated MIL-53(Cr) shows the minute CH4 uptake (0.2 mmol∙g−1 at 20 bar), compared to its dehydrated counterpart (4.6 mmol∙g−1). This behavior is projected to have wide applications in gas storage and separation, such as the selective adsorption of CO2 over CH4 [88,89]. An unprecedented adsorption behavior is observed as the hydrated MIL-53s are exposed to CO2 molecules. The hydrated MIL-53s remain the np form when CO2 pressure is below 5 bar. Further increasing the gas pressure turns the framework into the coexistence of np and lp forms. The framework turns into a pure lp form when the gas pressure reaches above 15 bar (about 7.2 mmol∙g−1), shown in Figure 10. The unit cell volume expands from 1012.8 Å3 (hydrated) to 1522.5 Å3 (hydrated + CO2). This behavior is different from their dehydrated counterparts that exhibit two-step structure transition upon CO2 adsorption (lp → np → lp), during which the first plateau occurs at around 3 mmol∙g−1 while the second plateau starts evolving around 8 bar (7.7 mmol∙g−1). The lag of the structure transition might be due to the steric hindrance in the presence of H2O molecules prior to the CO2 adsorption. Furthermore, the hydrated MIL-53(Cr) shows the minute CH4 uptake (0.2 mmol∙g−1 at 20 bar), compared to its dehydrated counterpart (4.6 mmol∙g−1). This behavior is projected to have wide applications in gas storage and separation, such as the selective adsorption of CO2 over CH4 [88,89]. Materials 2014, 7 3214 Figure 10. (a) The illustration of structural change of hydrated MIL-53(Cr) using in situ X-ray diffraction with respect to the pressure cycle of CO2; (b) Schematic diagram of corresponding breathing behavior of MIL-53(Cr). Reprinted with permission from [86]. Copyright 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. Figure 10. (a) The illustration of structural change of hydrated MIL-53(Cr) using in situ X-ray diffraction with respect to the pressure cycle of CO2; (b) Schematic diagram of corresponding breathing behavior of MIL-53(Cr). Reprinted with permission from [86]. Copyright 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. 3.1. Porosity Measurements 3.1. Porosity Measurements The structural transformation of MOFs, displaying the breathing behavior, is induced by the dipole moment interaction between guest molecules and the host framework. The transition affects the gas adsorption capacity that can be quantified using porosity measurements. Measurements using powder X-ray diffraction (PXRD), and 129Xe NMR are conducted to unveil the structure transition at the atomic level, providing a comprehensive porosity information of the breathing dynamics [90]. The gas adsorption isotherms are classified by International Union of Pure and Applied Chemistry (IUPAC) into six types, shown in Figure 11 [91]. Type I is the Langmuir isotherm commonly observed in microporous materials, of which the steep increase of adsorbed quantity at low relative pressure indicates that the available microporous volume is occupied. Type-II, and -IV isotherms are possible similar materials, in which mesopores and macropores are present. In addition, the multi-layer adsorption might occur in the middle relative pressure range. The subtle difference between these two types of isotherms is the presence of hysteresis loop in Type IV, in which implies the occurrence of capillary condensation within mesopores, due to the strength of adsorbate–adsorbent and adsorbate–adsorbate interactions [24]. Type-III and -V isotherms indicate the weak adsorbate-adsorbent interactions. Type-VI isotherm might associate with layer-by-layer adsorption on a uniform surface. As aforementioned, dehydrated MIL-53 (Al, Cr) demonstrate the lp form. The N2 sorption measurements show Type I isotherms, indicating their microporosity [26,92,93]. However, dehydrated MIL-53 (Fe) and MIL-53 (Sc) display no microporosity towards the specific stimuli, as indicated in Figure 12 [94–96]. This is attributed to the presence of closed-pore (cp) form or very-narrow-pore (vnp) form of these two MIL-53s [97]. Similar Type I isotherm can be observed in the dehydrated MIL-47(V) [98], which is an isostructural analogue of MIL-53, formed by connecting V4+O6 octahedra with 1,4-benzenedicarboxylic (BDC) acids [99]. Materials 2014, 7 Materials 2014, 7 3215 Figure 11. Six types of gas physisorption isotherms. Reprinted with permission from [100]. Copyright 1985 IUPAC. Figure 11. Six types of gas physisorption isotherms. Reprinted with permission from [100]. Copyright 1985 IUPAC. Figure 12. Adsorption isotherms of n-hexane of various modified MIL-53(Fe). MIL-53(Fe)-4H stands for the non-modified MIL-53(Fe). Reprinted with permission from [101]. Copyright 2011 American Chemical Society. Figure 12. Adsorption isotherms of n-hexane of various modified MIL-53(Fe). MIL-53(Fe)-4H stands for the non-modified MIL-53(Fe). Reprinted with permission from [101]. Copyright 2011 American Chemical Society. However, if CO2 is used as the adsorbate, a Type VI isotherm is noticed in Figure 13, in which a step or an inflection point can be observed. The presence of the inflection point in isotherms could be evolved from changes in MOFs’ framework [102] or from the electrostatic interactions between adsorbate molecules [103]. Theoretical approaches from thermodynamic viewpoints support that the However, if CO2 is used as the adsorbate, a Type VI isotherm is noticed in Figure 13, in which a step or an inflection point can be observed. The presence of the inflection point in isotherms could be evolved from changes in MOFs’ framework [102] or from the electrostatic interactions between adsorbate molecules [103]. Theoretical approaches from thermodynamic viewpoints support that the 3216 Materials 2014, 7 Materials 2014, 7 step likely associates with the structural transition between np and lp forms [99]. As a result, the structural flexibility and breathing behaviors are characterized with the presence of the step in the adsorption isotherm and are commonly noticed in MIL-53s (Al, Cr, Fe) [96,99,104]. However, this step is absent from the CO2 adsorption isotherm of MIL-47, possibly due to the presence of the μ2-oxo group of the metal center [99], yielding its rigid framework. Figure 13. Isotherms at 304 K for the adsorption of CH4 and CO2. (a) MIL-53(Al); and (b) MIL-53(Cr). Reprinted with permission from [99]. Copyright 2005 American Chemical Society. Figure 13. Isotherms at 304 K for the adsorption of CH4 and CO2. (a) MIL-53(Al); and (b) MIL-53(Cr). Reprinted with permission from [99]. Copyright 2005 American Chemical Society. More complex breathing behaviors can be observed from the adsorption of long-chain alkanes on MIL-53s [101,105,106]. For MIL-53(Al, Cr), the stepped adsorption isotherms can be observed when propane or longer alkanes are used as adsorbates, shown in Figure 14. The occurrence of stepped adsorption isotherm shifts towards a higher relative pressure when Al is the metal node as well as longer alkanes. It was speculated that the presence of stepped isotherms was caused by host-guest interaction, entropic effects, and confinement effects. F h h i Fi 13 d 15 h b f d d i i h h More complex breathing behaviors can be observed from the adsorption of long-chain alkanes on MIL-53s [101,105,106]. For MIL-53(Al, Cr), the stepped adsorption isotherms can be observed when propane or longer alkanes are used as adsorbates, shown in Figure 14. The occurrence of stepped adsorption isotherm shifts towards a higher relative pressure when Al is the metal node as well as longer alkanes. It was speculated that the presence of stepped isotherms was caused by host-guest interaction, entropic effects, and confinement effects. More complex breathing behaviors can be observed from the adsorption of long-chain alkanes on MIL-53s [101,105,106]. For MIL-53(Al, Cr), the stepped adsorption isotherms can be observed when propane or longer alkanes are used as adsorbates, shown in Figure 14. 3.2. X-ray Diffraction In addition to gas adsorption measurements, the structural transition associating with the breathing behavior can be characterized using in situ X-ray diffraction [26,27,105,106]. From powder XRD patterns (PXRDs), the crystallographic refinement of MIL-53(Cr) crystals, under the hydration-dehydration cycle, yield the unit cell dimensions of MIL-53as (Cr) are as follows: orthorhombic system, Pnam, a = 17.340(1), b = 12.178(1), c = 6.822(1) Å, V = 1440.6(1) Å3; for MIL-53ht (Cr), orthorhombic system, Imcm, a = 16.733(1), b = 13.038(1), c = 6.812(1) Å, V = 1486.2(2) Å3; for MIL-53lt (Cr), monoclinic system, C2/c, a = 19.685(4), b = 7.849(1), c = 6.782(1) Å, V = 1012.8(1) Å3, β = 104.90(2)° [108]. The result clearly shows the volume contraction and expansion during the breathing cycle while using heat as the external stimulus. Similarly, the structural transitions of MIL-53(Al, Fe) can be observed using the same technique [26,109]. The structural transition upon adsorptions of CO2 was also analyzed using in situ PXRD patterns as: MIL-53-CO2 (Cr): monoclinic system, C2/c, a = 19.713(1), b = 8.310(1), c = 6.806(1) Å, V = 1072.5(1) Å3, β = 105.85(2)° [27]. Structural transitions were also analyzed for MIL-53s upon adsorptions of hydrocarbons [101,105,106]. Analyses of unit cell parameters of MIL-53 (Al, Fe, Cr, Ga, Sc) upon applying various external stimuli are summarized in Table 3. Materials 2014, 7 3219 The occurrence of stepped adsorption isotherm shifts towards a higher relative pressure when Al is the metal node as well as longer alkanes. It was speculated that the presence of stepped isotherms was caused by host-guest interaction, entropic effects, and confinement effects. Furthermore, shown in Figures 13 and 15, the absence of stepped adsorption isotherms when methane and ethane were used as the adsorbates suggested that the adsorption enthalpy of these adsorbates cannot overcome the threshold of adsorption enthalpy above which the breathing occurs (−20 kJ mol−1). In contrast, MIL-53(Fe, Sc) exhibits stepped isotherms for all tested linear alkanes, including methane and ethane. The adsorption isotherms exhibit multi-steps at various relative pressures, which have been attributed to the existence of four discrete pore openings changing from very narrow pore (vnp) → intermediate pore (int) → narrow pore (np) → large pore (lp). The schematic diagram is shown in Figure 16. It is noted that the stepped adsorption isotherms are absent in any hydrocarbon adsorption of MIL-47(V), further confirming the rigidity of its framework [107]. Materials 2014, 7 Materials 2014, 7 3217 Figure 14. Adsorption isotherms of n-alkanes at 313 K. (a) MIL-53(Al); (b) MIL-53(Cr). Δ: n-hexane; ○: n-heptane; □: n-nonane; ◊: n-octane. Reprinted with permission from [105]. Copyright 2008 American Chemical Society. Figure 14. Adsorption isotherms of n-alkanes at 313 K. (a) MIL-53(Al); (b) MIL-53(Cr). Δ: n-hexane; ○: n-heptane; □: n-nonane; ◊: n-octane. Reprinted with permission from [105]. Copyright 2008 American Chemical Society. (a) (b) (a) (a) (b) (b) Materials 2014, 7 3218 Materials 2014, 7 Figure 15. Adsorption isotherms of C1-C4 alkanes at 303 K. (a) MIL-53(Fe); (b) MIL-53(Cr). Reprinted with permission from [106]. Copyright 2009 American Chemical Society. Figure 16. Schematic diagram of structural evolutions of MIL-53(Fe) and MIL-53(Al, Cr) upon adsorption of n-alkanes. Reprinted with permission from [106]. Copyright 2009 American Chemical Society. (a) (b) (a) (a) (b) (b) Figure 16. Schematic diagram of structural evolutions of MIL-53(Fe) and MIL-53(Al, Cr) upon adsorption of n-alkanes. Reprinted with permission from [106]. Copyright 2009 American Chemical Society. Materials 2014, 7 3.3. 129Xe Nuclear Magnetic Resonance Xenon adsorption studies combined with 129Xe NMR spectroscopy are favorable methods for characterizations of porosity and framework flexibility of MOFs, due to the high polarizability of Xe atoms upon interacting with its environment [90,104,110]. For MIL-53(Al), adsorption isotherms of Xe below 300 K show steps and hysteresis loops, implying the structural transition from lp to np forms, shown in Figure 17. Figure 17. Xe adsorption (solid symbols) and desorption (open symbols) isotherms measured at 195 (lozenges), 220 (squares), 273 (triangles), 292 (circles), and 323 K (stars). Inset: Low-pressure region of the isotherm at 220 K. Reprinted with permission from [110]. Copyright 2010 American Chemical Society. 3220 Materials 2014, 7 Table 3. Summary Physical Properties of “Breathing” MIL-53s. Table 3. Summary Physical Properties of “Breathing” MIL-53s. MOF class Absorbent inside channels Channel Diameter (Å) Space group Unit cell dimension (Å) Bravis Lattice Surface Area (m2∙g−1) Unit Cell Volume (Å3) Reference a b c Al-MIL-53 as Disordered templating BDC 7.3 × 7.7 Pnma 17.129 6.628 12.182 Orthorhombic – 1383.1 [26,111,112] Al-MIL-53 ht Empty 8.5 × 8.5 Imma 6.608 16.675 12.813 Orthorhombic – 1411.9 [26,112,113] Al-MIL-53 lt H2O 2.6 × 13.6 Cc 19.513 7.612 6.576 Monoclinic 1592 (Langmuir) 1140 (BET) 946.8 [26,111,112] Al-MIL-53 CO2 – – 6.59 18.14 10.38 – – 1234.2 [113] Fe-MIL-53 ht Empty 6.759 C2/c 21.269 6.759 6.884 Monoclinic – 899.6 [114] Fe-MIL-53 lt H2O 7.518 C2/c 19.320 15.036 6.835 Monoclinic – 1973.5 [106,115] Cr-MIL-53 as Disordered templating BDC 8 Pnam 17.340 12.178 6.822 Orthorhombic 1400 (Langmuir) 1440.6 [27,116,117] Cr-MIL-53 ht Empty 8 Imcm 16.733 13.038 6.812 Orthorhombic – 1486.1 [27,116,117] Cr-MIL-53 lt H2O – C2/c 19.685 7.849 6.782 Monoclinic –– 1012.0 [27,116,117] Cr-MIL-53 LP CO2 – C2/c 19.713 8.310 6.806 Monoclinic – 1072.0 [27,116,117] C2H5OH 19.621 9.301 6.811 1185.1 [118] CH3OH 19.641 9.151 6.811 1168.1 Cr-MIL-53 HP CH3OH – Imcm 16.131 13.961 6.831 Orthorhombic – 1538.1 [118] CO2 16.439 13.500 1527.3 [27,116,117] C2H5OH 16.151 13.971 1541.1 [118] Ga-MIL-53 ht H2O – Imma 6.717 16.678 13.209 Orthorhombic 1462 (Langmuir) 1140 (BET) 1479.7 [119] Ga-MIL-53 lt H2O – C2c 19.833 6.856 6.714 Monoclinic – 886.3 [119] Sc-MIL-53-cp (100 K) Empty – P21/c 20.298 7.331 11.691 Monoclinic – 1680.8 [120] Sc-MIL-53-cp (573 K) Empty – P21/c 20.538 7.299 12.560 Monoclinic – 1804.9 [120] Sc-MIL-53-vnp (623 K) Empty – C2c 21.505 6.630 7.274 Monoclinic – 950.83 [120] 3221 4. Thermodynamic Viewpoints Adsorptions of various adsorbates (H2O, CO2, Xe, and n-alkanes) induce framework structural transitions due to adsorbate–adsorbent interactions, which is recognized as the breathing effect [124]. The striking breathing phenomena of MOFs lead to eminent steps and hysteresis in adsorption and desorption isotherms. Researchers also speculate that rationales of this unprecedented behavior via thermodynamic viewpoints can unveil strategies to “tailor” the MOFs’ gas sorption capacity. Coudert et al., recently developed thermodynamic viewpoints on framework transitions [125] in which they proposed that framework transitions are determined by five parameters: the free energy difference between the empty host structures (1, ΔFhost), the pore volumes (2, , and ), and the Henry constants (2, K1, and K2). These parameters consist of the widely-recognized osmotic potential ensemble model: (1) (1) whereas means the osmotic potential of the different solid phases during the guest-induced structural transitions of the host material. The guest-host interaction of CO2 adsorption evolves from the quadrupole moment of CO2 molecules. The CO2 adsorption on MIL-53(Al, Cr) exhibits two transition steps, changing from lp to np at low pressure (0.3 bar) and from np to lp at high pressure (5 bar). At low pressure region, the free energy difference was estimated about 2.5 kJ/mole. The energy barrier of the first transition is very comparable with kT at room temperature. The energy barrier of the second framework transition is about 20 kJ/mole at high pressure. This framework transition relies on: (1) the affinity between adsorbate-adsorbent; (2) pore volumes; and (3) the free energy difference between the narrow-pore (np) and the large-pore (lp) states. Using the model, the authors predicted four scenarios shown in Figure 19: (a) SCSC single transition ( ): unit cell volume expands as the pressure increases; (b) no framework transition ( and ): as the empty structure is the most stable phase; (c) SCSC double transition ( and ): unit cell volume contracts at the low pressure and expands at the high pressure region; (d) no framework transition ( and ): as van der Waals force interaction is too weak to offset the free energy of framework transition (2–6 kJ/mol). The model successfully explains the presence of dual framework transitions of MIL-53(Al, Cr) upon the CO2 adsorption, and the absence of framework transition upon CH4 adsorption [125]. Materials 2014, 7 129Xe NMR spectra show a single isotropic line at the chemical shift range of 60–70 ppm (region a) under low pressure in Figure 18. The presence of isotropic line indicates the framework adopts the lp form. As the adsorbed Xe quantity increases, the single isotropic line exhibits higher chemical shift and less intensity. Upon reaching a threshold pressure, a reversible anisotropic line with the chemical shift range of 120–160 ppm (region b) appears, indicating the framework becomes the np form. The structural transition could be induced by the interaction between Xe atoms and organic parts of MOFs, yielding the rotation of phenyl rings during shrinkage. Detailed NMR analyses indicate that the extent of structural transition is not complete at room temperature, comparable to the “forceps” effect observed in MIL-53(Fe) [121]. 129Xe NMR technique was used to characterize the framework flexibility of DUT-8(Ni), a MOF with Ni2(2,6-NDC)2(DABCO) (DUT = Dresden University of Technology, 2,6-NDC = 2,6-naphthalenedicarboxylate, DABCO = 1,4-diazabicyclo[2.2.2]octane) [122]. Figure 18. 129Xe NMR spectra measured under various xenon pressures. (A) Adsorption; (B) Desorption. Reprinted with permission from [110]. Copyright 2010 American Chemical Society. 3.4. Electron Paramagnetic Resonance (EPR) Using Nitroxide as Probes 3.4. Electron Paramagnetic Resonance (EPR) Using Nitroxide as Probes Recently, Bagryanskaya’s group developed an alternative approach to elucidating the generic breathing behavior of MIL-53(Al) using electron paramagnetic resonance (EPR) [123]. A very low concentration of (2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO) was introduced as the probe (1 molecule/1,000 unit cells). Its tumbling correlation times (τc) were analyzed at various temperatures, showing that the probe molecules become immobile as the framework turns into the np form at 80 K. The probe molecules become freely rotating as the framework changes into the lp form at room temperature. The obtained reversible breathing behavior of MIL-53(Al) was comparable with those obtained using aforementioned techniques. Materials 2014, 7 Materials 2014, 7 3222 Materials 2014, 7 Materials 2014, 7 Figure 19. The four potential scenarios of adsorbents exhibiting the structural transition between Phases 1 (red) and 2 (blue), using osmotic potential ensemble model. (Top) the adsorption isotherms; (Middle) the Langmuir isotherms of each phase in respective scenario; (Bottom) the osmotic potential of each phase. Green arrows indicate the occurrence of structural transitions. Reprinted with permission from [125]. Copyright 2008 American Chemical Society. American Chemical Society. The absence of framework transition upon CH4 adsorption was explicitly predicted using the concept of relative affinity, Knp/Klp, which is a function of pore volumes and the free energy difference between np and lp states [126]. As the relative affinity is much larger than unity, the osmotic potential difference switches sign twice, leading to the presence of double framework transitions between lp and np forms upon long-chain hydrocarbon adsorptions, shown in Figure 20. Due to the low relative affinity of CH4 (almost equal to unity), the lp form is thermodynamically stable, yielding no step of its adsorption isotherm [126]. The absence of framework transition upon CH4 adsorption was explicitly predicted using the concept of relative affinity, Knp/Klp, which is a function of pore volumes and the free energy difference between np and lp states [126]. As the relative affinity is much larger than unity, the osmotic potential difference switches sign twice, leading to the presence of double framework transitions between lp and np forms upon long-chain hydrocarbon adsorptions, shown in Figure 20. Due to the low relative affinity of CH4 (almost equal to unity), the lp form is thermodynamically stable, yielding no step of its adsorption isotherm [126]. The absence of framework transition upon CH4 adsorption was explicitly predicted using the concept of relative affinity, Knp/Klp, which is a function of pore volumes and the free energy difference between np and lp states [126]. As the relative affinity is much larger than unity, the osmotic potential difference switches sign twice, leading to the presence of double framework transitions between lp and np forms upon long-chain hydrocarbon adsorptions, shown in Figure 20. Due to the low relative affinity of CH4 (almost equal to unity), the lp form is thermodynamically stable, yielding no step of its adsorption isotherm [126]. 4. Thermodynamic Viewpoints For MIL-53, where the lp form is intrinsically more stable than the np form at room temperature, it was predicted that either the occurrence of two structural transitions upon gas adsorption or the absence of any transition is determined by a balance between intrinsic stability of the crystal structures, adsorption affinities, and accessible volume. With respect to alkanes adsorption of MIL-53, adsorption isotherms exhibit that the variation of pore volume for both forms is small [126] and the main factor is resulted from the change in adsorption affinities of different guests [126]. 3223 Materials 2014, 7 The absence of framework transition upon CH4 adsorption was explicitly predicted using the concept of relative affinity, Knp/Klp, which is a function of pore volumes and the free energy difference between np and lp states [126]. As the relative affinity is much larger than unity, the osmotic potential difference switches sign twice, leading to the presence of double framework transitions between lp and np forms upon long-chain hydrocarbon adsorptions, shown in Figure 20. Due to the low relative affinity of CH4 (almost equal to unity), the lp form is thermodynamically stable, yielding no step of its adsorption isotherm [126]. The affinity between adsorbate (Xe)-adsorbent (MIL-53(Al)) was measured using the Henry’s constant (KH) at np and lp forms, which in turn was used to estimate the adsorption stress (σs) during the structural transition [127,128]. At the low pressure region when adsorbate molecules are introduced, the framework starts contracting, yielding decreasing framework stress. Once the critical stress is reached ( ), the lp form becomes unstable and the framework turns into the np form. During the period, MOFs are experiencing the “abnormal breathing” behavior—the sample contracts while inhaling. This results in a sharp uptake step on the adsorption isotherm; the sample volume decreases 3224 Materials 2014, 7 Materials 2014, 7 Figure 21. (a) Experimental adsorption (filled) and desorption (empty) isotherms of Xe in MIL-53(Al) at 220 K (in blue), and the fitting of Langmuir isotherms using the lp (black dash) and np (red dash) forms, respectively; (b) simulated structural transition stresses of the lp (black) and np (red) forms. σnp and σlp denote the critical stresses of the structural transitions upon adsorption (A1 and A2) and desorption (D1 and D2), respectively. Reprinted with permission from [128]. Copyright 2010 American Chemical Society. Figure 21. (a) Experimental adsorption (filled) and desorption (empty) isotherms of Xe in MIL-53(Al) at 220 K (in blue), and the fitting of Langmuir isotherms using the lp (black dash) and np (red dash) forms, respectively; (b) simulated structural transition stresses of the lp (black) and np (red) forms. σnp and σlp denote the critical stresses of the structural transitions upon adsorption (A1 and A2) and desorption (D1 and D2), respectively. Reprinted with permission from [128]. Copyright 2010 American Chemical Society. At the low pressure region in Figure 22a, it was shown that a dual-interaction between the O atoms of CO2 and two hydrogen atoms of μ2 hydroxyl groups on the opposing pore walls presents in both the lp and np forms of MIL-53(Al) [129,130]. The distances of OCO2-Hμ2-OH are comparable with the hydrogen bonds of clathrated water interacting with the carboxylate groups in MIL-53as (Al) [26]. While the adsorption takes place at high pressure region (>5 bars), in Figure 22b, all four of the hydroxyl groups present are involved in forming this dual-interaction. In addition, the adsorbate-adsorbate interaction becomes more prominent, resulting in the formation the single-interaction between CO2 adsorbate and both organic linker and μ2 hydroxyl group, which subsequently leads to a more open porosity. This initiates the structural transition from the np form to the lp form. As the pores are further filled, the adsorbate molecules mainly interacts with the hydrogen atoms on the organic groups, as well as the inorganic part [129]. (a) (b) (a) (b) At the low pressure region in Figure 22a, it was shown that a dual-interaction between the O atoms of CO2 and two hydrogen atoms of μ2 hydroxyl groups on the opposing pore walls presents in both the lp and np forms of MIL-53(Al) [129,130]. The distances of OCO2-Hμ2-OH are comparable with the hydrogen bonds of clathrated water interacting with the carboxylate groups in MIL-53as (Al) [26]. Materials 2014, 7 Materials 2014, 7 by 40% (line A1 in Figure 21). As the pressure increases, the np-form framework starts expanding, which yields the increasing framework stress. Once the critical stress is reached ( ), the np form becomes unstable. The framework turns into the lp form, during which the sample expands while “inhaling”. The sample volume increases by 40%, displayed by the second step on the adsorption isotherm (line A2). As the adsorption occurs further in the lp phase, the solvation pressure increases, causing the elastic expansion of sample, which remains stable [128]. This stress-based model implies that two distinctive states of framework exist, which cannot explain the coexistence of np and lp forms during the contracting state [110,127]. From molecular dynamics (MD) simulations, translational (τ) and orientational (S) orders of confined CO2 molecules were used to monitor the framework transition. As the structural transition (lp → np form) upon the adsorption occurs at low pressure region, the orientational order of CO2 increased drastically due to the confinement of np form. Similar change can be noticed at the translation order loss along the channel direction, indicating the rearrangement of CO2 molecules while the framework contracts. It is suggested that the heat of adsorption via the guest-host interaction can compensate for the entropy loss of confined CO2 molecules, making the framework transition becomes an enthalpy-driven step [127]. Figure 20. (a) Adsorption isotherms of CH4 (red) and C4H10 (blue) in MIL-53(Cr), in a Langmuir model; (b) the osmotic potential change during the structural transition between the lp and np forms. Vertical dotted lines indicate the structural transition induced by C4H10 molecules. Reprinted with permission from [126]. Copyright 2009 American Chemical Society. Figure 20. (a) Adsorption isotherms of CH4 (red) and C4H10 (blue) in MIL-53(Cr), in a Langmuir model; (b) the osmotic potential change during the structural transition between the lp and np forms. Vertical dotted lines indicate the structural transition induced by C4H10 molecules. Reprinted with permission from [126]. Copyright 2009 American Chemical Society. (a) (b) (a) (b) 3225 Materials 2014, 7 While the adsorption takes place at high pressure region (>5 bars), in Figure 22b, all four of the hydroxyl groups present are involved in forming this dual-interaction. In addition, the adsorbate-adsorbate interaction becomes more prominent, resulting in the formation the single-interaction between CO2 adsorbate and both organic linker and μ2 hydroxyl group, which subsequently leads to a more open porosity. This initiates the structural transition from the np form to the lp form. As the pores are further filled, the adsorbate molecules mainly interacts with the hydrogen atoms on the organic groups, as well as the inorganic part [129]. At the low pressure region in Figure 22a, it was shown that a dual-interaction between the O atoms of CO2 and two hydrogen atoms of μ2 hydroxyl groups on the opposing pore walls presents in both the lp and np forms of MIL-53(Al) [129,130]. The distances of OCO2-Hμ2-OH are comparable with the hydrogen bonds of clathrated water interacting with the carboxylate groups in MIL-53as (Al) [26]. While the adsorption takes place at high pressure region (>5 bars), in Figure 22b, all four of the hydroxyl groups present are involved in forming this dual-interaction. In addition, the adsorbate-adsorbate interaction becomes more prominent, resulting in the formation the single-interaction between CO2 adsorbate and both organic linker and μ2 hydroxyl group, which subsequently leads to a more open porosity. This initiates the structural transition from the np form to the lp form. As the pores are further filled, the adsorbate molecules mainly interacts with the hydrogen atoms on the organic groups, as well as the inorganic part [129]. Materials 2014, 7 3226 Materials 2014, 7 Figure 22. Simulation of CO2 arrangement of the adsorption on MIL-53(Al). (a) at the initial stage of loading in MIL-53np (Al); (b) at the high stage of loading in MIL-53lp (Al). Reprinted with permission from [130]. Copyright 2008 American Chemical Society. Figure 22. Simulation of CO2 arrangement of the adsorption on MIL-53(Al). (a) at the initial stage of loading in MIL-53np (Al); (b) at the high stage of loading in MIL-53lp (Al). Reprinted with permission from [130]. Copyright 2008 American Chemical Society. (a) (b) An inflection point was observed in the simulated energetic profile during the framework transition, indicating that an energetic unfavorable intermediate state exists during the framework transition of CO2 adsorption on MIL-53(Cr). Materials 2014, 7 This observation implies that the framework transition could be a progressive type instead of a distinct one, meaning that some part of pores are open while the remaining are closed [131]. Shown in Figure 23, the free energy model that incorporates host free energy, guest-guest interactions, and guest-host interaction is introduced successfully to reproduce the CO2 and CH4 adsorption isotherms of MIL-53(Cr) [132]. Under the low and high pressure regions, a global minimum of free energy is observed (red dots), indicating the lp form is the single stable framework structure (plots 1, 2, 10, and 11). Upon changing the pressure, a local minimum of free energy (black dots) presents, showing that the np form coexists with the lp form (plots 3, 4, 8, and 9). Further changing the gas pressure leads to the disappearance of lp form, and the framework becomes the pure np form (plots 5, 6, and 7). The model also rationalizes the framework transition depends on the type and the size of guest molecules [127,132,133]. The observations are corroborated with experimental data [110]. (a) (b) (a) (b) (b) (a) An inflection point was observed in the simulated energetic profile during the framework transition, indicating that an energetic unfavorable intermediate state exists during the framework transition of CO2 adsorption on MIL-53(Cr). This observation implies that the framework transition could be a progressive type instead of a distinct one, meaning that some part of pores are open while the remaining are closed [131]. Shown in Figure 23, the free energy model that incorporates host free energy, guest-guest interactions, and guest-host interaction is introduced successfully to reproduce the CO2 and CH4 adsorption isotherms of MIL-53(Cr) [132]. Under the low and high pressure regions, a global minimum of free energy is observed (red dots), indicating the lp form is the single stable framework structure (plots 1, 2, 10, and 11). Upon changing the pressure, a local minimum of free energy (black dots) presents, showing that the np form coexists with the lp form (plots 3, 4, 8, and 9). Further changing the gas pressure leads to the disappearance of lp form, and the framework becomes the pure np form (plots 5, 6, and 7). The model also rationalizes the framework transition depends on the type and the size of guest molecules [127,132,133]. The observations are corroborated with experimental data [110]. Materials 2014, 7 An inflection point was observed in the simulated energetic profile during the framework transition, indicating that an energetic unfavorable intermediate state exists during the framework transition of CO2 adsorption on MIL-53(Cr). This observation implies that the framework transition could be a progressive type instead of a distinct one, meaning that some part of pores are open while the remaining are closed [131]. Shown in Figure 23, the free energy model that incorporates host free energy, guest-guest interactions, and guest-host interaction is introduced successfully to reproduce the CO2 and CH4 adsorption isotherms of MIL-53(Cr) [132]. Under the low and high pressure regions, a global minimum of free energy is observed (red dots), indicating the lp form is the single stable framework structure (plots 1, 2, 10, and 11). Upon changing the pressure, a local minimum of free energy (black dots) presents, showing that the np form coexists with the lp form (plots 3, 4, 8, and 9). Further changing the gas pressure leads to the disappearance of lp form, and the framework becomes the pure np form (plots 5, 6, and 7). The model also rationalizes the framework transition depends on the type and the size of guest molecules [127,132,133]. The observations are corroborated with experimental data [110]. Other than guest molecules, the framework transition can be triggered by temperature variation, leading to the dehydration-rehydration cycle. It has been shown that MIL-53 exhibits the lp form under high temperature (>300 K) and the np form under low temperature (<300 K). With the absence of van der Waals force interactions between adsorbate and adsorbent, the framework transition occurs through (1) twisted benzene groups of benzenedicarboxylate (BDC) ligands, also called as π flipping; and (2) distortion mode from the corner-sharing octahedral MO6 (M = Al, Cr) clusters, displayed in Figure 24 [112,134]. Similar temperature-dependent framework transition is also observed for MOF-5 [135]. Materials 2014, 7 3227 Materials 2014, 7 Figure 23. Free energy simulation of CO2 adsorption on MIL-53(Cr) at 300 K under various pressure ranges. The blue-colored contour represents the lowest free energy, while the red-colored contour means the highest free energy. The under panel shows the expected framework structure during adsorption, equilibrium and desorption processes. Reprinted with permission from [132]. Copyright 2013 American Chemical Society. Figure 24. Neutron powder diffraction of MIL-53(Al) collected at (a) an elevated temperature (lp form); (b) room temperature (np form). The π–flipping of BDC ligands and the distortion of octahedral metal nodes can be visualized in the inset and (c). Reprinted with permission from [112]. Copyright 2008 American Chemical Society. [27]. Copyright 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. Figure 24. Neutron powder diffraction of MIL-53(Al) collected at (a) an elevated temperature (lp form); (b) room temperature (np form). The π–flipping of BDC ligands and the distortion of octahedral metal nodes can be visualized in the inset and (c). Reprinted with permission from [112]. Copyright 2008 American Chemical Society. [27]. Copyright 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. 3228 Materials 2014, 7 Figure 24. Cont. 5. Functionalization of MOFs to Modulate Breathing Behaviors For gas storage and separation applications, it is widely recognized that the gas uptake cap gas adsorption selectivity are correlated with the pore size and pore shape of MOFs. Altering size and pore shape of MOFs can be achieved by designing new MOFs using new organic l Figure 24. Cont. Figure 24. Cont. g 5. Functionalization of MOFs to Modulate Breathing Behaviors For gas storage and separation applications, it is widely recognized that the gas uptake capacity and gas adsorption selectivity are correlated with the pore size and pore shape of MOFs. Altering the pore size and pore shape of MOFs can be achieved by designing new MOFs using new organic linkers or metal centers prior to synthesis. The approach enjoys the advantage of one-pot synthesis, to tailor the pore size, adsorption affinity, and separation selectivity. However, the approach of incorporating predetermined or pre-functionalized organic ligands may be not impeccable, due to their poor stability, reactivity, and solubility under reaction conditions. This drawback has driven the needs to modify the MOF framework functionality using post synthetic modifications (PSMs) [136–141]. PSMs are chemical treatments of parent MOFs, while maintaining the structure intact. Almost all PSMs are 5. 5.1. Pre-Functionalized Organic Linkers Tuning the breathing effect of MOFs has recently been achieved by designing the structure of organic linkers at the pre-synthesis stage. MIL-53 structure contains linkages of AlO4(OH)2-octahedra with 1,4-benzenedicarboxylic acid (BDC), resulting in one-dimensional channels in c direction. The inclusion of amine-bearing organic linkers in MIL-53s leads to the slight expansion of unit cell volume [148]. However, it exerts no change on the MOF structure nor the breathing behaviors. The expansion of unit cell volume could be originated from the occupation of amine groups inside channels. The occupation of amine groups also result in less adsorption capacity and less thermal stability, compared with non-functionalized MIL-53. A similar effect can be observed for NH2-MIL-53 on the adsorption capacity of CO2 which decreases from 40 wt% to 30 wt%, due to the less available micropore volume. However, NH2-MIL-53 exhibits a higher separation factor of CO2 over CH4, due to the coupling contribution from the strong interaction between NH2 groups and CO2 molecules and the flexible framework of MIL-53 [149]. It was determined experimentally that the zero coverage adsorption enthalpy of CO2 is increased from 20.1 kJ/mol to 38.4 kJ/mol when MIL-53 framework is decorated with amino groups. Amino groups of NH2-MOFs are considered as tags for creating versatile functionalities using PSMs [136,150]. In addition to the amino group, various pre-functionalized organic linkers are adopted in synthesizing Al- and Fe-based MIL-53s [111,115]. Stock’s group solvothermally synthesized five flexible functionalized MIL-53(Al) using BDC-X as linkers, whereas BDC = 1,4-benzenedicarboxy-late; X = –Cl, –Br, –CH3, –NO2, and –(OH)2. The breathing behaviors of the MOFs were altered based on the appended functional groups. During the hydration-dehydration cycle, all materials demonstrated the np-to-lp transition at various temperatures, in which Cl-MIL-53(Al) occurs at 110 °C, Br-MIL-53(Al) is at 80 °C, CH3-MIL-53(Al) is at 120 °C, and NO2-MIL-53(Al) is at 130 °C. This phenomenon was attributed to the hydrogen bonding interactions between the adsorbate and absorbents. However, the lp form was originally observed in (OH)2-MIL-53(Al) below 120 °C and was transformed into a pure np form above 140 °C, due to the removal of water molecules. This work has recently been extended to investigate the effect of fluorinated BDC on the sorption capacity of F-MIL-47 and F-MIL-53(Al) [151,152]. F-MIL-47 exhibits better n-hexane uptake capacity than its parent materials, possibly due to the enhanced hydrophobicity of its framework. Materials 2014, 7 Materials 2014, 7 based on amine-bearing organic linkers. Requirements for successful PSMs includes: accessible porosity or surface, available functional groups, and stability under reaction conditions and environment. The covalent PSM on MOFs was first implemented by Kim’s group, on which free pyridyl groups of organic moieties was treated [142]. There have been several excellent review articles published specifically on this topic [143], and is not detailed in this work. Other approaches adopted include using: (1) pre-functionalized organic linkers [76,115,144]; (2) mixed organic linkers [145,146]; (3) different metal nodes [147]; (4) mixed metal ions as nodes in MOFs [96]. Materials 2014, 7 Functionalization of MOFs to Modulate Breathing Behaviors 5. Functionalization of MOFs to Modulate Breathing Behaviors For gas storage and separation applications, it is widely recognized that the gas uptake capacity and gas adsorption selectivity are correlated with the pore size and pore shape of MOFs. Altering the pore size and pore shape of MOFs can be achieved by designing new MOFs using new organic linkers or metal centers prior to synthesis. The approach enjoys the advantage of one-pot synthesis, to tailor the pore size, adsorption affinity, and separation selectivity. However, the approach of incorporating predetermined or pre-functionalized organic ligands may be not impeccable, due to their poor stability, reactivity, and solubility under reaction conditions. This drawback has driven the needs to modify the MOF framework functionality using post synthetic modifications (PSMs) [136–141]. PSMs are chemical treatments of parent MOFs, while maintaining the structure intact. Almost all PSMs are 3229 5.1. Pre-Functionalized Organic Linkers The F-MIL-53(Al) shows the decreased n-hexane adsorption capacity, possibly originated from its framework rigidity. The functionalized MIL-53(Al)s showed smaller accessible micropore volumes compared with their parent material, due to the occupancy of functional groups inside the one-dimensional channels. The accessible micropore volumes demonstrate the decreasing trend in the following order: 3230 Materials 2014, 7 NO2-MIL-53(Al) > Cl-MIL-53(Al) = CH3-MIL-53(Al) > Br-MIL-53(Al) > (OH)2-MIL-53(Al), due to the steric hindrance of functional groups. Their CO2 adsorption capacities exhibited the similar trend; however, all of them were smaller than NH2-MIL-53(Al), shown in Figure 25. This trend is owing to the availability and affinity of amino groups towards CO2 molecules. It is noted that the CO2 adsorption capacity of –NO2 is higher than that of NH2-MIL-53(Al) when CO2 pressure was above 1 bar, due to the presence of np form of NH2-MIL-53(Al). Figure 25. Adsorption isotherms of functionalized MIL-53s: (a) CO2; (b) H2O. Symbols: : Cl-MIL-53(Al); : Br-MIL-53(Al); : CH3-MIL-53(Al); : NO2-MIL-53(Al); : (OH)2-MIL-53(Al); : NH2-MIL-53(Al). Reprinted with permission from [111]. Copyright 2011 American Chemical Society. Figure 25. Adsorption isotherms of functionalized MIL-53s: (a) CO2; (b) H2O. Symbols: : Cl-MIL-53(Al); : Br-MIL-53(Al); : CH3-MIL-53(Al); : NO2-MIL-53(Al); : (OH)2-MIL-53(Al); : NH2-MIL-53(Al). Reprinted with permission from [111]. Copyright 2011 American Chemical Society. Furthermore, several derivatives of 1,4-benzenedicarboxylic acid (BDC)-based organic linkers with fl ibl id h i d h h ifi i i f i 2 3 dih d l BDC d Furthermore, several derivatives of 1,4-benzenedicarboxylic acid (BDC)-based organic linkers with flexible side chains were prepared through esterification reactions, forming 2,3-dihydroxyl-BDC and 2,5-dihydroxyl-BDC [76]. The breathing behavior of the pillar-layered MOFs [Zn2(functionalized-BDC)2 (DABCO)]n was tailored with various functional groups, illustrated in Figure 26. The dangling substituted groups form stronger attraction forces (van der Waal, hydrogen bonding, or π–π interactions), yielding more contraction of unit cell volume, ranging from 86% to 72% (Figure 27). However, if the linker is substituted at Positions 2 and 3 of BDC ligands, the frameworks become much rigid, compared with those formed from 2,5-di-substituted ones. Furthermore, several derivatives of 1,4-benzenedicarboxylic acid (BDC)-based organic linkers with flexible side chains were prepared through esterification reactions, forming 2,3-dihydroxyl-BDC and 2,5-dihydroxyl-BDC [76]. The breathing behavior of the pillar-layered MOFs [Zn2(functionalized-BDC)2 (DABCO)]n was tailored with various functional groups, illustrated in Figure 26. The dangling substituted groups form stronger attraction forces (van der Waal, hydrogen bonding, or π–π interactions), yielding more contraction of unit cell volume, ranging from 86% to 72% (Figure 27). However, if the linker is substituted at Positions 2 and 3 of BDC ligands, the frameworks become much rigid, compared with those formed from 2,5-di-substituted ones. Materials 2014, 7 e, several derivatives of 1,4-benzenedicarboxylic acid (BDC)-based organic linkers with Furthermore, several derivatives of 1,4-benzenedicarboxylic acid (BDC)-based organic linkers with flexible side chains were prepared through esterification reactions, forming 2,3-dihydroxyl-BDC and 2,5-dihydroxyl-BDC [76]. The breathing behavior of the pillar-layered MOFs [Zn2(functionalized-BDC)2 (DABCO)]n was tailored with various functional groups, illustrated in Figure 26. The dangling substituted groups form stronger attraction forces (van der Waal, hydrogen bonding, or π–π interactions), yielding more contraction of unit cell volume, ranging from 86% to 72% (Figure 27). However, if the linker is substituted at Positions 2 and 3 of BDC ligands, the frameworks become much rigid, compared with those formed from 2,5-di-substituted ones. Furthermore, several derivatives of 1,4-benzenedicarboxylic acid (BDC)-based organic linkers with flexible side chains were prepared through esterification reactions, forming 2,3-dihydroxyl-BDC and 2,5-dihydroxyl-BDC [76]. The breathing behavior of the pillar-layered MOFs [Zn2(functionalized-BDC)2 (DABCO)]n was tailored with various functional groups, illustrated in Figure 26. The dangling substituted groups form stronger attraction forces (van der Waal, hydrogen bonding, or π–π interactions), yielding more contraction of unit cell volume, ranging from 86% to 72% (Figure 27). However, if the linker is substituted at Positions 2 and 3 of BDC ligands, the frameworks become much rigid, compared with those formed from 2,5-di-substituted ones. The chemical treatment can decorate the channels of MOF frameworks with controlled and various functionalities. The approach is very suitable for applications in molecule recognition via several types of interaction, such as van der Waals force, hydrogen bonds between MOF frameworks and substrates [153]. The authors proposed three approaches to tune the functionalities of MOFs in differential microenvironment: (1) the formation of cubic nets as interpenetrating frameworks, by combining paddle-wheel clusters M2(COO)4 and dicarboxylic acid (BDC), and pillar bi-dentate organic linkers. By changing the organic linker from 4,4-bipyridine (4,4′-Bipy) to trans-bis(4-pyridyl)-ethylene (4,4′-Bpe), the resulting MOFs exhibit selective sorption towards H2 and CO2 and selective separation of linear and branched isomers (Figure 28) [154]; (2) immobilization of open metal centers (Cu2+) within channels showing extraordinarily high acetylene storage at room temperature, making the 3231 Materials 2014, 7 Materials 2014, 7 Materials 2014, 7 3233 Materials 2014, 7 Materials 2014, 7 Materials 2014, 7 transportation of high density acetylene storage for industrial processes (Figure 29) [155]; (3) immobilization of open metal centers using coordinating ligands, such as Schiff base, for separation. Figure 26. Chemical structure of derivatives of 2,3-dihydroxyl-BDC and 2,5-dihydroxyl-BDC. Reprinted with permission from [76]. Copyright 2012 American Chemical Society. e 26. Chemical structure of derivatives of 2,3-dihydroxyl-BDC and 2,5-dihydroxyl-BDC. nted with permission from [76]. Copyright 2012 American Chemical Society. Figure 27. The volume contraction of various functionalized MOFs [Zn2(functionalized- BDC)2(DABCO)]n. Reprinted with permission from [76]. Copyright 2012 American Chemical Society. Figure 27. The volume contraction of various functionalized MOFs [Zn2(functionalized- BDC)2(DABCO)]n. Reprinted with permission from [76]. Copyright 2012 American Chemical Society. 3232 Materials 2014, 7 Figure 28. MOF using M2(COO)4 as the metal node, and 4,4′-Bpe as the organic linker. The MOF exhibits selective adsorptions towards (a) H2 and (b) CO2. Reprinted with permission from [154]. Copyright 2007 American Chemical Society. Figure 28. MOF using M2(COO)4 as the metal node, and 4,4′-Bpe as the organic linker. The MOF exhibits selective adsorptions towards (a) H2 and (b) CO2. Reprinted with permission from [154]. Copyright 2007 American Chemical Society. Figure 28. MOF using M2(COO)4 as the metal node, and 4,4′-Bpe as the organic linker. The MOF exhibits selective adsorptions towards (a) H2 and (b) CO2. Reprinted with permission from [154]. Copyright 2007 American Chemical Society. Figure 29. (Left) crystal structures of (a) HKUST-1; (b) MOF-505; (c) MOF-508; (d) MIL-53; (e) MOF-5; and (f) ZIF-8, in which (a) and (b) show open Cu2+ sites (green). (Right) the C2H2 adsorption isotherms of several MOFs, HKUST-1 (black); MOF-505 (red); MOF-508 (green); MIL-53 (blue); MOF-5 (cyan); and ZIF-8 (magenta). Reprinted with permission from [155]. Copyright 2009 American Chemical Society. Figure 29. (Left) crystal structures of (a) HKUST-1; (b) MOF-505; (c) MOF-508; (d) MIL-53; (e) MOF-5; and (f) ZIF-8, in which (a) and (b) show open Cu2+ sites (green). (Right) the C2H2 adsorption isotherms of several MOFs, HKUST-1 (black); MOF-505 (red); MOF-508 (green); MIL-53 (blue); MOF-5 (cyan); and ZIF-8 (magenta). Reprinted with permission from [155]. Copyright 2009 American Chemical Society. 5.2. Mixed Organic Linkers The breathing behavior, adsorption capacity, selectivity, and thermo-mechanical property of MOFs can be tuned by mixing various organic linkers. Yaghi’s group synthesized 18 one-phase multivariate (MTV) MOF-5s that contains distinct functionalities, using 8 types of functionalized 1,4-benzenedicarboxylate (BDC) as organic linkers, shown in Figure 30 [144]. Compared with un-functionalized MOF-5, the resulted MOFs exhibits enhanced H2 uptake capacities (84%, maximum), and improved CO2/CO selectivity (400%), shown in Figure 31. Similarly, an improved CO2/N2 selectivity was reported by Costantino’s group using mixed organic linkers, in which a water-stable isoreticular phosphonate MOFs are successfully formed [156]. Farrusseng’s group synthesized a series of MIL-53s using various ratio of BDC and amino-substituted BDC [124]. The resulted MOFs show smaller micropore volume as the adopted functionalized BDC content is increased. The higher content of amino groups leads to higher Henry constants, yielding lower lp-to-np phase transition pressure and higher np-to-lp phase transition pressure, as well as the gas adsorption capacity. It is noted that the altered phase-transition temperature of MOFs synthesized from the mixed-linker approach also affects their thermo-mechanical properties, recently reported by Fischer’s group, depending on the chain length and the hydrophobicity of the functionalized groups [146]. Figure 30. MTV-MOF-5 equipped with distinct functionalities in one phase. Reprinted with permission from [144]. Copyright 2010, American Association for the Advancement of Science. 3234 Materials 2014, 7 Materials 2014, 7 Figure 31. Several MTV-MOF-5s exhibit enhanced CO2/CO selectivity. E: NO2-BDC; H: (C3H5O)2-BDC; I: (C7H7O)2-BDC. Reprinted with permission from [144]. Copyright 2010, American Association for the Advancement of Science. 5.3. Different Metal Nodes 5.3. Different Metal Nodes 5.3. Different Metal Nodes Gascon’s group investigated the effect of metal node on the breathing behaviors of MOFs [147]. In their study, Al, Ga, and In were successfully used as various metal nodes to synthesize amino-functionalized MIL-53s. Measurements of CO2 adsorption on these MOFs showed that NH2-MIL-53(Al) displayed a Langmuir isotherm (Type I) while its Ga- and In-derivatives exhibited stepped isotherms respectively, shown in Figure 32. The presence of stepped isotherm was attributed to the existence of observable vnp form in NH2-MIL-53(Ga) and NH2-MIL-53(In), which are oriented from the electropositivity of the metals. The distance between amine groups of organic liners and Oμ2 of a metal node decreases in the order of Al < Ga < In, showing a stronger acidity of μ2-OH in NH2-MIL-53(In) which in turn modulated the resulted breathing behaviors. MIL-47 and MIL-53 are iso-structural MOFs with different metal centers. The former has VIV as its metal centers, and the later has AlIII or CrIII as its metal nodes. The existence of corner-sharing μ2-OH groups in MIL-53 framework contribute to it framework flexibility. However, the presence of μ2-oxo groups in MIL-47 framework yields the rigidity of its framework. Hence, MIL-47 only shows as a microporous material and exhibits no breathing behaviors [47,99]. How does the framework functionality affect their gas adsorption behaviors and breathing phenomena? NH2-MIL-47 shows less CH4 and CO2 adsorption capacities compared with MIL-47 at 30 °C within test pressure range [157]. The declined gas adsorption capacity is owing to the presence of dangling amino-groups in the pores, yielding a decrease in the pore volume (from 0.46 mL/g to 0.40 mL/g). However, the presence of NH2 groups doesn’t affect significantly the affinity towards CO2 and CH4 molecules. This feature results in a similar separation factor of CO2 and CH4 between NH2-MIL-47 and MIL-47. The separation factor of CO2 and CH4 is defined as the ratio of Henry adsorption constants of CO2 over CH4. The absence of significant change on the CO2 and CH4 affinities could be 3235 Materials 2014, 7 Materials 2014, 7 Materials 2014, 7 3236 Materials 2014, 7 Materials 2014, 7 due to the rigidity of NH2-MIL-47 framework. DFT analyses indicate that the adsorption of CO2 molecules is preferred towards NH2 groups in NH2-MIL-47, due to strong electrostatic interactions between the adsorbate carbon and the negative charge of the nitrogen. Similar results can be noticed on F-MIL-47 framework [151], and NH2-MOF-5 (IRMOF-3) [158]. Figure 32. CO2 adsorption isotherms at 273 K. (a) 120 kPa; (b) 3.5 MPa. Red: NH2-MIL-53(Al); Blue: NH2-MIL-53(Ga); Black: NH2-MIL-53(In). Reprinted with permission from [147]. Copyright 2012 American Chemical Society. NH2-MIL-53(Al); Blue: NH2-MIL-53(Ga); Black: NH2-MIL-53(In). Reprinted with permission from [147]. Copyright 2012 American Chemical Society. Similarly, Loiseau’s group [119] reported the effect of metal identity on the breathing behavior of MIL-53s, in which MIL-53(Ga) exhibits the breathing behavior upon dehydration-hydration cycle, due to the presence of μ2-hydroxo linkages in its framework. An intermediary phase showing two types of closed channels was observed, in which one type of channels has a strong hydrogen-bond guest-host interaction, while the other type has a weak hydrogen-bond interaction. These interactions contribute to the thermal stability of np-form of MIL-53(Ga) and MIL-53(Fe). MIL-53(Al) demonstrates better thermal stability of lp framework without the presence of the intermediary np form. The effect of metal identity on the framework thermal stability follows: Al > Ga > Fe. Similarly, Loiseau’s group [119] reported the effect of metal identity on the breathing behavior of MIL-53s, in which MIL-53(Ga) exhibits the breathing behavior upon dehydration-hydration cycle, due to the presence of μ2-hydroxo linkages in its framework. An intermediary phase showing two types of closed channels was observed, in which one type of channels has a strong hydrogen-bond guest-host interaction, while the other type has a weak hydrogen-bond interaction. These interactions contribute to the thermal stability of np-form of MIL-53(Ga) and MIL-53(Fe). MIL-53(Al) demonstrates better thermal stability of lp framework without the presence of the intermediary np form. The effect of metal identity on the framework thermal stability follows: Al > Ga > Fe. 5.4. Mixing Metal Nodes Tuning the breathing behavior of MOFs can be achieved via mixing cations or altering the valence of metal cations. As aforementioned, MIL-47(VIV) is a rigid MOF. Its framework can demonstrate the flexibility or even the breathing behavior using biphenyl-4,4′-dicarboxylate as the organic linker [159] or using high mechanical pressure [160]. Clet’s group reported to activate the flexibility of MIL-47(V) by changing the metal valence from VIV to VIII [161]. The framework of hydrated MIL-47(VIII) adopts the np form, which is in contrast to the lp form of MIL-47(VIV). The reversible thermal response of MIL-47(VIII) is similar with MIL-53(Fe), following the framework transition of np → cp → int → lp as the temperature increases. The breathing behavior of MIL-47(VIII) is exhibited by the adsorption of CO2, in which presents a stepped adsorption isotherm shown in Figure 33. This flexibility is owing to the presence of μ2-OH groups in MIL-47(VIII) and is gradually diminished as the fraction of VIV in MIL-47(VIII/VIV) increases. The flexibility of MIL47(VIV) can be also induced by doping various amount of AlIII ions into its framework [162]. The presence of vanadyl units in pseudo-octahedral or square-pyramidal geometry might contribute to the breathing behavior of MIL-47(AlIII/VIV) upon the CO2 adsorption. Figure 33. CO2 adsorption isotherms at 303 K. (a) MIL-47(VIII); (b) MIL-47(VIV). Symbols: ■—adsorbed CO2; ◊—fraction of narrow pores in MOF determined by Raman spectra. Reprinted with permission from [161]. Copyright 2011 American Chemical Society. Figure 33. CO2 adsorption isotherms at 303 K. (a) MIL-47(VIII); (b) MIL-47(VIV). Symbols: ■—adsorbed CO2; ◊—fraction of narrow pores in MOF determined by Raman spectra. Reprinted with permission from [161]. Copyright 2011 American Chemical Society. Using a mixture of Fe-Cr as cations, Serre’s group was able to tune the breathing behaviors by taking advantage of the nature of cation ions upon the hydration-dehydration cycle [96]. While those solids are hydrated, they all adopted the np form; whereas MIL-53(Cr) and MIL-53(Fe) change to the lp form and a closed pore (cp) form respectively, under dehydrated conditions. This is caused by the departure of H2O molecules, yielding the cp-form structure followed by the lp form. However, the thermal response behavior of MIL-53(FeII/III/VIII/IV) becomes different, depending on the Fe/V ratio [163]. For Fe-rich MIL-53(FeII/VIII), its framework transforms reversibly from np → int → lp upon the hydration and dehydration cycle (Figure 34a). 5.4. Mixing Metal Nodes However, Using a mixture of Fe-Cr as cations, Serre’s group was able to tune the breathing behaviors by taking advantage of the nature of cation ions upon the hydration-dehydration cycle [96]. While those solids are hydrated, they all adopted the np form; whereas MIL-53(Cr) and MIL-53(Fe) change to the lp form and a closed pore (cp) form respectively, under dehydrated conditions. This is caused by the departure of H2O molecules, yielding the cp-form structure followed by the lp form. However, the thermal response behavior of MIL-53(FeII/III/VIII/IV) becomes different, depending on the Fe/V ratio [163]. For Fe-rich MIL-53(FeII/VIII), its framework transforms reversibly from np → int → lp upon the hydration and dehydration cycle (Figure 34a). However, Using a mixture of Fe-Cr as cations, Serre’s group was able to tune the breathing behaviors by taking advantage of the nature of cation ions upon the hydration-dehydration cycle [96]. While those solids are hydrated, they all adopted the np form; whereas MIL-53(Cr) and MIL-53(Fe) change to the lp form and a closed pore (cp) form respectively, under dehydrated conditions. This is caused by the departure of H2O molecules, yielding the cp-form structure followed by the lp form. However, the thermal response behavior of MIL-53(FeII/III/VIII/IV) becomes different, depending on the Fe/V ratio [163]. For Fe-rich MIL-53(FeII/VIII), its framework transforms reversibly from np → int → lp upon the hydration and dehydration cycle (Figure 34a). However, Using a mixture of Fe-Cr as cations, Serre’s group was able to tune the breathing behaviors by taking advantage of the nature of cation ions upon the hydration-dehydration cycle [96]. While those solids are hydrated, they all adopted the np form; whereas MIL-53(Cr) and MIL-53(Fe) change to the lp form and a closed pore (cp) form respectively, under dehydrated conditions. This is caused by the departure of H2O molecules, yielding the cp-form structure followed by the lp form. However, the thermal response behavior of MIL-53(FeII/III/VIII/IV) becomes different, depending on the Fe/V ratio [163]. For Fe-rich MIL-53(FeII/VIII), its framework transforms reversibly from np → int → lp upon the hydration and dehydration cycle (Figure 34a). However, 3237 Materials 2014, 7 the framework shows irreversible transition into the lp form upon heating when the Fe/V ratio is 1, shown in Figure 34b. The modulation of breathing behaviors by mixing cations is also reported by Biradha’s group in {[M(L)2(H2O)2]·(ClO4)2·2(CH3OH)·2(CHCl3)·4(H2O)}n (M = Zn(II), Cd(II); L = benzene-1,3,5-triyltriisonicotinate) [164]. 5.4. Mixing Metal Nodes The MOFs using a mixture of Zn and Cd show higher N2 adsorption capacity compared with each parent material. Figure 34. (a) Temperature variation of PXRD patterns of MIL-53(FeII/VIII) upon heating; (b) room temperature PXRD patterns of MIL-53(FeII/VIII) calcined at 573 K. Reprinted with permission from [163]. Copyright 2013 American Chemical Society. p [ ] py g y 6. Outlooks In addition to the dimensionality of organic ligands and metal clusters, it is clearly noticeable that those MOFs exhibiting the breathing behavior have two common features: (1) non-rigid areas, or weak points, appear in the frameworks and (2) a free space within the framework to accommodate the steric hindrance of the movements of weak points [7]. Hence, some empirical rules have been developed to predict the presence of weak points in MOFs: (1) inorganic clusters might possess a mirror plane with the organic ligands in symmetrical position; (2) O–O axes of carboxylates are perpendicular to the elongation axis. Similar requirement can be noticed very recently [159]; (3) include even cycles at the level of the cluster and/or at the level of the topology of the skeleton; (4) the ratio C/M should be ≥2, whereas C is the number of carbons of the carboxylate surrounding the cluster, and M is the number of metallic atoms within the cluster; (5) breathing effects can only occur with ditopic carboxylates, but not on MOFs with tri- or tetratopic ligands. In addition, it is reported that the metal centers must possess μ2-hydroxyl groups that are preferred interacting sites for adsorbates, leading to the flexibility point. The presence of μ2-oxo groups in MIL-47 (VIV) still contribute to the rigidity of framework even though all above conditions are satisfied [161]. Furthermore, more flexible MOFs showing phenomenal breathing behaviors are synthesized via tritopic ligands [81,165] or tetratopic ligands [11,166]. The flexibility is evolved from even cycles in their skeleton topologies, despite of forming odd cycles 6. Outlooks Materials 2014, 7 Materials 2014, 7 Materials 2014, 7 at the cluster level. An alternative approach is to create additional “kneecaps” (rotating moiety) within a bulky organic ligand (semi-rigid ligand), shown in Scheme Ι. Upon exposing to external stimuli, the twisted organic ligands lead to the framework flexibility. Scheme Ι. The illustration of “breathing” MOFs synthesized using semi-rigid ligands. Green spheres mean metal nodes, gray rods mean rigid organic ligands, and blue components mean flexible or rotating ligands. Reprinted with permission from [166]. Copyright 2014 American Chemical Society. Recently, Huang et al. [4] reported a breathing MOFs [Zn3(OH)2(btca)2]·DMF·4H2O (H2btca = benzotriazole-5-carboxylic acid). The detail structure analysis indicates that all above conditions are satisfied; however an isostructural MOF [Co3(OH)2(btca)2]·3.7H2O doesn’t exhibit the breathing effect upon dehydration [167], suggesting the breathing phenomenon of MOFs is a synergetic effect from the coordination of metal centers, and the strength of host–guest and guest–guest interactions. The author claimed that the formation of zigzag rod structure in parallel with the channel direction should be taken into the consideration, which was supported by O’Keeffe’s prediction [63]. Recently, Huang et al. [4] reported a breathing MOFs [Zn3(OH)2(btca)2]·DMF·4H2O (H2btca = benzotriazole-5-carboxylic acid). The detail structure analysis indicates that all above conditions are satisfied; however an isostructural MOF [Co3(OH)2(btca)2]·3.7H2O doesn’t exhibit the breathing effect upon dehydration [167], suggesting the breathing phenomenon of MOFs is a synergetic effect from the coordination of metal centers, and the strength of host–guest and guest–guest interactions. The author claimed that the formation of zigzag rod structure in parallel with the channel direction should be taken into the consideration, which was supported by O’Keeffe’s prediction [63]. 6. Outlooks In addition to the dimensionality of organic ligands and metal clusters, it is clearly noticeable that those MOFs exhibiting the breathing behavior have two common features: (1) non-rigid areas, or weak points, appear in the frameworks and (2) a free space within the framework to accommodate the steric hindrance of the movements of weak points [7]. Hence, some empirical rules have been developed to predict the presence of weak points in MOFs: (1) inorganic clusters might possess a mirror plane with the organic ligands in symmetrical position; (2) O–O axes of carboxylates are perpendicular to the elongation axis. Similar requirement can be noticed very recently [159]; (3) include even cycles at the level of the cluster and/or at the level of the topology of the skeleton; (4) the ratio C/M should be ≥2, whereas C is the number of carbons of the carboxylate surrounding the cluster, and M is the number of metallic atoms within the cluster; (5) breathing effects can only occur with ditopic carboxylates, but not on MOFs with tri- or tetratopic ligands. In addition, it is reported that the metal centers must possess μ2-hydroxyl groups that are preferred interacting sites for adsorbates, leading to the flexibility point. The presence of μ2-oxo groups in MIL-47 (VIV) still contribute to the rigidity of framework even though all above conditions are satisfied [161]. Furthermore, more flexible MOFs showing phenomenal breathing behaviors are synthesized via tritopic ligands [81,165] or tetratopic ligands [11,166]. The flexibility is evolved from even cycles in their skeleton topologies, despite of forming odd cycles 3238 Materials 2014, 7 Materials 2014, 7 function theory (DFT) and molecular dynamic simulations. Both approaches indicate that the breathing behaviors of MOFs evolve from synergetic effects of coordination symmetry of metal nodes and rotatable axis of organic ligands, as well as strong guest-host interactions. Post-synthetic modifications and pre-functionalized organic ligands are accommodated to rationally design the extent of breathing behaviors. Syntheses of MOFs are mainly conducted in solvothermal conditions, in which the cost and recovery of organic solvents becomes the major hurdle for large-scale production. Exploring viable hydrothermal synthesis conditions of MOFs are commensurate with the design strategies towards the breathing of MOFs. Author Contributions The authors contributed equally to this work. The authors contributed equally to this work. Acknowledgments The authors are grateful for the financial support from the Department of Chemical Engineering at Ryerson University, Faculty of Engineering and Architectural Science (FEAS) at Ryerson University, Ryerson Library Open Access Fund of the Library at Ryerson University, and Natural Sciences and Engineering Research Council of Canada-Discovery Grant (NSERC-DG). Conflicts of Interest The authors declare no conflict of interest. The authors declare no conflict of interest. 7. Conclusions Metal-organic frameworks (MOFs) have gained intense research attention in applications of gas adsorption/storage, gas separation, catalysts for fine chemicals, and biomedical imaging, among others, due to their unprecedented large surface area, tunable framework functionality, decent thermal stability, and the most striking phenomenon of framework flexibility (breathing). The breathing behaviors are characterized using in situ X-ray diffraction, gas adsorption, and nuclear magnetic resonance spectroscopy upon the exposure of various external stimuli, such as heat, pressure, and gases. Theoretical calculations are used to elucidate the principles of breathing behaviors using density 3239 References Effects of solvation on the framework of a breathing copper MOF employing a semirigid linker. Inorg. Chem. 2013, 42, 2182–2187. 12. Murdock, C.R.; Hughes, B.C.; Lu, Z.; Jenkins, D.M. Approaches for synthesizing breathing MOFs by exploiting dimensional rigidity. Coord. Chem. Rev. 2014, 258–259, 119–136. 13. MacGillivray, L.R. Metal-Organic Frameworks: Design and Application; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2010. 14. Rayner, J.H.; Powell, H.M. Structure of molecular compounds. Part X. Crystal structure of the compound of benzene with an ammonia-nickel cyanide complex. J. Chem. Soc. 1952, 319–328. 15. Hoskins, B.F.; Robson, R. Infinite polymeric frameworks consisting of three dimensionally linked rod-like segments. J. Am. Chem. Soc. 1989, 111, 5962–5964. 16. Hoskins, B.F.; Robson, R. Design and construction of a new class of scaffolding-like materials comprising infinite polymeric frameworks of 3D-linked molecular rods. A reappraisal of the Zn(CN)2 and Cd(CN)2 structures and the synthesis and structure of the diamond-related frameworks [N(CH3)4][CuIZnII(CN)4] and CuI[4,4′,4′′,4′′′-tetracyanotetraphenylmethane]BF4 xC6H5NO2. J. Am. Chem. Soc. 1990, 112, 1546–1554. 17. James, S.L. Metal-organic framework. Chem. Soc. Rev. 2003, 32, 276–288. 18. Kondo, M.; Yoshitomi, T.; Seki, K.; Matsuzaka, H.; Kitagawa, S. Three-dimensional framework with channeling cavities for small molecules: {M2(4,4'-bpy)3(NO3)4]∙xH2O}n (M = Co, Ni, Zn). Angew. Chem. Int. Ed. 1997, 36, 1725–1727. 19. Noro, S.-I.; Kitaura, R.; Kondo, M.; Kitagawa, S.; Ishii, T.; Matsuzaka, H.; Yamashita, M. Framework engineering by anions and porous functionalities of Cu(II)/4,4'-bpy coordination polymers. J. Am. Chem. Soc. 2002, 124, 2568–2583. 20. Pan, L.; Adams, K.M.; Hernandez, H.E.; Wang, X.; Zheng, C.; Hattori, Y.; Kaneko, K. Porous lanthanide-organic frameworks: Synthesis, characterization, and unprecedented gas adsorption properties. J. Am. Chem. Soc. 2003, 125, 3062–3067. 21. Kepert, C.J.; Rosseinsky, M.J. Zeolite-like crystal structure of an empty microporus molecular framework. Chem. Commun. 1999, 375–376. 22. Zeng, M.-H.; Feng, X.-L.; Chen, X.-M. Crystal-to-crystal transformations of a microporous metal-organic laminated framework triggered by guest exchange, dehydration and readsorption. Dalton Trans. 2004, 2217–2223. 23. Biradha, K.; Fujita, M. A springlike 3D-coordination network that shrinks or swells in a crystal-to-crystal manner upon guest removal or readsorption. Angew. Chem. Int. Ed. 2002, 41, 3392–3395. 24. Fletcher, A.J.; Thomas, K.M.; Rosseinsky, M.J. Flexibility in metal-organic framework materials: Impact on sorption properties. J. Solid State Chem. 2005, 178, 2491–2510. 25. Maspoch, D.; Ruiz-Molina, D.; Wurst, K.; Domingo, N.; Cavallini, M.; Biscarini, F.; Tejada, J.; Rovira, C.; Veciana, J. A nanoporous molecular magnet with reversible solvent-induced mechanical and magnetic properties. Nat. Mater. References 1. Lee, J.; Farha, O.K.; Roberts, J.; Scheidt, K.A.; Nguyen, S.T.; Hupp, J.T. Metal-organic framework materials as catalysts. Chem. Soc. Rev. 2009, 38, 1450–1459. 1. Lee, J.; Farha, O.K.; Roberts, J.; Scheidt, K.A.; Nguyen, S.T.; Hupp, J.T. Metal-organic framework materials as catalysts. Chem. Soc. Rev. 2009, 38, 1450–1459. 2. Mueller, U.; Schubert, M.; Teich, F.; Puetter, H.; Schierle-Arndt, K.; Pastré, J. Metal-organic frameworks—Prospective industrial applications. J. Mater. Chem. 2006, 16, 626–636. 2. Mueller, U.; Schubert, M.; Teich, F.; Puetter, H.; Schierle-Arndt, K.; Pastré, J. Metal-organic frameworks—Prospective industrial applications. J. Mater. Chem. 2006, 16, 626–636. 3. Beurroies, I.; Boulhout, M.; Llewellyn, P.L.; Kuchta, B.; Férey, G.; Serre, C.; Denoyel, R. Using pressure to provoke the structural transition of metal-organic frameworks. Angew. Chem. Int. Ed. 2010, 49, 7526–7529. 3. Beurroies, I.; Boulhout, M.; Llewellyn, P.L.; Kuchta, B.; Férey, G.; Serre, C.; Denoyel, R. Using pressure to provoke the structural transition of metal-organic frameworks. Angew. Chem. Int. Ed. 2010, 49, 7526–7529. 4. Xiao, J.; Wu, Y.; Li, M.; Liu, B.-Y.; Huang, X.-C.; Li, D. Crystalline structural intermediates of a breathing metal-organic framework that functions as a luminescent sensor and gas reservior. Chem. Eur. J. 2013, 19, 1891–1895. 4. Xiao, J.; Wu, Y.; Li, M.; Liu, B.-Y.; Huang, X.-C.; Li, D. Crystalline structural intermediates of a breathing metal-organic framework that functions as a luminescent sensor and gas reservior. Chem. Eur. J. 2013, 19, 1891–1895. 5. Khan, N.A.; Hasan, Z.; Jhung, S.H. Adsorptive removal of hazardous materials using metal-organic frameworks (MOFs): A review. J. Hazard. Mater. 2013, 244–245, 444–456. 5. Khan, N.A.; Hasan, Z.; Jhung, S.H. Adsorptive removal of hazardous materials using metal-organic frameworks (MOFs): A review. J. Hazard. Mater. 2013, 244–245, 444–456. 6. Li, J.-R.; Sculley, J.; Zhou, H.-C. Metal-organic frameworks for separations. Chem. Rev. 2012, 112, 869–912. 7. Férey, G.; Serre, C. Large breathing effects in three-dimensional porous hybrid matter: Facts, analyses, rules and consequences. Chem. Soc. Rev. 2009, 38, 1380–1399. 8. Yaghi, O.M.; O’Keeffe, M.; Ockwig, N.W.; Chae, H.K.; Eddaoudi, M.; Kim, J. Reticular synthesis and the design of new materials. Nature 2003, 423, 705–714. Materials 2014, 7 Materials 2014, 7 3240 9. Kitagawa, S.; Kitaura, R.; Noro, S.-I. Functional porous coordination polymers. Angew. Chem. Int. Ed. 2004, 43, 2334–2375. 10. Kitagawa, S.; Uemura, K. Dynamic porous properties of coordination polymers inspired by hydrogen bonds. Chem. Soc. Rev. 2005, 34, 109–119. 11. Murdock, C.R.; Lu, Z.; Jenkins, D.M. References 2003, 2, 190–195. Materials 2014, 7 Materials 2014, 7 3241 26. Loiseau, T.; Serre, C.; Huguenard, C.; Fink, G.; Taulelle, F.; Henry, M.; Bataille, T.; Férey, G. A rationale for the large breathing of the porous aluminum terephthalate (MIL-53) upon hydration. Chem. Eur. J. 2004, 10, 1373–1382. 27. Serre, C.; Bourrelly, S.; Vimont, A.; Ramsahye, N.A.; Maurin, G.; Llewellyn, P.L.; Daturi, M.; Filinchuk, Y.; Leynaud, O.; Barnes, P.; et al. An explanation for the very large breathing effect of a metal-organic framework during co2 adsorption. Adv. Mater. 2007, 19, 2246–2251. 28. Stock, N.; Biswas, S. Synthesis of metal-organic frameworks (MOFs): Routes to various MOF topologies, morphologies, and composites. Chem. Rev. 2012, 112, 933–969. 29. Meetk, S.T.; Greathouse, J.A.; Allendorf, M.D. Metal-organic frameworks: A rapidly growing class of versatile nanoporous materials. Adv. Mater. 2011, 23, 249–267. 30. Plater, M.J.; St. J. Foreman, M.R.; Howie, R.A.; Lachowski, E.E. Hydrothermal synthesis and characterization of M(pdc)∙3H2O (pdc = 2,5-pyridinedicarboxylate); M = Co, Ni, CoxNiy (x = 0.4 – 0.6, y = 0.6 – 0.4). J. Chem. Res. Synop. 1998, 1998, 754–755. 31. Gaab, M.; Trukhan, N.; Maurer, S.; Gummaraju, R.; Müller, U. The progression of al-based metal-organic-frameworks—From academic research to industrial production and applications. Microporous Mesoporous Mater. 2012, 157, 131–136. 32. Yilmaz, B.; Trukhan, N.; Müller, U. Industrial outlook on zeolites and metal organic frameworks. Chin. J. Catal. 2012, 33, 3–10. 33. Demazeau, G. Solvothermal processes: A route to the stabilization of new materials. J. Mater. Chem. 1999, 9, 15–18. 34. Laye, R.H.; McInnes, E.J.L. Solvothermal synthesis of paramagnetic molecular clusters. Eur. J. Inorg. Chem. 2004, 2811–2818. 35. Rabenau, A. The role of hydrothermal synthesis in preparative chemistry. Angew. Chem. Int. Ed. 1985, 24, 1026–1040. 36. Takako, N.; Takuji, I.; Chie, A.; Yasuhisa, H.; Yoshimichi, K.; Takaako, H. Solvothermal synthesis of -lit-type zeolite. Cryst. Growth Des. 2012, 12, 1752–1761. 37. Walton, R.I. Subcritical solvothermal synthesis of condensed inorganic materials. Chem. Soc. Rev. 2002, 31, 230–238. 38. Eddaoudi, M.; Kim, J.; Rosi, N.L.; Vodak, D.; Wachter, J.; O’Keeffe, M.; Yaghi, O.M. Systematic design of pore size and functionality in isoreticular MOFs and their application in methane storage. Science 2002, 298, 469–472. 39. Zhang, J.; Wojtas, L.; Larsen, R.W.; Eddaoudi, M.; Zaworotko, M.J. Temperature and concentration control over interpenetration in a metal-organic material. J. Am. Chem. Soc. 2009, 131, 17040–17041. 40. References Volkringer, C.; Loiseau, T.; Guillou, N.; Férey, G.; Haouas, M.; Taulelle, F.; Elkaim, E.; Stock, N. High-throughput aided synthesis of the porous metal-organic framework-type aluminum pyromellitate, MIL-121, with extra carboxylic acid functionalization. Inorg. Chem. 2010, 49, 9852–9862. 41. Rowsell, J.L.C.; Yaghi, O.M. Metal-organic frameworks: A new class of porous materials. Microporous Mesoporous Mater. 2004, 73, 3–14. 42. Bauer, S.; Serre, C.; Devic, T.; Horcajada, P.; Marrot, J.; Férey, G. High-throughput assisted rationalization of the formation of metal organic frameworks in the iron(III) aminoterephthalate solvothermal system. Inorg. Chem. 2008, 47, 7568–7576. Materials 2014, 7 Materials 2014, 7 3242 43. Long, P.; Wu, H.; Zhao, Q.; Wang, Y.; Dong, J.; Li, J. Solvent effect on the synthesis of MIL-96(Cr) and MIL-100(Cr). Microporous Mesoporous Mater. 2011, 142, 489–493. 44. Stavitski, E.; Goesten, M.; Juan-Alcañiz, J.; Martinez-Joaristi, A.; Serra-Crespo, P.; Petukhov, A.V.; Gascon, J.; Kapteijn, F. Kinetic control of metal-organic framework crystallization investigated by time-resolved in situ X-ray scattering. Angew. Chem. Int. Ed 2011, 50, 9624–9628. 45. Seo, Y.-K.; Yoon, J.W.; Lee, J.S.; Lee, U.-H.; Hwang, Y.K.; Jun, C.-H.; Horcajada, P.; Serre, C.; Chang, J.-S. Large scale fluorine-free synthesis of hierarchically porous iron (III) trimesate MIL-100(Fe) with a zeolite mtn topology. Microporous Mesoporous Mater. 2012, 157, 137–145. 46. Wang, Q.M.; Shen, D.; M. B.; Lau, M.L.; Deng, S.; Fitch, F.R.; Lemocoff, N.O.; Semanscin, J. Metallo-organic molecular sieve for gas separation and purification. Microporous Mesoporous Mater. 2002, 55, 217–230. 47. Barthelet, K.; Marrot, J.; Riou, D.; Férey, G. A breathing hybrid organic-inorganic solid with very large pores and high magnetic characteristics. Angew. Chem. Int. Ed. 2002, 41, 281–284. 48. Loiseau, T.; Mellot-Draznieks, C.; Muguerra, H.; Férey, G.; Haouas, M.; Taulelle, F. Hydrothermal synthesis and crystal structure of a new three-dimensional aluminum-organic framework MIL-69 with 2,6-naphthalenedicarboxylate (ndc), Al(OH)(ndc)∙H2O. C. R. Chimie 2005, 8, 765–772. 49. Loiseau, T.; Lecroq, L.; Volkringer, C.; Marrot, J.; Férey, G.; Haouas, M.; Taulelle, F.; Bourrelly, S.; Llewellyn, P.L.; Latroche, M. MIL-96, a porous aluminum trimesate 3D structure constructed from a hexagonal network of 18-membered rings and μ3-oxo-centered trinuclear units. J. Am. Chem. Soc. 2006, 128, 10223–10230. 50. Férey, G.; Serre, C.; Mellot-Draznieks, C.; Millange, F.; Surblé, S.; Dutour, J.; Margiolaki, I. A hybrid solid with giant pores prepared by a combination of targeted chemistry, simulation and powder diffraction. Angew. Chem. Int. Ed. 2004, 43, 6296–6301. 51. Materials 2014, 7 3243 Materials 2014, 7 56. Volkringer, C.; Loiseau, T.; Guillou, N.; Férey, G.; Elkaim, E. Syntheses and structures of the MOF-type series of metal 1,4,5,8,-naphthalenetetracarboxylates M2(OH)2[C14O8H4] (Al, Ga, In) with infinite trans-connected M–OH–M chain (MIL-122). Solid State Sci. 2009, 11, 1507–1512. 57. Wang, S.; Li, L.; Zhang, J.; Yuan, X.; Su, C.-Y. Anion-tuned sorption and catalytic properties of a soft metal-organic solid with polycatenated frameworks. J. Mater. Chem. 2011, 21, 7098–7104. 58. Kondo, A.; Kajiro, H.; Noguchi, H.; Carlucci, L.; Proserpio, D.M.; Ciani, G.; Kato, K.; Takata, M.; Seki, H.; Sakamoto, M.; et al. Super flexibility of a 2D Cu-based porous coordination framework on gas adsorption in comparison with a 3D framework of identical composition: Framework dimensionality-dependent gas adsorptivities. J. Am. Chem. Soc. 2011, 133, 10512–10522. 59. Kitaura, R.; Seki, K.; Akiyama, G.; Kitagawa, S. Porous coordination-polymer crystals with gated channels specific for supercritical gases. Angew. Chem. Int. Ed. 2003, 42, 428–431. 60. Huang, L.; Wang, H.; Chen, J.; Wang, Z.; Sun, J.; Zhao, D.; Yan, Y. Synthesis, morphology control, and properties of porous metal-organic coordination polymers. Microporous Mesoporous Mater. 2003, 58, 105–114. 61. Li, H.; Eddaoudi, M.; O'Keeffe, M.; Yaghi, O.M. Design and synthesis of an exceptionally stable and highly porous metal-organic framework. Nature 1999, 402, 276–279. 62. Tranchemontagne, D.J.; Hunt, J.R.; Yaghi, O.M. Room temperature synthesis of metal-organic frameworks: MOF-5, MOF-74, MOF-177, MOF-199, and IRMOF-0. Tetrahedron 2008, 64, 8553–8557. 63. Rosi, N.L.; Kim, J.; Eddaoudi, M.; Chen, B.; O’Keeffe, M.; Yaghi, O.M. Rod packings and metal-organic frameworks constructed from rod-shaped secondary building units. J. Am. Chem. Soc. 2005, 127, 1504–1518. 64. Chae, H.K.; Siberio-Pérez, D.Y.; Kim, J.; Go, Y.; Eddaoudi, M.; Matzger, A.J.; O'Keeffe, M.; Yaghi, O.M. A route to high surface area porosity and inclusion of large molecules in crystals. Nature 2004, 417, 523–527. 65. Chui, S.S.-Y.; Lo, S.M.-F.; Charmant, J.P.H.; Orpen, A.G.; Williams, I.D. A chemically functionalizable nanoporous material [Cu3(TMA)2(H2O)3]n. Science 1999, 283, 1148–1150. 66. Rowsell, J.J.C.; Yaghi, O.M. Effects of functionalization, catenation, and variation of the metal oxide and organic linking units on the low-pressure hydrogen adsorption properties of metal-organic frameworks. J. Am. Chem. Soc. 2006, 128, 1304–1315. 67. Ma, M.; Bétard, A.; Weber, I.; Al-Hokbany, N.S.; Fischer, R.A.; Metzler-Nolte, N. Iron-based metal-organic frameworks MIL-88b and NH2-MIL-88b: High quality microwave synthesis and solvent-induced lattice “Breathing”. Cryst. Growth Des. 2013, 13, 2286–2291. 68. Surblé, S.; Serre, C.; Mellot-Draznieks, C.; Millange, F.; Férey, G. References Yoon, J.W.; Seo, Y.-K.; Hwang, Y.K.; Chang, J.-S.; Leclerc, H.; Wuttke, S.; Bazin, P.; Vimont, A.; Daturi, M.; Bloch, E.; et al. Controlled reducibility of a metal-organic framework with coordinatively unsaturated sites for preferential gas sorption. Angew. Chem. Int. Ed. 2010, 49, 5949–5952. 52. Volkringer, C.; Popov, D.; Loiseau, T.; Férey, G.; Burghammer, M.; Riekel, C.; Haouas, M.; Taulelle, F. Synthesis, single-crystal x-ray microdiffraction, and nmr characterizations of the giant pore metal-organic framework aluminum trimesate MIL-100. Chem. Mater. 2009, 21, 5695–5697. 53. Volkringer, C.; Popov, D.; Loiseau, T.; Guillou, N.; Férey, G.; Haouas, M.; Taulelle, F.; Mellot-Draznieks, C.; Burghammer, M.; Riekel, C. A microdiffraction set-up for nanoporous metal-organic framework type solids. Nat. Mater. 2007, 6, 760–764. 54. Volkringer, C.; Loiseau, T.; Guillou, N.; Férey, G.; Haouas, M.; Taulelle, F.; Audebrand, N.; Margiolaki, I.; Popov, D.; Burghammer, M.; et al. Structural transitions and flexibility during dehydration-rehydration process in the MOF-type aluminum pyromellitate Al2(OH)2[C10O8H2] (MIL-118). Cryst. Growth Des. 2009, 9, 2927–2936. 55. Volkringer, C.; Loiseau, T.; Haouas, M.; Taulelle, F.; Popov, D.; Burghammer, M.; Riekel, C.; Zlotea, C.; Cuevas, F.; Latroche, M.; et al. Occurrence of uncommon infinite chains consisting of edge-sharing octahedra in a porous metal organic framework-type aluminum pyromellitate Al4(OH)8[C10O8H2] (MIL-120): Synthesis, structure, and gas sorption properties. Chem. Mater. 2009, 21, 5783–5791. Materials 2014, 7 A new isoreticular class of metal-organic frameworks with the MIL-88 topology. Chem. Commun. 2006, 284–286. 69. Salles, F.; Maurin, G.; Serre, C.; Llewellyn, P.L.; Knöfel, C.; Choi, H.J.; Filinchuk, Y.; Oliviero, L.; Vimont, A.; Long, J.R.; et al. Multistep N2 breathing in the metal-organic framework Co(1,4-benzenedipyrazolate). J. Am. Chem. Soc. 2010, 132, 13782–13788. 70. Grzywa, M.; Geβner, C.; Denysenko, D.; Bredenkötter, B.; Gschwind, F.; Fromm, K.M.; Nitek, W.; Klemm, E.; Volkmer, D. CFA-2 and CFA-3 (coordination framework augsburg university-2 and -3); novel MOFs assembled from trinuclear Cu(I)/Ag(I) secondary building units and 3,3',5,5'-tetraphenyl-bipyrazolate ligands. Dalton Trans. 2013, 42, 6909–6921. Materials 2014, 7 3244 71. Thallapally, P.K.; Tian, J.; Kishan, M.R.; Fernandez, C.A.; Dalgarno, S.J.; McGrail, P.B.; Warren, J.E.; Atwood, J.L. Flexible (breathing) interpenetrated metal-organic frameworks for CO2 separation applications. J. Am. Chem. Soc. 2008, 130, 16842–16843. 71. Thallapally, P.K.; Tian, J.; Kishan, M.R.; Fernandez, C.A.; Dalgarno, S.J.; McGrail, P.B.; Warren, J.E.; Atwood, J.L. Flexible (breathing) interpenetrated metal-organic frameworks for CO2 separation applications. J. Am. Chem. Soc. 2008, 130, 16842–16843. 72. Dybtsev, D.N.; Chun, H.; Kim, K. Rigid and flexible: A highly porous metal-organic framework with unusual guest-dependent dynamic behavior. Angew. Chem. Int. Ed. 2004, 43, 5033–5036. 73. Uemura, K.; Yamasaki, Y.; Komagawa, Y.; Tanaka, K.; Kita, H. Two-step adsorption/desorption on a jungle-gym-type porous coordination polymer. Angew. Chem. Int. Ed. 2007, 46, 6662–6665. 74. Tanaka, K.; Nakagawa, K.; Higuchi, M.; Horike, S.; Kubota, Y.; Kobayashi, T.C.; Takata, M.; Kitagawa, S. Kinetic gate-opening process in a flexible porous coordination polymer. Angew. Chem. Int. Ed. 2008, 47, 3914–3918. 75. Mu, B.; Li, F.; Huang, Y.; Walton, K.S. Breathing effects of CO2 adsorption on a flexible 3D lanthanide metal-organic framework. J. Mater. Chem. 2012, 22, 10172–10178. 76. Henke, S.; Schneemann, A.; Wütscher, A.; Fischer, R.A. Directing the breathing behavior of pillared-layered metal-organic frameworks via a systematic library of functionalized linkers bearing flexible substituents. J. Am. Chem. Soc. 2012, 134, 9464–9474. 77. Hijikata, Y.; Horike, S.; Sugimoto, M.; Sato, H.; Matsuda, R.; Kitagawa, S. Relationship between channel and sorption properties in coordination polymers with interdigitated structures. Chem. Eur. J. 2011, 17, 5138–5144. 78. Yang, C.; Wang, X.; Omary, M.A. Fluorous metal-organic frameworks for high-density gas adsorption. J. Am. Chem. Soc. 2007, 129, 15454–15455. 79. Yang, C.; Wang, X.; Omary, M.A. Crystallographic observation of dynamic gas adsorption sites and thermal expansion in a breathable fluorous metal-organic framework. Angew. Chem. Int. Ed. 2009, 48, 2500–2505. 80. Materials 2014, 7 Materials 2014, 7 3245 86. Llewellyn, P.L.; Bourrelly, S.; Serre, C.; Filinchuk, Y.; Férey, G. How hydration drastically improves adsorption selectivity for CO2 over CH4 in the flexible chromium terephthalate MIL-53. Angew. Chem. Int. Ed. 2006, 45, 7751–7754. 86. Llewellyn, P.L.; Bourrelly, S.; Serre, C.; Filinchuk, Y.; Férey, G. How hydration drastically improves adsorption selectivity for CO2 over CH4 in the flexible chromium terephthalate MIL-53. Angew. Chem. Int. Ed. 2006, 45, 7751–7754. 87. Paesani, F. Molecular mechanisms of water-mediated proton transport in MIL-53 metal-organic frameworks. J. Phys. Chem. C 2013, 117, 19508–19516. 88. Li, J.; Kuppler, R.J.; Zhou, H.-C. Selective gas adsorption and separation in metal-organic frameworks. Chem. Soc. Rev. 2009, 38, 1477–1504. 89. Murray, L.J.; Dincă, M.; Long, J.R. Hydrogen storage in metal-organic frameworks. Chem. Soc. Rev. 2009, 38, 1294–1315. 90. Klein, N.; Hoffmann, H.C.; Cadiau, A.; Getzschmann, J.; Lohe, M.R.; Paasch, S.; Heydenreich, T.; Adil, K.; Senkovska, I.; Brunner, E.; et al. Structural flexibility and intrinsic dynamics in the M2(2,6-ndc)2(dabco) (M = Ni, Cu, Co, Zn) metal-organic frameworks J. Mater. Chem. 2012, 22, 10303–10312. 91. Rouquerol, F.; Rouquerol, J.; Sing, K. Adsorption by Powders and Porous Solids: Principles, Methodology and Applications; Academic Press: Waltham, MA, USA, 1999. 92. Do, X.-D.; Hoang, V.-T.; Serge, K. MIL-53(Al) mesostructured metal-organic frameworks. Microporous Mesoporous Mater. 2011, 141, 135–139. 93. Rallapalli, P.; Prasanth, K.P.; Patil, D.; Somanti, R.S.; Jasra, R.V.; Bajaj, H.C. Sorption studies of CO2, CH4, N2, CO, O2, and ar on nanoporous aluminum terephthalate [MIL-53(Al)]. J. Porous Mater. 2011, 18, 205–210. 94. Jhung, S.H.; Khan, N.A.; Hasan, Z. Analogous porous metal-organic frameworks: Synthesis, stability, and application in adsorption. CrystEngComm 2012, 14, 7099–7109. 95. Mowat, J.P.S.; Miller, S.R.; Slawin, A.M.Z.; Seymour, V.R.; Ashbrook, S.E.; Wright, P.A. Synthesis, characterization, and adsorption properties of microporous scandium carboxylates with rigid and flexible frameworks. Microporous Mesoporous Mater. 2011, 142, 322–333. 96. Nouar, F.; Devic, T.; Chevreau, H.; Guillou, N.; Gibson, E.; Clet, G.; Daturi, M.; Vimont, A.; Grenèche, J.M.; Breeze, M.I.; et al. Tuning the breathing behavior of MIL-53 by cation mixing. Chem. Commun. 2012, 48, 10237–10239. 97. Chen, L.; Mowat, J.P.S.; Fairen-Jimenez, D.; Morrison, C.A.; Thompson, S.P.; Wright, P.A.; Düren, T. Elucidating the breathing of the metal-organic framework MIL-53(Sc) with ab initio molecular dynamics simulations and in situ X-ray powder diffraction experiments. J. Am. Chem. Soc. 2013, 135, 15763–15773. 98. Khan, N.A.; Jun, J.W.; Jeong, J.H.; Jhung, S.H. Materials 2014, 7 Chen, Z.; Ling, Y.; Yang, H.; Guo, Y.; Weng, L.; Zhou, Y. Two-step synthesis, structure and adsorption property of a dynamic zinc phosphonocarboxylate framework. CrystEngComm 2011, 13, 3378–3383. 81. Kyprianidou, E.J.; Papaefstathiou, G.S.; Manos, M.J.; Tasiopoulos, A.J. A flexible Cd2+ metal organic framework with a unique (3,3,6)-connected topology, unprecedented secondary building units and single crystal to single crystals solvent exchange properties. CrystEngComm 2012, 14, 8368–8373. 82. Fernandez, C.A.; Thallapally, P.K.; Motkuri, R.K.; Nune, S.K.; Sumrak, J.C.; Tian, J.; Liu, J. Gas-induced expansion and contraction of a fluorinated metal-organic framework. Cryst. Growth Des. 2010, 10, 1037–1039. 83. Gustafsson, M.; Su, J.; Yue, H.; Yao, Q.; Zou, X. A family of flexible lanthanide bipyridinedicarboxylate metal-organic frameworks showing reversible single-crystal to single-crystal transformations. Cryst. Growth Des. 2012, 12, 3423–3429. 84. Procopio, E.Q.; Fukushima, T.; Barea, E.; Navarro, J.A.R.; Horike, S.; Kitagawa, S. A soft copper (II) porous coordination polymer with unprecedented aqua bridge and selective adsorption properties. Chem. Eur. J. 2012, 18, 13117–13125. 85. Jiang, Z.-Q.; Jiang, G.-Y.; Wang, F.; Zhao, Z.; Zhang, J. Controlling state of breathing of two isoreticular microporous metal-organic frameworks with triazole homologues. Chem. Eur. J. 2012, 18, 10525–10529. Materials 2014, 7 Remarkable adsorptive performance of a metal-organic framework, vanadium-benzenedicarboxylate (MIL-47), for benzothiophene. Chem. Commun. 2011, 47, 1306–1308. 99. Bourrelly, S.; Llewellyn, P.L.; Serre, C.; Millange, F.; Loiseau, T.; Férey, G. Different adsorption behaviors of methane and carbon dioxide in the isotypic nanoporous metal terephthalates MIL-53 and MIL-47. J. Am. Chem. Soc. 2005, 127, 13519–13521. 100. Sing, K.S.W.; Everett, D.H.; Haul, R.A.W.; Moscou, L.; Pierotti, R.A.; Rouquerol, F.; Siemieniewska, T. Reporting physisorption data for gas/solid systems with special reference to the determination of surface area and porosity. Pure Appl. Chem. 1985, 57, 603619. Materials 2014, 7 Materials 2014, 7 3246 101. Ramsahye, N.A.; Trung, T.K.; Bourrelly, S.; Yang, Q.; Devic, T.; Maurin, G.; Horcajada, P.; Llewellyn, P.L.; Yot, P.; Serre, C.; et al. Influence of the organic lignand functionalization on the breathing of the porous iron terephthalate metal organic framework type material upon hydrocarbon adsorption. J. Phys. Chem. C 2011, 115, 18683–18695. 102. Fletcher, A.J.; Cussen, E.J.; Prior, T.J.; Rosseinsky, M.J.; Kepert, C.J.; Thomas, K.M. Adsorption dynamics of gases and vapors on the nanoporous metal organic framework material Ni2(4,4'-bipyridine)3(NO3)4: Guest modification of host sorption behavior. J. Am. Chem. Soc. 2001, 123, 10001–10011. 103. Walton, K.S.; Millward, A.R.; Dubbeldam, D.; Frost, H.; Low, J.J.; Yaghi, O.M.; Snurr, R.Q. Understanding inflections and steps in carbon dioxide adsorption isotherms in metal-organic frameworks. J. Am. Chem. Soc. 2008, 130, 406–407. 104. Boutin, Z.; Springuel-Huet, M.-A.; Nossov, A.; Gédéon, A.; Loiseau, T.; Volkringer, C.; Férey, G.; Goudert, F.-X.; Fuchs, A.H. Breathing transitions in MIL-53(Al) metal-organic framework upon xenon adsorption. Angew. Chem. Int. Ed. 2009, 48, 8314–8317. 105. Trung, T.K.; Trens, P.; Tanchoux, N.; Bourrelly, S.; Llewellyn, P.L.; Lorea-Serna, S.; Serre, C.; Loiseau, T.; Fajula, F.; Férey, G. Hydrocarbon adsorption in the flexible metal organic frameworks MIL-53(Al, Cr). J. Am. Chem. Soc. 2008, 130, 16926–16932. 106. Llewellyn, P.L.; Horcajada, P.; Maurin, G.; Devic, T.; Rosenbach, N.; Bourrelly, S.; Serre, C.; Vincent, D.; Lorea-Serna, S.; Filinchuk, Y.; et al. Complex adsorption of short linear alkanes in the flexible metal-organic framework MIL-53 (Fe). J. Am. Chem. Soc. 2009, 131, 13002–13008. 107. Trung, T.K.; Déroche, I.; Rivera, A.; Yang, Q.; Yot, P.; Ramsahye, N.A.; Vinot, S.D.; Devic, T.; Horcajada, P.; Serre, C.; et al. Hydrocarbon adsorption in the isostructural metal organic frameworks MIL-53(Cr) and MIL-47(V). Microporous Mesoporous Mater. 2011, 140, 114–119. 108. Serre, C.; Millange, F.; Thouvenot, C.; Nogués, M.; Marsolier, G.; Louër, D.; Férey, G. Materials 2014, 7 Very large breathing effect in the first nanoporous chromium(III)-based solids: MIL-53 or CrIII(OH)∙{O2C–C6H4–CO2}∙{HO2C–C6H4–CO2H}x∙H2Oy. J. Am. Chem. Soc. 2002, 124, 13519–13526. 109. Férey, G.; Millange, F.; Morcrette, M.; Serre, C.; Doublet, M.-L.; Grenèche, J.M.; Tarascon, J.-M. Mixed-valence Li/Fe-based metal-organic framework with both reversible redox and sorption properties. Angew. Chem. Int. Ed. 2007, 46, 3259–3263. 110. Springuel-Huet, M.-A.; Nossov, A.; Adem, Z.; Guenneau, F.; Volkringer, C.; Loiseau, T.; Férey, G.; Gédéon, A. 129Xe NMR study of the framework flexibility of the porous hybrid MIL-53(Al). J. Am. Chem. Soc. 2010, 132, 11599–11607. 111. Biswas, S.; Ahnfeldt, T.; Stock, N. New functionalized flexible Al-MIL-53-X (X = -Cl, -Br, -CH3, -NO2, -(OH)2) solids: Syntheses, characterization, sorption, and breathing behavior. Inorg. Chem. 2011, 50, 9518–9526. 112. Liu, Y.; Her, J.-H.; Dailly, A.; Ramirez-Cuesta, A.J.; Neumann, D.A.; Brown, C.M. Reversible structural transition in MIL-53 with large temperature hysteresis. J. Am. Chem. Soc. 2008, 130, 11813–11818. 113. Stavitski, E.; Pidko, E.A.; Couck, S.; Remy, T.; Hensen, E.J.M.; Weckhuysen, B.M.; Denayer, J.F.M.; Gascon, J.; Kapteijn, F. Complexity behind CO2 capture on NH2-MIL-53(Al). Langmuir 2011, 27, 3970–3976. Materials 2014, 7 3247 Materials 2014, 7 114. Millange, F.; Guillou, N.; Walton, R.I.; Grenèche, J.M.; Margiolaki, I.; Férey, G. Effect of the nature of the metal on the breathing steps in MOFs with dynamic frameworks. Chem. Commun. 2008, 39, 4732–4734. 114. Millange, F.; Guillou, N.; Walton, R.I.; Grenèche, J.M.; Margiolaki, I.; Férey, G. Effect of the nature of the metal on the breathing steps in MOFs with dynamic frameworks. Chem. Commun. 2008, 39, 4732–4734. 115. Devic, T.; Horcajada, P.; Serre, C.; Salles, F.; Maurin, G.; Moulin, B.; Heurtaux, D.; Clet, G.; Vimont, A.; Grenèche, J.M.; et al. Functionalization in flexible porous solids: Effects on the pore opening and the host-guest interactions. J. Am. Chem. Soc. 2010, 132, 1127–1136. 116. Férey, G.; Latroche, M.; Serre, C.; Millange, F.; Loiseau, T.; Percheron-Guégan, A. Hydrogen adsorption in the nanoporous metal-benzenedicarboxylate M(OH)(O2C–C6H4–CO2) (M = Al3+, Cr3+), MIL-53. Chem. Commun. 2003, 24, 2976–2977. 117. Llewellyn, P.L.; Maurin, G.; Devic, T.; Lorea-Serna, S.; Rosenbach, N.; Serre, C.; Bourrelly, S.; Horcajada, P.; Filinchuk, Y.; Férey, G. Prediction of the conditions for breathing of metal organic framework materials using a combination of X-ray powder diffraction, microcalorimetry, and molecular simulation. J. Am. Chem. Soc. 2008, 130, 12808–12814. 118. Bourrelly, S.; Moulin, B.; Rivera, A.; Maurin, G.; Devautour-Vinot, S.; Serre, C.; Devic, T.; Horcajada, P.; Vimont, A.; Clet, G.; et al. Materials 2014, 7 Explanation of the adsorption of polar vapors in the highly flexible metal organic framework MIL-53(Cr). J. Am. Chem. Soc. 2010, 132, 9488–9498. 119. Volkringer, C.; Loiseau, T.; Guillou, N.; Férey, G.; Elkaim, E.; Vimont, A. XRD and IR structural investigations of a particular breathing effect in the MOF-type gallium terephthalate MIL-53(Ga). Dalton Trans. 2009, 12, 2241–2249. 120. Mowat, J.P.S.; Seymour, V.R.; Griffin, J.M.; Thompson, S.P.; Slawin, A.M.Z.; Fairen-Jimenez, D.; Düren, T.; Ashbrook, S.E.; Wright, P.A. A novel structural form of MIL-53 observed for the scandium analogue and its response to temperature variation and CO2 adsorption. Dalton Trans. 2012, 41, 3937–3942. 121. Millange, F.; Serre, C.; Guillou, N.; Férey, G.; Walton, R.I. Structural effects of solvents on the breathing of metal-organic frameworks: An in situ diffraction study. Angew. Chem. Int. Ed. 2008, 47, 4100–4105. 122. Hoffmann, H.C.; Assfour, B.; Epperlein, F.; Klein, N.; Paasch, S.; Senkovska, I.; Kaskel, S.; Seifert, G.; Brunner, E. High-pressure in situ 129Xe NMR spectroscopy and computer simulations of breathing transitions in the metal-organic framework Ni2(2,6-ndc)2(dabco) (DUT-8(Ni)). J. Am. Chem. Soc. 2011, 133, 8681–8690. 123. Sheveleva, A.M.; Kolokolov, D.I.; Gabrienko, A.A.; Stepanov, A.G.; Gromilov, S.A.; Shundrina, I.K.; Sagdeev, R.Z.; Fedin, M.V.; Bagryanskaya, E.G. Structural dynamics in a “Breathing” Metal-organic framework studied by electron paramagnetic resonance of nitroxide spin probe. J. Phys. Chem. Lett. 2014, 5, 20–24. 124. Lescouet, T.; Kockrick, E.; Bergeret, G.; Pera-Titus, M.; Farrusseng, D. Engineering MIL-53(Al) flexibility by controlling amino tags. Dalton Trans. 2011, 40, 11359–11361. 125. Coudert, F.-X.; Jeffroy, M.; Fuchs, A.H.; Boutin, A.; Mellot-Draznieks, C. Thermodynamics of guest-induced structural transitions in hybrid organic-inorganic frameworks. J. Am. Chem. Soc. 2008, 130, 14294–14302. 126. Coudert, F.-X.; Mellot-Draznieks, C.; Fuchs, A.H.; Boutin, A. Double structural transition in hybrid material MIL-53 upon hydrocarbon adsorption: The thermodynamics behind the scenes. J. Am. Chem. Soc. 2009, 131, 3442–3443. Materials 2014, 7 3248 Materials 2014, 7 127. Ghoufi, A.; Maurin, G.; Férey, G. Physics behind the guest-assisted structural transitions of a porous metal-organic framework material. J. Phys. Chem. Lett. 2010, 1, 2810–2815. 128. Neimark, A.V.; Coudert, F.-X.; Boutin, A.; Fuchs, A.H. Stress-based model for the breathing of metal-organic frameworks. J. Phys. Chem. Lett. 2010, 1, 445–449. 129. Ramsahye, N.A.; Maurin, G.; Bourrelly, S.; Llewellyn, P.L.; Loiseau, T.; Serre, C.; Férey, G. On the breathing effect of a metal-organic framework upon CO2 adsorption: Monte carlo compared to microcalorimetry experiments. Chem. Commun. 2007, 3261–3263. 130. Materials 2014, 7 Materials 2014, 7 Materials 2014, 7 3249 144. Deng, H.; Doonan, C.J.; Furukawa, H.; Ferreira, R.B.; Towne, J.; Knobler, C.B.; Wang, B.; Yaghi, O.M. Multiple functional groups of varying ratios in metal-organic frameworks. Science 2010, 327, 846–851. 145. Lescouet, T.; Kockrick, E.; Bergeret, G.; Pera-Titus, M.; Aguado, S.; Farrusseng, D. Homogeneity of flexible metal-organic frameworks containing mixed linkers. J. Mater. Chem. 2012, 22, 10287–10293. 146. Henke, S.; Schneemann, A.; Fischer, R.A. Massive anisotropic thermal expansion and thermo-responsive breathing in metal-organic frameworks modulated by linker functionalization. Adv. Funct. Mater. 2013, 23, 5990–5996. 147. Serra-Crespo, P.; Gobechiya, E.; Ramos-Fernandez, E.V.; Juan-Alcañiz, J.; Martinez-Joaristi, A.; Stavitski, E.; Kirschhock, C.E.A.; Martens, J.A.; Kapteijn, F.; Gascon, J. Interplay of metal node and amine functionality in NH2-MIL-53: Modulating breathing behavior through intra-framework interactions. Langmuir 2012, 28, 12916–12922. 148. Ahnfeldt, T.; Gunzelmann, D.; Loiseau, T.; Hirsemann, D.; Senker, J.; Férey, G.; Stock, N. Synthesis and modification of a functionalized 3D open-framework structure with MIL-53 topology. Inorg. Chem. 2009, 48, 3057–3064. 149. Couck, S.; Denayer, J.F.M.; Baron, G.V.; Rémy, T.; Gascon, J.; Kapteijn, F. An amine-functionalized MIL-53 metal-organic framework with large separation power for CO2 and CH4. J. Am. Chem. Soc. 2009, 131, 6326–6327. 150. Wang, Z.; Cohen, S.M. Modulating metal-organic frameworks to breathe: A postsynthetic covalent modification approach. J. Am. Chem. Soc. 2009, 131, 16675–16677. 151. Biswas, S.; Rémy, T.; Couck, S.; Denysenko, D.; Rampelberg, G.; Denayer, J.F.M.; Volkmer, D.; Detavernier, C.; van der Voort, P. Partially fluorinated MIL-47 and Al-MIL-53 frameworks: Influence of functionalization on sorption and breathing properties. Phys. Chem. Chem. Phys. 2013, 15, 3552–3561. 152. Biswas, S.; Couck, S.; Denysenko, D.; Bhunia, A.; Grzywa, M.; Denayer, J.F.M.; Volkmer, D.; Janiak, C.; van der Voort, P. Sorption and breathing properties of difluorinated MIL-47 and Al-MIL-53 frameworks. Microporous Mesoporous Mater. 2013, 181, 175–181. 153. Chen, B.; Xiang, S.; Qian, G. Metal-organic frameworks with functional pores for regonition of small molecules. Acc. Chem. Res. 2010, 43, 1115–1124. 154. Chen, B.; Ma, S.; Zapata, F.; Fronczek, F.R.; Lobkovsky, E.B.; Zhou, H.-C. Rationally designed micropores within a metal-organic framework for selective sorption of gas molecules. Inorg. Chem. 2007, 46, 1233–1236. 155. Xiang, S.; Zhou, W.; Gallegos, J.M.; Liu, Y.; Chen, B. Exceptionally high acetylene uptake in a microporous metal-organic framework with open metal sites. J. Am. Chem. Soc. 2009, 131, 12415–12419. 156. Taddei, M.; Costantino, F.; Ienco, A.; Comotti, A.; Dau, P.V.; Cohen, S.M. Materials 2014, 7 Ramsahye, N.A.; Maurin, G.; Bourrelly, S.; Llewellyn, P.L.; Serre, C.; Loiseau, T.; Devic, T.; Férey, G. Probing the adsorption sites for CO2 in metal organic frameworks materials MIL-53 (Al, Cr) and MIL-47 (V) by density functional theory. J. Phys. Chem. C 2008, 112, 514–520. 131. Coudert, F.-X.; Boutin, A.; Fuchs, A.H.; Neimark, A.V. Adsorption deformation and structural transitions in metal-organic frameworks: From the unit cell to the crystal. J. Phys. Chem. Lett. 2013, 4, 3198–3205. 132. Ghysels, A.; Vanduyfhuys, L.; Vandichel, M.; Waroquier, M.; van Speybroeck, V.; Smit, B. On the thermodynamics of framework breathing: A free energy model for gas adsorption in MIL-53. J. Phys. Chem. C 2013, 117, 11540–11554. 133. Bousquet, D.; Coudert, F.-X.; Fossati, A.G.J.; Neimark, A.V.; Fuchs, A.H.; Boutin, A. Adsorption induced transitions in soft porous crystals: An osmotic potential approach to multistability and intermediate structures. J. Chem. Phys. 2013, 138, doi:10.1063/1.4802888. 134. Kolokolov, D.I.; Jobic, H.; Stepanov, A.G.; Guillerm, V.; Devic, T.; Serre, C.; Férey, G. Dynamics of benzene rings in MIL-53(Cr) and MIL-47(V) frameworks studies by 2H NMR spectroscopy. Angew. Chem. Int. Ed. 2010, 49, 4791–4794. 135. Greathouse, J.A.; Allendorf, M.D. Force field validation for molecular dynamics simulations of IRMOF-1 and other isoreticular zinc carboxylate coordination polymers. J. Phys. Chem. C 2008, 112, 5795–5802. 136. Wang, Z.; Cohen, S.M. Postsynthetic covalent modification of a neutral metal-organic framework. J. Am. Chem. Soc. 2007, 129, 12368–12369. 137. Bae, Y.-S.; Farha, O.K.; Hupp, J.T.; Snurr, R.Q. Enhancement of CO2/N2 selectivity in a metal-organic framework by cavity modification. J. Mater. Chem. 2011, 19, 2131–2134. 138. Park, H.J.; Cheon, Y.E.; Suh, M.P. Post-synthetic reversible incorporation of organic linkers into porous metal-organic frameworks through single-crystal-single-crystal transformations and modification of gas-sorption properties. Chem. Eur. J. 2010, 16, 11662–11669. 139. Savonnet, M.; Bazer-Bachi, D.; Bats, N.; Perez-Pellitero, J.; Jeanneau, E.; Lecocq, V.; Pinel, C.; Farrusseng, D. Generic postfunctionalization route from amino-derived metal-organic frameworks. J. Am. Chem. Soc. 2010, 132, 4518–4519. 140. Cohen, S.M. Modifying MOFs: New chemistry, new materials. Chem. Sci. 2010, 1, 32–36. 141. Wang, Z.; Cohen, S.M. Postsynthetic modification of metal-organic frameworks. Chem. Soc. Rev. 2009, 38, 1315–1329. 142. Seo, J.S.; Whang, D.; Lee, H.; Jun, S.I.; Oh, J.; Jeon, Y.J.; Kim, K. A homochiral metal-organic porous material for enantioselective separation and catalysis. Nature 2000, 404, 982–986. 143. Cohen, S.M. Postsynthetic methods for the functionalization of metal-organic frameworks. Chem. Rev. 2012, 112, 970–1000. Materials 2014, 7 Synthesis, breathing, and gas sorption study of the first isoreticular mixed-linker phosphonate based metal-organic frameworks. Chem. Commun. 2013, 49, 1315–1317. 157. Leus, K.; Couck, S.; Vandichel, M.; Vanhaelweyn, G.; Liu, Y.-Y.; Marin, G.B.; van Driessche, I.; Depla, D.; Waroquier, M.; van Speybroeck, V.; et al. Synthesis, characterization and sorption properties of NH2-MIL-47. Phys. Chem. Chem. Phys. 2012, 14, 15562–15570. Materials 2014, 7 3250 Materials 2014, 7 158. Karra, J.R.; Walton, K.S. Molecular simulations and experimental studies of CO2, CO, and N2 adsorption in metal-organic frameworks. J. Phys. Chem. C 2010, 114, 15735–15740. 159. Liu, Y.-Y.; Couck, S.; Vandichel, M.; Grzywa, M.; Leus, K.; Biswas, S.; Volkmer, D.; Gascon, J.; Kapteijn, F.; Denayer, J.F.M.; et al. New V(IV)-based metal-organic framework having framework flexibility and high CO2 adsorption capacity. Inorg. Chem. 2013, 52, 113–120. 160. Yot, P.G.; Ma, Q.; Haines, J.; Yang, Q.; Ghoufi, A.; Devic, T.; Serre, C.; Dmitriev, V.; Férey, G.; Zhong, C.; et al. Large breathing of the MOF MIL-47(VIV) under mechanical pressure: A joint experimental-modelling exploration. Chem. Sci. 2012, 3, 1100–1104. 161. Leclerc, H.; Devic, T.; Devautour-Vinot, S.; Bazin, P.; Audebrand, N.; Férey, G.; Daturi, M.; Vimont, A.; Clet, G. Influence of the oxidation state of the metal center on the flexibility and adorption properties of a porous metal organic framework: MIL-47(V). J. Phys. Chem. C 2011, 115, 19828–19840. 162. Kozachuk, O.; Meilikhov, M.; Yusenko, K.; Schneemann, A.; Jee, B.; Kuttatheyil, A.V.; Bertmer, M.; Sternemann, C.; Pöppl, A.; Fischer, R.A. A solid-solution approach to mixed-metal metal-organic frameworks–detailed characterization of local structures, defects and behavior of al/v frameworks. Eur. J. Inorg. Chem. 2013, 4546–4557. 163. Breeze, M.I.; Clet, G.; Campo, B.C.; Vimont, A.; Daturi, M.; Grenèche, J.-M.; Dent, A.J.; Millange, F.; Walton, R.I. Isomorphous substitution in a flexible metal-organic framework: Mixed-metal, mixed-valent MIL-53 type materials. Inorg. Chem. 2013, 52, 8171–8182. 164. Mukherjee, G.; Biradha, K. Modulation of breathing behavior of layered coordination polymers via a solid solution approach: The influence of metal ions on sorption behavior. Chem. Commun. 2014, 50, 670–672. 165. Wei, Y.-S.; Chen, K.-J.; Liao, P.-Q.; Zhu, B.-Y.; Lin, R.-B.; Zhou, H.-L.; Wang, B.-Y.; Xue, W.; Zhang, J.-P.; Chen, X.-M. Turning on the flexibility of isoreticular porous coordination frameworks for drastically tunable framework breathing and thermal expansion. Chem. Sci. 2013, 4, 1539–1546. 166. Murdock, C.R.; McNutt, N.W.; Keffer, D.J.; Jenkins, D.M. Rotating phenyl rings as a guest-dependent switch in two-dimensional metal-organic frameworks. J. Am. Chem. Soc. 2014, 136, 671–678. 167. Zhang, X.-M.; Hao, Z.-M.; Zhang, W.-X.; Chen, X.-M. Dehydration-induced conversion from a single-chain magnet into a metamagnet in a homometallic nanoporous metal-organic framework. Angew. Chem. Int. Ed. 2007, 46, 3456–3459. © 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). Materials 2014, 7 © 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). © 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/).
https://openalex.org/W2952216057	https://academic.oup.com/database/article-pdf/doi/10.1093/database/baw015/8222933/baw015.pdf	English	null	Wikidata as a semantic framework for the Gene Wiki initiative	Database	2,016	cc-by	6,411	Original article Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/2630183 by guest on 24 October 2024 Wikidata as a semantic framework for the Gene Wiki initiative Sebastian Burgstaller-Muehlbacher1, Andra Waagmeester2, Elvira Mitraka3, Julia Turner1, Tim Putman1, Justin Leong4, Chinmay Naik5, Paul Pavlidis4, Lynn Schriml3, Benjamin M Good1,* and Andrew I Su1,* 1The Scripps Research Institute, La Jolla, CA, USA, 2micelio.be, Antwerp, Belgium, 3University of Maryland Baltimore, Baltimore, MD, USA, 4The University of British Columbia, Vancouver, British Columbia, Canada and, 5Bangalore Inst. Of Technology, Visvesvaraya Technological University, Bangalore, Karnataka Corresponding author: Email: bgood@scripps.edu, asu@scripps.edu Corresponding author: Email: bgood@scripps.edu, asu@scripps.edu Citation details: Burgstaller-Muehlbacher,S., Waagmeester,A., Mitraka,E. et al. Wikidata as a semantic framework for the gene wiki initiative. Database (2016) Vol. 2016: article ID baw015; doi:10.1093/database/baw015 tion details: Burgstaller-Muehlbacher,S., Waagmeester,A., Mitraka,E. et al. Wikidata as a semantic framework for th e wiki initiative. Database (2016) Vol. 2016: article ID baw015; doi:10.1093/database/baw015 Received 7 November 2015; Revised 30 December 2015; Accepted 1 February 2016 V C The Author(s) 2016. Published by Oxford University Press. Page 1 of 10 This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Database, 2016, 1–10 doi: 10.1093/database/baw015 Original article V C The Author(s) 2016. Published by Oxford University Press. e Author(s) 2016. Published by Oxford University Press. is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licen stricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Database URL: https://www.wikidata.org/ Introduction triples. The subject of the triple is the Wikidata item on which the claim appears, the predicate is a Wikidata prop- erty, and the object is a date, a string, a quantity, a URL or another Wikidata item. For example, the claim that Rosalind Franklin received her Ph.D. at the University of Cambridge is represented as Q7474 (Rosalind Franklin) - P69 (educated at) - Q35794 (University of Cambridge). Claims can be further amended with qualifiers (to indicate the context in which the triple is valid), and references can be added to indicate the provenance of the claim. The over- all combination of a claim and references is referred to as a ‘statement’. A full description of the Wikidata data model can be found at https://www.mediawiki.org/wiki/ Wikibase/DataModel/Primer. Wikipedia (www.wikipedia.org) is a well-established en- cyclopedia and collection of free form text, operated by the Wikimedia Foundation and edited by thousands of volun- teer editors. As the seventh most-visited site on the Internet (http://www.alexa.com/topsites), Wikipedia has articles on a broad range of topics. With respect to molecular biology articles, at least two systematic efforts have been described, both initiated in 2007. The RNA Wikiproject created 600 new Wikipedia articles on non-coding RNA families (1). In parallel, our Gene Wiki team created a collection of 8000 Wikipedia articles on human genes (2). Since its in- ception, the Gene Wiki has grown into an integral and strongly interlinked part of the English Wikipedia, now counting >11 000 articles (3, 4). The Gene Wiki articles have been expanded by the Wikipedia community and are highly accessed by users of Wikipedia, collectively viewed >4 million times per month (according to http://stats.grok. se/en/). Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/2630183 by guest on 24 October 2024 Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/b A primary motivation for creating Wikidata was ena- bling easy accessibility by all language-specific Wikipedias. Now, statements about any Wikidata item can be dis- played in the context of any Wikipedia item. This process is facilitated by ‘interwiki links’ that establish connections between structured Wikidata items and the Wikipedia art- icles they are most closely related to. All major projects from the Wikimedia Foundation, including the language- specific Wikipedias, are linked to Wikidata using interwiki links. These links can be established between any existing Wikidata item and other MediaWiki content pages, e.g. Wikipedia articles. Introduction Currently, a single Wikidata item can have an interwiki link to many different language-specific Wikipedia articles on the same topic. In return, this setup allows for all linked language-specific Wikipedia articles to access data from the central Wikidata item. Importantly, a Wikipedia article in a certain language can only have one interwiki link to a Wikidata item and a Wikidata item can only be linked to one single Wikipedia article in a certain language. Of note, the interwiki links are also used to or- ganize the links to other language versions of a Wikipedia article, visible in the sidebar of many Wikipedia articles. This system has the advantage that data are stored once in Wikidata, and then made available for reuse across the en- tire Wikimedia universe. Gene Wiki articles consist of two central parts, the free text and the Gene Wiki infobox. The free text represents a review of a gene’s function, biological role and impact on human health and disease. The Gene Wiki infobox pro- vides structured data on the human gene and protein, and the orthologous mouse gene and protein. The data in the infobox comprises standardized identifiers and chromo- some coordinates as well as functional annotation with Gene Ontology terms (5), structural information from the Protein Data Bank (PDB) (6) and tissue-specific gene ex- pression (7) (e.g. https://en.wikipedia.org/wiki/Reelin). Wikipedia has proven a highly effective medium for col- laboratively capturing unstructured text, but is technically lacking facilities for authoring structured data. Several at- tempts have been made to better represent structured data within Wikipedia (8, 9). In late 2012, the Wikidata project (www.wikidata.org) was launched with the goal of creat- ing an open, structured knowledge repository to comple- ment and facilitate the unstructured content in Wikipedia (10). Like all other Wikimedia projects, Wikidata can be edited by anyone, and maintains an extensive version his- tory for every item to allow for easy comparisons or rever- sions to past states. All content in Wikidata is licensed under CC0 (https://creativecommons.org/about/cc0) and therefore can be used by anyone without restrictions. Wikidata enables programmatic querying and access outside of the Wikipedia context. Specifically, Wikidata offers a Representational State Transfer (REST) API to eas- ily perform structured data queries and retrieve Wikidata statements in structured formats. Abstract Open biological data are distributed over many resources making them challenging to in- tegrate, to update and to disseminate quickly. Wikidata is a growing, open community database which can serve this purpose and also provides tight integration with Wikipedia. In order to improve the state of biological data, facilitate data management and dissemin- ation, we imported all human and mouse genes, and all human and mouse proteins into Wikidata. In total, 59 721 human genes and 73 355 mouse genes have been imported from NCBI and 27 306 human proteins and 16 728 mouse proteins have been imported from the Swissprot subset of UniProt. As Wikidata is open and can be edited by anybody, our cor- pus of imported data serves as the starting point for integration of further data by scien- tists, the Wikidata community and citizen scientists alike. The ﬁrst use case for these data is to populate Wikipedia Gene Wiki infoboxes directly from Wikidata with the data inte- grated above. This enables immediate updates of the Gene Wiki infoboxes as soon as the data in Wikidata are modiﬁed. Although Gene Wiki pages are currently only on the English language version of Wikipedia, the multilingual nature of Wikidata allows for usage of the data we imported in all 280 different language Wikipedias. Apart from the Gene Wiki info- box use case, a SPARQL endpoint and exporting functionality to several standard formats (e.g. JSON, XML) enable use of the data by scientists. In summary, we created a fully open and extensible data resource for human and mouse molecular biology and biochemistry data. This resource enriches all the Wikipedias with structured information and serves as a new linking hub for the biological semantic web. Database URL: https://www.wikidata.org/ Database URL: https://www.wikidata.org/ V C The Author(s) 2016. Published by Oxford University Press. Page 1 of 10 This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. (page number not for citation purposes) (page number not for citation purposes) Database, Vol. 2016, Article ID baw015 Page 2 of 10 Data import and data maintenance In order to implement a data model in Wikidata and to populate this data model with data, two types of commu- nity processes need to be passed. The properties required for the data model need to be proposed to and discussed by the community and if consensus can be reached, the prop- erty will be created in Wikidata. Furthermore, large scale data import to Wikidata can only be done with a bot user account, which has enhanced write permissions compared to a standard user account. Bot user accounts also go through a community approval process, in order to deter- mine if they are performing beneficial tasks and if they are operating reliably. For the present project, we successfully proposed a series of properties required for our data model and also reached community consensus on the usefulness of our bots task. Introduction In addition, more com- plex queries are possible via a SPARQL Protocol and RDF Query Language (SPARQL) endpoint (https://query.wiki data.org) and a custom-built WikiData Query (WDQ) tool (http://wdq.wmflabs.org/wdq/). Wikidata consists of two entity types – items (e.g. https://www.wikidata.org/wiki/Q7474 for Rosalind Frank- lin) and properties (e.g. https://www.wikidata.org/wiki/ Property:P351 for NCBI Entrez gene ID)—and every entity is assigned a unique identifier. Wikidata items and proper- ties all have a label, a description and aliases (Figure 1). Every item record contains a list of claims in the form of In this work, we describe our efforts to migrate our Gene Wiki bot from English Wikipedia to Wikidata. This system offers significant advantages with respect to main- tainability of the data, accessibility within the Wikipedia Database, Vol. 2016, Article ID baw015 Page 3 of 10 and HGNC IDs were also added to each gene item (24). For mouse gene nomenclature, the Jackson Laboratory Mouse Genome Informatics (MGI) data were used (13). For protein items, we used UniProt as the primary data source. All protein items received the ‘subclass of’ (P279) property value ‘protein’ (Q8054). A range of protein anno- tations were also added, including NCBI RefSeq Protein IDs (P637), Ensembl Protein IDs (P705), and PDB IDs (P638). Gene Ontology terms were introduced as separate Wikidata items, and annotations were added to protein items using three separate properties for Molecular Function (P680), Cell Component (P681) and Biological Process (P682). ecosystem, and programmatic integration with other resources. and HGNC IDs were also added to each gene item (24). For mouse gene nomenclature, the Jackson Laboratory Mouse Genome Informatics (MGI) data were used (13). For protein items, we used UniProt as the primary data source. All protein items received the ‘subclass of’ (P279) property value ‘protein’ (Q8054). A range of protein anno- tations were also added, including NCBI RefSeq Protein IDs (P637), Ensembl Protein IDs (P705), and PDB IDs (P638). Gene Ontology terms were introduced as separate Wikidata items, and annotations were added to protein items using three separate properties for Molecular Function (P680), Cell Component (P681) and Biological Process (P682). Database construction and usage In collaboration with the Wikidata community, we decided to implement the representation of genes and proteins in Wikidata as separate Wikidata items. These gene and pro- tein items are linked by the reciprocal properties ‘encodes’ (P688) and ‘encoded by’ (P702) carried by genes and pro- teins, respectively (Figure 2). Furthermore, orthologous genes between species are reciprocally linked by the prop- erty ortholog (P684) and also link out to NCBI HomoloGene (11) with the HomoloGene ID (P593). Homologous genes in the HomoloGene database and therefore also on Wikidata gene items share the same ID and can also be associated this way. The community dis- cussion and decision process to establish this model, a very crucial mechanism in Wikidata, as well as Wikipedia, can be viewed here: https://www.wikidata.org/wiki/Wikidata_ talk:WikiProject_Molecular_biology. Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/2630183 by guest on 24 October 2024 Data model and data sources for Wikidata 2016, Article ID baw015 Page 4 of 10 Figure 1 Wikidata item and data organization. Wikidata items can be added or edited by anyone manually. A Wikidata item consists of: (1) a language- speciﬁc label, (2) its unique identiﬁer, (3) language speciﬁc aliases, (4) interwiki links to the different language Wikipedia articles or other Wikimedia pro- jects and (5) a list of statements. For this speciﬁc example, the human protein Reelin was used (https://www.wikidata.org/wiki/Q13569356) Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/26301 Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/2630183 by guest on 24 October 2024 Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/b m/database/article/doi/10.1093/database/baw015/2630183 by guest on 24 October 2024 Figure 1 Wikidata item and data organization. Wikidata items can be added or edited by anyone manually. A Wikidata item consists of: (1) a language- speciﬁc label, (2) its unique identiﬁer, (3) language speciﬁc aliases, (4) interwiki links to the different language Wikipedia articles or other Wikimedia pro- jects and (5) a list of statements. For this speciﬁc example, the human protein Reelin was used (https://www.wikidata.org/wiki/Q13569356) challenge to our data model so far. To prevent and resolve it, we implemented a bot which can detect and undo these merges. Since we also unified the interwiki links on the human genes, these merges have now decreased to zero, showing the importance of consistent interwiki links. SPARQL queries (https://bitbucket.org/sulab/wikidatas parqlexamples/src). These queries detect inconsistencies in our data model and, as an ongoing effort, we implemented bots which execute these SPARQL queries and report and resolve the inconsistencies detected. In view of Wikidata being completely open, the number of inconsistencies we experienced so far is low. Data model and data sources for Wikidata We populated Wikidata with items for all Homo sapiens (human) genes, Homo sapiens proteins, Mus musculus (mouse) genes and Mus musculus proteins (Table 1). As described above, each gene and each protein was repre- sented as a single Wikidata item. A full list of Wikidata properties used on gene and protein items is provided in Table 2. Briefly summarized, for gene items, we imported data from the latest annotation releases from NCBI (Homo sa- piens release 107, Mus musculus release 105) and created statements using many properties, including Entrez Gene IDs, RefSeq RNA IDs and chromosomal positions (11). Ensembl Gene IDs and Ensembl Transcript IDs were also imported and added to each gene item in Wikidata (12). Genes were categorized according to eight subclasses. A generic subclass was used to identify a Wikidata item as a gene (Q7187), for increased granularity, the subclasses protein coding gene (Q20747295), ncRNA gene (Q27087), snRNA gene (Q284578), snoRNA gene (Q284416), rRNA gene (Q215980), tRNA gene (Q201448) and pseudo gene (Q277338) were added (Table 2). Genomic coordinates were encoded using the properties chromosome (P1057), genomic start (P644), and genomic end (P645), and the qualifier property ‘GenLoc assembly’ (P659) was used to indicate the corres- ponding assembly version—GRCh37 (Q21067546) or GRCh38 (Q20966585). Gene symbols were added based on the HUGO Gene Nomenclature Committee (HGNC) We implemented the data importing process using Python (www.python.org) scripts, colloquially termed as bots by the Wikidata community. We run these bots with the Wikidata user account ProteinBoxBot (https://www. wikidata.org/wiki/User:ProteinBoxBot). The source code for the bots is available under GNU AGPLv3 (http://www. gnu.org/licenses/agpl.html) on our Bitbucket repository (https://bitbucket.org/sulab/wikidatabots/). As our aim was to establish and also to maintain a high quality data source, the bots are designed to update the data from their original sources on a regular basis. Bots are run every few days to weeks, depending on the data source, to ensure data being up-to-date. Before our bots write data to a Wikidata item or create a new item, thorough checks are being performed to ensure that the correct item has been selected for writing or an item does not yet exist and so a new one should be created. Data on Wikidata can be edited by anybody. In order to keep our data model consistent and to detect accidental er- roneous modification or potential vandalism of data, we implemented a set of consistency tests based on Database, Vol. Populating gene wiki infoboxes with data from wikidata As mentioned in the introduction, interwiki links are organized in Wikidata. Thus, users from the different lan- guage Wikipedias aim to link their articles to the same Wikidata item. The data model in Wikidata therefore needs to be designed in a way that all interwiki links unam- biguously are present on the correct Wikidata item. Otherwise, Wikipedia users might be encouraged to merge items which are separate things in one’s data model. The several dozens of merges we experienced where the biggest As a first use case of the data, we focused on using the gene and protein data imported into Wikidata to populate Wikipedia Gene Wiki infoboxes. In our data model, we connected Wikipedia Gene Wiki pages to Wikidata human gene items with interwiki links (Figure 2). Four Wikidata items are required to fully represent one Gene Wiki info- box on a Wikipedia article. For this paper, we chose the Page 5 of 10 Database, Vol. 2016, Article ID baw015 Figure 2 Gene Wiki data model in Wikidata. Each entity (human gene, human protein, mouse gene, mouse protein) is represented as a separate Wikidata item. Arrows represent direct links between Wikidata statements. The English language interwiki link on the human gene item points to the corresponding Gene Wiki article on the English Wikipedia. Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/2630183 by guest on 24 October 2024 Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/b Figure 2 Gene Wiki data model in Wikidata. Each entity (human gene, human protein, mouse gene, mouse protein) is represented as a separate Wikidata item. Arrows represent direct links between Wikidata statements. The English language interwiki link on the human gene item points to the corresponding Gene Wiki article on the English Wikipedia. gene RELN (protein Reelin) as an example. Specifically, the English language Wikipedia page (https://en.wikipedia. org/wiki/Reelin) of Reelin is directly linked to the human gene Wikidata item (Q414043) with an interwiki link. The Wikidata human gene item in turn links to the human pro- tein (Q13561329), mouse gene (Q14331135), and mouse protein (Q14331165) using the data model described in Figure 2. at largely replicating the content of the previous, template based infoboxes. By clicking the Wikidata link in the side- bar or the infobox of a Gene Wiki page, a user can directly access the Wikidata item and view or modify the entire data available on the item. Populating gene wiki infoboxes with data from wikidata In order to display additional data in the infoboxes, slight adaptations of the Lua code are required. Of note, the Lua code can be reused to dis- play the same infobox on any language version of Wikipedia. In order for Wikipedia pages to retrieve data from Wikidata, we used the MediaWiki extension module Scribunto (https://www.mediawiki.org/wiki/Extension:S cribunto), which integrates scripting capabilities based on the programming language Lua (http://www.lua.org). We created a new module in Lua code which generates the en- tire Wikipedia Gene Wiki infobox based on Wikidata data (https://en.wikipedia.org/wiki/Module:Infobox_gene). The new system has now been deployed on several Gene Wiki pages within Wikipedia, and we will complete the migration after extensive testing and community con- sensus on all implementational details has been reached. Data usage beyond Wikipedia Gene Wiki infoboxes Using this new Wikidata-based infrastructure, a Gene Wiki infobox can be added to any Wikipedia page for a human gene by simply adding the markup code ‘{{infobox gene}}’, provided the Wikipedia page has an interwiki link to a Wikidata human gene item. Not all data present on the Wikidata gene and protein items are currently also repre- sented in the Gene Wiki info boxes, at this stage, we aimed Data from Wikidata can be widely used in any application of interest. For example, the Gene Wiki infobox could now be rendered on any website on the Internet. As described in the introduction, a SPARQL endpoint, WDQ, the sophisti- cated Mediawiki API and a free text search engine consti- tute the main ways of querying data. These facilities can easily be integrated in downstream analysis using Python, Page 6 of 10 Database, Vol. 2016, Article ID baw015 R or any other data analysis language which supports eamless integration of data from the web. The SPARQL endpoint should be particularly useful for dynamic data in- egration. For example, for annotation of deletions in a cancer genome, it would be useful to know: ‘Which genes, encoding for proteins, are located on chromosome 9 be- through the Wikidata endpoint facilitates a single query that can answer this question immediately (Figure 4). As the data are richly referenced, data origin and valid- ity of statements can be reviewed instantly. Furthermore, the unique multilinguality of Wikidata, enabled by multi- lingual item labels, descriptions, and interwiki links to all Figure 3 GeneWiki infobox populated with data from Wikidata, using data from Wikidata items Q414043 for the human gene, Q13561329 for human protein, Q14331135 for the mouse gene and Q14331165 for the mouse protein. Three dots indicate that there is more information in the real Gene Wiki infobox for Reelin (https://en.wikipedia.org/wiki/Reelin). Page 6 of 10 Database, Vol. 2016, Article ID baw015 Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/2630183 by guest on 24 October 2024 Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/b Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/2630183 by guest on 24 October 2024 Figure 3 GeneWiki infobox populated with data from Wikidata, using data from Wikidata items Q414043 for the human gene, Q13561329 for human protein, Q14331135 for the mouse gene and Q14331165 for the mouse protein. Three dots indicate that there is more information in the real Gene Wiki infobox for Reelin (https://en.wikipedia.org/wiki/Reelin). Discussion We created an open, community editable structured re- source for all human and mouse genes and proteins using Wikidata as the technical platform. As a first use case of this data corpus, we demonstrated a remodelled Wikipedia Gene Wiki infobox which retrieves its data entirely from Wikidata, greatly simplifying the maintenance of these infoboxes. Until now, each of the 280 language-specific Wikipedias had to independently manage their infobox data in the context of MediaWiki templates meant for managing information display, not for storage and re- trieval. This created significant redundancy, a great risk for errors and inconsistencies, and out of date data, as the same data needed to exist in every language-specific Wikipedia. Now, not only do the Wikipedias benefit from higher data quality when based on a centralized data re- pository, Wikidata also benefits from the focused human effort in the global Wikipedia community. An important aspect of the broad applicability and reus- ability of Wikidata is its connection to the Semantic Web and Linked Open Data (Figure 4). Wikidata IDs give genes and proteins stable Uniform Resource Identifiers (URIs) in the Semantic Web, which in turn link to other common identifiers used in the biomedical research community. Moreover, Wikidata provides perhaps the simplest inter- face for anyone to edit the Semantic Web, which is other- wise limited by high technical barriers to contribute. Similar or complementary efforts for integration of bio- medical data are Bio2RDF and SNPedia. Bio2RDF inte- grates 37 different, well-established biomedical resources by defining an ontology for each data source and so con- verting them into normalized RDF documents (18, 19). The Bio2RDF data integration effort currently is consider- ably larger than our project. Apart from that, the data models in Wikidata are not based on ontologies, but they are a product of community discussion. Only the basic Wikidata data model exists as an ontol- ogy (https://www.mediawiki.org/wiki/Wikibase/Indexing/ RDF_Dump_Format), which allows all data and data models in Wikidata to exist within this one ontology, which simplifies creation of new data models but might complicate semantic reasoning. Another major difference is that Wikidata can be edited by any user whereas for Bio2RDF, additions or corrections need to be performed In addition to these benefits to the Wikipedia commu- nity, the Gene Wiki effort in Wikidata also offers many data integration advantages to the biomedical research community. Data usage beyond Wikipedia Gene Wiki infoboxes Figure 3 GeneWiki infobox populated with data from Wikidata, using data from Wikidata items Q414043 for the human gene, Q13561329 for human protein, Q14331135 for the mouse gene and Q14331165 for the mouse protein. Three dots indicate that there is more information in the real Gene Wiki infobox for Reelin (https://en.wikipedia.org/wiki/Reelin). through the Wikidata endpoint facilitates a single query that can answer this question immediately (Figure 4). R or any other data analysis language which supports seamless integration of data from the web. The SPARQL endpoint should be particularly useful for dynamic data in- tegration. For example, for annotation of deletions in a cancer genome, it would be useful to know: ‘Which genes, encoding for proteins, are located on chromosome 9 be- tween 21 and 30 megabase pairs?’. Content accessible R or any other data analysis language which supports seamless integration of data from the web. The SPARQL endpoint should be particularly useful for dynamic data in- tegration. For example, for annotation of deletions in a cancer genome, it would be useful to know: ‘Which genes, encoding for proteins, are located on chromosome 9 be- tween 21 and 30 megabase pairs?’. Content accessible As the data are richly referenced, data origin and valid- ity of statements can be reviewed instantly. Furthermore, the unique multilinguality of Wikidata, enabled by multi- lingual item labels, descriptions, and interwiki links to all of the Wikipedias allows the data to be used globally. Database, Vol. 2016, Article ID baw015 Page 7 of 10 to the output of medium-sized biocuration efforts, to indi- vidual statements added by individual users. We consider integration of data on human and mouse genes and pro- teins as the basis also for our own efforts with Wikidata. Figure 4 illustrates how the data could be used by the sci- entific community. The query returns all protein coding genes on a certain region of human chromosome 9, the chromosomal coordinates being based on the human gen- ome build GRCh38. As NCBI gene nor a different resource provide a SPARQL endpoint with access to genomic co- ordinates, to the best of our knowledge, a similar query cannot be achieved with any other semantic web resource currently available. As the Gene Ontology terms are at- tached to human proteins, a slightly modified query would also allow functional annotation of the gene products, e.g. Data usage beyond Wikipedia Gene Wiki infoboxes for copy number aberrations of cancer genomes, through executing a SPARQL query. Currently, tools like ANNOVAR (17) provide richer annotation for variants and mutations, but Wikidata could e.g. prove useful for fast functional annotation of copy number aberrations in hereditary disease and cancer sequencing. With Wikidata growing, also the number of (annotation) use cases will in- crease. In addition to extending data in the biology and molecular biology space (e.g. cell lines, non-coding RNAs), we are currently importing data on human diseases and drugs. Di i Table 1 Overview on Homo sapiens and Mus musculus data imported to Wikidata Data source Item count Homo sapiens genes (NCBI release 107) 59 721 Homo sapiens proteins (Uniprot) 27 306 Mus musculus genes (NCBI release 105) 73 355 Mus musculus proteins (Uniprot) 16 728 Gene Ontology terms 17 098 Table 1 Overview on Homo sapiens and Mus musculus data imported to Wikidata Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/b Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/2630183 by guest on 24 October 2024 Discussion Biomedical knowledge is fragmented across many resources and databases, and small-scale data inte- gration is an often-repeated exercise in almost every data analysis project. As for the language-specific Wikipedias, these small-scale integration efforts are incomplete, ineffi- cient and error-prone. Instead, users now have the option of accessing and querying a central biomedical resource within Wikidata that is already pre-populated with many key resources and identifiers. While we certainly recognize that our effort does not yet include every resource in the biomedical space, Wikidata does empower any user to con- tribute data from their resource of interest. For example, any user could easily contribute data from other third party resources [e.g. International Union of Basic and Clinical Pharmacology (IUPHAR) (14), DECIPHER (15), COSMIC (16)] with minimal effort. These contributions can range from programmatic addition of large databases, Database, Vol. 2016, Article ID baw015 Page 8 of 10 Table 2 Wikidata properties used in this study Label Target data type Property ID Description Wikidata gene items Subclass of Wikidata item P279 Deﬁnes to what category this item belongs to. Every gene item carries the value ‘gene’ (Q7187). Further subcategories are protein coding gene (Q20747295), ncRNA gene (Q27087), snRNA gene (Q284578), snoRNA gene (Q284416), rRNA gene (Q215980), tRNA gene (Q201448) and pseudogene (Q277338). Discussion Entrez Gene ID String P351 The NCBI gene ID as in annotation release 107 Found in taxon Wikidata item P703 The taxon, either Homo sapiens (Q5) or Mus musculus (Q83310) Ensembl Gene ID String P594 Gene ID from the Ensembl database Ensembl Transcript ID String P704 Transcript IDs from the Ensembl database Gene symbol String P353 Human gene symbol according to HUGO Gene Nomenclature Committee HGNC ID String P354 HUGO Gene Nomenclature Committee ID HomoloGene ID String P593 Identiﬁer for the HomoloGene database NCBI RefSeq RNA ID String P639 Chromosome Wikidata item P1057 Chromosome a gene is residing on Ortholog Wikidata item P684 Ortholog based on the Homologene database Genomic start String P644 Genomic start according to GRCh37 and GRCh38, sourced from NCBI Genomic stop String P645 Genomic stop according to GRCh37 and GRCh38, sourced from NCBI Mouse Genome Informatics ID String P671 Jackson lab mouse genome informatics database encodes Wikidata item P688 Protein item a gene encodes Wikidata protein items Subclass of Wikidata item P279 protein (Q8054) UniProt ID String P352 PDB ID String P638 Protein structure IDs from PDB.org RefSeq protein ID String P637 NCBI RefSeq Protein ID encoded by Wikidata item P702 Gene item a protein is encoded by Ensembl protein ID String P705 EC number String P591 Enzyme Category number Protein structure image Wiki Commons Media File P18 Prefered protein structure image retrieved from PDB.org Cell component Wikidata item P681 Gene ontology term items for cell components Biological process Wikidata item P682 Gene ontology term items for biological processes Molecular function Wikidata item P680 Gene ontology term items for molecular function Column one contains the description as in Wikidata, column two the data type, column three the property number and column four a short description of the nature of the content. Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/2630183 b Table 2 Wikidata properties used in this study Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/2630183 by guest on 24 October 2024 genes (21). Compared to Wikidata, SNPedia is restricted to SNPs and lacks a SPARQL endpoint or similar query capabil- ities. We do not have immediate plans to integrate SNP data, but we consider SNPedia a valuable resource which comple- ments our efforts (22). Should SNP data be integrated into Wikidata, SNPedia would be either a data source or a re- source Wikidata SNP items should be linking to. by the maintainers. Discussion Furthermore, integration of smaller biomedical databases, which exist in large numbers (20), seems to be easier with Wikidata, whereas Bio2RDF cur- rently might be better suited for dealing with very large numbers of triples. Although Bio2RDF users who would like to integrate their data sources of interest in their ana- lysis still have the option to create their own Bio2RDF compatible RDF conversions, we think that having the data ready to go in a triple store should considerably sim- plify data analysis. Nevertheless, as both projects attempt to open biology to the semantic web, exploration of syner- gies between them seems worthwhile. by the maintainers. Furthermore, integration of smaller biomedical databases, which exist in large numbers (20), seems to be easier with Wikidata, whereas Bio2RDF cur- rently might be better suited for dealing with very large numbers of triples. Although Bio2RDF users who would like to integrate their data sources of interest in their ana- lysis still have the option to create their own Bio2RDF compatible RDF conversions, we think that having the data ready to go in a triple store should considerably sim- plify data analysis. Nevertheless, as both projects attempt to open biology to the semantic web, exploration of syner- gies between them seems worthwhile. Other wiki-based initiatives resembling our project, un- fortunately seem to have ceased to exist (23), which might be due to difficulties establishing an active community. One major advantage of Wikidata is a large community with many contributors being experts in the field. Currently, Wikidata has >16 000 active users (https:// www.wikidata.org/wiki/Special:Statistics). SNPedia is a semantic web enabled wiki for single nucleo- tide polymorphisms (SNPs) and also provides information on Page 9 of 10 Database, Vol. 2016, Article ID baw015 Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/2630183 by guest on 24 October 2024 Figure 4 An example SPARQL query, using the Wikidata SPARQL endpoint (query.wikidata.org). It retrieves all Wikidata (WD) items which are of sub- class protein-coding gene (Q840604), which have a chromosomal start position (P644) according to human genome build GRCh38 and reside on human chromosome (P659) 9 (Q20966585) and a chromosomal end position (P645) also on chromosome 9. Furthermore, the region of interest is re- stricted to a chromosomal start position between 21 and 30 megabase pairs. Colors: Red indicates SPARQL commands, blue represents variable names, green represents URIs and brown are strings. Arrows point to the source code the description applies to. References 1. Daub,J., Gardner,P.P., Tate, J. et al. (2008) The RNA WikiProject: community annotation of RNA families. RNA, 14, 2462–2464. 1. Daub,J., Gardner,P.P., Tate, J. et al. (2008) The RNA WikiProject: community annotation of RNA families. RNA, 14, 2462–2464. 2. Huss,J.W. 3rd,Orozco, C., Goodale, J. et al. (2008) A gene wiki for community annotation of gene function. PLoS Biol., 6, e175. 2. Huss,J.W. 3rd,Orozco, C., Goodale, J. et al. (2008) A gene wiki for community annotation of gene function. PLoS Biol., 6, e175. 3. Huss,J.W. 3rd, Lindenbaum, P., Martone, M. et al. (2010) The Gene Wiki: community intelligence applied to human gene anno- tation. Nucleic Acids Res., 38, D633–D639. 4. Good,B.M., Clarke,E.L., de Alfaro, L. et al. (2012) The Gene Wiki in 2011: community intelligence applied to human gene an- notation. Nucleic Acids Res., 40, D1255–D1261. 5. Ashburner,M., Ball,C.A., Blake,J.A. et al. (2000) Gene ontology: tool for the uniﬁcation of biology. The Gene Ontology Consortium. Nat. Genet, 25, 25–29. 6. Berman,H.M., Westbrook, J., Feng, Z. et al. (2000) The Protein Data Bank. Nucleic Acids Res., 28, 235–242. This work describes our initial effort to seed Wikidata with data from several key genomics resources. While this action has direct value to our Gene Wiki project, we hope and expect this first step to nucleate further growth of sci- entific data in Wikidata. With sufficient contribution and participation by the community, Wikidata can evolve into the most comprehensive, current and collaborative know- ledge base for biomedical research. 7. Su,A.I., Wiltshire, T., Batalov, S. et al. (2004) A gene atlas of the mouse and human protein-encoding transcriptomes. Proc. Natl. Acad. Sci. USA, 101, 6062–6067. 8. Good,B.M., Clarke,E.L., Loguercio, S. et al. (2012) Building a biomedical semantic network in Wikipedia with Semantic Wiki Links. Database, 2012, bar060. 9. Kr€otzsch,M., Vrandecic, D., V€olkel, M. et al. (2007) Semantic Wikipedia. Web Semantics: Science, Services and Agents on the World Wide Web, 5, 251–261. 10. Vrandecic,D. and Kr€otzsch, M. (2014) Wikidata. A free collab- orative knowledgebase. Commun. ACM, 57, 78–85. Discussion Finally, Wikidata comes with a number of challenges rele- vant to the scientific community. The data model needs to be checked and protected against erroneous alterations, but true vandalism seems to be very rare. An one time import of data from a source is most likely not sustainable but the import of a certain data source should be a long term commitment to syn- chronization with the original source using bots. The commu- nity process might appear intimidating at first, but critical questions by the community help to improve one’s approach. Moreover, who contributes, automatically becomes part of the community with the right to vote. Licensing of the original data source could also pose challenging for an import to Wikidata, because Wikidata content is licensed as public do- main (CC0). Acknowledgements 11. NCBI Resource Coordinators (2015) Database resources of the National Center for Biotechnology Information. Nucleic Acids Res., 43, D6–17. We would like to thank the Wikipedia user RexxS for providing sub- stantial help with the Gene Wiki infobox Lua code. Furthermore, we would like to thank the Wikidata community for valuable discus- sions and suggestions. 12. Cunningham,F., Amode,M.R., Barrell, D. et al. (2015) Ensembl 2015. Nucleic Acids Res., 43, D662–D669. 13. Eppig,J.T., Blake,J.A., Bult,C.J. et al. Mouse Genome Database Group (2015) The Mouse Genome Database (MGD): facilitating mouse as a model for human biology and disease. Nucleic Acids Res., 43, D726–D736. Funding This work is supported by the National Institutes of Health under grant GM089820, GM083924, GM114833 and DA036134. 14. Southan, C., Sharman,J.L., Benson,H.E. et al. NC-IUPHAR (2015) The IUPHAR/BPS Guide to PHARMACOLOGY in Conﬂict of interest. None declared. Database, Vol. 2016, Article ID baw015 Page 10 of 10 2016: towards curated quantitative interactions between 1300 protein targets and 6000 ligands. Nucleic Acids Res. research and knowledge discovery. J. Biomed. Semantics, 5, 14 20. Galperin,M.Y., Rigden,D.J. and Fernandez-Suarez,X.M. (2015) The 2015 Nucleic Acids Research Database Issue and molecular biology database collection. Nucleic Acids Res., 43, D1–D5. 15. Bragin,E., Chatzimichali,E.A., Wright,C.F. et al. (2014) DECIPHER: database for the interpretation of phenotype-linked plausibly pathogenic sequence and copy-number variation. Nucleic Acids Res., 42, D993–D1000. 21. Cariaso,M. and Lennon, G. (2012) SNPedia: a wiki supporting personal genome annotation, interpretation and analysis. Nucleic Acids Res., 40, D1308–D1312. 16. Forbes,S.A., Beare, D., Gunasekaran, P. et al. (2015) COSMIC: exploring the world’s knowledge of somatic mutations in human cancer. Nucleic Acids Res., 43, D805–D811. 22. Good,B.M., Clarke,E.L., Loguercio, S. et al. (2012) Linking genes to diseases with a SNPedia-Gene Wiki mashup. J. Biomed. Semantics, 3, S6. Downloaded from https://academic.oup.com/database/article/doi/10.1093/database/baw015/2630183 by guest on 24 October 2024 17. Yang,H. and Wang, K. (2015) Genomic variant annotation and prioritization with ANNOVAR and wANNOVAR. Nat. Protoc., 10, 1556–1566. 23. Mons,B., Ashburner, M., Chichester, C. et al. (2008) Calling on a million minds for community annotation in WikiProteins. Genome Biol., 9, R89. 18. Belleau,F., Nolin,M.A., Tourigny,N. et al. (2008) Bio2RDF: to- wards a mashup to build bioinformatics knowledge systems. J. Biomed. Inform., 41, 706–716. 24. Gray,K.A., Yates, B., Seal,R.L. et al. (2015) Genenames.org: the HGNC resources in 2015. Nucleic Acids Res., 43, D1079–D1085. 19. Dumontier,M., Baker,C.J., Baran, J. et al. (2014) The Semanticscience Integrated Ontology (SIO) for biomedical
https://openalex.org/W2062016229	https://escholarship.org/content/qt2hj9t73j/qt2hj9t73j.pdf?t=nvwr42	English	null	Variations in the Distributions of Marine Birds with Water Mass in the Northern Bering Sea	The condor	1,993	cc-by	7,921	UC Irvine UC Irvine Previously Published Works Title Variations in the Distributions of Marine Birds with Water Mass in the Northern Bering Sea Permalink https://escholarship.org/uc/item/2hj9t73j Journal The Condor, 95(1) ISSN 0010-5422 Authors Elphick, Chris S Hunt,, George L Publication Date 1993-02-01 DOI 10.2307/1369384 Copyright Information This work is made available under the terms of a Creative Commons Attribution License, availalbe at https://creativecommons.org/licenses/by/4.0/ Peer reviewed UC Irvine UC Irvine Previously Published Works Title Variations in the Distributions of Marine Birds with Water Mass in the Northern Bering Sea Permalink https://escholarship.org/uc/item/2hj9t73j Journal The Condor, 95(1) ISSN 0010-5422 Authors Elphick, Chris S Hunt,, George L Publication Date 1993-02-01 DOI 10.2307/1369384 Copyright Information This work is made available under the terms of a Creative Commons Attribution License, availalbe at https://creativecommons.org/licenses/by/4.0/ Peer reviewed Permalink https://escholarship.org/uc/item/2hj9t73j Copyright Information This work is made available under the terms of a Creative Commons Attribution License, availalbe at https://creativecommons.org/licenses/by/4.0/ Peer reviewed Peer reviewed UC Irvine UC Irvine Previously Published Works Title Title Variations in the Distributions of Marine Birds with Water Mass in the Northern Bering Sea INTRODUCTION lower end of the mesoscale range (50-150 km). Second, in the Chirikov Basin of the northern Bering Sea, where three separate water masses occur in close proximity, investigations of the distribution of nesting colonies of marine birds have shown that colonies located in Alaska Coastal Water support primarily piscivorous species, whereas colonies near Anadyr Water are dominated by planktivorous auklets (Sowls et aJ. 1978, Springer and Roseneau 1985). These patterns indicate that differences exist in the use of different water masses by different species (Springer ct al. 1987). This supposition was strengthened by limited transect data (Drury et al. 198 1; summarized by Hunt et al. 1981) that showed considerable differences in use of the eastern and western portions of the Chirikov Ba- sin by birds. Patterns in the distribution and abundance of seabirds at sea vary as a function of the spatial and temporal scales at which they occur (Hunt and Schneider 198 7). According to these authors, macroscale ( 1,000-3,000 km; after Haury et al. 1978) patterns of seabird distribution most likely reflect variations in primary and secondary pro- duction, whereas rnesoscale patterns (I 00-1,000 km) involve variations in avian species com- position in response to variations in the com- position of prey communities. At still smaller scales (1-100 km). the abundances of individual species often reflect opportunities to forage at local concentrations of prey (Hunt and Schneider 1987). In this generalized scheme, variations in the avian community are alternately character- ized by changes in either biomass or species com- position, and reflect a similar alternation in the prey community. This paper investigates the foraging distribu- tions of seabirds in the Chirikov Basin with re- spect to water masses and their constituent prey communities. We test whether a variety of spe- cies show habitat specificity when several habi- tats occur over a small area (on the order of the birds' daily flight ranges). The study, however, was conducted at a spatial scale generally asso- ciated with responses to changes in local prey concentrations (Schneider 1982, Schneider and Piatt 1986, Schneider et al. 1987), rather than to different prey communities. In addition to our findings for the Chirikov Basin, we discuss the effect of the boundary gradient between habitat patches on a bird's ability to respond to habitat changes and the need to examine the combined effects of boundary conditions and scale on hab- itat selection by birds. Powered by the California Digital Library University of California eScholarship.org eScholarship.org The Condor 95:33-44 C The Cooper Omitholocjcal Society 1993 CHRIS s. ELPlilCK2 AND GEORGE L. HUNT, JR.3 CHRIS s. ELPlilCK2 AND GEORGE L. HUNT, JR.3 CHRIS s. ELPlilCK2 AND GEORGE L. HUNT, JR.3 Department of Ecology and El•o/utionary Biology, University of California, Irvine, CA 92 Department of Ecology and El•o/utionary Biology, University of California, Irvine, CA 927 17 . Abstr~ct . We examined the pelagic distributions of 12 species of northern Bering Sea birds wtth respect to the water masses in which they were observed during the summers of 1984, 198:5 and .1986. Despite the prediction of earlier work that differences in community structure are unhkely to occur at small spatial scales, we found significant habitat preferences for all but one of the species studied. We suggest that the strength of the gradient between habitat types is the cause of this discrepancy, and that boundary conditions should also be considered when discussing the influence of spatial scales on community processes. Ker .words: Seabird distributions; Bering Sea; habitat preferences; spatial scale; boundary co11d11 ions. ' Received 13 March 1992. Accepted 9 September 1992. dd i l l i 2 Present address: Program in Ecology, Evolution, and .Conservation Biology, 1000 Valley Road, Uni- versity of Nevada, Reno, NV 89512. 1 Corresponding author. ' Received 13 March 1992. Accepted 9 September 1992. 2 Present address: Program in Ecology, Evolution, and .Conservation Biology, 1000 Valley Road, Uni- versity of Nevada, Reno, NV 89512. 1 Corresponding author. STUDY AREA The Chirikov Basin lies between Siberia and Alaska and is bounded by St. Lawrence Island to the south and Bering Strait to the north (Fig. 1). The oceanography of the area is influenced by three water masses: to the east flows Alaska Coastal Water, to the west Anadyr Water, and between lies Bering Shelf Water (Coachman et al. 1975). These three water masses are distinct both in their physical and biological properties and can be readily distinguished by their salinity characteristics (Coachman et al. 1975). In the analysis, we divided transects, which varied in length and ofien traversed water mass boundaries, into 15-min (2.3-5.6 km long, de- pending on ship speed) intervals. For each in- terval, we calculated the position of the central point, the area over which birds were recorded, and the number of birds of each species seen during that period. To reduce data loss, intervals ofl2.5-17.5 min from theendsoftransectswere included as samples. In the analysis, we ignored rarely encountered species and combined counts for all "dark-bellied shearwaters" [primarily Short-tailed (Puffinus tenuirostris), but including some Sooty Shearwaters (P. griseus)] and murres (Uria spp.) because there were high proportions of unidentified individuals for these taxa. For the 1985 data, however, separate tests were Possible for each murre species because a high propQrtion (42.8%) of those seen were identified to species. The shearwater data from 1984 and 1985 were not analyzed because few were seen in either year. For Red Phalaropes (Phalaropus fulicaria) and the alcids we excluded from the analysis indi- viduals that were first observed in flight. Because these species feed only from the water surface, this allowed us to exclude birds that were just flying to or from a feeding site. In all instances, numbers of birds were converted into densities (birds/km2) to allow comparisons between tran- sect intervals covered at different speeds. Alaska Coastal Water originates from the In- ner Domain of the southeastern Bering Sea shelf and moves northward along the coast, where it is diluted by extensive freshwater input in Nor- ton Sound, primarily from the Yukon River (Coachman et al. I 97 5, Kinder and Schumacher 1981 ). This is the warmest of the three water masses. To the west is the colder, more saline Anadyr Water that enters from the Gulf of Ana- dyr, where the endemic cold water mixes with water flowing in from the southern Bering Sea (Coachman et al. STUDY AREA 1975). The third water mass, Bering Shelf Water, is characterized by temper- atures and salinities intermediate between those of Alaska Coastal Water and Anadyr Water, and by extensive stratification in summer (Coach- man et al. 1975, Hunt and Harrison 1990). This water mass enters the Chirikov Basin around both ends of St. Lawrence Island and is formed by the mixing of cold water from the northern Bering Sea shelf and oceanic water from the Be- ring Sea (Coachman et al. 197 5). Distribution maps were created by dividing the region into blocks with dimension of 0.1 de- grees of latitude and 0.2 degrees of longitude. This choice of block size was a compromise be- tween obtaining a fine-grain resolution in the geographical distribution of birds and the sample size available for most blocks; with this block size, 48%, 46%, and 33% of the blocks had at least four counts in 1984, 1985, and 1986, re- spectively. For each block, we calculated the mean density of each species for all 15-min samples whose midpoint fell within that block. CHRIS S. ELPHICK AND GEORGE L. HUNT, JR. 34 METHODS each flock sighting was recorded to the nearest 0.1 min, and details of the ship's Position and speed were noted at regular intervals to facilitate the calculation of bird densities. (For details of the methodology see Hunt and Harrison 1990, Hunt et al. I 990a.) INTRODUCTION Despite the prediction, based on the above scheme, that differences in the species compo- sition of seabird communities are unlikely to ex- ist over relatively small spatial scales, at least two lines of evidence have indicated that such patterns exist. First, Haney ( 1986) has shown significant differences in the densities of several bird species over four water masses associated with Gulf Stream eddies while working at the [33] [33] CHRIS S. ELPHICK AND GEORGE L. HUNT, JR. CHRIS S. ELPHICK AND GEORGE L. HUNT, JR. DATA COLLECTION AND ANALYSIS Distributions of seabird species were determined using data collected over the periods 1-7 July 1984, 26 July- 13 August 1985, and 11-21 Au- gust 1986, when numerous ship transects were made near St. Lawrence and King Islands (Fig. 2). For 1984, our data comes only from transects near King Island; during 1985 and 1986, the sur- veys covered much of the southern Chirikov Ba- sin. Data collection along these transects entailed identifying, counting, and recording the behavior of all birds seen within a 300 m arc from directly in front, to 90" to the side, of the ship. This method is suitable for obtaining estimates of birds on the water, but tends to overestimate densities of flying birds (Tasker et al. 1984). The time of The positions of the boundaries between water masses were determined using salinity (Coach- man et aJ. 1975). ln 1985 and 1986, Bering Shelf Water was defined as 31.8-32.5 ppt (Coachman 1986, Walsh et al. 1989), with Anadyr Water MARINE BIRDS AND WATER MASSES 35 175° I I I I I I I I I I I I ) , I ,.~ / I , 170° ' ~ 30 m I 18.5Km 1---1 10 nautical miles 65° 64° 63° FIGURE I. Study area, showing generalized movements and origins of the major water masses flowing north through Bering Strait. Inner box shows area covered by distribution maps. FIGURE I. Study area, showing generalized movements and origins of the major water masses flowing north through Bering Strait. Inner box shows area covered by distribution maps. cinity of King Island was determined by Hunt and Harrison ( 1990). more saline, and Alaska Coastal Water less so. Although variations occur in the salinities of these water masses, they are typically small on a year to year time-scale (Coachman et al. 197 5, Coach- man and Shigaev, in prep.); hence, the same def- initions were applied to the 1984 data. For 1985 and 1986, the locations of boundaries between water masses follow Coachman (l 986:fig. 5; re- produced in part by Walsh et al. 1989), who produced maps based on extensive water sam- pling for the periods directly before and after the bird surveys. In addition, physical oceanograph- ic data collected during bird surveys were used to determine more precisely the positions of wa- ter mass boundaries (see salinity profiles in Hunt and Harrison 1990 and Hunt et al. l 990a). RESULTS Significant relationships between bird distribu- tions and water masses were detected, in at least one year, for each of the species considered ex- cept for the separate analysis for Thick-billed Murres (Uria lomvia) in 1985 (Table I). These relationships varied in their extent and nature between species and in a few cases between years for the same species. Further, the types of rela- tionships differed between species that are pre- dominantly planktivorous and those that are pi- scivorous. 66.0 .-------....,,,,,,.-------7.:'"'---i 65.5 64.0 0 0 000 CXD 0 e o <D 0 ~ 0000 BSW ~ (n•l66) 635 173 172 171 170 169 166 167 166 0 LONGITUDE DATA COLLECTION AND ANALYSIS Water mass coverage for each of the three years of the study: a) 1984, b) 1985, c) 1986. AW = Anadyr Water; BSW = Bering Shelf Water; ACW = Alaska Coastal Water. 36 a significant relationship, we used Dunn's mul- tiple comparison test for nonparametric data of unequal sample sizes (Zar 1984). This test de- termines which differences in water mass use contribute to the significant result. In each case, a null hypothesis of no difference in use of the water masses was assumed. 64.0 Samples/Block 0 1-3 a> 4-6 e 7-9 • 10+ 63·5173 172 171 170 169 166 167 166 Because these tests assume that samples are independent, we tested for autocorrelation with- in the data set. These tests were conducted for each species on each transect and yielded 35 (8.3%) out of 422 tests in which the lag-I coef- ficient lay outside the critical region and the null hypothesis of independence could not be ac- cepted. Even with independent samples, how- ever, one would expect I out of every 20 tests (5.0%) to produce a value that lies outside this region (Chatfield 1980). Consequently, we chose not to reject data from nine cases that had coef- ficients that were only just outside the critical region. The data from the remaining 26 cases with significant coefficients of autocorrelation were not used in the analysis. Tests were carried out using SYST AT (Wilkinson 1990) and em- ployed a significance level of 5%. DATA COLLECTION AND ANALYSIS Less- detailed information is available for 1984, al- though the position of the front between Alaska Coastal Water and Bering Shelf Water in the vi- more saline, and Alaska Coastal Water less so. Although variations occur in the salinities of these water masses, they are typically small on a year to year time-scale (Coachman et al. 197 5, Coach- man and Shigaev, in prep.); hence, the same def- initions were applied to the 1984 data. For 1985 and 1986, the locations of boundaries between water masses follow Coachman (l 986:fig. 5; re- produced in part by Walsh et al. 1989), who produced maps based on extensive water sam- pling for the periods directly before and after the bird surveys. In addition, physical oceanograph- ic data collected during bird surveys were used to determine more precisely the positions of wa- ter mass boundaries (see salinity profiles in Hunt and Harrison 1990 and Hunt et al. l 990a). Less- detailed information is available for 1984, al- though the position of the front between Alaska Coastal Water and Bering Shelf Water in the vi- We then determined the water mass in which each 15-min interval was obtained. In cases where an interval intersected a boundary between water masses, the water mass over which its central point lay was taken to be that of the sample. This method of assigning samples to water masses is problematic in that it uses "average" positions for boundaries that fluctuate over time spans as short as a few days (Coachman 1986, Hunt and Harrison 1990). Misclassifications, however, are most likely to reduce, or have no affect on, the apparent significance of differences in bird den- sities in each water mass. Consequently, the P-values obtained are probably conservative. During 1984 and 1986, data were collected in only two of the three water masses (Fig. 2); hence, for these years, we used Mann-Whitney U-tests CHRIS S. ELPHICK AND GEORGE L. HUNT, JR. 36 CHRIS S. ELPHICK AND GEORGE L. HUNT lLJ a ::> I- ~ _J 0 64.0 Samples/Block 0 1-3 a> 4-6 e 7-9 • 10+ 63·5173 172 171 170 169 166 167 166 64.0 635 173 172 171 170 169 166 167 166 66.0 .-------....,,,,,,.-------7.:'"'---i 65.5 64.0 0 0 000 CXD 0 e o <D 0 ~ 0000 BSW ~ (n•l66) 635 173 172 171 170 169 166 167 166 0 LONGITUDE FIGURE 2. PLANKTIVOROUS BIRDS The planktivorous species [shearwaters, Red Phalarope, Least (Aethia pusilla) and Crested (A. cristatella) auklets], and the omnivorous North- ern Fulmar (Fulmarus g/acialis), occurred in higher densities in Bering Shelf and Anadyr Wa- ter than in Alaska Coastal Water (Fig. 3, Tables 1, 2). Within this group, further differences ex- isted between species with different feeding strat- egies. The procellarids showed consistent pref- erences for the most westerly water mass sampled (i.e., that with the strongest Anadyr Water influ- ence; Figs. 4a, b). Red Phalaropes showed a strong preference for Anadyr Water in 1985, but they FIGURE 2. Water mass coverage for each of the three years of the study: a) 1984, b) 1985, c) 1986. AW = Anadyr Water; BSW = Bering Shelf Water; ACW = Alaska Coastal Water. FIGURE 2. Water mass coverage for each of the three years of the study: a) 1984, b) 1985, c) 1986. AW = Anadyr Water; BSW = Bering Shelf Water; ACW = Alaska Coastal Water. (Zar 1984) to determine whether a bird species occurred in each of the water masses equally. In 1985, all three water bodies were sampled, so we used Kruskal-Wallis tests to determine differ- ences in habitat use. For those species that showed MARINE BIRDS AND WATER MASSES 37 TABLE 1. Results of statistical comparisons of use of water masses by species commonly found in the Chirikov Basin. The Mann-Whitney U-statistic is given for 1984 and 1986, and the Kruskal-Wallis H-statistic is given for 1985. Values in parentheses are sample sizes for AW, BSW, and ACW, respectively. Sample sizes vary among species, within years, due to the need to exclude transects with excessive levels of autocorrelation between samples. 1984 Species u p Northern Fulmar 193.5 <0.001 (-, 21 , 41) 'Dark-bellied' Shearwater Black-legged Kittiwake 941.5 0.325 (-, 38, 44) Red Phalarope 616.0 <0.001 (-, 38, 44) Parakeet Auklet 575.0 0.014 (-, 38, 44) Least AukJet 166.0 <0.001 (-, 38, 44) Crested AukJet 627.5 0.004 (-, 38, 44) All murres 1,327.5 <0.001 (-, 38, 44) Common Murre Thick-billed Murre Horned Puffin 1,135.5 <0.001 (-, 38, 44) Tufted Puffin 1,312.0 <0.001 (-. PLANKTIVOROUS BIRDS Consequently, we carried out a mul- tiple comparison test for this species which showed significantly higher densities in Alaska Coastal Water, than in Bering Shelf Water, al- though not when compared with Anadyr Water (Table 2). This weak relationship is of interest given the absence of any differences between did not in 1986 (Fig. 3, Tables 1, 2). In all years, however, phalaropes were most numerous in ar- eas where Gray Whales (Eschrichtius robustus) were seen (Fig. 4c; also see Obst and Hunt 1990). The two Aethia auklets occurred in highest den- sities in Bering Shelf Water in 1984 and 1985 (Fig. 3). In 1986, Least Auklets apparently fa- vored Anadyr Water (Fig. 3, Table 1). Inspection of Fig. 4d, however, also shows high densities in stratified Bering Shelf Water north of St. Law- rence Island and near the front to the west of King Island. No preference for any water mass was detected for Crested Auklets in 1986. Par- akeet Auklets occurred in highest densities in Bering Shelf Water in 1984, avoided that water in 1985, and exhibited no preference in I 986. did not in 1986 (Fig. 3, Tables 1, 2). In all years, however, phalaropes were most numerous in ar- eas where Gray Whales (Eschrichtius robustus) were seen (Fig. 4c; also see Obst and Hunt 1990). The two Aethia auklets occurred in highest den- sities in Bering Shelf Water in 1984 and 1985 (Fig. 3). In 1986, Least Auklets apparently fa- vored Anadyr Water (Fig. 3, Table 1). Inspection of Fig. 4d, however, also shows high densities in stratified Bering Shelf Water north of St. Law- rence Island and near the front to the west of King Island. No preference for any water mass was detected for Crested Auklets in 1986. Par- akeet Auklets occurred in highest densities in Bering Shelf Water in 1984, avoided that water in 1985, and exhibited no preference in I 986. PLANKTIVOROUS BIRDS 38, 44) Year 1985 1986 H p u p 111.1 <0.00 I 3,911.5 0.009 (119, 124, 71) (67, 149, -) 4,499.0 0.009 (69, 166, -) 26.3 <0.001 5,255.5 0.196 (104, 144, 40) (64, 148, -) 12.6 0.002 5,456.5 0.086 (106, 139, 65) (69, 166, -) 7.5 0.024 5,135.5 0.145 ( 117, 148, 46) (69, 166, -) 86.6 <0.001 1,804.5 <0.001 ( 114, 118, 34) (53, 116, -) 59.5 <0.001 5,041.5 0.900 (127, 143, 71) (67, 149, -) 3.9 0.142 5,461.0 0.263 (118, 123, 59) (67, 149, -) 5.8 0.056 (118, 147, 59) 2.1 0.355 (116, 104, 59) 27.l <0.001 5,417.5 0.187 (127, 153, 71) (69, 166, -) 7.9 0.020 5,955.0 0.386 (127, 153, 71) (69, 166, -) Kittiwakes (Rissa tridactyla), murres, and puffins (Fratercula spp.)] either occurred at highest den- sities in Alaska Coastal Water or exhibited no preference for any water mass (Fig. 5, Tables I, 2). In 1986, when coastal water was not sampled, none of these species showed a significant pref- erence for either of the other water masses (Table 1). The comparison of the two murres' distri- butions during 1985 revealed differences in their pattern of habitat use. Although Kruskal-Wallis tests failed to show a significant preference for either species (Table I), the result for Common Murres (U. aalge) was very close to the 0.05 level (P = 0.056). Consequently, we carried out a mul- tiple comparison test for this species which showed significantly higher densities in Alaska Coastal Water, than in Bering Shelf Water, al- though not when compared with Anadyr Water (Table 2). This weak relationship is of interest given the absence of any differences between Kittiwakes (Rissa tridactyla), murres, and puffins (Fratercula spp.)] either occurred at highest den- sities in Alaska Coastal Water or exhibited no preference for any water mass (Fig. 5, Tables I, 2). In 1986, when coastal water was not sampled, none of these species showed a significant pref- erence for either of the other water masses (Table 1). The comparison of the two murres' distri- butions during 1985 revealed differences in their pattern of habitat use. Although Kruskal-Wallis tests failed to show a significant preference for either species (Table I), the result for Common Murres (U. aalge) was very close to the 0.05 level (P = 0.056). PISCIVOROUS BIRDS 64.0 BirdolKm' 28.01-5 ; ®5.01·10 •>10 172 171 170 169 168 167 166 w C> :::> I- I- <{ -' 0 - 66.0 { ' '· 65.5 \ ' BirdolKm' £Ko1 - 10 110.01 -30 >30 170 169 168 167 166 0 LON 65.5 64.5 64.0 63.5 / BirdlJKln' ~801 - 10 ~~:r.:~ @ 100.01 . 500 •>500 173 172 171 170 169 168 167 166 0 LONGITUDE FIGURE 4. Pelagic distributions of some planktivorous species across the Chirikov Basin: a) Northern Fulmar, 1985; b) "dark-bellied shearwaters," 1986; c) Red Phalarope, 1985; d) Least Auklet, 1986. The polygon of small dots on a) and c) shows the region in which Gray Whales were commonly seen. FIGURE 4. Pelagic distributions of some planktivorous species across the Chirikov Basin: a) Northern Fulmar, 1985; b) "dark-bellied shearwaters," 1986; c) Red Phalarope, 1985; d) Least Auklet, 1986. The polygon of small dots on a) and c) shows the region in which Gray Whales were commonly seen. ing a primary influence on where puffins forage, then one would expect them to be equally dis- tributed around their colonies. Inspection of the data, however, shows that their distribution ex- tends much farther to the east of King Island than to the west, where their occurrence ends near the boundary between Alaska Coastal Wa- ter and Bering Shelf Water (Fig. 6). two broad categories, with planktivorous bird species using Anadyr Water and Bering Shelf Water most frequently and piscivorous species primarily using Alaska Coastal Water. For some bird species, the boundaries between water mass- es appeared to set the limits of foraging distri- butions (e.g., puffins; Fig. 6), whereas for others the frontal areas were a focus of foraging activity [e.g., Least Auklets (Hunt and Harrison 1990) and shearwaters (Fig. 4b)]. PISCIVOROUS BIRDS In contrast to the planktivores, those seabird spe- cies that forage primarily on fish [Black-legged CHRIS S. ELPHICK AND GEORGE L. HUNT, JR. 38 ~ i 10 g 10 i ';,, l6 u. ! ! 5 5 ~ 0 0 198.4 1985 1986 1984 1985 1986 5 150 ! 50 198-4 1985 1986 1984 1985 1986 1984 1985 1986 1984 1985 1988 AG URE 3. Mean densities of planktivorous species in each water mass, for 1984-1986. Open bars = Al Coastal Water; solid bars= Bering Shelf Water; cross-hatched bars= Anadyr Water. Error bars give stan errors. Sample sizes are as given in Table I. AG URE 3. Mean densities of planktivorous species in each water mass, for 1984-1986. Open bars = Alaska Coastal Water; solid bars= Bering Shelf Water; cross-hatched bars= Anadyr Water. Error bars give standard errors. Sample sizes are as given in Table I. water masses in the densities of Thick-billed Murres. in the densities of puffins in each water mass was particularly striking in 1984 (Fig. 6). These two species often forage close to their colonies (Sealy 1973, Hunt et al. 1981) which creates the poten- tial for a confounding effect of colony position within our analysis. If colony situation were hav- Homed (Fratercula corniculata) and Tufted (F cirrhata) puffins both showed preferences for Alaska Coastal Water in the two years when it was sampled (Fig. 5, Tables l, 2). The contrast MARINE BIRDS AND WATER MASSES 39 64.0 ___. 64.0 BirdolKm' w £8.01-5 28.01-5 C> ,, :s.01-10 ; ®5.01·10 >10 •>10 :::> I- 63.5 173 172 171 170 169 168 167 166 172 171 170 169 168 167 166 I- 66.0 <{ -' '· 0 65.5 \ ' 65.5 64.5 BirdolKm' 64.0 £Ko1 - 10 110.01 -30 >30 63.5 / BirdlJKln' ~801 - 10 ~~:r.:~ @ 100.01 . 500 •>500 170 169 168 167 166 173 172 171 170 169 168 167 166 0 LONGITUDE FIGURE 4. Pelagic distributions of some planktivorous species across the Chirikov Basin: a) Northern Fulmar, 1985; b) "dark-bellied shearwaters," 1986; c) Red Phalarope, 1985; d) Least Auklet, 1986. The polygon of small dots on a) and c) shows the region in which Gray Whales were commonly seen. DISCUSSION Our results support the hypothesis that the pe- lagic distributions of seabirds across the Chiri- kov Basin reflect preferences for one or more of the three water masses, as predicted by Drury et al. ( 1981) and Springer et al. (1987). The vari- ation in these preferences, however, indicates that factors other than water mass also have some influence. Preferences for water masses fell into The concentrations of shearwaters in Anadyr Water, and in particular near the front between Anadyr and Bering Shelf Waters (Fig. 4b), may reflect a dependence on frontal systems for the availability of prey. These shearwaters also con- centrate near the Inner Front in the southeastern Bering Sea (Schneider 1982) and near a front to the north of St. Paul Island that is physically CHRIS S. ELPHICK AND GEORGE L. HUNT, JR. 40 TABLE 2. Results of Dunn's multiple comparison tests between each pair of water masses for 1985. For each comparison, the water mass with the highest density of birds, and Q, are given. Significance levels: * = P < 0.05, •• = p < 0.01, ••• = p < 0.001. TABLE 2. Results of Dunn s multiple comparison tests between each pair of water masses for 1985. For each comparison, the water mass with the highest density of birds, and Q, are given. Significance levels: * = P < 0.05, •• = p < 0.01, ••• = p < 0.001. Species Northern Fulmar Black-legged Kittiwake Red Phalarope Parakeet Auklet Least Auklet Crested Auklet Common Murre Homed Puffin Tufted Puffin 1984 1984 1985 1985 AW/BSW AW 8.65* AW 2.12 AW 3.06 AW 1.85 BSW 7.57* BSW 4.11* AW 0.79 BSW 0.62 BSW 2.80 1986 1986 Water mass preferred in the comparison of BSW/ ACW BSW 1.77 ACW 5.10* ACW 0.53 ACW 2.51* BSW 6.63* BSW 7.62* ACW 2.41* ACW 4.55* ACW 0.97 1984 1984 1985 1985 AW/ACW AW 9.19 ACW 3.43 AW 3.01 ACW 1.14 AW 3.42 AW 4.09* ACW I. 71 ACW 4.91* ACW 1.34 1986 1986 FIGURE 5. Mean densities of piscivorous species in each water mass, for 1984-1986. Symbols as described for Figure 3. comparison, the water mass with the highest density of birds, and Q, are given. Significance levels: = P < 0.05, •• = p < 0.01, ••• = p < 0.001. DISCUSSION Species Northern Fulmar Black-legged Kittiwake Red Phalarope Parakeet Auklet Least Auklet Crested Auklet Common Murre Homed Puffin Tufted Puffin 1984 1984 1985 1985 AW/BSW AW 8.65* AW 2.12 AW 3.06 AW 1.85 BSW 7.57* BSW 4.11* AW 0.79 BSW 0.62 BSW 2.80 1986 1986 Water mass preferred in the comparison of BSW/ ACW BSW 1.77 ACW 5.10* ACW 0.53 ACW 2.51* BSW 6.63* BSW 7.62* ACW 2.41* ACW 4.55* ACW 0.97 1984 1984 1985 1985 AW/ACW AW 9.19 ACW 3.43 AW 3.01 ACW 1.14 AW 3.42 AW 4.09* ACW I. 71 ACW 4.91* ACW 1.34 1986 1986 FIGURE 5. Mean densities of piscivorous species in each water mass, for 1984-1986. Symbols as described for Figure 3. Species Northern Fulmar Black-legged Kittiwake Red Phalarope Parakeet Auklet Least Auklet Crested Auklet Common Murre Homed Puffin Tufted Puffin AW/BSW AW 8.65* AW 2.12 AW 3.06 AW 1.85 BSW 7.57* BSW 4.11* AW 0.79 BSW 0.62 BSW 2.80 Water mass preferred in the comparison of BSW/ ACW BSW 1.77 ACW 5.10* ACW 0.53 ACW 2.51* BSW 6.63* BSW 7.62* ACW 2.41* ACW 4.55* ACW 0.97 AW/ACW AW 9.19 ACW 3.43 AW 3.01 ACW 1.14 AW 3.42 AW 4.09* ACW I. 71 ACW 4.91* ACW 1.34 1984 1984 1985 1985 1986 1986 1984 1984 1985 1985 1986 1986 FIGURE 5. Mean densities of piscivorous species in each water mass, for 1984-1986. Symbols as describ for Figure 3. FIGURE 5. Mean densities of piscivorous species in each water mass, for 1984-1986. Symbols as described for Figure 3. MARINE BIRDS AND WATER MASSES 41 41 65.5 65.0 ~ BSW 1 ACW 64.5 64.0 _,,, w 1~0• · 2 0 Z.Ot • 5 >5 => 63.5 t- 173 172 171 170 169 168 167 166 t- <t _J '· 0 65.5 65.0 64.5 64.0 _,,,. igo, .2 2.01 . 5 •5 63.5173 172 171 170 169 168 167 166 0 LONGITUDE FIGURE 6. Pelagic distributions of a) Homed and b) Tufted puffins around King Island in 1984. equivalent to the Inner Front(Coyleand Cooney, in press). At this latter front, shearwaters forage on Thysanoessa raschii (Hunt et al., unpubl.). We do not know the diets of these shearwaters in the Chirikov Basin, but if they are taking euphau- siids, the strong vertical fluxes associated with the edge of the Anadyr Current (Haney 1991) could aid in concentrating and transporting this prey upward in the water column. DISCUSSION During this study, Least Auklets were wide- spread in both Anadyr and Bering Shelf Waters. Other analyses have shown that within these wa- ter masses they tended to concentrate in frontal regions (Hunt and Harrison 1990, Harrison et al. 1990) and over stratified water (Hunt et al. l 990a, Hunt and Harrison 1990). These conclu- sions, based on the analysis of individual tran- sects, are borne out by our Figure 4d which shows highest densities close to the front between Ana- dyr and Bering Shelf Waters and offshore to the north of St. Lawrence lsland, where water tends to be well stratified (Hunt et al. l 990a; Coach- man and Shigaev, in prep.). Inspection of the data from the earlier analyses showed that Least Auklets used stratified water only when Bering Shelf Water was present in the upper mixed layer (e.g., Hunt and Harrison l 990:fig. 8). When Alas- ka Coastal Water overlay Bering Shelf Water, Least Auklets were largely absent. This obser- vation may explain the results of Haney ( 1991 ), who found that Least Auklets avoided strongly stratified water near shore at the western end of St. Lawrence lsland. He did not identify the or- igin of this water or the composition of its plank- ton community, but ifthe upper mixed layer was derived from terrestrial runoff, the copepods that are the principal prey of Least Auklets may have been absent. w 0 => t- t- <t _J 0 0 LONGITUDE FIGURE 6. Pelagic distributions of a) Homed and b) Tufted puffins around King Island in 1984. curred in highest densities in the region where Gray Whales were most commonly seen (Figs. 4a, c). The other two species (Black-legged K.it- tiwake and Thick-billed Murre) did not exhibit similar water mass preferences, and in 1985 the former preferred Alaska Coastal Water, a pref- erence similar to other piscivores. These differ- ences suggest that whales act to concentrate avian predators once they have already chosen their foraging habitat, rather than being the primary attraction. As the area in which Gray Whales were most frequently seen during this study ex- tends across all three water masses (Fig. 7), one might therefore expect different assemblages of birds associating with whales in each water mass. DISCUSSION To test the validity of Haney's argument for the piscivorous alcids, we will have to learn the distribution by depth and water mass of the prey species identified by Springer et al. ( 1987). Recently, both marine (e.g., Steele 1978, 1989; Schneider and Duffy 1985) and terrestrial (e.g., Wiens 1976) ecologists have focused on the im- portance of the scale of habitat heterogeneity for ecological processes. An underlying concept is that there may be characteristic spatial (and tem- poral) scales at which certain processes occur (e.g., Haury et al. 1978, Steele 1978, Hunt and Schnei- der 1987). Throughout these discussions the spa- tial scales associated with different processes are explicitly addressed. The biological importance of the boundary conditions that define habitat regions, however, have seldom been described. Kotliar and Wiens (1990) have recently pro- posed that the nature of patch boundaries may change with patch size; large-scale patches may have more ambiguous bounds than small patch- es. They also emphasized a continuum in the degree of contrast between a patch and the matrix in which it exists. We suggest that the degree of contrast is not necessarily a function of the scale of the patch but may be independent of patch size. We hypothesize that a predator's ability to recognize a habitat patch will depend on both the size of the patch, and the sharpness of its contrast with the surrounding matrix .. For ex- ample, the differences in seabird species distri- butions described for the Chirikov Basin were detectable over the small scales at which one might have expected to find differences in the biomass of avian predators, but not in species composition (Hunt and Schneider 1987). We suggest that this segregation of habitat use de- pended on the strong fronts (steep property gra- dients) bounding the three water masses that cre- ated sharp gradients in the distribution and abundance of prey to which the predators re- sponded (Hunt and Harrison 1990). Had these gradients been weaker, one would expect that the dramatic differences in use of water masses over short distances west of King Island would have been much reduced or absent. Other evidence from marine birds suggests that at scales below the mesoscale, birds largely ignore weakly de- fined habitats (e.g., Hunt et al. l 990b), whereas when definition is strong, habitat patches are used FIGURE 7. DISCUSSION The preference oflargely planktivorous species for Bering Shelf Water and An ad yr Water is un- derstandable because of the presence of larger prey species (e.g., Neoca/anus p/umchrus, N. cris- curred in highest densities in the region where Gray Whales were most commonly seen (Figs. 4a, c). The other two species (Black-legged K.it- tiwake and Thick-billed Murre) did not exhibit similar water mass preferences, and in 1985 the former preferred Alaska Coastal Water, a pref- erence similar to other piscivores. These differ- ences suggest that whales act to concentrate avian predators once they have already chosen their foraging habitat, rather than being the primary attraction. As the area in which Gray Whales were most frequently seen during this study ex- tends across all three water masses (Fig. 7), one might therefore expect different assemblages of birds associating with whales in each water mass. Our results for Crested Auklets indicate that, like Least Auklets, they prefer Bering Shelf Wa- ter, suggesting that they may also favor stratified water. In contrast, Parakeet Auklets exhibit no consistent preference for any water mass. This suggests that factors determining Parakeet Auk- let distributions are independent of water mass and, hence, differ from those influencing the dis- tribution of the other auklets (Harrison 1987). Obst and Hunt (1990) showed that four of the species discussed here tend to associate with Gray Whales in the Chirikov Basin. Our analysis shows that two of these (Northern Fulmar and Red Phalarope) also showed a preference for Anadyr Water and that within this water mass they oc- The preference oflargely planktivorous species for Bering Shelf Water and An ad yr Water is un- derstandable because of the presence of larger prey species (e.g., Neoca/anus p/umchrus, N. cris- tatus and Eucalanus bungii) there (Springer et al. CHRIS S. ELPHICK AND GEORGE L. HUNT, JR. 42 65.5 65.0 64.5 64.0 635 '----""-'--'-""""'""'-'-"""---------~ 173 172 17 1 170 169 168 167 166 65.5 65.0 64.5 64.0 635 '----""-'--'-""""'""'-'-"""---------~ 173 172 17 1 170 169 168 167 166 FIGURE 7. Distrilbution of Gray Whales in the Chi- rikov Basin in 1985. Filled circles represent sightings of at least one individual; open circles represent blocks through which transects were made. calJy. This argument, however, would not ex- plain the preference of surface-feeding kittiwakes for Alaska Coastal Water. DISCUSSION Distrilbution of Gray Whales in the Chi- rikov Basin in 1985. Filled circles represent sightings of at least one individual; open circles represent blocks through which transects were made. 1989, Hunt and Harrison 1990, Hunt et al. l 990a). These prey are not numerous in Alaska Coastal Water, and are presumed to be suffi- ciently more profitable as food items than the more numerous small zooplankton. In contrast, we do not know why piscivores apparently avoid Bering Shelf Water and Anadyr Water, but little is known about the abundance of fish species in the different water masses. Some of the region's most important species of forage fish [e.g., sand- lance (Ammodytes hexapterns), capelin (Mallo- tus vil/osus); Drury et al. 1981, Springer et al. 1987] spawn in shallow water or the intertidal zone (Hart 1973) and are therefore at least pe- riodically abundant in coastal waters. Much of the Alaska Coastal Water is more than 90 km from shore, however, and there are no data of which we are aware on the abundance of forage fish in offshore areas. Based on the distributions of the piscivorous seabird species, we can only hypothesize that fishes must be more available in Alaska Coastal Water than in the other water masses of the Chirikov Basin. An alternative explanation for the preference of the piscivores for Alaska Coastal Water is that they prefer to forage in shallow water. Haney ( 1991) found that large alcids near the western end of St. Lawrence Island occurred in higher numbers in shallower water. He hypothesized that these birds may take a large proportion of epibenthic fish in their diets and that foraging in shallow water would be more efficient energeti- MARINE BIRDS AND WATER MASSES 43 differentially (e.g., Haney 1986, Hunt and Har- rison 1990, Veit and Hunt 199 1). In contrast, at mesoscaJes and above, even weakly-defined hab- itats differ in use (e.g., Schneider et al. 1986, Hunt and Schneider 1987, Hunt et al. 1990b). Thus, the scheme described by Hunt and Schnei- der ( 1987} may need modification. It may not be possible to characterize the types of relation- ships between seabirds and their prey solely by spatial scale, but rather by a combination of both the scale and the strength of the boundaries that define a habitat type. HAURY, LR., J. A. McGOWAN, AND P. H. WIEBE. 1978. LITERATURE CITED CHATFIELD, C. 1980. The analysis of time series: an introduction. Chapman and Hall, London. COACHMAN, L. K. 1986. Hydrography of the ISH- TAR study area, 1985 and 1986. ISHTAR prog- ress report. Univ. of Washington, Seattle. KoTUAR, N. B., AND J. A. WrENs. 1990. Multiple scales of patchiness and patch structure: a hier- archical framework for the study of heterogeneity. Oikos 59:253-260. CoACHMAN, L. K., K. AAGARO, AND R. B. TRIPP. 1975. Bering Strait: the regional physical oceanography. Univ. Washington Press, Seattle. OBST, B. S., AND G. L. HUNT, JR. 1990. Marine birds feed at gray whale mud plumes in the Bering Sea. Auk 107:678-688. CoYLE, K. 0., AND R. T. CooNEY. In press. Water column sound scattering and hydrography around the Pribilof Islands, Bering Sea. Cont. Shelf Res. c 5cHNE1DER, D. C. 1982. Fronts and seabird aggre- gations in the southeastern Bering Sea. Mar. Ecol. Prog. Ser. 10:101-103. DRURY, W. H., c. RAMSDELL, AND J.B. FRENCH, JR. 1981. Ecological studies in the Bering Strait Re- gion, p. 175-488. In Environmental assessment of the Alaskan Continental Shelf. U.S. Dept. Comm. NOAA, OCSEAP, Final Report, Vol. 2. ScHN-ErDER, D. C., AND D. C. OtJFFY. 1985. Scale- dependent variability in seabird abundance. Mar. Ecol. Prog. Ser. 25:211-218. HANEY, J. C. 1986. Seabird segregation at Gulf Stream frontal eddies. Mar. Ecol. Prog. Ser. 28:279-285. 5cHNEIDER, D. C., G. L. HUNT, JR., AND N. M. HARRISON. 1986. Mass and energy transfer to seabirds in the southeastern Bering Sea. Cont. Shelf Res. 5:241- 257. HANEY, J. C. 1991- Influence of pycnocline topog- raphy and water-column structure on marine dis- tributions ofalcids (Aves: Alcidae) in Anadyr Strait, northern Bering Sea, Alaska. Mar. Biol. 110:419- 435. ScHNEmER, D. C., AND J. F. PIATT. 1986. Scale-de- pendent correlation of seabirds with schooling fish in a coastal ecosystem. Mar. Ecol. Prog. Ser. 32: 237- 246. HARRISON, N. M. 1987. Foraging behavior and co- existence of seabirds in the Bering Sea. Ph.D.diss., Univ. of California, Irvine, CA. ScHNEJDER, D. c., N. M. HARRISON, AND G. L. HUNT, JR. 1987. Variation in the occurrence of marine birds at fronts in the Bering Sea. Estuar. Coast. ShelfSci. 25:135-141. HARRISON, N. M., G. L. HUNT, JR., AND R. T. CooNEY. 1990. Front affecting the distribution of seabirds in the northern Bering Sea. Pol. Res. 8:29-31. J L HART, J. L. 1973. Pacific fishes of Canada. Fish. Res. Board Can. Bun. 180, Ottawa. ACKNOWLEDGMENTS We thank Z. Eppley, N. Harrison, B. Obst, C. Williams and B. Young for carrying out bird observations for this research. The captain, crew, and marine techni- cians of the RIV Alpha Helix were most helpfu~ with- out their help and cooperation, the work would not have been possible. R. H. Day, M. Rubega, D. Schnei- der, E. Woehler, members of Ecology and Evolutionary Biology 227 at UC Irvine, and two anonymous re- viewers provided helpful comments on earlier drafts of this manuscript. The research was supported in pan by NSF Grant DPP-8308232 to GLH. CSE was sup- ported while at UC Irvine, by the BIO/USA program of the University of East Anglia, England. HUNT, G. L., JR., AND N. M. HARRJSON. 1990. For- aging habits and prey taken by Least Auklets at Kfog Island, Alaska. Mar. Ecol. Prog. Series 65: 141-150. HUNT, G. L.,JR., N. M. HARRisoN, ANDR. T.CooNEY. I 990a. The influence of hydrographic structure and prey abundance on foraging of Least Auklets. Stud. Avian Biol. 14:7-22. HUNT, G. L , JR., 0 . HEINEMANN, R. R. VEIT, R. B. HEYWOOD, AND I. EVERSON. I 990b. The distri- bution, abundance and community structure of marine birds in southern Drake Passage and Bransfield Strait, Antarctica. Cont. Shelf. Res. 10: 243-257. K!NDER, T. H., AND J. D. ScHuMACHER. 1981. Cir- culation over the continental shelf of the south- eastern Bering Sea, p. 53-75. In D. W. Hood and J. A. Calder [eds.], The eastern Bering Sea shelf: oceanography and resources. Vol. I. NOAA/ OMPA, Washington, DC. DISCUSSION Patterns and processes in time-space scales of plankton distributions, p. 277-327. In J. H. Steele [ed.], Spatial pattern in plankton commu- nities. Plenum Press, New York, NY. HUNT, G. L., JR., P. J. GOULD, D. J. FORSELL, AND H. PETERSON. 1981 . Pelagic distribution of marine birds in the eastern Bering Sea, p. 689-718. In D. W. Hood and J. A. Calder [eds.), The eastern Be- ring Sea shelf: oceanography and resources. Vol. 2. NOAA/OMPA, Washington, DC. HUNT, G. L., JR., AND D. C. 5cHNEJDER. 1987. Scale- dependent processes in the physical and biological environment of marine birds, p. 7-41. In J. P. Croxall [ed.], Seabirds: feeding ecology and role in marine ecosystems. Cambridge Univ. Press, Cambridge, England. LITERATURE CITED SEALY, S. G. 1973. Breeding biology of the homed puffin on St. Lawrence Island, Bering Sea, with 44 CHRISS. ELPHICK AND GEORGE L. HUNT, JR. TASKER, M. L., P. HOPE JONES, T. DIXON, AND B. F. BI.AKE. 1984. Counting seabirds at sea from shlps: a review of methods employed and a suggestion for a standardized approach. Auk 101:567-577. zoogeographical notes on the North Pacific puffins. Pac. Sci. 27:99-119. zoogeographical notes on the North Pacific puffins. Pac. Sci. 27:99-119. SoWLS, A. L., s. A. HATCH, AND c. J. LENSfNK. 1978. Catalog of Alaskan seabird colonjes. Biologkal Services Program, U.S. Fish and Wildlife Service, Anchorage, AK. VEIT, R. R ., AND G . L. Hum, JR. 1991. Broadscale density and aggregation of pelagic birds from a circumnavigational survey of the Antarctic Ocean. Auk 108:790-800. SPRINGER, A. M., AND D. G. RosENEAU. 1985. Co- pepod-based food webs: auklets and oceanography in the Bering Sea. Mar. Ecol. Prog. Ser. 21 :229- 237. WALSH, J. J., C. P. McRov, L. K. COACHMAN, J. J. GOERING, J. J. NIHOUL, T. E. WHITLEDGE, T. H. BLACKBURN, P. L. PAR.KER, c. D. WIRICK, P. G. SHUERT, J. M. GREBMEIER, A. M. SPRJNGER, R. D. TRJPP, D. A. HANSELL, s. DJENIDI, E. DELEERS- NUDER, K. HENRIKSEN, B. A. LUND, P. ANDERSEN, F. E. MOLLER-KARGER, AND K. DEAN. 1989. Carbon and nitrogen cycling within the Bering/ Chukchi seas: source regions for organic matter effecting AOU demands of the Arctic Ocean. Prog. Oceanog. 22:277-359. SPRJNGER, A. M., E. c. MURPHY, D. G. ROSENEAU, c. P. McRov, AND B. A. COOPER. 1987. The par- adox of pelagic food webs in the northern Bering Sea: I. Seabird food habits. Cont. Shelf Res. 7:895- 911. SPRJNGER, A. M., C. P. McRov, AND K. R. TuRco. 1989. The paradox of pelagic food webs in the northern Bering Sea: ll. Zooplankton communi- ties. Cont. Shelf Res. 9:359-386. WIENS, J. A. 1976. Population responses to patchy environments. Ann. Rev. Ecol. Syst. 7:81-120. STEELE, J. H. 1978. Some comments on plankton patches, p. 1-20. Jn J. H. Steele [ed.], Spatial pat- tern in plankton communities. Plenum Press, New York. W11.KJNsoN, L. 1990. SYST AT: the system for sta- tistics. SYST AT, Inc., Evanston, IL. STEELE, J. H. 1989. The ocean 'landscape.' Landscape Ecol. 3:185-192. ZAR, J. H. 1984. Biostatistical analysis. 2nd ed. Pren- tice-Hall, Englewood Cliffs, NJ.
https://openalex.org/W2948774368	https://www.e3s-conferences.org/articles/e3sconf/pdf/2019/24/e3sconf_wri-162018_06014.pdf	English	null	Climate change impact assessment on water security in South Africa: A case study in a semi-arid river basin	E3S web of conferences	2,019	cc-by	1,769	ERROR: type should be string, got "https://doi.org/10.1051/e3sconf/20199806014 https://doi.org/10.1051/e3sconf/20199806014 E3S Web of Conferences 98, 06014 (2019) \nWRI-16 Climate change impact assessment on \nwater security in South Africa: A case \nstudy in a semi-arid river basin Yali Woyessa1,* 1Department of Civil Engineering, Central University of Technology, Free State, Private \nBag X20539, Bloemfontein 9300, South Africa Abstract. The main aim of this paper is to assess the impact of regional \nclimate change scenarios on the availability of water resources in a semi-\narid river basin in South Africa using a hydrological model called Soil and \nWater Assessment Tool (SWAT). In this paper, climate change data was \nderived from two downscaling approaches, namely statistical downscaling \nexperiment (SDE) and dynamic downscaling (CORDEX). These were \nderived from the GCM simulations of the Coupled Model Inter-\ncomparison Project Phase-5 (CMIP5) and across two greenhouse gas \nemission scenarios known as Representative Concentration Pathways \n(RCP) 4.5 and 8.5. The spatial resolution of the dataset for the SDE \nmethod is 25 km × 25 km and 50 km × 50 km for the CORDEX method. Six GCM models were used for SDE set of data and four for the CORDEX \nset of data. SWAT model was run using these data for a period of up to \nmid-century (2020 – 2050) for SDE and for a period of up to the end of \nthis century (2020 – 2100) for CORDEX data. The results were then \ncompared with long-term historical data (1975-2005). Comparison of \nmeasured data with simulated historical data showed strong correlation (R2 \n= 0.95 for SDE data and R2 = 0.92 for CORDEX data), which is indicative \nof the reliability of projected future climate. 1 Introduction Water security is understood to be the availability of an acceptable quantity and quality \nof water for health, livelihoods, ecosystems and production, coupled with an acceptable \nlevel of water-related risks to people, the environment and economies [1]. The problem of \nwater scarcity is likely to deepen according to some projected climate change scenarios. It \nis projected that, by 2050, the Southern African region will be generally drier [2]. In \ngeneral, for South Africa there is a tendency towards a rise in temperature as evidenced in \nthe past five decades [3]. Maximum and minimum temperatures have also shown \nsignificant increases annually, and in almost all seasons. High temperature extremes have \nincreased significantly in frequency, and low temperatures have decreased significantly in \nfrequency annually and in most seasons across the country [2]. 2 Methodology The research focused on a semi-arid river basin in the central region of South Africa, \nknown as the Modder river basin. The Modder river basin is a large basin with a total area \nof 1.73 million hectares and divided into three sub-basins, namely the Upper Modder, the \nMiddle Modder and the Lower Modder. It is located within the Upper Orange Water \nManagement Area to the east of the city of Bloemfontein. The water supply to the middle \nand lower reaches of the Modder River is stabilized by two dams known as the Rustfontein \nand Mockes dams in the east, and by Krugersdrift Dam in the west of the city of \nBloemfontein. Climate change data was derived from two downscaling approaches, namely statistical \ndownscaling experiment (SDE) and dynamic downscaling (CORDEX). The downscaled \ndata were derived from the GCM simulations of the Coupled Model Inter-comparison \nProject Phase-5 (CMIP5) and across two greenhouse gas emission scenarios known as \nRepresentative Concentration Pathways (RCP) 4.5 and 8.5. The spatial resolution of the \ndataset for the SDE method is 25 km × 25 km and for the CORDEX method it is 50 km × \n50 km. Six GCM models were used for SDE set of data and four for the CORDEX set of \ndata. SWAT model was run using these data for a period of up to mid-century (2020 – \n2050) for SDE and for a period of up to the end of this century (2020 – 2100) for CORDEX \ndata. E3S Web of Conferences 98, 06014 (2019) \nWRI-16 Recent climate projections suggest a drop of up to 10% in precipitation in most of \nsouthern Africa by 2050 [4]. It is estimated that by the year 2025 almost one-half of the \nworld population will be living in water stressed regions [5], with rapid population increase, \nindustrialization and pollution putting further strain on the available and diminishing fresh \nwater resources. Within this context, there is a general consensus that development has to \nbe sustainable, adaptable and resilient to these global changes such as climate change. Resilient development involves initiatives that will deliver benefits under all potential \nfuture circumstances related to environmental changes [1]. Therefore, this calls for \ninnovative strategic approaches to deliver water security and resilient development. The \nmain aim of this paper is to investigate impact of climate change scenarios on water \nresources and water security. * Corresponding author: ywoyessa@cut.ac.za © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons \nAttribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1051/e3sconf/20199806014 E3S Web of Conferences 98, 06014 (2019) \nWRI-16 3 Results and Discussion Comparison of measured and simulated historical values are shown in Figures 1a and 1b for \nthe SDE approach and the CORDEX approach, respectively, for the period of 1975 to 2005. Figures 2a and 2b show the correlation between these values for SDE and CORDEX \napproaches over the same period. In Figures 2a and 2b, the blue points represent the \nobserved versus simulated values whereas the blue broken line shows the linear fit between \nthe observed and simulated values. The solid black line represents the one-to-one line (the \nperfect fit scenario). In both cases, the relationship between these two variables (measured \nand simulated historical values) are found to be strong with R2 > 0.92. The purpose of \nsimulating the historical data was to compare with climate observations over a common \nperiod and use information derived from the comparison to adjust future climate projections \nso that they are (progressively) more consistent with the historical climate records and, \npresumably, more realistic for the spatial domain of interest. 2 2 E3S Web of Conferences 98, 06014 (2019) E3S Web of Conferences 98, 06014 (2019) \nWRI-16 https://doi.org/10.1051/e3sconf/20199806014 ig. 1. Comparison of measured and simulated average historical monthly values of rain \nsing multimodel average: (a) SDE; (b) CORDEX. Fig. 1. Comparison of measured and simulated average historical monthly values of rain \nusing multimodel average: (a) SDE; (b) CORDEX. Fig. 1. Comparison of measured and simulated average historical monthly values of rain \nusing multimodel average: (a) SDE; (b) CORDEX. The projected decrease in rainfall is expected to have an impact on water balance \ncomponents of the river basin. The assessment of impact of climate change on water \nbalance components of the river basin showed varied results depending on the type of \nclimate model used and the downscaling approach, but generally, the trends were similar in \nmost cases (Figure 3a and 3b). In the case of the SDE approach, the multimodel average \nshowed a possible decrease in precipitation (by 14 %), a decrease in water yield (by 15 %) \nand an increase in potential evapotranspiration (by 10 %). A similar trend was observed for \nthe CORDEX approach, with a decrease in rainfall (by 3 %) and a decrease in water yield \n(by 13%) and an increase in evaporation (by 22 %). The increase in evaporation in both \ncases is indicative of possible drought spells between rainy events. 4 Conclusion The hydrological simulation as well as analysis of climate change data showed that there \nwill be a reduction in rainfall and other river basin water balance components under both \ngreenhouse gas emission scenarios, but the effect will be more pronounced under the \nRCP8.5. The expected decrease in rainfall coupled with increase in PET are going to be the \ntwo major drivers in exacerbating the existing problem of water scarcity in the region. For \nexample, decrease in rainfall (according to SDE’s approach) will result in the reduction of \nwater storage in a reservoir (Rustfontein Dam) by about 28% of its full capacity. If the 3 E3S Web of Conferences 98, 06014 (2019) https://doi.org/10.1051/e3sconf/20199806014 E3S Web of Conferences 98, 06014 (2019) \nWRI-16 increase in PET is added to this scenario, the effect will be much more pronounced and \nthere will be much more reduction in water storage. 2. Observed versus simulated average monthly values of rain using multimodel \nage: (a) SDE; (b) CORDEX. Climate change has direct impact on water resources. Similarly, the manageme\ner resources affects the vulnerability of ecosystems, socio-economic activities 2. Observed versus simulated average monthly values of rain using multimodel Fig. 2. Observed versus simulated average monthly values of rain using multimodel \naverage: (a) SDE; (b) CORDEX. Climate change has direct impact on water resources. Similarly, the management of \nwater resources affects the vulnerability of ecosystems, socio-economic activities and \nhuman health. Climate change is projected to lead to major changes in water availability \nwith increasing water scarcity and droughts mainly in the central region of South Africa. Water management is, therefore, expected to play an increasingly important role \nin adaptation to climate change. It is expected that the results of this research will assist in \nformulation of adaptation strategies that will minimize the negative impact of climate \nchange in the region. 4 E3S Web of Conferences 98, 06014 (2019) E3S Web of Conferences 98, 06014 (2019) \nWRI-16 https://doi.org/10.1051/e3sconf/20199806014 Fig. 3. Percentage change of future water balance components: (a) mid-century – SDE, (b) \nend of century – CORDEX. Fig. 3. Percentage change of future water balance components: (a) mid-century – SDE, (b) \nend of century – CORDEX. Fig. 3. Percentage change of future water balance components: (a) mid-century – SDE, (b) \nend of century – CORDEX. References 1. AMCOW, Water security and climate resilient development: Technical Background \nDocument (2012) 1. AMCOW, Water security and climate resilient development: Technical Background \nDocument (2012) 2. South African Department of Environmental Affairs, Climate trends and scenarios for \nSouth Africa, in Long Term Adaptation Scenarios (2013) 2. South African Department of Environmental Affairs, Climate trends and scenarios for \nSouth Africa, in Long Term Adaptation Scenarios (2013) 3. N. MacKellar, et al., South Afr J Sci, 110, 1-13 (2014) 3. N. MacKellar, et al., South Afr J Sci, 110, 1-13 (2014) 4. E. Levina, Domestic Policy Frameworks for Adaptation to Climate Change in the \nWater Sector Part II: Non-Annex I Countries. Lessons Learned from Mexico, India, \nArgentina and Zimbabwe, OECD, Paris (2006) 4. E. Levina, Domestic Policy Frameworks for Adaptation to Climate Change in the \nWater Sector Part II: Non-Annex I Countries. Lessons Learned from Mexico, India, \nArgentina and Zimbabwe, OECD, Paris (2006) 5. World Water Council, World Water Vision: Making Water Everybody’s Business, \nEarthscan Publications, London (2000) 5"
https://openalex.org/W4312354588	https://file.techscience.com/ueditor/files/or/TSP_OR-29-6/TSP_OR_26236/TSP_OR_26236.pdf	English	null	LncRNA PRRT3-AS1 exerts oncogenic effects on nonsmall cell lung cancer by targeting microRNA-507/homeobox B5 axis	Oncology research	2,021	cc-by	8,764	LncRNA PRRT3-AS1 exerts oncogenic effects on nonsmall cell lung cancer by targeting microRNA-507/homeobox B5 axis RUI ZHOU#; JIANYANG XU#; LINGWEI WANG; JIANXIN LI Department of Thoracic Surgery, Afﬁliated Zhongshan Hospital of Dalian University, Dalian, 116011, China Key words: Targeted therapy, lncRNA, microRNA, NSCLC Abstract: Long noncoding RNAs (lncRNAs) act as key regulators controlling complex cellular behaviors in nonsmall cell lung cancer (NSCLC). We investigated the expression of lncRNA PRRT3 antisense RNA 1 (PRRT3-AS1) in paired samples of NSCLC and adjacent normal tissues from a patient cohort in our hospital using real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR) and found that it was signiﬁcantly higher in NSCLC tissue than in normal tissue, consistent with The Cancer Genome Atlas database. Furthermore, functional investigation revealed that lncRNA PRRT3-AS1 depletion inhibited NSCLC-cell proliferation, colony formation, invasion, and migration, whereas its overexpression exerted the opposite effects. Moreover, PRRT3-AS1 knockdown suppressed in vivo NSCLC growth. Investigation of downstream mechanisms using RNA immunoprecipitation and luciferase reporter assay revealed that lncRNA PRRT3-AS1 acted as a competing endogenous RNA by adsorbing microRNA-507 (miR-507) and enhanced the expression of its target gene, homeobox B5 (HOXB5), in NSCLC. Furthermore, miR-507 downregulation or HOXB5 upregulation eliminated the cancer-inhibiting effects of lncRNA PRRT3-AS1 depletion in NSCLC cells. To conclude, the lncRNA PRRT3-AS1/miR-507/HOXB5 pathway acts as a promoter of malignant characteristics in NSCLC, and this newly identiﬁed competing endogenous RNA pathway may be an effective diagnostic, prognostic, and therapeutic target in NSCLC. NSCLC pathogenesis is vital to identify effective targets for NSCLC diagnosis, prevention, and management. ONCOLOGY RESEARCH 2021 29(6): 411-423 ONCOLOGY RESEARCH 2021 29(6): 411-423 ONCOLOGY RESEARCH 2021 29(6): 411-423 ech T Press Science Introduction Long noncoding RNAs (lncRNAs) are a family of RNA molecules comprising 200 or more nucleotides with no protein-coding ability [5,6]. In the past, lncRNAs were considered “transcriptional noise” or “dark matter” [7–9]. However, recent studies have shown that lncRNAs are crucial for regulating physiological and pathological processes, such as chromatin remodeling, cell differentiation, and carcinogenesis [10–13], as well as cancer progression [14–16]. Recently, several lncRNAs were reported to be dysregulated in NSCLC [17–19], with upregulated lncRNAs playing an oncogenic role to promote NSCLC progression, whereas weakly expressed lncRNAs playing an anticancer role [20–22]. Among all cancers in humans, lung cancer has the highest incidence and mortality rate globally [1], with approximately 2.1 million lung cancer cases and over 1.8 million mortalities annually. Nonsmall cell lung cancer (NSCLC) is the predominant type of lung cancer, accounting for approximately 85% of all lung cancer cases [2]. Although advances in therapeutics have signiﬁcantly improved the prognosis of NSCLC patients in recent decades, the therapeutic efﬁcacy remains unsatisfactory, and many NSCLC patients still die [3]. The NSCLC pathogenesis and development is a complicated multistep process involving many risk factors and genes, and the underlying molecular mechanisms are poorly understood [4]. Therefore, determining the mechanisms underlying MicroRNAs (miRNAs) are a family of short (approximately 17–23 nucleotides) endogenous noncoding RNA molecules [23]. MiRNA dysregulation, widely reported in NSCLC, affects malignant properties [24–26]. During the past decade, the theory of competing endogenous RNAs (ceRNAs) has been proposed [27], notably showing that lncRNAs can sequester miRNAs, thereby increasing the downstream targets of miRNAs [28]. Therefore, Address correspondence to: Lingwei Wang, wanglw@tongji.edu.cn; Jianxin Li, Lijianxin1966@sina.com #Contributed equally Received: 25 August 2022; Accepted: 13 October 2022 RUI ZHOU et al. qPCR) to determine miR-507 expression by using the Mir-X miRNA quantitative reverse transcription PCR (qRT-PCR) TB Green Kit (TaKaRa). U6 served as the endogenous control for miR-507. To measure PRRT3-AS1 and HOXB5 expression, we used the PrimeScript RT Reagent Kit for cDNA synthesis and SYBR Premix Ex TaqTM II (TaKaRa) for qPCR. PRRT3-AS1 and HOXB5 levels were normalized to GAPDH the internal control, and reported as log-fold changes using the 2−ΔΔCq formula. characterizing the functions of novel NSCLC-associated lncRNAs or miRNAs is critical for developing promising therapeutic targets. p g Several lncRNAs are known to be key regulators of NSCLC malignancy. We investigated the expression and detailed functions of PRRT3 antisense RNA 1 (PRRT3-AS1) in NSCLC, in addition to the downstream mechanisms. Cell cultures Cell cultures We purchased human NSCLC cell lines A549, H460, SK- MES-1, and H1299 from the American Type Culture Collection (ATCC; Manassas, VA, USA). Bronchial epithelial cell growth medium (Lonza; Walkersville, MD, USA) was adopted for culturing a human nontumorigenic bronchial epithelial cell line BEAS-2B (ATCC). Cell lines A549 and H1299 were grown in RPMI-1640 medium (Gibco). SK-MES-1 and H460 cell lines were grown in F- 12K and MEM (Gibco), respectively. Each of the above culture media were supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (Gibco). All cells were grown in a humidiﬁed atmosphere of 5% CO2 at 37°C. Transwell cell migration and invasion assays Transwell cell migration and invasion assays We explored the invasive ability of NSCLC cells using Transwell inserts (BD Biosciences). We added 100 µL of Matrigel (BD Biosciences) into the inner sides of the Transwell inserts and performed polymerization by incubating them at 37°C for 2 h. Next, we loaded the upper chambers with 100 μL of FBS-free culture medium containing 4 × 104 of transfected cells, and plated 600 μL of culture medium supplemented with 10% FBS into the lower chambers. We allowed the NSCLC cells to penetrate the pores in the submembrane surface for 24 h, then ﬁxed them with 100% methanol, and stained them with 0.1% crystal violet. The cells that invaded the lower chambers were counted in ﬁve randomly chosen regions (in each well) under an inverted microscope. For the migration assay, we did not apply Matrigel, but the remaining steps were the same as those in the Matrigel invasion assay. Clinical tissue samples p We obtained NSCLC tissues from 53 patients in our hospital who had not undergone chemotherapy or radiotherapy before surgery. Immediately after tissue excision, all tissue samples were immersed and stored in liquid nitrogen until further analysis. This research was approved by the ethics committee of the Afﬁliated Zhongshan Hospital of Dalian University. In addition, all participating patients provided written informed consent. For colonies, a cell suspension was prepared, and 2 ml of the cell suspension containing 600 cells was added into each well of 6-well plates. After 14 days of cultivation, the newly formed colonies were ﬁxed with 100% methanol and stained with 0.1% crystal violet. The cells in the colonies were manually counted using an inverted microscope (x200 magniﬁcation). Doi: 10.32604/or.2022.026236 Doi: 10.32604/or.2022.026236 Doi: 10.32604/or.2022.026236 www.techscience.com/journal/or This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 412 RUI ZHOU et al. Materials and Methods Cell Counting Kit-8 and colony formation assays We inoculated transfected NSCLC cells into 96-well plates with a density of 2000 cells per well and incubated them at 37°C in a 5% CO2 atmosphere. After incubating with 10 µL of the cell counting kit-8 (CCK-8) solution (Beyotime), we estimated the proliferation rate by the measurement of optical density at 450 nm. We performed the assay every day until day 4 and plotted the results as a growth curve. THE ACTIONS OF PRRT3-AS1 IN NSCLC 413 mice were anesthetized by means of cervical dislocation, and the tumor xenografts were completely detached and weighed. Western blotting We performed total protein extraction and quantiﬁcation using the Pierce RIPA lysis buffer and the PierceTM bicinchoninic acid (BCA) Kit (both from Thermo Fisher Scientiﬁc; MA, USA), respectively. Protein samples were separated by 10% SDS-PAGE and blotted onto PVDF membranes. After blocking them with 5% defatted milk powder, the membranes were incubated at 4°C overnight with the primary antibodies against Hoxb5 (ab109375) or Gapdh (ab128915), followed by incubation with the HRP-conjugated anti-rabbit secondary antibody (ab6721; all from Abcam, USA). Finally, the bands of target proteins were developed by treatment with an enhanced chemiluminescence reagent (Beyotime). All animal experiments were approved by the Animal Care and Use Committee of Afﬁliated Zhongshan Hospital of Dalian University. The humane endpoints were tumor diameter >1.5 cm, tumor ulceration, abnormal feeding, weight loss, ascites and cachexia. Meanwhile, no anesthesia was applied in the animal experiments. Fluorescence in situ hybridization The subcellular location of PRRT3-AS1 in NSCLC cells was examined using the Fluorescent In Situ Hybridization (FISH) Kit (RiboBio Co., Ltd.,). In detail, after washing with phosphate-buffered saline, H460 and A549 cells were ﬁxed with 4% paraformaldehyde, followed by the hybridization treatment speciﬁcally targeting PRRT3-AS1 (RiboBio Co., Ltd.) at 37°C without light. On the next day, DNA staining was performed using Hoechst solution. The cells were, imaged under a confocal laser scanning microscope (Leica; Solms, Germany). Statistical analysis All experiments were independently repeated three times. Results are reported as the mean ± 1 SD. The between- group differences were compared using Student’s t-test or ANOVA. p < 0.05 was considered statistically signiﬁcant. Transfection We purchased small interfering RNAs (siRNAs) against PRRT3-AS1 (si-PRRT3-AS1) and homeobox B5 (HOXB5; si-HOXB5) and negative control (NC) siRNA (si-NC) from GenePharma (Shanghai, China). In addition, we used the miR-507 mimic and miR-507 inhibitor (RiboBio Co., Ltd. Guangzhou, China) to knock down and upregulate endogenous miR-507 levels, respectively. We used miRNA mimic (miR-NC) and inhibitor NCs as controls. The PRRT3-AS1 overexpression plasmid pcDNA3.1-PRRT3-AS1 (pc-PRRT3-AS1) and HOXB5 overexpression plasmid pcDNA3.1-HOXB5 (pc-HOXB5) were designed and synthesized by GenePharma. NSCLC cells were seeded into 6-well plates, and cell transfection was executed when cell growth reached 80% conﬂuence. Cellular transfection was done by applying Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) and the vectors to the wells. Xenograft tumor assay The short hairpin RNA (shRNA) targeting PRRT3-AS1 (sh-PRRT3-AS1), and NC shRNA (sh-NC) were designed and chemically generated by GenePharma. We inserted the shRNAs into a lentiviral vector to transfect the shRNAs into H460 cells. The PRRT3-AS1-knockdown H460 cells were selected by cultivating them with puromycin. Next, a total of 8 BALB/c nude mice (age, 4–6 weeks) were obtained from HFK Bioscience (Beijing, China) and randomly divided into sh-PRRT3-AS1 and sh-NC groups. A total of 2 × 106 PRRT3-AS1-knockdown H460 cells were subcutaneously inoculated into each mouse in the sh-PRRT3-AS1 group. We monitored the width and length of tumor xenografts from 1 week after cell inoculation and then every 5 days. Thirty-two days later, all Quantitative reverse transcription polymerase chain reaction We used the TRIzol reagent (Invitrogen) to extract total RNA. Total RNA was reverse-transcribed by using the Mir- X miRNA First-Strand Synthesis Kit (TaKaRa; Dalian, China). With the obtained cDNA as a template, we performed quantitative polymerase chain reaction (PCR; Bioinformatics prediction Bioinformatics prediction We searched the miRNA Target Prediction DataBase (miRDB) (http://mirdb.org/) database for PRRT3-AS1’s target miRNAs. To predict the downstream target of miR- 507, we used Targetscan (https://www.targetscan.org/vert_ 80/) applied to miRDB. PRRT3-AS1 promotes the aggressive behaviors of NSCLC cells PRRT3-AS1 expression in human cancers was ﬁrst investigated by analyzing The Cancer Genome Atlas dataset. We found that this gene was upregulated in nearly all human cancer types (Fig. 1A). Additionally, PRRT3-AS1 was the 47th overexpressed lncRNA in lung adenocarcinoma (LUAD; Fig. 1B). Also, relative to control samples, PRRT3-AS1 was obviously overexpressed in lung squamous cell carcinoma and LUAD samples in the TCGA database (Fig. 1C). Next, we collected paired samples of NSCLC and adjacent normal tissues from 53 patients and determined the lncRNA PRRT3-AS1 levels in both sets. The qRT-PCR analysis revealed that PRRT3-AS1 was strongly expressed in NSCLC tissues (Fig. 1D). In addition, all four NSCLC cell lines manifested relatively higher lncRNA PRRT3-AS1 levels than the BEAS-2B (non-tumor control) cells (Fig. 1E). RNA immunoprecipitation assay (B) PRRT3-AS1 ranks the 47th overexpressed lncRNA in LUAD. (C) PRRT3-AS1 level in LUAD and LUSC tissues from the TCGA database. (D) PRRT3-AS1 level in NSCLC tissues from our cohort (N = 3). (E) PRRT3-AS1 level in NSCLC cell lines (N = 3). p < 0.001. by anti-Ago2 antibody (Fig. 3G). Overall, these results indicate that lncRNA PRRT3-AS1 functioned as a miR-507 sponge. by anti-Ago2 antibody (Fig. 3G). Overall, these results indicate that lncRNA PRRT3-AS1 functioned as a miR-507 sponge. PRRT3-AS1 acts as a miR-507 sponge in NSCLC PRRT3-AS1 acts as a miR-507 sponge in NSCLC To determine downstream mechanisms of PRRT3-AS1 activation, we detected the subcellular localization of PRRT3-AS1 in NSCLC cells. First, to predict lncRNA PRRT3-AS1’s subcellular location, we used lncLocator (http:// www.csbio.sjtu.edu.cn/bioinf/lncLocator/), which indicated that it was mostly located in the cytoplasm (Fig. 3A). We veriﬁed the prediction by using FISH, conﬁrming that PRRT3-AS1 was indeed a cytoplasmic lncRNA (Fig. 3B). This observation suggested that lncRNA PRRT3-AS1 may act as a ceRNA of miRNAs. To follow up on this lead, we performed a bioinformatic analysis of the miRDB database revealing that lncRNA PRRT3-AS1 had potential binding sites for 20 miRNAs (Fig. 3C). A search of the literature identiﬁed six miRNAs, including miR-494-3p [29], miR-146b-5p [30], miR-146a-5p [31], miR-1827 [32], miR-507 [33], and miR-136-5p [34], which were shown to be signiﬁcantly correlated with NSCLC progression. Therefore, we selected these six miRNAs for further analysis. We investigated the regulatory effects of lncRNA PRRT3-AS1 on the six miRNAs and found that miR-507 was the one that changed the most (Fig. 3D). Luciferase reporter assay was adopted to conﬁrm the binding between lncRNA PRRT3-AS1 and miR-507 (Fig. 3E) and indicated that upregulated miR-507 expression decreased the luciferase activity of WT-PRRT3-AS1, while MUT-PRRT3- AS1 activity remained unaffected after miR-507 upregulation (Fig. 3F). Additionally, PRRT3-AS1 and miR-507 were abundant in the RNA immunoprecipitated HOXB5 is directly targeted by miR-507 in NSCLC cells and controlled by the lncRNA PRRT3-AS1/miR-507 axis Next, we determined the roles of miR-507 overexpression (Fig. 4A) in NSCLC cells. We found that miR-507 enhanced tumor-suppressing activities in NSCLC cells, affecting multiple aggressive behaviors (Figs. 4B–4E). Since HOXB5 (Fig. 5A) was predicted as a potential target of miR-507, we next analyzed whether miR-507 directly targeted HOXB5 in NSCLC cells. RNA immunoprecipitation assay In addition, si-PRRT3-AS1- induced inhibition of HOXB5 expression in H460 cells recovered after cotransfection with an miR-507 inhibitor, FIGURE 1. PRRT3-AS1 is upregulated in NSCLC. (A) PRRT3- AS1 level in all human cancer types from TCGA dataset (N = 3). (B) PRRT3-AS1 ranks the 47th overexpressed lncRNA in LUAD. (C) PRRT3-AS1 level in LUAD and LUSC tissues from the TCGA database. (D) PRRT3-AS1 level in NSCLC tissues from our cohort (N = 3). (E) PRRT3-AS1 level in NSCLC cell lines (N = 3). p < 0.001. PRRT3-AS1 acts as a miR-507 sponge in NSCLC by anti-Ago2 antibody (Fig 3G) Overall these results FIGURE 1. PRRT3-AS1 is upregulated in NSCLC. (A) PRRT3- AS1 level in all human cancer types from TCGA dataset (N = 3). (B) PRRT3-AS1 ranks the 47th overexpressed lncRNA in LUAD. (C) PRRT3-AS1 level in LUAD and LUSC tissues from the TCGA database. (D) PRRT3-AS1 level in NSCLC tissues from our cohort (N = 3). (E) PRRT3-AS1 level in NSCLC cell lines (N = 3). p < 0.001. FIGURE 1. PRRT3-AS1 is upregulated in NSCLC. (A) PRRT3- AS1 level in all human cancer types from TCGA dataset (N = 3). (B) PRRT3-AS1 ranks the 47th overexpressed lncRNA in LUAD. (C) PRRT3-AS1 level in LUAD and LUSC tissues from the TCGA database. (D) PRRT3-AS1 level in NSCLC tissues from our cohort (N = 3). (E) PRRT3-AS1 level in NSCLC cell lines (N = 3). p < 0.001. FIGURE 1. PRRT3-AS1 is upregulated in NSCLC. (A) PRRT3- AS1 level in all human cancer types from TCGA dataset (N = 3). (B) PRRT3-AS1 ranks the 47th overexpressed lncRNA in LUAD. (C) PRRT3-AS1 level in LUAD and LUSC tissues from the TCGA database. (D) PRRT3-AS1 level in NSCLC tissues from our cohort (N = 3). (E) PRRT3-AS1 level in NSCLC cell lines (N = 3). p < 0.001. FIGURE 1. PRRT3-AS1 is upregulated in NSCLC. (A) PRRT3- AS1 level in all human cancer types from TCGA dataset (N = 3). (B) PRRT3-AS1 ranks the 47th overexpressed lncRNA in LUAD. (C) PRRT3-AS1 level in LUAD and LUSC tissues from the TCGA database. (D) PRRT3-AS1 level in NSCLC tissues from our cohort (N = 3). (E) PRRT3-AS1 level in NSCLC cell lines (N = 3). p < 0.001. FIGURE 1. PRRT3-AS1 is upregulated in NSCLC. (A) PRRT3- AS1 level in all human cancer types from TCGA dataset (N = 3). RNA immunoprecipitation assay We performed a RNA immunoprecipitation (RIP) assay using the Magna RIP RNA-Binding Protein Immunoprecipitation kit (Merck-Millipore; Bedford, MA, USA). Brieﬂy, we harvested 80% conﬂuent NSCLC cells and lysed them using the RIP lysis buffer. Next, we performed Argonaute 2 (Ago2) immunoprecipitation by incubating the cell lysate with magnetic beads conjugated with anti-Ago2 antibody or normal immunoglobulin G (IgG) (control; Merck Millipore). Ago2 is an important element of RNA-induced silencing complex, and it can promote the degradation of target mRNAs via its catalytic activity in gene silencing processes induced by miRNAs. After overnight incubation at 4°C, we extracted the immunoprecipitated RNA and evaluated the abundance of lncRNA PRRT3-AS1, miR-507, and HOXB5 mRNA using qRT-PCR. We knocked down PRRT3-AS1 in H460 cells by transfecting them with si-PRRT3-AS1. To avoid the off- target effect, we used two siRNAs and veriﬁed that both manifested effective PRRT3-AS1 knockdown. Meanwhile, we applied pc-PRRT3-AS1 to overexpress PRRT3-AS1 in A549 cells (Fig. 2A), and then examined the biological effects of PRRT3-AS1 in the cells. PRRT3-AS1 knockdown clearly inhibited the proliferation of H460 cells, whereas PRRT3-AS1 overexpression promoted A549 cell proliferation (Fig. 2B). In addition, colony formation was clearly restricted in H460 cells following PRRT3-AS1 ablation, while PRRT3-AS1 upregulation caused the opposite tendency in A549 cells (Fig. 2C). Furthermore, PRRT3-AS1 knockdown decreased the motility (Figs. 2D and 2E) of H460 cells, whereas transfection with pc-PRRT3- AS1 resulted in the enhancement of A549 cell motility (Figs. 2D and 2E). Together, these results indicate that lncRNA PRRT3-AS1 performed oncogenic actions during NSCLC progression. Luciferase reporter assay The PRRT3-AS1 and HOXB5 wild-type (WT) sequences, both enclosing the predicted miR-507-binding sequences, were ampliﬁed by GenePharma, and fused to the psiCHECKTM-2 luciferase reporter vector (Promega), generating the WT-PRRT3-AS1 and WT-HOXB5 reporter vectors. Similarly, the mutant (MUT) luciferase reporter vectors MUT-PRRT3-AS1 and MUT-HOXB5 were built by inserting PRRT3-AS1 and HOXB5 mutant sequences into the psiCHECKTM-2 vector. Next, NSCLC cells were transfected with WT or matched MUT reporter vectors alongside the miR-507 mimic or miR-NC. On post- transfection day 2, we detected luciferase activity by applying the dual-luciferase reporter assay (Promega). RUI ZHOU et al. RUI ZHOU et al. RUI ZHOU et al. 414 PRRT3-AS1 acts as a miR-507 sponge in NSCLC To determine downstream mechanisms of PRRT3-AS1 activation, we detected the subcellular localization of PRRT3-AS1 in NSCLC cells. RNA immunoprecipitation assay First, to predict lncRNA PRRT3-AS1’s subcellular location, we used lncLocator (http:// www.csbio.sjtu.edu.cn/bioinf/lncLocator/), which indicated that it was mostly located in the cytoplasm (Fig. 3A). We veriﬁed the prediction by using FISH, conﬁrming that PRRT3-AS1 was indeed a cytoplasmic lncRNA (Fig. 3B). This observation suggested that lncRNA PRRT3-AS1 may act as a ceRNA of miRNAs. To follow up on this lead, we performed a bioinformatic analysis of the miRDB database revealing that lncRNA PRRT3-AS1 had potential binding sites for 20 miRNAs (Fig. 3C). A search of the literature identiﬁed six miRNAs, including miR-494-3p [29], miR-146b-5p [30], miR-146a-5p [31], miR-1827 [32], miR-507 [33], and miR-136-5p [34], which were shown to be signiﬁcantly correlated with NSCLC progression. Therefore, we selected these six miRNAs for further analysis. We investigated the regulatory effects of lncRNA PRRT3-AS1 on the six miRNAs and found that miR-507 was the one that changed the most (Fig. 3D). Luciferase reporter assay was adopted to conﬁrm the binding between lncRNA PRRT3-AS1 and miR-507 (Fig. 3E) and indicated that upregulated miR-507 expression decreased the luciferase activity of WT-PRRT3-AS1, while MUT-PRRT3- AS1 activity remained unaffected after miR-507 upregulation (Fig. 3F). Additionally, PRRT3-AS1 and miR-507 were abundant in the RNA immunoprecipitated by anti-Ago2 antibody (Fig. 3G). Overall, these results indicate that lncRNA PRRT3-AS1 functioned as a miR-507 sponge. HOXB5 is directly targeted by miR-507 in NSCLC cells and controlled by the lncRNA PRRT3-AS1/miR-507 axis Next, we determined the roles of miR-507 overexpression (Fig. 4A) in NSCLC cells. We found that miR-507 enhanced tumor-suppressing activities in NSCLC cells, affecting multiple aggressive behaviors (Figs. 4B–4E). Since HOXB5 (Fig. 5A) was predicted as a potential target of miR-507, we next analyzed whether miR-507 directly targeted HOXB5 in NSCLC cells. We found that miR-507 overexpression decreased the luciferase activity of the WT-HOXB5 vector in NSCLC cells, whereas no change was observed for the MUT-HOXB5 luciferase vector (Fig. 5B). In addition, our data showed a signiﬁcant decrease in HOXB5 levels (Figs. 5C and 5D) in miR-507-overexpressed-NSCLC cells. All these results together demonstrate that HOXB5 is a downstream target of miR-507 in NSCLC cells. Given that lncRNA PRRT3-AS1 sequestered miR-507 in NSCLC, we hypothesized that lncRNA PRRT3-AS1 indirectly modulated HOXB5 expression through sequestering miR-507. In fact, HOXB5 levels were decreased in H460 cells after lncRNA PRRT3-AS1 depletion, whereas HOXB5 levels were increased in pc-PRRT3-AS1-transfected A549 cells (Figs. 5E–5H). RNA immunoprecipitation assay We found that miR-507 overexpression decreased the luciferase activity of the WT-HOXB5 vector in NSCLC cells, whereas no change was observed for the MUT-HOXB5 luciferase vector (Fig. 5B). In addition, our data showed a signiﬁcant decrease in HOXB5 levels (Figs. 5C and 5D) in miR-507-overexpressed-NSCLC cells. All these results together demonstrate that HOXB5 is a downstream target of miR-507 in NSCLC cells. y We investigated the regulatory effects of lncRNA PRRT3-AS1 on the six miRNAs and found that miR-507 was the one that changed the most (Fig. 3D). Luciferase reporter assay was adopted to conﬁrm the binding between lncRNA PRRT3-AS1 and miR-507 (Fig. 3E) and indicated that upregulated miR-507 expression decreased the luciferase activity of WT-PRRT3-AS1, while MUT-PRRT3- AS1 activity remained unaffected after miR-507 upregulation (Fig. 3F). Additionally, PRRT3-AS1 and miR-507 were abundant in the RNA immunoprecipitated g Given that lncRNA PRRT3-AS1 sequestered miR-507 in NSCLC, we hypothesized that lncRNA PRRT3-AS1 indirectly modulated HOXB5 expression through sequestering miR-507. In fact, HOXB5 levels were decreased in H460 cells after lncRNA PRRT3-AS1 depletion, whereas HOXB5 levels were increased in pc-PRRT3-AS1-transfected A549 cells (Figs. 5E–5H). In addition, si-PRRT3-AS1- induced inhibition of HOXB5 expression in H460 cells recovered after cotransfection with an miR-507 inhibitor, THE ACTIONS OF PRRT3-AS1 IN NSCLC 415 ile treatment with a miR-507 mimic lead to recovery from increased HOXB5 expression induced by pc PRRT3 AS1 in products immunoprecipitated by the anti-Ago2 antib (Fig 5I) Together these results showed that lncR GURE 2. PRRT3-AS1 facilitates the malignant properties of NSCLC cells. (A) Inhibition efﬁciency of two siRNAs targeting PRRT3-AS 60 cells. The efﬁciency of pc-PRRT3-AS1 was also explored in A549 cells (N = 3). (B) Cell proliferation of PRRT3-AS1-knockdown H s and PRRT3-AS1-overexpressed A549 cells (N = 3). (C) The colony formation of si-PRRT3-AS1-transfected H460 cells and pc-PRR 1-transfected A549 cells (N = 3). (D) The migration of si-PRRT3-AS1-transfected H460 cells and pc-PRRT3-AS1-transfected A549 cells ) 100 magniﬁcation. (E) The invasion of si-PRRT3-AS1-transfected H460 cells and pc-PRRT3-AS1-transfected A549 cells (N = 3) 10 gniﬁcation. p < 0.01 and p < 0.001. FIGURE 2. PRRT3-AS1 facilitates the malignant properties of NSCLC cells. (A) Inhibition efﬁciency of two siRNAs targeting PRRT3-AS1 in H460 cells. The efﬁciency of pc-PRRT3-AS1 was also explored in A549 cells (N = 3). (B) Cell proliferation of PRRT3-AS1-knockdown H460 cells and PRRT3-AS1-overexpressed A549 cells (N = 3). RNA immunoprecipitation assay (C) The colony formation of si-PRRT3-AS1-transfected H460 cells and pc-PRRT3- AS1-transfected A549 cells (N = 3). (D) The migration of si-PRRT3-AS1-transfected H460 cells and pc-PRRT3-AS1-transfected A549 cells (N = 3) 100 magniﬁcation. (E) The invasion of si-PRRT3-AS1-transfected H460 cells and pc-PRRT3-AS1-transfected A549 cells (N = 3) 100 magniﬁcation. p < 0.01 and p < 0.001. while treatment with a miR-507 mimic lead to recovery from the increased HOXB5 expression induced by pc-PRRT3-AS1 in A549 cells (Figs. 5E–5H). The RIP assay validated that lncRNA PRRT3-AS1, miR-507, and HOXB5 were abundant in products immunoprecipitated by the anti-Ago2 antibody (Fig. 5I). Together, these results showed that lncRNA PRRT3-AS1 decoyed miR-507 in NSCLC cells and thus, positively modulate HOXB5 expression. RUI ZHOU et al. 416 FIGURE 3. PRRT3-AS1 is an miR-507 sponge in NSCLC cells. (A) Predicted location of PRRT3-AS1 according to lncLocator. (B) Subcellular distribution of PRRT3-AS1 in NSCLC cells. (C) miRDB online database to predict potential miRNA-targeting PRRT3-AS1. (D) Levels of the six candidates in H460 cells after PRRT3-AS1 depletion. Also, the levels of six candidates in A549 cells after PRRT3-AS1 upregulation was also investigated (N = 3). (E) The WT and MUT binding sites of miR-507 within PRRT3-AS1. (F) Luciferase reporter assay implemented in NSCLC cells after cotransfection with the miR-507 mimic or miR-NC and WT-PRRT3-AS1 or MUT-PRRT3-AS1 (N = 3). (G) RIP assay to analyze the interaction between PRRT3-AS1 and miR-507 (N = 3). p < 0.001. 416 RUI ZHOU et al. FIGURE 3. PRRT3-AS1 is an miR-507 sponge in NSCLC cells. (A) Predicted location of PRRT3-AS1 according to lncLocator. (B) Subcellular distribution of PRRT3-AS1 in NSCLC cells. (C) miRDB online database to predict potential miRNA-targeting PRRT3-AS1. (D) Levels of the six candidates in H460 cells after PRRT3-AS1 depletion. Also, the levels of six candidates in A549 cells after PRRT3-AS1 upregulation was also investigated (N = 3). (E) The WT and MUT binding sites of miR-507 within PRRT3-AS1. (F) Luciferase reporter assay implemented in NSCLC cells after cotransfection with the miR-507 mimic or miR-NC and WT-PRRT3-AS1 or MUT-PRRT3-AS1 (N = 3). (G) RIP assay to analyze the interaction between PRRT3-AS1 and miR-507 (N = 3). p < 0.001. FIGURE 3. PRRT3-AS1 is an miR-507 sponge in NSCLC cells. (A) Predicted location of PRRT3-AS1 according to lncLocator. (B) Subcellular distribution of PRRT3-AS1 in NSCLC cells. (C) miRDB online database to predict potential miRNA-targeting PRRT3-AS1. RNA immunoprecipitation assay (B, C) Cell proliferation and colony formation of miR-507-overxressed NSCLC cells (N = 3). (D, E) The motility of miR-507-overexpressed NSCLC cells (N = 3) 100 magniﬁcation. p < 0.001. FIGURE 4. Overexpressed miR-507 inhibits the malignant process of NSCLC cells. (A) The efﬁciency of miR-507 mimic transfection in NSCLC cells (N = 3). (B, C) Cell proliferation and colony formation of miR-507-overxressed NSCLC cells (N = 3). (D, E) The motility of miR-507-overexpressed NSCLC cells (N = 3) 100 magniﬁcation. p < 0.001. FIGURE 4. Overexpressed miR-507 inhibits the malignant process of NSCLC cells. (A) The efﬁciency of miR-507 mimic transfection in NSCLC cells (N = 3). (B, C) Cell proliferation and colony formation of miR-507-overxressed NSCLC cells (N = 3). (D, E) The motility of miR-507-overexpressed NSCLC cells (N = 3) 100 magniﬁcation. p < 0.001. PRRT3-AS1-overexpresed A549 cells underwent cotransfection with the miR-507 mimic or si-HOXB5. The cell proliferation inhibited by si-PRRT3-AS1 was reversed by miR-507 inhibition or HOXB5 upregulation (Fig. 6C). Conversely, miR-507 overexpression or HOXB5 depletion recovered the increased cell proliferation in A549 cells transfected with pc-PRRT3-AS1 (Fig. 6C). Remarkably, interference with lncRNA PRRT3-AS1 hampered colony formation of H460 cells; conversely, inhibition of miR-507 or upregulation of HOXB5 prevented this inﬂuence (Fig. 7A). In summary, treatment with the miR-507 mimic or si-HOXB5 was capable of recovering the colony formation of A549 cells that was promoted by pc-PRRT3- AS1 (Fig. 7B). while those of cells transfected with miR-507 mimic or si- HOXB5 decreased this effect (Fig. 7D). To conclude, lncRNA PRRT3-AS1 worsened the oncogenicity of NSCLC, at least, in part, by targeting the miR-507/HOXB5 regulatory axis. PRRT3-AS1 depletion impairs growth of NSCLC cells in vivo Xenograft tumor assay was performed to illustrate whether lncRNA PRRT3-AS1 affected NSCLC tumor growth in vivo. In contrast to tumors in the sh-NC group, the sh-PRRT3- AS1-transfected subcutaneous tumors grew signiﬁcantly slower (Figs. 8A and 8B). Also, tumor weight was much lower in the sh-PRRT3-AS1 group in comparison with that in the sh-NC group (Fig. 8C). Furthermore, lncRNA PRRT3-AS1 (Fig. 8D) and HOXB5 (Fig. 8E) levels were reduced in tumor xenografts obtained from the sh-PRRT3- AS1 group, and lncRNA PRRT3-AS1-depleted tumor xenografts manifested clearly higher miR-507 levels (Fig. 8F). Taken together, these results indicated that In addition, the introduction of the miR-507 inhibitor or pc-HOXB5 offset the impaired H460 cell migration and invasion (Fig. 7C) induced by PRRT3-AS1 downregulation. RNA immunoprecipitation assay (D) Levels of the six candidates in H460 cells after PRRT3-AS1 depletion. Also, the levels of six candidates in A549 cells after PRRT3-AS1 upregulation was also investigated (N = 3). (E) The WT and MUT binding sites of miR-507 within PRRT3-AS1. (F) Luciferase reporter assay implemented in NSCLC cells after cotransfection with the miR-507 mimic or miR-NC and WT-PRRT3-AS1 or MUT-PRRT3-AS1 (N = 3). (G) RIP assay to analyze the interaction between PRRT3-AS1 and miR-507 (N = 3). p < 0.001. LncRNA PRRT3-AS1 worsens the oncogenicity of NSCLC cells via targeting the miR-507/HOXB5 axis We designed rescue experiments to determine whether lncRNA PRRT3-AS1 performed tumor-promoting actions in NSCLC cells by controlling the miR-507/HOXB5 axis. First, we tested the transfection efﬁciency of the miR-507 LncRNA PRRT3-AS1 worsens the oncogenicity of NSCLC cells via targeting the miR-507/HOXB5 axis LncRNA PRRT3-AS1 worsens the oncogenicity of NSCLC cells via targeting the miR-507/HOXB5 axis LncRNA PRRT3-AS1 worsens the oncogenicity of NSCLC cells via targeting the miR-507/HOXB5 axis We designed rescue experiments to determine whether lncRNA PRRT3-AS1 performed tumor-promoting actions in NSCLC cells by controlling the miR-507/HOXB5 axis. First, we tested the transfection efﬁciency of the miR-507 inhibitor in NSCLC cells, verifying that miR-507 signiﬁcantly decreased in miR-507 inhibitor-transfected cells (Fig. 6A). Meanwhile, the efﬁciency of pc-HOXB5 in H460 cells and si-HOXB5 in A549 cells was explored (Fig. 6B). Next, si-PRRT3-AS1 was combined with the miR-507 inhibitor or the pc-HOXB5 to transfect H460 cells. inhibitor in NSCLC cells, verifying that miR-507 signiﬁcantly decreased in miR-507 inhibitor-transfected cells (Fig. 6A). Meanwhile, the efﬁciency of pc-HOXB5 in H460 cells and si-HOXB5 in A549 cells was explored (Fig. 6B). We designed rescue experiments to determine whether lncRNA PRRT3-AS1 performed tumor-promoting actions in NSCLC cells by controlling the miR-507/HOXB5 axis. First, we tested the transfection efﬁciency of the miR-507 Next, si-PRRT3-AS1 was combined with the miR-507 inhibitor or the pc-HOXB5 to transfect H460 cells. 417 THE ACTIONS OF PRRT3-AS1 IN NSCLC FIGURE 4. Overexpressed miR-507 inhibits the malignant process of NSCLC cells. (A) The efﬁciency of miR-507 mimic transfection in NSCLC cells (N = 3). (B, C) Cell proliferation and colony formation of miR-507-overxressed NSCLC cells (N = 3). (D, E) The motility of miR-507-overexpressed NSCLC cells (N = 3) 100 magniﬁcation. p < 0.001. FIGURE 4. Overexpressed miR-507 inhibits the malignant process of NSCLC cells. (A) The efﬁciency of miR-507 mimic transfection in NSCLC cells (N = 3). RNA immunoprecipitation assay (B) Luciferase reporter assay to verify binding interaction between miR-507 and HOXB5 3′-UTR in NSCLC cells (N = 3). (C and D) HOXB5 levels in miR-507-overxressed NSCLC cells (N = 3). (E–H) PRRT3-AS1-silenced H460 cells were cotransfected with miR-507 inhibitor or NC inhibitor, while miR-507 mimic or miR-NC was transfected into PRRT3-AS1-overexrssed A549 cells. After transfection, HOXB5 mRNA and protein levels were examined (N = 3). (I) RIP assay to certify the interaction among PRRT3-AS1, miR-507, and HOXB5 (N = 3). p < 0.001 and ##p < 0.001. downregulation of PRRT3-AS1 hampered NSCLC tumor growth in vivo. several studies have highlighted other important roles played by lncRNAs in regulating complex cellular behaviors in NSCLC [35–37]. Therefore, ﬁnding novel cancer- associated lncRNAs and exploring their regulatory actions in NSCLC is necessary to identify promising targets for NSCLC treatment. To date, 33,829 lncRNAs have been veriﬁed in the human genome according to the ENCODE RNA immunoprecipitation assay Moreover, the cell migration and invasion abilities of cells transfected with pc-PRRT3-AS1 signiﬁcantly increased, 418 RUI ZHOU et al. downregulation of PRRT3-AS1 hampered NSCLC tumor several studies have highlighted other important roles FIGURE 5. HOXB5 is under the control of PRRT3-AS1/miR-507 axis. (A) The WT and MUT binding sites between miR-507 and HOXB5 3′- UTR. (B) Luciferase reporter assay to verify binding interaction between miR-507 and HOXB5 3′-UTR in NSCLC cells (N = 3). (C and D) HOXB5 levels in miR-507-overxressed NSCLC cells (N = 3). (E–H) PRRT3-AS1-silenced H460 cells were cotransfected with miR-507 inhibitor or NC inhibitor, while miR-507 mimic or miR-NC was transfected into PRRT3-AS1-overexrssed A549 cells. After transfection, HOXB5 mRNA and protein levels were examined (N = 3). (I) RIP assay to certify the interaction among PRRT3-AS1, miR-507, and HOXB5 (N = 3). p < 0.001 and ##p < 0.001. FIGURE 5. HOXB5 is under the control of PRRT3-AS1/miR-507 axis. (A) The WT and MUT binding sites between miR-507 and HOXB5 3′- UTR. (B) Luciferase reporter assay to verify binding interaction between miR-507 and HOXB5 3′-UTR in NSCLC cells (N = 3). (C and D) HOXB5 levels in miR-507-overxressed NSCLC cells (N = 3). (E–H) PRRT3-AS1-silenced H460 cells were cotransfected with miR-507 inhibitor or NC inhibitor, while miR-507 mimic or miR-NC was transfected into PRRT3-AS1-overexrssed A549 cells. After transfection, HOXB5 mRNA and protein levels were examined (N = 3). (I) RIP assay to certify the interaction among PRRT3-AS1, miR-507, and HOXB5 (N = 3). p < 0.001 and ##p < 0.001. FIGURE 5. HOXB5 is under the control of PRRT3-AS1/miR-507 axis. (A) The WT and MUT binding sites between miR-507 and HOXB5 3′ UTR. (B) Luciferase reporter assay to verify binding interaction between miR-507 and HOXB5 3′-UTR in NSCLC cells (N = 3). (C and D HOXB5 levels in miR-507-overxressed NSCLC cells (N = 3). (E–H) PRRT3-AS1-silenced H460 cells were cotransfected with miR-507 inhibitor or NC inhibitor, while miR-507 mimic or miR-NC was transfected into PRRT3-AS1-overexrssed A549 cells. After transfection HOXB5 mRNA and protein levels were examined (N = 3). (I) RIP assay to certify the interaction among PRRT3-AS1, miR-507, and HOXB5 (N = 3). p < 0.001 and ##p < 0.001. FIGURE 5. HOXB5 is under the control of PRRT3-AS1/miR-507 axis. (A) The WT and MUT binding sites between miR-507 and HOXB5 3′- UTR. Discussion Many lncRNAs have been considered as crucial promoters or inhibitors of NSCLC progression [25], but more recently, 419 THE ACTIONS OF PRRT3-AS1 IN NSCLC database [38]; however, for most of these, the biological role and working mechanisms are still unknown. Our aim in this study was to investigate the speciﬁc functions of the lncRNA PRRT3-AS1 in NSCLC and the underlying downstream proteins and regulate gene expression at the transcriptional level [40]. In contrast, one of the critical roles of lncRNAs distributed in the cell cytoplasm is to partake, as ceRNAs, in the post-transcriptional regulation of gene expression [41]. FIGURE 6. PRRT3-AS1 regulates NSCLC cell proliferation via targeting miR-507/HOXB5. (A) Inhibition efﬁciency of miR-507 inhibitor in NSCLC cells (N = 3). (B) The efﬁciency of pc-HOXB5 in H460 cells and si-HOXB5 in A549 cells (N = 3). (C) PRRT3-AS1- silenced H460 cells were cotransfected with miR-507 inhibitor or pc-HOXB5 while miR- 507 mimic or si-HOXB5 was introduced into PRRT3-AS1- overexrssed A549 cells. After transfection, cell proliferation was determined via CCK-8 assay (N = 3). p < 0.01, p < 0.001, #p < 0.01, and ##p < 0.001. FIGURE 6. PRRT3-AS1 regulates NSCLC cell proliferation via targeting miR-507/HOXB5. (A) Inhibition efﬁciency of miR-507 inhibitor in NSCLC cells (N = 3). (B) The efﬁciency of pc-HOXB5 in H460 cells and si-HOXB5 in A549 cells (N = 3). (C) PRRT3-AS1- silenced H460 cells were cotransfected with miR-507 inhibitor or pc-HOXB5 while miR- 507 mimic or si-HOXB5 was introduced into PRRT3-AS1- overexrssed A549 cells. After transfection, cell proliferation was determined via CCK-8 assay (N = 3). p < 0.01, *p < 0.001, #p < 0.01, and ##p < 0.001. FIGURE 6. PRRT3-AS1 regulates NSCLC cell proliferation via targeting miR-507/HOXB5. (A) Inhibition efﬁciency of miR-507 inhibitor in NSCLC cells (N = 3). (B) The efﬁciency of pc-HOXB5 in H460 cells and si-HOXB5 in A549 cells (N = 3). (C) PRRT3-AS1- silenced H460 cells were cotransfected with miR-507 inhibitor or pc-HOXB5 while miR- 507 mimic or si-HOXB5 was introduced into PRRT3-AS1- overexrssed A549 cells. After transfection, cell proliferation was determined via CCK-8 assay (N = 3). p < 0.01, p < 0.001, #p < 0.01, and ##p < 0.001. database [38]; however, for most of these, the biological role and working mechanisms are still unknown. Our aim in this study was to investigate the speciﬁc functions of the lncRNA PRRT3-AS1 in NSCLC and the underlying downstream mechanisms. Discussion proteins and regulate gene expression at the transcriptional level [40]. In contrast, one of the critical roles of lncRNAs distributed in the cell cytoplasm is to partake, as ceRNAs, in the post-transcriptional regulation of gene expression [41]. According to the ceRNA theory, lncRNAs adsorb speciﬁc miRNAs, disabling the interaction of miRNAs with their target mRNAs [41]. Using two different methods, lncLocator prediction and ﬂuorescence in situ hybridization, we determined that the subcellular localization of lncRNA PRRT3-AS1 was primarily in the cytoplasm of NSCLC cells, offering a theoretical basis for lncRNA PRRT3-AS1 functions as a ceRNA. PRRT3-AS1 is overexpressed in prostate cancer, and the knockdown of PRRT3-AS1 suppresses cell viability, migration, and invasion, while promoting cell apoptosis [39]. However, the expression pattern, clinical implications, and functions of PRRT3-AS1 in NSCLC are still unknown. Our RNA expression analysis results revealed that PRRT3- AS1 was strongly expressed in NSCLC tissues, consistent with the information about PRRT3-AS1 in the TCGA database. Loss- and gain-of-function assays to determine comprehensively the biological functions of lncRNA PRRT3-AS1 in NSCLC progression showed that it played a pro-oncogenic role in NSCLC; PRRT3-AS1 knockdown evidently inhibited NSCLC cell growth and motility in vitro, whereas PRRT3-AS1 overexpression exerted the opposite effects. These results suggest that PRRT3-AS1 could be considered as a possible target for NSCLC diagnosis, prognosis, and management. Bioinformatic analysis of downstream targets of lncRNA PRRT3-AS1 predicted miR-507 as a possible lncRNA PRRT3- AS1-sequestered miRNA. Subsequent luciferase reporter and RIP assays conﬁrmed that lncRNA PRRT3-AS1 acted as a miR-507 sponge in NSCLC cells. Mechanistic investigation also identiﬁed HOXB5 as a direct target of miR-507 in NSCLC cells, showing that lncRNA PRRT3-AS1 indirectly modulated HOXB5 in NSCLC cells by sponging miR-507. Therefore, we hypothesize that PRRT3-AS1, miR-507, and HOXB5 coexist in the same RNA-induced silencing complex. Overall, lncRNA PRRT3-AS1 acts as a ceRNA to directly interact with miR-507 and indirectly regulate HOXB5 expression in NSCLC cells, thereby forming the novel lncRNA PRRT3-AS1/miR-507/HOXB5 ceRNA pathway. Next, we determined the mechanisms underlying the effect of lncRNA PRRT3-AS1 on malignant properties. The extensive research on the potential mechanisms of lncRNA function has mostly been dependent on the cellular localization of lncRNAs. Nuclear lncRNAs directly bind to 420 RUI ZHOU et al. We have known that miR-507 is differentially expressed family, is a direct target of miR-507 in NSCLC. HOXB5 IGURE 7. PRRT3-AS1 promotes NSCLC cell colony formation and motility through controlling miR-507/HOXB5. Discussion (A) The si-PRRT3-AS1 ansfected H460 cells underwent cotransfection with miR-507 inhibitor or pc-HOXB5. Cell colony formation was then examined (N = 3). (B miR-507 mimic or si-HOXB5 alongside pc-PRRT3-AS1 was cotransfected into A549 cells, and then subjected to colony formation etermination (N = 3). (C) The si-PRRT3-AS1 together with miR-507 inhibitor or pc-HOXB5 was transfected into H460 cells. Cel motility was detected (N = 3) 100 magniﬁcation. (D) A549 cells were transfected with pc-PRRT3-AS1 alongside miR-507 mimic or si HOXB5. Cell motility was also measured (N = 3) 100 magniﬁcation. p < 0.001, ##p < 0.001 and #p < 0.001. We have known that miR-507 is differentially expressed l h i l di NSCLC [42] H family, is a direct target of miR-507 in NSCLC. HOXB l i i l b l ti th t t GURE 7. PRRT3-AS1 promotes NSCLC cell colony formation and motility through controlling miR-507/HOXB5. (A) The si-PRRT3-AS ansfected H460 cells underwent cotransfection with miR-507 inhibitor or pc-HOXB5. Cell colony formation was then examined (N = 3). ( iR-507 mimic or si-HOXB5 alongside pc-PRRT3-AS1 was cotransfected into A549 cells, and then subjected to colony formatio termination (N = 3). (C) The si-PRRT3-AS1 together with miR-507 inhibitor or pc-HOXB5 was transfected into H460 cells. C otility was detected (N = 3) 100 magniﬁcation. (D) A549 cells were transfected with pc-PRRT3-AS1 alongside miR-507 mimic or s OXB5. Cell motility was also measured (N = 3) 100 magniﬁcation. p < 0.001, ##p < 0.001 and #p < 0.001. FIGURE 7. PRRT3-AS1 promotes NSCLC cell colony formation and motility through controlling miR-507/HOXB5. (A) The si-PRRT3-AS1- transfected H460 cells underwent cotransfection with miR-507 inhibitor or pc-HOXB5. Cell colony formation was then examined (N = 3). (B) miR-507 mimic or si-HOXB5 alongside pc-PRRT3-AS1 was cotransfected into A549 cells, and then subjected to colony formation determination (N = 3). (C) The si-PRRT3-AS1 together with miR-507 inhibitor or pc-HOXB5 was transfected into H460 cells. Cel motility was detected (N = 3) 100 magniﬁcation. (D) A549 cells were transfected with pc-PRRT3-AS1 alongside miR-507 mimic or si- HOXB5. Cell motility was also measured (N = 3) 100 magniﬁcation. *p < 0.001, ##p < 0.001 and #p < 0.001. FIGURE 7. PRRT3-AS1 promotes NSCLC cell colony formation and motility through controlling miR-507/HOXB5. (A) The si-PRRT3-AS1- transfected H460 cells underwent cotransfection with miR-507 inhibitor or pc-HOXB5. Cell colony formation was then examined (N = 3). Discussion (B) miR-507 mimic or si-HOXB5 alongside pc-PRRT3-AS1 was cotransfected into A549 cells, and then subjected to colony formation determination (N = 3). (C) The si-PRRT3-AS1 together with miR-507 inhibitor or pc-HOXB5 was transfected into H460 cells. Cell motility was detected (N = 3) 100 magniﬁcation. (D) A549 cells were transfected with pc-PRRT3-AS1 alongside miR-507 mimic or si- HOXB5. Cell motility was also measured (N = 3) 100 magniﬁcation. *p < 0.001, ##p < 0.001 and #p < 0.001. We have known that miR-507 is differentially expressed in several human cancers, including NSCLC [42]. Here, we report that miR-507 is downregulated in NSCLC. Our functional experiments further unveiled that miR-507 inhibited NSCLC cell proliferation, migration, and invasion. Mechanistically, HOXB5, a member of the HOX gene family, is a direct target of miR-507 in NSCLC. HOXB5 plays a carcinogenic role by regulating growth, metastasis, epithelial–mesenchymal transition, and tumorigenesis [43]. Our data supports the notion that HOXB5 is negatively regulated by miR-507 and positively modulated by lncRNA PRRT3-AS1 in NSCLC. In fact, our rescue experiments 421 THE ACTIONS OF PRRT3-AS1 IN NSCLC FIGURE 8. PRRT3-AS1 depletion decreases tumor growth in vivo. (A) Representative photographs of harvested tumors. Each group contained four nude mice. (B) Growth curve of tumor xenografts. (C) The comparison of the weight in the tumor xenografts. (D, E) PRRT3-AS1 and HOXB5 expression in tumor xenografts. (F) miR-507 level in tumor xenografts. p < 0.001. FIGURE 8. PRRT3-AS1 depletion decreases tumor growth in vivo. (A) Representative photographs of harvested tumors. Each group contained four nude mice. (B) Growth curve of tumor xenografts. (C) The comparison of the weight in the tumor xenografts. (D, E) PRRT3-AS1 and HOXB5 expression in tumor xenografts. (F) miR-507 level in tumor xenografts. p < 0.001. proved that miR-507 downregulation or HOXB5 upregulation eliminated the cancer-inhibiting effects of si-PRRT3-AS1. Together, our results strongly suggest that the PRRT3-AS1/ miR-507/HOXB5 pathway functions as a promoter of malignant characteristics in NSCLC. Funding Statement: The authors received no speciﬁc funding for this study. Funding Statement: The authors received no speciﬁc funding for this study. Conﬂicts of Interest: The authors declare that they have no conﬂicts of interest to report regarding the present study. Our study had two limitations. Firstly, we did not explore the effect of PRRT3-AS1 on metastasis in vivo. Secondly, other ways may be involved in the mechanisms underlying the oncogenic roles of PRRT3-AS1 in NSCLC. Discussion We will resolve the two limitations in the near future. References H., Tsai, T. H. et al. (2020). miR-146b-5p enhances the sensitivity of NSCLC to EGFR tyrosine kinase inhibitors by regulating the IRAK1/ NF-kappaB pathway. Molecular Therapy. Nucleic Acids, 22, 471–483. DOI 10.1016/j.omtn.2020.09.015. 16. Farzaneh, M., Najaﬁ, S., Dari, M. A. G., Sheykhi-Sabzehpoush, M., Dayer, D. et al. (2022). Functional roles of long noncoding RNA MALAT1 in gynecologic cancers. Clinical & Translational Oncology. DOI 10.1007/s12094-022-02914-8. 31. Liu, X., Liu, B., Li, R., Wang, F., Wang, N. et al. (2020). miR- 146a-5p plays an oncogenic role in NSCLC via suppression of TRAF6. Frontiers in Cell and Developmental Biology, 2(8), 847. DOI 10.3389/fcell.2020.00847. 17. Cao, Z., Oyang, L., Luo, X., Xia, L., Hu, J. et al. (2022). The roles of long non-coding RNAs in lung cancer. Journal of Cancer, 13(1), 174–183. DOI 10.7150/jca.65031. 32. Fan, G., Xu, P., Tu, P. (2020). MiR-1827 functions as a tumor suppressor in lung adenocarcinoma by targeting MYC and FAM83F. Journal of Cellular Biochemistry, 121(2), 1675–1689. DOI 10.1002/jcb.29402. 18. Zeng, L., Luo, X., Zhang, Z., Wang, Z., Qian, J. (2022). Long non- coding FAM201A accelerates the tumorigenesis and progression of colorectal cancer through miR-3163/MACC1 axis. Translational Oncology, 25(2), 101490. DOI 10.1016/j. tranon.2022.101490. 33. Lv, H. L., Li, Z. L., Song, Y. B. (2019). MicroRNA-507 represses the malignant behaviors of non-small cell lung cancer via targeting zinc ﬁnger E-box binding homeobox 2. European Review for Medical and Pharmacological Sciences, 23(22), 9955–9964. DOI 10.26355/eurrev_201911_19562. 19. Hao, X., Zhang, M., Gu, M., Wang, Z., Zhou, S. et al. (2022). Long non-coding RNA BZRAP1-AS1 functions in malignancy and prognosis for non-small-cell lung cancer. PeerJ, 10(5), e13871. DOI 10.7717/peerj.13871. 34. Zhang, W., Song, C., Ren, X. (2021). Circ_0003998 regulates the progression and docetaxel sensitivity of DTX-resistant non-small cell lung cancer cells by the miR-136-5p/CORO1C axis. Technology in Cancer Research & Treatment, 20, 1–13. DOI 10.1177/1533033821990040. 20. Wei, M. M., Zhou, G. B. (2016). Long non-coding RNAs and their roles in non-small-cell lung cancer. Genomics, Proteomics & Bioinformatics, 14(5), 280–288. DOI 10.1016/j.gpb.2016. 03.007. 21. Kang, S., Ou, C., Yan, A., Zhu, K., Xue, R. et al. (2022). Long noncoding RNA SNHG5 induces the NF-κB pathway by regulating miR-181c-5p/CBX4 axis to promote the progression of non-small cell lung cancer. Archivos de Bronconeumologia. DOI 10.1016/j.arbres.2022.07.001. 35. Cheng, Z., Lu, C., Wang, H., Wang, N., Cui, S. et al. (2022). References Durendez-Saez, E., Torres-Martinez, S., Calabuig-Farinas, S., Meri-Abad, M., Ferrero-Gimeno, M. et al. (2021). Exosomal microRNAs in non-small cell lung cancer. Translational Cancer Research, 10(6), 3128–3139. DOI 10.21037/tcr-20-2815. 9. Policarpo, R., Sierksma, A., de Strooper, B., d’Ydewalle, C. (2021). From junk to function: LncRNAs in CNS health and disease. Frontiers in Molecular Neuroscience, 19(14), 714768. DOI 10.3389/fnmol.2021.714768. 25. Naeli, P., Youseﬁ, F., Ghasemi, Y., Savardashtaki, A., Mirzaei, H. (2020). The role of MicroRNAs in lung cancer: Implications for diagnosis and therapy. Current molecular medicine, 20(2), 90– 101. DOI 10.2174/1566524019666191001113511. 10. Zhang, H., Wang, Y., Liu, X., Li, Y. (2020). Progress of long noncoding RNAs in anti-tumor resistance. Pathology, Research and Practice, 216(11), 153215. DOI 10.1016/j.prp.2020.153215. 26. Cui, F., Luo, P., Wu, R., Meng, J. (2022). miR-381 inhibits proliferation and invasion of non-small-cell cancer cells by targeting USP39. Disease Markers, 2022(4), 2195393. DOI 10.1155/2022/2195393. 11. Xiao, Y., Xiao, T., Ou, W., Wu, Z., Wu, J. et al. (2020). LncRNA SNHG16 as a potential biomarker and therapeutic target in human cancers. Biomarker Research, 10(8), 41. DOI 10.1186/ s40364-020-00221-4. 27. Lin, W., Liu, H., Tang, Y., Wei, Y., Wei, W. et al. (2020). The development and controversy of competitive endogenous RNA hypothesis in non-coding genes. Molecular and cellular biochemistry, 476(1), 109–123. DOI 10.1007/s11010- 020-03889-2. 12. Liu, L., Zhang, Y., Lu, J. (2020). The roles of long noncoding RNAs in breast cancer metastasis. Cell Death & Disease, 11(9), 749. DOI 10.1038/s41419-020-02954-4. 13. Wang, T., Gan, X. (2022). Emerging roles of HMGB1-related lncRNA: From molecular biology to clinical application. American Journal of Physiology. Cell Physiology, 323(4), 1149– 1160. DOI 10.1152/ajpcell.00296.2022. 28. Qi, X., Zhang, D. H., Wu, N., Xiao, J. H., Wang, X. et al. (2015). ceRNA in cancer: Possible functions and clinical implications. Journal of Medical Genetics, 52(10), 710–718. DOI 10.1136/ jmedgenet-2015-103334. 14. Tan, H., Zhang, S., Zhang, J., Zhu, L., Chen, Y. et al. (2020). Long non-coding RNAs in gastric cancer: New emerging biological functions and therapeutic implications. Theranostics, 10(19), 8880–8902. DOI 10.7150/thno.47548. 29. Faversani, A., Amatori, S., Augello, C., Colombo, F., Porretti, L. et al. (2017). miR-494-3p is a novel tumor driver of lung carcinogenesis. Oncotarget, 8(5), 7231–7247. DOI 10.18632/ oncotarget.13933. 15. Sun, D. E., Ye, S. Y. (2020). Emerging roles of long noncoding RNA regulator of reprogramming in cancer treatment. Cancer Management and Research, 21(12), 6103–6112. DOI 10.2147/ CMAR.S253042. 30. Liu, Y. N., Tsai, M. F., Wu, S. G., Chang, T. References 1. Siegel, R. L., Miller, K. D., Jemal, A. (2020). Cancer statistics, 2020. CA: A Cancer Journal for Clinicians, 70(1), 7–30. DOI 10.3322/caac.21590. To conclude, PRRT3-AS1, which is overexpressed in NSCLC, upregulates HOXB5 expression by sequestering miR-507 and aggravates the oncogenicity of NSCLC. Our ﬁndings advance the understanding of NSCLC pathogenesis and strongly suggest the PRRT3-AS1/miR-507/HOXB5 pathway as a promising novel therapeutic target. 2. Yoon, S. M., Shaikh, T., Hallman, M. (2017). Therapeutic management options for stage III non-small cell lung cancer. World Journal of Clinical Oncology, 8(1), 1–20. DOI 10.5306/ wjco.v8.i1.1. 3. Xu-Welliver, M., Carbone, D. P. (2017). Blood-based biomarkers in lung cancer: prognosis and treatment decisions. Translational Lung Cancer Research, 6(6), 708–712. DOI 10.21037/ tlcr.2017.09.08. Availability of Data and Materials: Data available with the communication author and can be provided upon reasonable request. 4. Sun, X., Li, K., Zhao, R., Sun, Y., Xu, J. et al. (2021). Lung cancer pathogenesis and poor response to therapy were dependent on driver oncogenic mutations. Life Sciences, 15(265), 118797. DOI 10.1016/j.lfs.2020.118797. Author Contribution: The authors conﬁrm contribution to the paper as follows: study conception and design: Rui Zhou, Lingwei Wang and Jianxin Li; data collection: Rui Zhou, Jianyang Xu; analysis and interpretation of results: Rui Zhou, Jianyang Xu, Lingwei Wang, and Jianxin Li; draft manuscript preparation: Rui Zhou, and Lingwei Wang. All authors reviewed the results and approved the ﬁnal version of the manuscript. 5. Cech, T. R., Steitz, J. A. (2014). The noncoding RNA revolution- trashing old rules to forge new ones. Cell, 157(1), 77–94. DOI 10.1016/j.cell.2014.03.008. 6. Zeng, X., Xiao, J., Bai, X., Liu, Y., Zhang, M. et al. (2022). Research progress on the circRNA/lncRNA-miRNA-mRNA axis in gastric cancer. Pathology, Research and Practice, 19(238), 154030. DOI 10.1016/j.prp.2022.154030. Ethic Approval: This research was approved by the ethics committee of the Afﬁliated Zhongshan Hospital of Dalian University. All animal experiments were approved by the Animal Care and Use Committee of Afﬁliated Zhongshan Hospital of Dalian University. 7. Kung, J. T., Colognori, D., Lee, J. T. (2013). Long noncoding RNAs: Past, present, and future. Genetics, 193(3), 651–669. DOI 10.1534/genetics.112.146704. 8. Kapranov, P., St Laurent, G., Raz, T., Ozsolak, F., Reynolds, C. P. et al. (2010). The majority of total nuclear-encoded 422 RUI ZHOU et al. Developmental Cell, 11(4), 441–450. DOI 10.1016/j. devcel.2006.09.009. non-ribosomal RNA in a human cell is ‘dark matter’ un-annotated RNA. BMC Biology, 21(8), 149. DOI 10.1186/ 1741-7007-8-149. 24. References Long noncoding RNA LHFPL3-AS2 suppresses metastasis of non- small cell lung cancer by interacting with SFPQ to regulate TXNIP expression. Cancer Letters, 10(531), 1–13. DOI 10.1016/ j.canlet.2022.01.031. 22. Xu, L. M., Yuan, Y. J., Yu, H., Wang, S., Wang, P. (2022). LINC00665 knockdown confers sensitivity in irradiated non- small cell lung cancer cells through the miR-582-5p/UCHL3/ AhR axis. Journal of Translational Medicine, 20(1), 350. DOI 10.1186/s12967-022-03516-2. 36. Li, L., Li, Z., Qu, J., Wei, X., Suo, F. et al. (2022). Novel long non- coding RNA CYB561-5 promotes aerobic glycolysis and tumorigenesis by interacting with basigin in non-small cell lung cancer. Journal of Cellular and Molecular Medicine, 26(5), 1402–1412. DOI 10.1111/jcmm.17057. 37. Wan, L., Cheng, Z., Sun, Q., Jiang, K. (2022). LncRNA HOXC- AS3 increases non-small cell lung cancer cell migration and 23. Kloosterman, W. P., Plasterk, R. H. (2006). The diverse functions of microRNAs in animal development and disease. THE ACTIONS OF PRRT3-AS1 IN NSCLC 423 invasion by sponging premature miR-96. Expert Review of Respiratory Medicine, 16(5), 587–593. DOI 10.1080/ 17476348.2022.2030223. 41. Wang, L., Cho, K. B., Li, Y., Tao, G., Xie, Z. et al. (2019). Long noncoding RNA (lncRNA)-mediated competing endogenous RNA networks provide novel potential biomarkers and therapeutic targets for colorectal cancer. International Journal of Molecular Sciences, 20(22), 5758. DOI 10.3390/ijms20225758. 38. Bu, D., Yu, K., Sun, S., Xie, C., Skogerbo, G. et al. (2012). NONCODE v3.0: Integrative annotation of long noncoding RNAs. Nucleic Acids Research, 40(D1), 210–215. DOI 10.1093/ nar/gkr1175. 42. Feng, X., Yang, S. (2020). Long non-coding RNA LINC00243 promotes proliferation and glycolysis in non-small cell lung cancer cells by positively regulating PDK4 through sponging miR-507. Molecular and Cellular Biochemistry, 463(1–2), 127– 136. DOI 10.1007/s11010-019-03635-3. 39. Fan, L., Li, H., Wang, W. (2020). Long non-coding RNA PRRT3- AS1 silencing inhibits prostate cancer cell proliferation and promotes apoptosis and autophagy. Experimental Physiology, 105(5), 793–808. DOI 10.1113/EP088011. 43. Zhang, B., Li, N., Zhang, H. (2018). Knockdown of homeobox B5 (HOXB5) inhibits cell proliferation, migration, and invasion in non-small cell lung cancer cells through inactivation of the Wnt/beta-catenin pathway. Oncology Research, 26(1), 37–44. DOI 10.3727/096504017X14900530835262. 40. Zhang, X. Z., Liu, H., Chen, S. R. (2020). Mechanisms of long non-coding RNAs in cancers and their dynamic regulations. Cancers, 12(5), 1245. DOI 10.3390/cancers12051245.
https://openalex.org/W2758204185	http://dspace.vutbr.cz/bitstream/11012/70123/1/molecules2201626v2.pdf	English	null	Anticarcinogenic Effect of Spices Due to Phenolic and Flavonoid Compounds—In Vitro Evaluation on Prostate Cells	Molecules/Molecules online/Molecules annual	2,017	cc-by	10,166	Article Zuzana Lackova 1,2, Hana Buchtelova 1,2, Zaneta Buchtova 1, Borivoj Klejdus 1, Zbynek Heger 1,2, Martin Brtnicky 2,3, Jindrich Kynicky 2,3, Ondrej Zitka 1,2 and Vojtech Adam 1,2,* 1 Department of Chemistry and Biochemistry, Mendel University in Brno, Zemedelska 1, Brno CZ-613 00, Czech Republic; lackova14@seznam.cz (Z.L.); hanabuchtelova@seznam.cz (H.B.); ZanetaBurianova@email.cz (Z.B.); klejdusb@seznam.cz (B.K.); heger@mendelu.cz (Z.H.); zitkao@seznam.cz (O.Z.) 1 Department of Chemistry and Biochemistry, Mendel University in Brno, Zemedelska 1, Brno CZ-613 00, Czech Republic; lackova14@seznam.cz (Z.L.); hanabuchtelova@seznam.cz (H.B.); ZanetaBurianova@email.cz (Z.B.); klejdusb@seznam.cz (B.K.); heger@mendelu.cz (Z.H.); zitkao@seznam.cz (O.Z.) 2 1 Department of Chemistry and Biochemistry, Mendel University in Brno, Zemedelska 1, Brno CZ-613 00, Czech Republic; lackova14@seznam.cz (Z.L.); hanabuchtelova@seznam.cz (H.B.); ZanetaBurianova@email.cz (Z.B.); klejdusb@seznam.cz (B.K.); heger@mendelu.cz (Z.H.); zitkao@seznam.cz (O.Z.) 2 Central European Institute of Technology, Brno University of Technology, Technicka 3058/10, Brno CZ-616 00, Czech Republic; martin.brtnicky@mendelu.cz (M.B.); jindrich.kynicky@mendelu.cz (J.K.) 3 Department of Geology and Pedology, Mendel University in Brno, Zemedelska 1, Brno CZ-613 00, C h R bli Brno CZ-616 00, Czech Republic; martin.brtnicky@mendelu.cz (M.B.); jindrich.kynicky@mendelu.cz (J.K.) 3 Department of Geology and Pedology, Mendel University in Brno, Zemedelska 1, Brno CZ-613 00, Czech Republic 3 Department of Geology and Pedology, Mendel University in Brno, Zemedelska 1, Brno CZ-613 00, Czech Republic p * Correspondence: vojtech.adam@mendelu.cz; Tel.: +420-5-4513-3350; Fax: +420-5-4521-2044 Received: 2 September 2017; Accepted: 24 September 2017; Published: 28 September 2017 ceived: 2 September 2017; Accepted: 24 September 2017; Published: 28 September 2017 Abstract: This study shows the effects of spices, and their phenolic and ﬂavonoid compounds, on prostate cell lines (PNT1A, 22RV1 and PC3). The results of an MTT assay on extracts from eight spices revealed the strongest inhibitory effects were from black pepper and caraway seed extracts. The strongest inhibitory effect on prostatic cells was observed after the application of extracts of spices in concentration of 12.5 mg·mL−1. An LC/MS analysis identiﬁed that the most abundant phenolic and ﬂavonoid compounds in black pepper are 3,4-dihydroxybenzaldehyde and naringenin chalcone, while the most abundant phenolic and ﬂavonoid compounds in caraway seeds are neochlorogenic acid and apigenin. Using an MTT assay for the phenolic and ﬂavonoid compounds from spices, we identiﬁed the IC50 value of ~1 mmol·L−1 PNT1A. The scratch test demonstrated that the most potent inhibitory effect on PNT1A, 22RV1 and PC3 cells is from the naringenin chalcone contained in black pepper. Article From the spectrum of compounds assessed, the naringenin chalcone contained in black pepper was identiﬁed as the most potent inhibitor of the growth of prostate cells. Keywords: apigenin; clonogenic assay; 3,4-dihydroxybenzaldehyde; MTT assay; naringenin chalcone; neochlorogenic acid; prostate cancer; scratch test; spices molecules molecules molecules Molecules 2017, 22, 1626; doi:10.3390/molecules22101626 molecules molecules 1. Introduction At present, the inﬂuence of dietary habits and food quality, in terms of substance content, on the development of cancer is being increasingly studied [1]. There are a number of studies pointing to the positive effects of fruit and vegetables, prevalently due to compounds in their contents, such as phenols, ﬂavonoids, vitamins and mineral substances [2,3]. On the other hand, spices have been involved in the human diet for plenty of years, and used as essential additional ingredients for much cooking, and as seasoning. Could they be considered to be one of the major sources of anticarcinogenic compounds because they contain antioxidants and other biologically active molecules? Thus, it is not surprising that numerous species have been studied in the context of their effects on human health. Antibacterial effects are demonstrated for sweet peppers, peppers and caraway seeds [4–6]. Marjoram, cinnamon and caraway seeds are also reported to have anti-inﬂammatory effects [6–8]. Cinnamon and caraway seeds are shown to have anticarcinogenic effects [6,7]. Thyme, pepper and oregano are used Molecules 2017, 22, 1626; doi:10.3390/molecules22101626 www.mdpi.com/journal/molecules Molecules 2017, 22, 1626 2 of 13 as antifungals [5,9,10]. The anticarcinogenic effects of spices are attributed to them containing phenolic compounds [11]. ungals [5,9,10]. The anticarcinogenic effects of spices are attributed to them containing phenolic unds [11]. Therefore, we focused on the effects of the most prevalent polyphenols on prostate cells, with regards to the fact that the urogenital tract is the most exposed tissue upon which the effect of the chemical substances present in ﬂuids passing through it should have the greatest impact [12]. To date, a number of studies show the anticarcinogenic effects of piperine [13], capsaicin [14] and curcumin [15–17] on prostate cancer cells. Of the eight kinds of spice tested in this work, studies on the anticarcinogenic effect on prostate cancer cells have been performed on black pepper only [13,18]. For oregano, marjoram, anise, thyme, sweet pepper, cinnamon and caraway seeds, no studies have yet been published on their effects on prostate cells [7,11,19–24]. For our experiments, we directed our attention to phenolic (neochlorogenic acid and 3,4-dihydroxybenzaldehyde) and ﬂavonoid (apigenin and naringenin chalcone) compounds. Apigenin is a ﬂavonoid compound that is abundantly present in fruits and vegetables. 1. Introduction Apigenin reduces low density lipoprotein and cholesterol levels; stimulates PPAR-γ; augments the endogenous antioxidants; regulates the death-signaling of reactive oxygen species [25]; regulates inﬂammatory mediators, including IL-1β and TNF-α [26]; inhibits tumor growth and angiogenesis induced by different cancer cells; and has antiproliferative and antitumor properties in the colon, pancreas and prostate cancer cells [27]. In addition, it was revealed that apigenin can disrupt cancer cell motility by suppressing the focal adhesion kinase/Src signaling [28], which is a key step in the development of tumors and, ultimately, metastasis. Naringenin chalcone is ﬂavonoid compound and its inhibitory effects are demonstrated in U87MG cells [29]. Neochlorogenic acid is a phenolic acid, which exhibits antioxidant and chemopreventive activity in colon and breast cancer, and in U937 leukemia cells; it protects cells from oxidative stress by scavenging reactive oxidative stress (ROS), and suppressing the proliferation of breast and colon cancer cells [30]. The study [31] showed strong inhibition of growth on a breast cancer cell line (MDA-MB-435) and low toxic effect on a normal breast cell line (MCF-10A). The phenolic compound 3,4-dihydroxybenzaldehyde has antioxidant and anti-inﬂammatory effects: it decreases the proliferation of human breast cancer, and induces apoptosis with inhibition of casein kinase II activity in leukemia cells [32]. However, the potential anticancer mechanisms of phenolic (neochlorogenic acid and 3,4-dihydroxybenzaldehyde) and ﬂavonoid (apigenin and naringenin chalcone) compounds have not been elucidated, and the effects of prostate cancer cells have not been tested so far. The aim of this experiment was to determine the effect of eight selected spice species on three prostate cell lines. The proﬁle representation of the phenolic and ﬂavonoid compounds of the selected spices was performed by liquid chromatography/mass spectrometry (LC/MS). The most representative phenolic and ﬂavonoid compounds were used for the rest of the ﬁnal evaluation using an MTT assay, a scratch test and a clonogenic assay. 2. Results and Discussion 2.1. Cell-Line Proliferative Activity Testing (MTT Assay) for Extract from Eight Spices 2.1. Cell-Line Proliferative Activity Testing (MTT Assay) for Extract from Eight Spices An MTT assay was used to evaluate the cells’ metabolic activity and indicate the cytotoxicity of the tested compounds. Figure 1 shows the results of the effects of the extracts from eight kinds of spice. The results indicate that the best and most stable inhibitory effects were due to treatment with extracts from caraway seeds and black pepper. The strongest inhibitory effect on prostatic cells was observed when we applied a concentration of 12.5 mg·mL−1 to all cell lines used. If we take a look at the results in greater detail, those of the extracts from anise, thyme, marjoram, oregano and sweet pepper were very similar. A similar effect to those of other extracts for all cell lines used from an extract of cinnamon is shown in Figure 1A. However, difference was observed in the extract from cinnamon (Figure 1B,C), where it was found that the extract from cinnamon was supportive of cell growth here. The inhibitory effect of cinnamon in 29 types of human cancer cells was previously conﬁrmed in vitro, where the vast majority of the antitumorigenic effects of cinnamon 3 of 13 Molecules 2017, 22, 1626 extracts could be attributed to cinnamaldehydes [33], the main component of the essential oil. Further, it was described that biologically active substances in cinnamon also cause cell inhibition in prostate cancer cells [33,34]. Other studies describe the inhibitory effect of the selected compound or oil from cinnamon [35–38], but our study evaluated the effect of cinnamon extract, not the selected components contained in cinnamon, which may explain the deviations from the available literature. The reason for the different effect of cinnamon on prostate cell lines may be that in the aforementioned literature, only the selected cinnamon content has been evaluated, whereas in our study we used a whole-spice extract containing a wide range of substances. component of the essential oil. Further, it was described that biologically active substances in cinnamon also cause cell inhibition in prostate cancer cells [33,34]. Other studies describe the inhibitory effect of the selected compound or oil from cinnamon [35–38], but our study evaluated the effect of cinnamon extract, not the selected components contained in cinnamon, which may explain the deviations from the available literature. 2.2. Profiling of the Extracts of the Tested Spices Using LC/MS 2.2. Proﬁling of the Extracts of the Tested Spices Using LC/MS Subsequently, an LC/MS analysis was performed to determine which phenolic and flavonoid compounds are contained in selected spices. Determination of the presence and content of phenolic and flavonoid compounds was done using high-performance liquid chromatography with mass detection. The results have been recalculated per 1 g of spice. Subsequently, an LC/MS analysis was performed to determine which phenolic and ﬂavonoid compounds are contained in selected spices. Determination of the presence and content of phenolic and ﬂavonoid compounds was done using high-performance liquid chromatography with mass detection. The results have been recalculated per 1 g of spice. p g p For the extract from caraway seeds (Figure 2A), the most commonly found phenolic compound was neochlorogenic acid, for which the concentration reached 110 ± 5 μg·g−1, with extraction using 80% methanol. The other phenolic compounds determined did not exceed a concentration of 10 μg·g−1. Of the flavonoid compounds observed, the highest occurrences recorded in the extract from caraway seeds were for apigenin and naringenin chalcone. The highest concentration of naringenin chalcone had a value of 13 ± 1 μg·g−1, with extraction using 100% methanol. A very interesting concentration was found for apigenin when using extraction with 80% methanol, which was 16 ± 1 μg·g−1, compared to other types of extraction used. The other flavonoid compounds revealed did not have a concentration greater than 2 μg·g−1. Figure 2B describes the preparation of samples for the LC/MS analysis, and more experimental details can be found in Section 3.5. For the extract from caraway seeds (Figure 2A), the most commonly found phenolic compound was neochlorogenic acid, for which the concentration reached 110 ± 5 µg·g−1, with extraction using 80% methanol. The other phenolic compounds determined did not exceed a concentration of 10 µg·g−1. Of the ﬂavonoid compounds observed, the highest occurrences recorded in the extract from caraway seeds were for apigenin and naringenin chalcone. The highest concentration of naringenin chalcone had a value of 13 ± 1 µg·g−1, with extraction using 100% methanol. A very interesting concentration was found for apigenin when using extraction with 80% methanol, which was 16 ± 1 µg·g−1, compared to other types of extraction used. The other ﬂavonoid compounds revealed did not have a concentration greater than 2 µg·g−1. Figure 2B describes the preparation of samples for the LC/MS analysis, and more experimental details can be found in Section 3.5. 2.1. Cell-Line Proliferative Activity Testing (MTT Assay) for Extract from Eight Spices The reason for the different effect of cinnamon on prostate cell lines may be that in the aforementioned literature, only the selected cinnamon content has been evaluated, whereas in our study we used a whole-spice extract containing a wide range of substances. Figure 1. Results of the MTT assay for spice extracts: (A) for PNT1A cells; (B) for 22RV1 cells; and (C) for PC3 cells. Figure 1. Results of the MTT assay for spice extracts: (A) for PNT1A cells; (B) for 22RV1 cells; and (C) for PC3 cells. Figure 1. Results of the MTT assay for spice extracts: (A) for PNT1A cells; (B) for 22RV1 cells; and (C) for PC3 cells. Figure 1. Results of the MTT assay for spice extracts: (A) for PNT1A cells; (B) for 22RV1 cells; and (C) for PC3 cells. 2.2. Profiling of the Extracts of the Tested Spices Using LC/MS 2.2. Proﬁling of the Extracts of the Tested Spices Using LC/MS y p The predominant phenolic compound in the extract from black pepper was 3,4-dihydroxybenzaldehyde. As shown in Figure 3A, the highest concentration was measured using 80% methanol, with the concentration being 55 ± 5 μg·g−1. For the other phenolic compounds determined, the concentration was not higher than 16 μg·g−1. Contents of protocatechuic acid, caffeic acid, ferulic acid and vanillic acid were also demonstrated in the extract from black pepper in the The predominant phenolic compound in the extract from black pepper was 3,4-dihydroxybenzaldehyde. As shown in Figure 3A, the highest concentration was measured using 80% methanol, with the concentration being 55 ± 5 µg·g−1. For the other phenolic compounds determined, the concentration was not higher than 16 µg·g−1. Contents of protocatechuic acid, caffeic acid, ferulic acid and vanillic acid were also demonstrated in the extract from black pepper in the study [39], 4 of 13 Molecules 2017, 22, 1626 despite the different preparation of the sample. The black pepper extract has the highest content of naringenin chalcone compared to other flavonoid compounds. The highest concentration of naringenin chalcone was 13 ± 1 µg·g−1, and was measured in the extract obtained using 80% methanol. Comparable results were obtained in the case of using 100% methanol. Using extraction with 60% methanol, the lowest concentration of naringenin chalcone was recorded, that is, 5 ± 1 µg·g−1. The results showed that there was less than half the amount of release when using 60% methanol, compared to extractions using 80% and 100% methanol. No significant differences were observed for other flavonoid compounds using extraction with 60%, 80% and 100% methanol. The average concentration of these compounds was 0.8 ± 0.1 µg·g−1. In a study by authors Chandra et al. [39], the presence of apigenin and quercetin in black pepper was demonstrated. However, other flavonoid compounds were not investigated here. Figure 3B,D gives a chromatogram of 3,4-dihydroxybenzaldehyde and naringenin chalcone. Figure 3C,E shows a fragmentation spectrum for 3,4-dihydroxybenzaldehyde and naringenin chalcone. Molecules 2017, 22, 1626 4 of 12 methanol. Comparable results were obtained in the case of using 100% methanol. Using extraction with 60% methanol, the lowest concentration of naringenin chalcone was recorded, that is, 5 ± 1 μg·g−1. The results showed that there was less than half the amount of release when using 60% methanol, compared to extractions using 80% and 100% methanol. 2.2. Profiling of the Extracts of the Tested Spices Using LC/MS 2.2. Proﬁling of the Extracts of the Tested Spices Using LC/MS No significant differences were observed for other flavonoid compounds using extraction with 60%, 80% and 100% methanol. The average concentration of these compounds was 0.8 ± 0.1 μg·g−1. In a study by authors Chandra et al. [39], the presence of apigenin and quercetin in black pepper was demonstrated. However, other flavonoid compounds were not investigated here. Figure 3B,D gives a chromatogram of 3,4-dihydroxybenzaldehyde and naringenin chalcone. Figure 3C,E shows a fragmentation spectrum for 3,4-dihydroxybenzaldehyde and naringenin chalcone. Molecules 2017, 22, 1626 4 of 12 methanol. Comparable results were obtained in the case of using 100% methanol. Using extraction with 60% methanol, the lowest concentration of naringenin chalcone was recorded, that is, 5 ± 1 μg·g−1. The results showed that there was less than half the amount of release when using 60% methanol, compared to extractions using 80% and 100% methanol. No significant differences were observed for other flavonoid compounds using extraction with 60%, 80% and 100% methanol. The average concentration of these compounds was 0.8 ± 0.1 μg·g−1. In a study by authors Chandra et al. [39], the presence of apigenin and quercetin in black pepper was demonstrated. However, other flavonoid compounds were not investigated here. Figure 3B,D gives a chromatogram of 3,4-dihydroxybenzaldehyde and naringenin chalcone. Figure 3C,E shows a fragmentation spectrum Figure 2. (A) Determination of the concentration of selected phenolic and flavonoid compounds in the extract from caraway seeds; (B) scheme for preparing the sample for LC/MS analysis. Figure 2. (A) Determination of the concentration of selected phenolic and ﬂavonoid compounds in the extract from caraway seeds; (B) scheme for preparing the sample for LC/MS analysis. for 3,4-dihydroxybenzaldehyde and naringenin chalcone. Figure 2. (A) Determination of the concentration of selected phenolic and flavonoid compounds in the extract from caraway seeds; (B) scheme for preparing the sample for LC/MS analysis Figure 2. (A) Determination of the concentration of selected phenolic and flavonoid compounds in the extract from caraway seeds; (B) scheme for preparing the sample for LC/MS analysis. Figure 2. (A) Determination of the concentration of selected phenolic and ﬂavonoid compounds in the extract from caraway seeds; (B) scheme for preparing the sample for LC/MS analysis. Figure 2. (A) Determination of the concentration of selected phenolic and flavonoid compounds in Figure 3. 2.2. Profiling of the Extracts of the Tested Spices Using LC/MS 2.2. Proﬁling of the Extracts of the Tested Spices Using LC/MS 5 of 13 of 12 Molecules 2017, 22, 1626 Molecules 2017 22 1626 2.3. Cell-Line Proliferative Activity Testing (MTT Assay) for Phenolic and Flavonoid Compounds from Spices 2.3. Cell-Line Proliferative Activity Testing (MTT Assay) for Phenolic and Flavonoid Compounds from Spices 2.3. Cell-Line Proliferative Activity Testing (MTT Assay) for Phenolic and Flavonoid Compounds from Spices 2.3. Cell-Line Proliferative Activity Testing (MTT Assay) for Phenolic and Flavonoid Compounds from Spices In the following experiment, we selected caraway seeds and black pepper only, for which we evaluated the presence of selected phenolic and ﬂavonoid compounds (Figure 4). The aim of the MTT assay was to determine the half-maximal inhibitory concentration (IC50). According to the obtained results, an IC50 value of ~1 mmol·L−1 was determined for all cell lines used. The IC50 value is too high compared to other studies. However, the aim of this study is not to develop an anticancer drug, but to supplement it and to achieve a better availability of the drug to the organism. The IC50 value is 1 mmol·L−1 because the substances do not have such toxicity. Studies [40–44] also conﬁrm that ﬂavonoids can only be used to treat cancer as an adjunct to an anticancer drug, not as an anticancer drug alone. Flavonoids thus should serve as a low-dose prevention, not as an acute drug. The identiﬁed IC50 concentration was then used in further experiments. In PNT1A (Figure 4A), 22RV1 (Figure 4B) and PC3 (Figure 4C) cells, the results were comparable, except for lesser variations in neochlorogenic acid. The most potent inhibitory effect was for the naringenin chalcone for all cell lines used. These results were conﬁrmed in the scratch test, which is described in the results and discussion in Section 2.4. If we compare the clonogenic assay (Table S1, Supplementary Materials) and the MTT assay (Figure 4) for the selected phenolic and ﬂavonoid compounds, we have compliance for the naringenin chalcone and neochlorogenic acid for PC3 cells, and apigenin for PNT1A cells. Studies [27,28] also demonstrate the inhibitory effect of apigenin on prostate cancer cell lines and other cancer cell lines. The inhibitory effect is also demonstrated in naringenin chalcone [29], but not in prostate cancer cell lines. A decreased proliferation of cells is demonstrated when using 3,4-dihydroxybenzaldehyde [32], but not in prostate cancer cell lines. 2.2. Profiling of the Extracts of the Tested Spices Using LC/MS 2.2. Proﬁling of the Extracts of the Tested Spices Using LC/MS Neochlorogenic acid exhibits chemopreventive activity in colon cancer, and in breast cancer [30], but not in prostate cancer cell lines, where this study is the ﬁrst one to show these effects. In the following experiment, we selected caraway seeds and black pepper only, for which we evaluated the presence of selected phenolic and flavonoid compounds (Figure 4). The aim of the MTT assay was to determine the half-maximal inhibitory concentration (IC50). According to the obtained results, an IC50 value of ~1 mmol·L−1 was determined for all cell lines used. The IC50 value is too high compared to other studies. However, the aim of this study is not to develop an anticancer drug, but to supplement it and to achieve a better availability of the drug to the organism. The IC50 value is 1 mmol·L−1 because the substances do not have such toxicity. Studies [40–44] also confirm that flavonoids can only be used to treat cancer as an adjunct to an anticancer drug, not as an anticancer drug alone. Flavonoids thus should serve as a low-dose prevention, not as an acute drug. The identified IC50 concentration was then used in further experiments. In PNT1A (Figure 4A), 22RV1 (Figure 4B) and PC3 (Figure 4C) cells, the results were comparable, except for lesser variations in neochlorogenic acid. The most potent inhibitory effect was for the naringenin chalcone for all cell lines used. These results were confirmed in the scratch test, which is described in the results and discussion in Section 2.4. If we compare the clonogenic assay (Table S1, Supplementary Materials) and the MTT assay (Figure 4) for the selected phenolic and flavonoid compounds, we have compliance for the naringenin chalcone and neochlorogenic acid for PC3 cells, and apigenin for PNT1A cells. Studies [27,28] also demonstrate the inhibitory effect of apigenin on prostate cancer cell lines and other cancer cell lines. The inhibitory effect is also demonstrated in naringenin chalcone [29], but not in prostate cancer cell lines. A decreased proliferation of cells is demonstrated when using 3,4-dihydroxybenzaldehyde [32], but not in prostate cancer cell lines. Neochlorogenic acid exhibits chemopreventive activity in colon cancer, and in breast cancer [30], but not in prostate cancer cell lines, where this study is the first one to show these effects. Figure 4. Results of the MTT assay for selected phenolic and flavonoid compounds: (A) for PNT1A cells; (B) for 22RV1 cells; and (C) for PC3 cells. 2.2. Profiling of the Extracts of the Tested Spices Using LC/MS 2.2. Proﬁling of the Extracts of the Tested Spices Using LC/MS Figure 4. Results of the MTT assay for selected phenolic and ﬂavonoid compounds: (A) for PNT1A cells; (B) for 22RV1 cells; and (C) for PC3 cells. Figure 4. Results of the MTT assay for selected phenolic and flavonoid compounds: (A) for PNT1A cells; (B) for 22RV1 cells; and (C) for PC3 cells. Figure 4. Results of the MTT assay for selected phenolic and ﬂavonoid compounds: (A) for PNT1A cells; (B) for 22RV1 cells; and (C) for PC3 cells. 2.2. Profiling of the Extracts of the Tested Spices Using LC/MS 2.2. Proﬁling of the Extracts of the Tested Spices Using LC/MS Extract from black pepper: (A) determination of concentration of selected phenolic and flavonoid compounds; (B) Multiple reaction monitoring (MRM) chromatograms of 3,4-dihydroxybenzaldehyde; (C) fragmentation spectrum for 3,4-dihydroxybenzaldehyde (108.0 is the product ion and 173.1 is the precursor ion); (D) Multiple reaction monitoring (MRM) chromatograms for naringenin chalcone; and (E) fragmentation spectrum for naringenin chalcone (151.0 is the product ion and 271.0 is the precursor ion). Figure 3. Extract from black pepper: (A) determination of concentration of selected phenolic and flavonoid compounds; (B) Multiple reaction monitoring (MRM) chromatograms of 3,4-dihydroxybenzaldehyde; (C) fragmentation spectrum for 3,4-dihydroxybenzaldehyde (108.0 is the product ion and 173.1 is the precursor ion); (D) Multiple reaction monitoring (MRM) chromatograms for naringenin chalcone; and (E) fragmentation spectrum for naringenin chalcone (151.0 is the product ion and 271.0 is the precursor ion). Figure 3. Extract from black pepper: (A) determination of concentration of selected phenolic and flavonoid compounds; (B) Multiple reaction monitoring (MRM) chromatograms of 3,4-dihydroxybenzaldehyde; (C) fragmentation spectrum for 3,4-dihydroxybenzaldehyde (108.0 is the product ion and 173.1 is the precursor ion); (D) Multiple reaction monitoring (MRM) chromatograms for naringenin chalcone; and (E) fragmentation spectrum for naringenin chalcone (151.0 is the product ion and 271.0 is the precursor ion). Figure 3. Extract from black pepper: (A) determination of concentration of selected phenolic and flavonoid compounds; (B) Multiple reaction monitoring (MRM) chromatograms of 3,4-dihydroxybenzaldehyde; (C) fragmentation spectrum for 3,4-dihydroxybenzaldehyde (108.0 is the product ion and 173.1 is the precursor ion); (D) Multiple reaction monitoring (MRM) chromatograms for naringenin chalcone; and (E) fragmentation spectrum for naringenin chalcone (151.0 is the product ion and 271.0 is the precursor ion). Figure 3. Extract from black pepper: (A) determination of concentration of selected phenolic and flavonoid compounds; (B) Multiple reaction monitoring (MRM) chromatograms of 3,4-dihydroxybenzaldehyde; (C) fragmentation spectrum for 3,4-dihydroxybenzaldehyde (108.0 is the product ion and 173.1 is the precursor ion); (D) Multiple reaction monitoring (MRM) chromatograms for naringenin chalcone; and (E) fragmentation spectrum for naringenin chalcone (151.0 is the product ion and 271.0 is the precursor ion). Figure 3. Extract from black pepper: (A) determination of concentration of selected phenolic and flavonoid compounds; (B) Multiple reaction monitoring (MRM) chromatograms of 3,4-dihydroxybenzaldehyde; (C) fragmentation spectrum for 3,4-dihydroxybenzaldehyde (108.0 is the product ion and 173.1 is the precursor ion); (D) Multiple reaction monitoring (MRM) chromatograms for naringenin chalcone; and (E) fragmentation spectrum for naringenin chalcone (151.0 is the product ion and 271.0 is the precursor ion). 2.4. Wound-Healing Assay (Scratch Test) 2.4. Wound-Healing Assay (Scratch Test) The scratch test is one of the most widely used, fastest and most effective methods for obtaining the critical loads that are related to the adhesion properties of a coating. A scratch test assesses how power is used in inhibitory compounds for a specific cell. Cells PNT1A, 22RV1 and PC3 showed the strongest inhibitory effects with naringenin chalcone, which had the smallest cell growth over time compared to the control at a time of 0 h (Figure 5A–C and Tables 1–3, respectively). The second-strongest inhibitory effect was observed with apigenin in PNT1A, 22RV1 and PC3 cells compared to the control at a time of 0 h (Tables 1–3, respectively). The results for the 3,4-dihydroxybenzaldehyde and neochlorogenic acid in PNT1A, 22RV1 and PC3 cells were almost The scratch test is one of the most widely used, fastest and most effective methods for obtaining the critical loads that are related to the adhesion properties of a coating. A scratch test assesses how power is used in inhibitory compounds for a speciﬁc cell. Cells PNT1A, 22RV1 and PC3 showed the strongest inhibitory effects with naringenin chalcone, which had the smallest cell growth over time compared to the control at a time of 0 h (Figure 5A–C and Tables 1–3, respectively). The second-strongest inhibitory effect was observed with apigenin in PNT1A, 22RV1 and PC3 cells compared to the control at a time of 0 h (Tables 1–3, respectively). The results for the 3,4-dihydroxybenzaldehyde and neochlorogenic acid in PNT1A, 22RV1 and PC3 cells were almost comparable to the control at a time of 0 h (Tables 1–3, 6 of 13 of 12 Molecules 2017, 22, 1626 Molecules 2017, 22, 1626 respectively). The results of the scratch test corroborate the MTT assay results (Figure 4). The microscopy results for other selected phenolic and ﬂavonoid compounds are shown in Figures S1–S3. p ( p y) corroborate the MTT assay results (Figure 4). The microscopy results for other selected phenolic and flavonoid compounds are shown in Figures S1–S3. Figure 5. The effect of naringenin chalcone on all three cell lines in the scratch test: (A) for PNT1A cells; (B) for 22RV1 cells; and (C) for PC3 cells. Figure 5. The effect of naringenin chalcone on all three cell lines in the scratch test: (A) for PNT1A cells; (B) for 22RV1 cells; and (C) for PC3 cells. Figure 5. 2.4. Wound-Healing Assay (Scratch Test) 2.4. Wound-Healing Assay (Scratch Test) The effect of naringenin chalcone on all three cell lines in the scratch test: (A) for PNT1A cells; (B) for 22RV1 cells; and (C) for PC3 cells Figure 5. The effect of naringenin chalcone on all three cell lines in the scratch test: (A) for PNT1A cells; (B) for 22RV1 cells; and (C) for PC3 cells. Table 1. Results of the scratch test for PNT1A cells. Table 1. Results of the scratch test for PNT1A cells. Table 1. Results of the scratch test for PNT1A cells. Compounds 0 h 6 h 12 h 24 h Measured Value * % Measured Value * % Measured Value * % Measured Value * % Control 518.12 ± 6.58 100 443.24 ± 6.52 86 377.50 ± 2.13 73 80.39 ± 20.27 16 3,4-Dihydroxybenzaldehyde 505.10 ± 5.10 100 441.11 ± 6.82 97 398.04 ± 7.94 79 147.72 ± 2.92 29 Naringenin chalcone 492.72 ± 5.38 100 446.06 ± 8.28 91 431.77 ± 7.32 88 269.33 ± 3.78 55 Apigenin 490.69 ± 3.78 100 397.23 ± 6.59 81 360.21 ± 6.22 73 176.18 ± 9.94 36 Neochlorogenic acid 519.38 ± 7.00 100 436.78 ± 7.17 84 303.23 ± 6.49 58 71.92 ± 2.37 14 * The average of the five measurements. Table 1. Results of the scratch test for PNT1A cells. Compounds 0 h 6 h 12 h 24 h Measured Value * % Measured Value * % Measured Value * % Measured Value * % Control 518.12 ± 6.58 100 443.24 ± 6.52 86 377.50 ± 2.13 73 80.39 ± 20.27 16 3,4-Dihydroxybenzaldehyde 505.10 ± 5.10 100 441.11 ± 6.82 97 398.04 ± 7.94 79 147.72 ± 2.92 29 Naringenin chalcone 492.72 ± 5.38 100 446.06 ± 8.28 91 431.77 ± 7.32 88 269.33 ± 3.78 55 Apigenin 490.69 ± 3.78 100 397.23 ± 6.59 81 360.21 ± 6.22 73 176.18 ± 9.94 36 Neochlorogenic acid 519.38 ± 7.00 100 436.78 ± 7.17 84 303.23 ± 6.49 58 71.92 ± 2.37 14 * The average of the ﬁve measurements. 2.4. Wound-Healing Assay (Scratch Test) 2.4. Wound-Healing Assay (Scratch Test) Compounds 0 h 6 h 12 h 24 h Measured Value * % Measured Value * % Measured Value * % Measured Value * % Control 474.72 ± 7.77 100 420.75 ± 9.88 89 371.84 ± 11.27 78 140.11 ± 10.24 30 3,4-Dihydroxybenzaldehyde 406.92 ± 12.63 100 349.11 ± 13.20 86 304.66 ± 8.34 75 152.38 ± 11.25 37 Naringenin chalcone 461.85 ± 9.27 100 436.98 ± 10.68 95 411.39 ± 8.32 89 398.17 ± 11.82 86 Apigenin 480.67 ± 5.35 100 415.30 ± 6.16 86 368.72 ± 8.68 77 300.77 ± 8.78 63 Neochlorogenic acid 365.15 ± 4.35 100 318.30 ± 7.54 87 300.08 ± 6.22 82 139.02 ± 15.20 38 * The average of the five measurements. Table 2. Results of the scratch test for 22RV1 cells. Compounds 0 h 6 h 12 h 24 h Measured Value * % Measured Value * % Measured Value * % Measured Value * % Control 474.72 ± 7.77 100 420.75 ± 9.88 89 371.84 ± 11.27 78 140.11 ± 10.24 30 3,4-Dihydroxybenzaldehyde 406.92 ± 12.63 100 349.11 ± 13.20 86 304.66 ± 8.34 75 152.38 ± 11.25 37 Naringenin chalcone 461.85 ± 9.27 100 436.98 ± 10.68 95 411.39 ± 8.32 89 398.17 ± 11.82 86 Apigenin 480.67 ± 5.35 100 415.30 ± 6.16 86 368.72 ± 8.68 77 300.77 ± 8.78 63 Neochlorogenic acid 365.15 ± 4.35 100 318.30 ± 7.54 87 300.08 ± 6.22 82 139.02 ± 15.20 38 * The average of the ﬁve measurements. 0 h 6 h 12 h Table 2. Results of the scratch test for 22RV1 cells. 7 of 13 Molecules 2017, 22, 1626 Table 3. Results of the scratch test for PC3 cells. Table 3. Results of the scratch test for PC3 cells. 2.4. Wound-Healing Assay (Scratch Test) 2.4. Wound-Healing Assay (Scratch Test) 0 h 6 h 12 h 24 h Compounds Measured Value * % Measured Value * % Measured Value * % Measured Value * % Control 494.87 ± 7.25 100 436.01 ± 2.03 88 369.92 ± 7.54 75 136.69 ± 22.52 28 3,4-Dihydroxybenzaldehyde 477.30 ± 3.73 100 408.49 ± 2.33 86 373.09 ± 3.78 78 102.62 ± 8.73 22 Naringenin chalcone 444.85 ± 7.66 100 410.23 ± 7.47 92 398.44 ± 3.23 90 360.25 ± 6.42 81 Apigenin 398.55 ± 8.46 100 370.29 ± 6.75 93 341.14 ± 8.87 86 270.30 ± 9.43 68 Neochlorogenic acid 448.47 ± 6.80 100 393.60 ± 8.87 88 356.98 ± 8.52 80 136.54 ± 10.86 30 * The average of the ﬁve measurements. Naringenin chalcone is a ﬂavonoid compound that engages in various plant defense roles, due to its antibacterial, antifungal and anti-inﬂammatory activities and cytotoxicity in carcinoma cells [45–47]. Flavonoids occur abundantly in fruits, vegetables, medicinal plants and beverages [43,48]. Flavonoids are present as aglycones, and glycosylated and methylated derivatives. Glycosylated ﬂavonoids are very rich in the human diet. After eating foods containing glycosylated ﬂavonoids, the body hydrolyzes these compounds in the gastrointestinal tract, liberating the aglycones, which are further extensively metabolized by glucuronidation, sulfation or methylation in the small intestine and liver. Flavonoids have poor oral bioavailability, the metabolites predominate in systemic circulation, while the plasma levels of parent ﬂavonoids are very low (below 2 µM or nM range) [40,41,49]. However, the data summarized in the studies [42,44] strongly support the view that ﬂavonoids are promising candidates for the enhancement of oral drug bioavailability and chemoprevention. This is thought to be mainly due to their antioxidant effects, anti-inﬂammatory properties and ability to modulate metabolism of carcinogens by inhibition of distinct phase 1 metabolic enzymes and activation of phase 2 detoxifying enzymes. Researchers have further demonstrated that methylation of the ﬂavonoids at their free hydroxyl groups or carbon atoms dramatically increases their metabolic stability and enhances membrane transport, leading to facilitated absorption and highly increased oral bioavailability [47]. membrane transport, leading to facilitated absorption and highly increased oral bioavailability [47]. The mechanism of anticancer action of naringenin chalcone against prostate cancer cells lies in its capability to scavenge free radicals [50], resulting in effects on cell proliferation, inhibition of angiogenesis, inhibition of subcellular signaling and stimulation of DNA repair enzymes [51–54]. 2.4. Wound-Healing Assay (Scratch Test) 2.4. Wound-Healing Assay (Scratch Test) Reactive oxygen species can cause oxidative damage to biological macromolecules including nucleic acids. If cell damage is excessive, cell death or apoptosis occurs. In cells, checkpoint pathways are activated to inhibit progression of cells through the G1 and G2 phases to permit removal of damage and re-entry into the cell cycle. If the DNA damage is not repaired, gene mutations occur at a high rate and can lead to malignant transformation. For repair of oxidatively damaged DNA bases, the base excision repair (BER) pathway is responsible. Removal of the damaged base is a result of increased 8-oxoguanine-DNA glycosylase 1 and apurinic/apyrimidinic endonuclease activities. DNA polymerase β then ﬁlls the gap created by the excision of 8-hydroxydeoxyguanosine. Stimulation of the growth of prostate cell carcinoma in prostate cell lines is due to higher production of reactive oxygen species due to the loss of glutathione-S-transferase P1 [50]. The cell-signaling pathways involve cooperation with transcription factors, anti-apoptotic proteins, pro-apoptotic proteins, protein kinases and cell-cycle proteins [55]. For study [55], the IC50 value was reported from 200 to 250 µM for naringenin, and growth inhibition was detected. A similar IC50 value was recorded in study [56]. 3.3. Preparation of the Spice Samples for Cell-Line Proliferative Activity Testing (MTT Assay) In this experiment, eight kinds of spice (marjoram, sweet pepper, black pepper, caraway seeds, anise, thyme, cinnamon and oregano) were used. To determine the phenolic and ﬂavonoid compounds from the spices, extraction with 80% (anise, black pepper and caraway seeds) or 100% (thyme, marjoram, sweet pepper, cinnamon and oregano) methanol was used. 1 g sample was weighed for each of eight kinds of spice (Analytical Weight EP 240A, Precisa, Stare Mesto, Czech Republic). The samples of the eight kinds of spice were homogenized in a friction bowl with 10 mL of 80% or 100% methanol, and 0.05 to 0.1 g of sea sand (until evaporation). The homogenization was repeated once more. After the homogenization, the samples were vortexed (Vortex Mixers, VELP Scientiﬁca, Usmate Velate MB, Italy) for 1–2 min, and centrifuged at 4500 rpm and 16 ◦C for 10 min (Centrifuge Z326K, Hermle, Gosheim, Germany). Subsequently, each sample was ﬁltered through a ﬁlter (LUT Syringe Filters Nylon, LABICOM s.r.o., Olomouc, Czech Republic). Samples of the extracts of eight kinds of spice were pipetted (2 mL) and concentrated by nitrogen evaporation at 60 ◦C. 3.4. Cell-Line Proliferative Activity Testing (MTT Assay) For the extracts from eight kinds of spice, and for the phenolic and ﬂavonoid compounds from the spices, the treatment was initiated after the cells reached ~60–80% conﬂuence. The cells were then harvested, washed four times with phosphate-buffered saline (PBS) (pH 7.4), and counted using the Countess II FL Automated Cell Counter (Life Technologies, Carlsbad, CA, USA). The cells’ proliferative activity was estimated using the MTT assay. Brieﬂy, the suspension of 5000 cells in 50 µL medium was added to each well in the microtiter plates (E-plates 96) used, followed by incubation for 24 h at 37 ◦C with 5% CO2 to ensure cell growth. A volume of 50 µL of the medium containing an extract from eight kinds of spice, and phenolic and ﬂavonoid compounds from the conjugated spices, was added to the cells. To determine the effects on cell proliferative activity, the extract from eight kinds of spice (at a concentration of 0.05–25.00 mg·mL−1), and phenolic and ﬂavonoid compounds from spices (at a concentration of 0.001–1 mmol·L−1) were employed. The treated cells were incubated for 24 h at 37 ◦C with 5% CO2. In addition, 10 µL of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT (5 mg·mL−1 in PBS)) was added to the cells and the mixture was incubated for 4 h at 37 ◦C. The MTT-containing medium was replaced by 100 µL of 99.9% dimethyl sulfoxide to dissolve MTT–formazan crystals and, after 5 min incubation, the absorbance of the samples was determined at 570 nm (VersaMax Microplate Reader, Molecular Devices, Sunnyvale, CA, USA). The experiments were performed in triplicate. 3.1. Chemicals The chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA) in ACS purity, unless noted otherwise. Apigenin standard and neochlorogenic acid standard were purchased from Extrasynthese (Genay, France). Naringenin chalcone standard was purchased from Phytolab (Vestenbergsgreuth, Germany). 8 of 13 Molecules 2017, 22, 1626 3.2. Cells The PNT1A (immortalization of normal, adult, prostatic, epithelial cells), PC3 (androgen-independent) and 22RV1 (androgen-dependent) prostatic cancer cell lines were obtained from the American Type Culture Collection (ATCC) (Manassas, VA, USA). The cells were cultured in a complete RPMI-1640 medium (Hyclone, Waltham, MA, USA) with 10% fetal bovine serum (FBS) (Hyclone, Waltham, MA, USA), supplemented with penicillin and streptomycin (0.1 mg·mL−1) at 37 ◦C and 5% CO2 in a humidiﬁed incubator. 3.6. Wound-Healing Assay (Scratch Test) The treatment was initiated after the cells reached ~100% conﬂuence. The cells were then harvested, washed four times with PBS (pH 7.4), and counted using the Countess II FL Automated Cell Counter (Life Technologies, Carlsbad, CA, USA). Brieﬂy, the suspension of 105 cells in the medium was added to each well in the microtiter plates (E-plates 6) used, followed by incubation after reaching 100% conﬂuence at 37 ◦C with 5% CO2. After 24 h of serum starvation, a wound was made in the cell monolayer using a 200 µL pipette tip. The cells were washed using PBS, and treated with 2 mL of the medium containing the phenolic and ﬂavonoid compounds from the spices. To determine the effects on cell proliferative activity, the phenolic and ﬂavonoid compounds from the conjugated spices were added to the cells, and a concentration of 1 mmol·L−1 was employed. Images at a 400 µm magniﬁcation were obtained using an EVOS FL Auto Cell Imaging system (ThermoFisher Scientiﬁc, Waltham, MA, USA) at 0, 6, 12 and 24 h of treatment. Within each wound, we analyzed ﬁve distance measurements using the EVOS FL Auto software (version 1.7). The experiments were performed in duplicate and each microtiter plate measured ﬁve times. 3.5. Preparation of the Samples of Spices and the Analysis of Spice Extracts Using LC/MS To determine the phenolic and ﬂavonoid compounds of the spices, extraction with 80% (anise, black pepper and caraway seeds) or 100% (thyme, marjoram, sweet pepper, cinnamon and oregano) methanol was used. 1 g sample was weighed for each of eight kinds of spice (Analytical Weight EP 240A, Precisa, Czech Republic). The samples of the eight kinds of spice were homogenized in a friction bowl with 10 mL of 80% or 100% methanol, and 0.05 to 0.1 g of sea sand (until evaporation). The homogenization 9 of 13 Molecules 2017, 22, 1626 was repeated once more. After homogenization, the samples were vortexed (Vortex Mixers, VELP Scientiﬁca, Usmate Velate MB, Italy) for 1–2 min, and centrifuged at 4500 rpm and 16 ◦C for 10 min (Centrifuge Z326K, Hermle, Gosheim, Germany). Subsequently, each sample was ﬁltered through a ﬁlter (LUT Syringe Filters Nylon, LABICOM s.r.o., Olomouc, Czech Republic). Samples of the extracts of the eight kinds of spice were pipetted (400 µL) and analyzed using LC/MS. To determine the selected phenolic and ﬂavonoid compounds, a high-performance liquid chromatograph (HPLC Agilent 1200 Series) with a diode array detector and a triple quadrupole mass detector (6460 Triple Quad) LC/MS was used. For the separation of the phenolic and ﬂavonoid compounds, a column, Zorbax EC 18 of dimensions 50 mm × 3.0 mm and a particle size of 2.7 µm, was used. The column was held at 45 ◦C. Mobile phase A consisted of 100% methanol, and mobile phase B was 0.2% acetic acid. The ﬂow rate of the mobile phase was 0.6 mL·min−1. The compounds were eluted with a linear upward gradient: 0.00 min (85% B), 0.17 min (85% B), 0.50 min (75% B), 1.70 min (70% B), 4.00 min (70% B), and 6.00 min (85% B). The triple quadrupole mass detector was operated in negative mode. The gas (nitrogen) temperature was 300 ◦C, the gas ﬂow rate was set to 12 L·min−1, the pressure nebulizer had a value of 45 psi, the temperature of the focusing gas was 250 ◦C, the ﬂow rate of the focusing gas was set at 11 L·min−1, and the voltage on the capillary tube amounted to 3500 V. Supplementary Materials: Supplementary Materials are available online, Figures S1–S3 and T Acknowledgments: The study was ﬁnancially supported by IGA IP_33/2017 and CEITEC 2020 (LQ1601). Author Contributions: Design of experiments: Z.H., O.Z., M.B. and Z.L.; performance of experiments with spices: Z.L.; performance of LC/MS analysis: B.K.; performance of experiments with cell cultures: Z.L., H.B. and Z.B.; data analysis: B.K., J.K., Z.L. and H.B.; 1st draft of paper: Z.L.; revision of paper Z.B., O.Z. and V.A. Conﬂicts of Interest: The authors declare no conﬂicts of interest. 3.7. Clonogenic Assay The cells were then harvested, washed four times with PBS (pH 7.4), and counted using the Countess II FL Automated Cell Counter (Life Technologies, Carlsbad, CA, USA). Brieﬂy, the suspension of 103 cells in the medium was added to each well in the microtiter plates (E-plates 6) used, followed by incubation for 24 h at 37 ◦C with 5% CO2. After 24 h, the medium was removed, and the cells treated with 2 mL of the medium containing the phenolic and ﬂavonoid compounds from the spices. To determine the effects on cell proliferative activity, the phenolic and ﬂavonoid compounds from the conjugated spices were added to the cells, and a concentration of 1 mmol·L−1 was employed. This was followed by incubation for 24 h at 37 ◦C with 5% CO2, then a medium change and incubation for 9–14 days at 37 ◦C with 5% CO2. The medium was removed and the cells were washed with PBS and ﬁxation was completed with methanol:acetic acid (3:1) for 5 min. After ﬁxation, the cells were colored with 0.5% crystal violet in methanol for 15 min. The cells were washed with Milli-Q water. Images were obtained using a Canon EOS 650D (Canon, Ota, Japan). The experiments were performed in duplicate. Molecules 2017, 22, 1626 10 of 13 10 of 13 References Pak. J. Pharm. Sci. 2015, 28, 969–972. [PubMed] 7. Shahwar, D.; Ullah, S.; Khan, M.A.; Ahmad, N.; Saeed, A. Anticancer activity of cinnamon tamala leaf constituents towards human ovarian cancer cells. Pak. J. Pharm. Sci. 2015, 28, 969–972. [PubMed] 8. Ramadan, G.; El-Beih, N.M.; Arafa, N.M.S.; Zahra, M.M. Preventive effects of egyptian sweet marjoram (Origanum majorana L.) leaves on haematological changes and cardiotoxicity in isoproterenol-treated albino rats. Cardiovasc. Toxicol. 2013, 13, 100–109. [CrossRef] [PubMed] 8. Ramadan, G.; El-Beih, N.M.; Arafa, N.M.S.; Zahra, M.M. Preventive effects of egyptian sweet marjoram (Origanum majorana L.) leaves on haematological changes and cardiotoxicity in isoproterenol-treated albino rats. Cardiovasc. Toxicol. 2013, 13, 100–109. [CrossRef] [PubMed] 9. Martucci, J.F.; Gende, L.B.; Neira, L.M.; Ruseckaite, R.A. Oregano and lavender essential oils as antioxidant and antimicrobial additives of biogenic gelatin ﬁlms. Ind. Crop. Prod. 2015, 71, 205–213. [CrossRef] 9. Martucci, J.F.; Gende, L.B.; Neira, L.M.; Ruseckaite, R.A. Oregano and lavender essential oils as antioxidant and antimicrobial additives of biogenic gelatin ﬁlms. Ind. Crop. Prod. 2015, 71, 205–213. [CrossRef] 10. Ben-Jabeur, M.; Ghabri, E.; Myriam, M.; Hamada, W. Thyme essential oil as a defense inducer of tomato against gray mold and fusarium wilt. Plant Physiol. Biochem. 2015, 94, 35–40. [CrossRef] [PubMed] 11 G i P E N tt G D G i L S G ti U ib J A M t T l tt S U Mi ti 10. Ben-Jabeur, M.; Ghabri, E.; Myriam, M.; Hamada, W. Thyme essential oil as a defense inducer of tomato against gray mold and fusarium wilt. Plant Physiol. Biochem. 2015, 94, 35–40. [CrossRef] [PubMed] 11. Garcia-Perez, E.; Noratto, G.D.; Garcia-Lara, S.; Gutierrez-Uribe, J.A.; Mertens-Talcott, S.U. Micropropagation effect on the anti-carcinogenic activitiy of polyphenolics from mexican oregano (Poliomintha glabrescens gray) in human colon cancer cells ht-29. Plant Food Hum. Nutr. 2013, 68, 155–162. [CrossRef] [PubMed] 11. Garcia-Perez, E.; Noratto, G.D.; Garcia-Lara, S.; Gutierrez-Uribe, J.A.; Mertens-Talcott, S.U. Micropropagation effect on the anti-carcinogenic activitiy of polyphenolics from mexican oregano (Poliomintha glabrescens gray) in human colon cancer cells ht-29. Plant Food Hum. Nutr. 2013, 68, 155–162. [CrossRef] [PubMed] 12. Flora, S.J.S. Structural, chemical and biological aspects of antioxidants for strategies against metal and metalloid exposure. Oxidative Med. Cell. Longev. 2009, 2, 191–206. [CrossRef] [PubMed] 12. Flora, S.J.S. Structural, chemical and biological aspects of antioxidants for strategies against metal and metalloid exposure. Oxidative Med. Cell. Longev. 2009, 2, 191–206. [CrossRef] [PubMed] 13. References 1. Cencic, A.; Chingwaru, W. The role of functional foods, nutraceuticals, and food supplements in intestinal health. Nutrients 2010, 2, 611–625. [CrossRef] [PubMed] 1. Cencic, A.; Chingwaru, W. The role of functional foods, nutraceuticals, and food supplements in intestinal health. Nutrients 2010, 2, 611–625. [CrossRef] [PubMed] 2. Pandey, K.B.; Rizvi, S.I. Plant polyphenols as dietary antioxidants in human health and disease. Oxidative Med. Cell. Longev. 2009, 2, 270–278. [CrossRef] [PubMed] 2. Pandey, K.B.; Rizvi, S.I. Plant polyphenols as dietary antioxidants in human health and disease. Oxidative Med. Cell. Longev. 2009, 2, 270–278. [CrossRef] [PubMed] 3. Slavin, J.L.; Lloyd, B. Health beneﬁts of fruits and vegetables. Adv. Nutr. 2012, 3, 506–516. [CrossRef] [PubMed] 3. Slavin, J.L.; Lloyd, B. Health beneﬁts of fruits and vegetables. Adv. Nutr. 2012, 3, 506–516. [CrossRef] [PubMed] 4. El Ksibi, I.; Ben Slama, R.; Faidi, K.; Ben Ticha, M.; M’Henni, M.F. Mixture approach for optimizing the recovery of colored phenolics from red pepper (Capsicum annum L.) by-products as potential source of natural dye and assessment of its antimicrobial activity. Ind. Crop. Prod. 2015, 70, 34–40. [CrossRef] 4. El Ksibi, I.; Ben Slama, R.; Faidi, K.; Ben Ticha, M.; M’Henni, M.F. Mixture approach for optimizing the recovery of colored phenolics from red pepper (Capsicum annum L.) by-products as potential source of natural dye and assessment of its antimicrobial activity. Ind. Crop. Prod. 2015, 70, 34–40. [CrossRef] 5. Hertwig, C.; Reineke, K.; Ehlbeck, J.; Knorr, D.; Schluter, O. Decontamination of whole black pepper using different cold atmospheric pressure plasma applications. Food Control 2015, 55, 221–229. [CrossRef] 6. Mnif, S.; Aifa, S. Cumin (Cuminum cyminum L.) from traditional uses to potential biomedical applications. Chem. Biodivers. 2015, 12, 733–742. [CrossRef] [PubMed] 5. Hertwig, C.; Reineke, K.; Ehlbeck, J.; Knorr, D.; Schluter, O. Decontamination of whole black pepper using different cold atmospheric pressure plasma applications. Food Control 2015, 55, 221–229. [CrossRef] 6. Mnif, S.; Aifa, S. Cumin (Cuminum cyminum L.) from traditional uses to potential biomedical applications. 5. Hertwig, C.; Reineke, K.; Ehlbeck, J.; Knorr, D.; Schluter, O. Decontamination of whole black pepper using different cold atmospheric pressure plasma applications. Food Control 2015, 55, 221–229. [CrossRef] 6. Mnif, S.; Aifa, S. Cumin (Cuminum cyminum L.) from traditional uses to potential biomedical appli Chem. Biodivers. 2015, 12, 733–742. [CrossRef] [PubMed] 7. Shahwar, D.; Ullah, S.; Khan, M.A.; Ahmad, N.; Saeed, A. Anticancer activity of cinnamon tamala leaf constituents towards human ovarian cancer cells. 4. Conclusions The MTT assay we performed on the extracts from eight spices revealed the strongest inhibitory effects of caraway seeds and black pepper on the tested cell lines. From the black pepper and caraway seeds, we analyzed selected phenolic and ﬂavonoid compounds using LC/MS. The most represented phenolic and ﬂavonoid compounds were, in caraway seeds, neochlorogenic acid, and in black pepper, 3,4-dihydroxybenzaldehyde and naringenin chalcone. The results of the MTT assay for these compounds determined the value of IC50 (1 mmol·L−1 for the PNT1A, 22RV1 and PC3 cells). The most potent inhibitory effect on the PNT1A, 22RV1 and PC3 cells was from the naringenin chalcone (concentration 1 mmol·L−1) contained in black pepper, found using the scratch test. The inhibitory effect of naringenin chalcone was conﬁrmed in other studies, but in cells other than prostatic cells. The results obtained serve as a pilot study for further experiments, where other cells lines and/or potency of some identiﬁed biologically active molecules will be tested. Supplementary Materials: Supplementary Materials are available online, Figures S1–S3 and Table S1. References Samykutty, A.; Shetty, A.V.; Dakshinamoorthy, G.; Bartik, M.M.; Johnson, G.L.; Webb, B.; Zheng, G.; Chen, A.X.; Kalyanasundaram, R.S.; Munirathinam, G. Piperine, a bioactive component of pepper spice 13. Samykutty, A.; Shetty, A.V.; Dakshinamoorthy, G.; Bartik, M.M.; Johnson, G.L.; Webb, B.; Zheng, G.; Chen, A.X.; Kalyanasundaram, R.S.; Munirathinam, G. Piperine, a bioactive component of pepper spice 11 of 13 Molecules 2017, 22, 1626 exerts therapeutic effects on androgen dependent and androgen independent prostate cancer cells. PLoS ONE 2013, 8. [CrossRef] [PubMed] exerts therapeutic effects on androgen dependent and androgen independent prostate cancer cells. PLoS ONE 2013, 8. [CrossRef] [PubMed] 14. Surh, Y.J.; Kundu, J.K. Molecular mechanisms of chemoprevention with capsaicinoids from chili peppers. In Vegetables, Whole Grains, and Their Derivatives in Cancer Prevention; Mutanen, M., Pajari, A.M., Eds.; Springer: Dordrecht, The Netherlands, 2011; Volume 2, pp. 123–142. 15. Wei, X.C.; Du, Z.Y.; Cui, X.X.; Verano, M.; Mo, R.Q.; Tang, Z.K.; Conney, A.H.; Zheng, X.; Zhang, K. Effects of cyclohexanone analogues of curcumin on growth, apoptosis and NF-kappa B activity in PC-3 human prostate cancer cells. Oncol. Lett. 2012, 4, 279–284. [PubMed] p 16. Chendil, D.; Ranga, R.S.; Meigooni, D.; Sathishkumar, S.; Ahmed, M.M. Curcumin confers radiosensitizing effect in prostate cancer cell line PC-3. Oncogene 2004, 23, 1599–1607. [CrossRef] [PubMed] 17. Nakamura, K.; Yasunaga, Y.; Segawa, T.; Ko, D.J.; Moul, J.W.; Srivastava, S.; Rhim, J.S. Curcumin down-regulates ar gene expression and activation in prostate cancer cell lines. Int. J. Oncol. 2002, 21, 825–830. [CrossRef] [PubMed] 18. Ouyang, D.Y.; Zeng, L.H.; Pan, H.; Xu, L.H.; Wang, Y.; Liu, K.P.; He, X.H. Piperine inhibits the proliferation of human prostate cancer cells via induction of cell cycle arrest and autophagy. Food Chem. Toxicol. 2013, 60, 424–430. [CrossRef] [PubMed] 19. Al Dhaheri, Y.; Eid, A.; AbuQamar, S.; Attoub, S.; Khasawneh, M.; Aiche, G.; Hisaindee, S.; Iratni, R. Mitotic arrest and apoptosis in breast cancer cells induced by origanum majorana extract: Upregulation of TNF-α and downregulation of survivin and mutant p53. PLoS ONE 2013, 8, 1–14. [CrossRef] [PubMed] 20. Abdel-Massih, R.M.; Fares, R.; Bazzi, S.; El-Chami, N.; Baydoun, E. The apoptotic and anti-proliferative activity of origanum majorana extracts on human leukemic cell line. Leuk. Res. 2010, 34, 1052–1056. [CrossRef] [PubMed] 21. Ahmad, R.A.; Abdullah, S.; Serati-Nouri, H.; Majid, F.A.A.; Sarmidi, M.R.; Aziz, R.A. Antiproliferative activity of coumarin and cinnamon water extracts on human ovarian cancer cells. Lat. Am. J. Pharm. 2014, 33, 960–965. 22. References Chuang, L.Y.; Guh, J.Y.; Chao, L.K.; Lu, Y.C.; Hwang, J.Y.; Yang, Y.L.; Cheng, T.H.; Yang, W.Y.; Chien, Y.J.; Huang, J.S. Anti-proliferative effects of cinnamaldehyde on human hepatoma cell lines. Food Chem. 2012, 133, 1603–1610. [CrossRef] 23. Koppikar, S.J.; Choudhari, A.S.; Suryavanshi, S.A.; Kumari, S.; Chattopadhyay, S.; Kaul-Ghanekar, R. Aqueous cinnamon extract (ACE-c) from the bark of cinnamomum cassia causes apoptosis in human cervical cancer cell line (SiHa) through loss of mitochondrial membrane potential. BMC Cancer 2010, 10, 1–12. [CrossRef] [PubMed] 24. Kim, J.E.; Son, J.E.; Jeong, H.; Kim, D.J.; Seo, S.G.; Lee, E.; Lim, T.G.; Kim, J.R.; Kimbung, Y.R.; Chen, H.Y.; et al. A novel cinnamon-related natural product with Pim-1 inhibitory activity inhibits leukemia and skin cancer. Cancer Res. 2015, 75, 2716–2728. [CrossRef] [PubMed] 25. Mahajan, U.B.; Chandrayan, G.; Patil, C.R.; Arya, D.S.; Suchal, K.; Agrawal, Y.O.; Ojha, S.; Goyal, S.N. The protective effect of apigenin on myocardial injury in diabetic rats mediating activation of the PPAR- gamma pathway. Int. J. Mol. Sci. 2017, 18, 756. [CrossRef] [PubMed] 26. Cardenas, H.; Arango, D.; Nicholas, C.; Duarte, S.; Nuovo, G.J.; He, W.; Voss, O.H.; Gonzalez-Mejia, M.E.; Guttridge, D.C.; Grotewold, E.; et al. Dietary apigenin exerts immune-regulatory activity in vivo by reducing nf-kappa b activity, halting leukocyte inﬁltration and restoring normal metabolic function. Int. J. Mol. Sci. 2016, 17, 323. [CrossRef] [PubMed] 27. He, J.; Xu, Q.; Wang, M.; Li, C.Y.; Qian, X.; Shi, Z.M.; Liu, L.Z.; Jiang, B.H. Oral administration of apigenin inhibits metastasis through AKT/P70S6K1/MMP-9 pathway in orthotopic ovarian tumor model. Int. J. Mol. Sci. 2012, 13, 7271–7282. [CrossRef] [PubMed] 28. Franzen, C.A.; Amargo, E.; Todorovic, V.; Desai, B.V.; Huda, S.; Mirzoeva, S.; Chiu, K.; Grzybowski, B.A.; Chew, T.L.; Green, K.J.; et al. The chemopreventive bioﬂavonoid apigenin inhibits prostate cancer cell motility through the focal adhesion kinase/src signaling mechanism. Cancer Prev. Res. 2009, 2, 830–841. [CrossRef] [PubMed] 29. Zhang, S.; Jiang, Z.F.; Pan, Q.; Song, C.Y.; Zhang, W.H. Anti-cancer effect of naringenin chalcone is mediated via the induction of autophagy, apoptosis and activation of PI3K/AKT signalling pathway. Bangladesh J. Pharmacol. 2016, 11, 684–690. [CrossRef] 12 of 13 Molecules 2017, 22, 1626 30. Banerjee, N.; Kim, H.; Talcott, S.T.; Turner, N.D.; Byrne, D.H.; Mertens-Talcott, S.U. Plum polyphenols inhibit colorectal aberrant crypt foci formation in rats: Potential role of the miR-143/protein kinase b/mammalian target of rapamycin axis. Nutr. Res. 2016, 36, 1105–1113. [CrossRef] [PubMed] 31. References Noratto, G.; Porter, W.; Byrne, D.; Cisneros-Zevallos, L. Identifying peach and plum polyphenols with chemopreventive potential against estrogen-independent breast cancer cells. J. Agric. Food Chem. 2009, 57, 5219–5226. [CrossRef] [PubMed] 32. Jeong, J.B.; Lee, S.H. Protocatechualdehyde possesses anti-cancer activity through downregulating cyclin D1 and HDAC2 in human colorectal cancer cells. Biochem. Biophys. Res. Commun. 2013, 430, 381–386. [CrossRef] [PubMed] 33. Hong, S.H.; Ismail, I.A.; Kang, S.M.; Han, D.C.; Kwon, B.M. Cinnamaldehydes in cancer chemotherapy. Phytother. Res. 2016, 30, 754–767. [CrossRef] [PubMed] 34. Kang, H.S.; Ock, J.; Lee, H.J.; Lee, Y.J.; Kwon, B.M.; Hong, S.H. Early growth response protein 1 upregulation and nuclear translocation by 2’-benzoyloxycinnamaldehyde induces prostate cancer cell death. Cancer Lett. 2013, 329, 217–227. [CrossRef] [PubMed] 35. Qi, G.Y.; Chen, J.; Shi, C.R.; Wang, Y.K.; Mi, S.S.; Shao, W.H.; Yu, X.Y.; Ma, Y.L.; Ling, J.P.; Huang, J. Cinnamic acid (cinn) induces apoptosis and proliferation in human nasopharyngeal carcinoma cells. Cell. Physiol. Biochem. 2016, 40, 589–596. [CrossRef] [PubMed] 36. Munirathinam, G.; Gray, K. Cinnamaldehye, a dietary ﬂavoring agent derived from cinnamon targets prostate cancer cells. Endocr. Rev. 2014, 35, 1–2. 37. Park, K.R.; Nam, D.; Yun, H.M.; Lee, S.G.; Jang, H.J.; Sethi, G.; Cho, S.K.; Ahn, K.S. Beta-caryophyllene oxide inhibits growth and induces apoptosis through the suppression of PI3K/AKT/MTOR/S6K1 pathways and ros-mediated mapks activation. Cancer Lett. 2011, 312, 178–188. [CrossRef] [PubMed] 38. Zu, Y.G.; Yu, H.M.; Liang, L.; Fu, Y.J.; Efferth, T.; Liu, X.; Wu, N. Activities of ten essential oils towards propionibacterium acnes and PC-3, A-549 and MCF-7 cancer cells. Molecules 2010, 15, 3200–3210. [CrossRef] [PubMed] 39. Chandra, P.; Pandey, R.; Srivastva, M.; Rameshkumar, K.B.; Kumar, B. Quantitative determination of chemical constituents of piper spp. Using uplc-esi-ms/ms. Ind. Crop. Prod. 2015, 76, 967–976. [CrossRef] 39. Chandra, P.; Pandey, R.; Srivastva, M.; Rameshkumar, K.B.; Kumar, B. Quantitative determination of chemical constituents of piper spp. Using uplc-esi-ms/ms. Ind. Crop. Prod. 2015, 76, 967–976. [CrossRef] 40. Morris, M.E.; Zhang, S.Z. Flavonoid-drug interactions: Effects of ﬂavonoids on abc transporters. Life Sci. 40. Morris, M.E.; Zhang, S.Z. Flavonoid-drug interactions: Effects of ﬂavonoids on abc transporters. 2006, 78, 2116–2130. [CrossRef] [PubMed] 1. Stevenson, D.E.; Scheepens, A.; Hurst, R.D. Bioavailability and Metabolism of Dietary Flavonoids—M Known—Much More to Discover; Nova Science Publishers, Inc.: Hauppauge, NY, USA, 2009; pp. 1–52. 42. Miron, A.; Aprotosoaie, A.C.; Trifan, A.; Xiao, J.B. Flavonoids as modulators of metabolic enzymes and drug transporters. Ann. N. Y. Acad. Sci. 2017, 1398, 152–167. References [CrossRef] [PubMed] 43. Moon, Y.J.; Wang, X.D.; Morris, M.E. Dietary ﬂavonoids: Effects on xenobiotic and carcinogen metabolism. Toxicol. In Vitro 2006, 20, 187–210. [CrossRef] [PubMed] 44. Venturelli, S.; Burkard, M.; Biendl, M.; Lauer, U.M.; Frank, J.; Busch, C. Prenylated chalcones and ﬂavonoids for the prevention and treatment of cancer Nutrition 2016 32 1171 1178 [CrossRef] [PubMed] 44. Venturelli, S.; Burkard, M.; Biendl, M.; Lauer, U.M.; Frank, J.; Busch, C. Prenylated chalcones and ﬂavonoids for the prevention and treatment of cancer. Nutrition 2016, 32, 1171–1178. [CrossRef] [PubMed] 45. Orjala, J.; Wright, A.D.; Behrends, H.; Folkers, G.; Sticher, O.; Ruegger, H.; Rali, T. Cytotoxic and antibacterial dihydrochalcones from piper aduncum. J. Nat. Prod. 1994, 57, 18–26. [CrossRef] [PubMed] 46. Tamogami, S.; Kodama, O. Coronatine elicits phytoalexin production in rice leaves (Oryza sativa L.) in the same manner as jasmonic acid. Phytochemistry 2000, 54, 689–694. [CrossRef] 47. Koirala, N.; Thuan, N.H.; Ghimire, G.P.; Thang, D.V.; Sohng, J.K. Methylation of ﬂavonoids: Chemical structures, bioactivities, progress and perspectives for biotechnological production. Enzyme Microb. Technol. 2016, 86, 103–116. [CrossRef] [PubMed] 48. Wu, J.C.; Lai, C.S.; Tsai, M.L.; Ho, C.T.; Wang, Y.J.; Pan, M.H. Chemopreventive effect of natural dietary compounds on xenobiotic-induced toxicity. J. Food Drug Anal. 2017, 25, 176–186. [CrossRef] [PubMed] 48. Wu, J.C.; Lai, C.S.; Tsai, M.L.; Ho, C.T.; Wang, Y.J.; Pan, M.H. Chemopreventive effect of natural dietary compounds on xenobiotic-induced toxicity. J. Food Drug Anal. 2017, 25, 176–186. [CrossRef] [PubMed] 49. Kumar, S.; Pandey, A.K. Chemistry and biological activities of ﬂavonoids: An overview. Sci. World J. 2013. [CrossRef] [PubMed] compounds on xenobiotic-induced toxicity. J. Food Drug Anal. 2017, 25, 176–186. [CrossRef] [PubMed] 49. Kumar, S.; Pandey, A.K. Chemistry and biological activities of ﬂavonoids: An overview. Sci. World J. 2013. [CrossRef] [PubMed] 49. Kumar, S.; Pandey, A.K. Chemistry and biological activities of ﬂavonoids: An overview. Sci. World J. 2013. [CrossRef] [PubMed] 50. Gao, K.; Henning, S.M.; Niu, Y.T.; Yousseﬁan, A.A.; Seeram, N.P.; Xu, A.L.; Heber, D. The citrus ﬂavonoid naringenin stimulates DNA repair in prostate cancer cells. J. Nutr. Biochem. 2006, 17, 89–95. [CrossRef] [PubMed] 51. Yang, C.S.; Landau, J.M.; Huang, M.T.; Newmark, H.L. Inhibition of carcinogenesis by dietary polyphenolic compounds. Annu. Rev. Nutr. 2001, 21, 381–406. [CrossRef] [PubMed] Molecules 2017, 22, 1626 13 of 13 52. Afaq, F.; Saleem, M.; Krueger, C.G.; Reed, J.D.; Mukhtar, H. Sample Availability: Samples of the compounds are available from the authors. References Anthocyanin- and hydrolyzable tannin-rich pomegranate fruit extract modulates MAPK and NF-kappa b pathways and inhibits skin tumorigenesis in CD-1 mice. Int. J. Cancer 2005, 113, 423–433. [CrossRef] [PubMed] 53. Williams, R.J.; Spencer, J.P.E.; Rice-Evans, C. Flavonoids: Antioxidants or signalling mo Free Radic. Biol. Med. 2004, 36, 838–849. [CrossRef] [PubMed] 54. Morel, I.; Abalea, V.; Cillard, P.; Cillard, J. Repair of oxidized DNA by the ﬂavonoid myricetin. Methods Enzymol. 2001, 335, 308–316. [PubMed] 54. Morel, I.; Abalea, V.; Cillard, P.; Cillard, J. Repair of oxidized DNA by the ﬂavonoid myricetin. Methods Enzymol. 2001, 335, 308–316. [PubMed] 55. Lewinska, A.; Siwak, J.; Rzeszutek, I.; Wnuk, M. Diosmin induces genotoxicity and apoptosis in DU145 prostate cancer cell line. Toxicol. In Vitro 2015, 29, 417–425. [CrossRef] [PubMed] 55. Lewinska, A.; Siwak, J.; Rzeszutek, I.; Wnuk, M. Diosmin induces genotoxicity and apoptosis in DU145 prostate cancer cell line. Toxicol. In Vitro 2015, 29, 417–425. [CrossRef] [PubMed] 56. Siwak, J.; Lewinska, A.; Wnuk, M.; Bartosz, G. Protection of ﬂavonoids against hypochlorite-induced protein modiﬁcations. Food Chem. 2013, 141, 1227–1241. [CrossRef] [PubMed] 56. Siwak, J.; Lewinska, A.; Wnuk, M.; Bartosz, G. Protection of ﬂavonoids against hypochlorite-induced protein modiﬁcations. Food Chem. 2013, 141, 1227–1241. [CrossRef] [PubMed] Sample Availability: Samples of the compounds are available from the authors. Sample Availability: Samples of the compounds are available from the authors. © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W3005181359	https://digibug.ugr.es/bitstream/10481/63560/1/2242-11588-2-PB.pdf	Spanish; Castilian	null	System error//Ctrl+Alt+Supr: Connecting bodies and tecnologies from transfeminist perspectives	Athenea digital	2,020	cc-by	10,947	Athenea Digital - 20(1): e2242 (marzo 2020) -ARTÍCULOS- Athenea Digital - 20(1): e2242 (marzo 2020) -ARTÍCULOS- ISSN: 1578-8946 Abstract Keywords Body Technology Queer Transfeminism It is my purpose in this article to point out the transfeminist connections between bodily and technological dissidence. Starting from the transformations analysis of the technologies of the body from foucaultian biopolitics to the technopolitics un- veiled by Donna Haraway and Paul Preciado, in this paper I gather some of the points raised in the discussions within cyberfeminist, queer and decolonial per- spectives that connect bodies and technologies, emphasizing in the non-normative corporalities and sexualities on the network. p Based on the previous critical perspectives of technologies and on the ethno- graphic fieldwork carried out in the transfeminist context in Barcelona, I explore transfeminist political practices which lead me to acknowledge and to conclude that the potential of new techno-corporal articulations are intended to resist and transform technoscience. Martínez Pozo, Lola (2020). Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías desde perspectivas transfeministas. Athenea Digital, 20(1), e2242. https://doi.org/10.5565/rev/athenea.2242 Martínez Pozo, Lola (2020). Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías desde perspectivas transfeministas. Athenea Digital, 20(1), e2242. https://doi.org/10.5565/rev/athenea.2242 Resumen Recibido: 24-04-2019 Primera revisión: 08-06-2019 Aceptado: 28-08-2019 Publicado: 07-02-2020 En el presente artículo mi objetivo es plantear las conexiones transfeministas en- tre disidencias corporales y tecnológicas. Partiendo del análisis de las transforma- ciones de las tecnologías del cuerpo desde la biopolítica foucaultiana hasta la tec- nopolítica desvelada por Donna Haraway y Paul Preciado, en este trabajo recojo algunos de los debates planteados desde perspectivas ciberfeministas, queer y de- coloniales que conectan los cuerpos y las tecnologías, enfatizando en las corporali- dades y sexualidades no normativas en la red. Palabras clave Cuerpo Tecnología Queer Transfeminismo y A partir de las anteriores perspectivas críticas de las tecnologías y la realización de trabajo de campo de carácter etnográfico en el contexto transfeminista de Bar- celona, indago en prácticas políticas transfeministas que me llevan a reflexionar y a concluir sobre la potencialidad de nuevas articulaciones tecno-corporales dirigi- das a resistir y transformar la tecnociencia. ERROR EN EL SISTEMA//CTRL+ALT+SUPR: CONECTAR CUERPOS Y TECNOLOGÍAS DESDE PERSPECTIVAS TRANSFEMINISTAS SYSTEM ERROR//CTRL+ALT+SUPR: CONNECTING BODIES AND TECNOLOGIES FROM TRANSFEMINIST PERSPECTIVES Lola Martínez Pozo Introducción En mi trabajo de investigación parto de una comprensión de los cuerpos como espa- cios políticos, construcciones y producciones socioculturales, ámbitos de reflexión so- cial y académica. Formulaciones de amplio calado inauguradas por Michel Foucault (1979; 1976/1987) y desarrolladas desde diversas corrientes críticas en el marco de la teoría social del cuerpo ( Csordas, 1993; Turner, 1989, 1994), la antropología del cuerpo (Citro, 2010; Esteban, 2004; Le Breton, 2002), los enfoques posfeministas (Butler, 1993/2002; De Lauretis, 1987/1989; Fausto-Sterling, 2000/2006; Haraway, 1991/1995; Preciado, 2008); las perspectivas queer (Córdoba, Sáez y Vidarte, 2005/2007; Romero, 1 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías García y Bargueiras, 2005) y los feminismos decoloniales (bell hooks et al., 2004; Suá- rez y Hernández, 2008). Nutriéndome de los anteriores desarrollos teóricos entiendo los cuerpos como materializaciones encarnadas de tecnologías de poder que se articulan en base a cate- gorías sociales como el sexo, el género, la sexualidad, la clase, la raza, la nacionalidad, las migraciones, la salud, la habilidad y la edad, para producir cuerpos disciplinariza- dos y viables, así como cuerpos patológicos y/o anormales. Coincidiendo con Paul Pre- ciado (2014) entiendo las corporalidades como un conjunto articulado y/o conflictivo de diversas ficciones históricas, culturales y geopolíticas. Desde esta comprensión del cuerpo, en primer lugar, en este texto planteo las co- nexiones que encuentro entre cuerpos, tecnologías y poder en las contemporáneas ci- ber-tecno sociedades1 del capitalismo tardío. Desde la biopolítica conceptualizada por Michel Foucault (1976/1987), la performatividad propuesta por Judith Butler (1993/2002) y la tecnopolítica desarticulada por Donna Haraway (1991/1995) y Paul Preciado (2008), entre otrxs2, destaco una mutación en las tecnologías de producción y regulación corporal donde la tecnociencia adquiere una relevancia central en la con- formación de los cuerpos, las subjetividades y las sociedades normalizadas. Con el objetivo de contemplar lecturas y prácticas de resistencia a la tecnociencia dominante, continúo indagando en las reapropiaciones feministas y queer de las tec- nologías a partir de la política cyborg propuesta por Haraway (1991/1995). En esta línea, enfatizo en debates procedentes de perspectivas ciberfeministas, tecnofeministas, feministas decoloniales, ciberqueer y tecnoqueer que posibilitaron la articulación entre cuerpos y tecnologías digitales. Las anteriores perspectivas críticas de las tecnologías me orientaron a la realiza- ción de una investigación etnográfica en el contexto transfeminista de Barcelona para indagar en las conexiones transfeministas con los cuerpos y las tecnologías. 1 Concepto propuesto a partir de los planteamientos en torno a la "cibersociedad" y “cibercultura" de Jordi Planella (2006) para destacar el impacto y la relevancia de las tecnologías en las sociedades. 2 Empleo la X para suprimir la identificación normativa de género. 2 Empleo la X para suprimir la identificación normativa de género. Introducción En este sentido, en la segunda parte del artículo destaco las aportaciones de lxs activistas en- trevistadxs y recojo algunos proyectos políticos que conectan disidencias corporales, sexuales y tecnológicas, como referentes que dan lugar a nuevos enfoques y fructíferas articulaciones feministas orientadas a subvertir las tecnologías dominantes de produc- ción corporal y subjetiva y a generar códigos alternativos para habitar nuestros cuer- pos y máquinas. 2 Lola Martínez Pozo El cuerpo como ficción: de la biopolítica a la tecnopolítica Con la intención de contemplar las articulaciones feministas dirigidas a la subversión de la tecnociencia mediante la conexión entre cuerpos y tecnologías, en este apartado destaco las aportaciones que me han permitido contextualizar la transformación de las tecnologías biopolíticas de producción y regulación corporal y subjetiva hacia nuevas formas de tecnopolítica dando lugar a nuevas configuraciones del cuerpo. El cuerpo se constituye como la articulación de ficciones políticas vivas y plurales, que nunca ha contado con una ontología natural, lo cual implica que la comprensión de su existencia esté vinculada a marcos culturales, históricos, políticos y epistemoló- gicos concretos. En este sentido, considero relevante aquello que Preciado (2015) deno- mina somateca, es decir, la perspectiva de pensar el cuerpo como un archivo de ficcio- nes políticas vivas, somatopolíticas. Comiendo recogiendo las aportaciones de Foucault (1976/1987) ya que teorizó so- bre los desplazamientos del poder en las sociedades occidentales de finales del siglo XVIII desde regímenes de poder soberano-necropolítico, basados en dar la muerte o permitir la vida, a regímenes de poder biopolítico donde las relaciones entre saber, verdad, poder y corporalidad se dirigen a administrar la vida. Yuxtaposiciones de regímenes de poder vienen a articularse directamente en los cuerpos a través de la red que Foucault (1979) denominó bio-poder o somato-poder. El autor centró su mirada en la sexualidad como elemento fundamental en el cru- ce de dos ejes, las disciplinas del cuerpo y la biopolítica de la población, a través de los cuales se desarrolló un gran dispositivo de tecnologías políticas de administración de la vida durante los siglos XVIII y XIX (Foucault, 1979/1987). Por otro lado, las aportaciones de Judith Butler (1993/2002), en cuanto a las rela- ciones entre poder y corporalidad, me permiten contar con un mayor alcance. La auto- ra parte de la noción foucaultiana acerca de que el poder produce los propios cuerpos que controla y desarrolla la noción de poder a través de la performatividad, como ac- tos y producciones reguladas, reiteradas y excluyentes que dibujan y establecen los lí- mites de aquello que se consideran cuerpos viables y aquello que no. Desde mi punto de vista, la autora comprende el poder no sólo como medio productivo normativo que conforma cuerpos inteligibles, sino que, también funciona produciendo un exterior de cuerpos abyectos. El cuerpo como ficción: de la biopolítica a la tecnopolítica 3 3 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías Desde mi punto de vista, el trabajo de Preciado (2008) también amplía las teoriza- ciones inauguradas por Foucault y viene a conceptualizar un nuevo régimen de sexua- lidad denominado farmacopornográfico, el cual desplaza a la sociedad disciplinarizada analizada por el autor. Para Preciado, este régimen emerge a partir de la segunda gue- rra mundial y, destacando un conjunto de transformaciones en las tecnologías del cuerpo, desvela cómo la tecnociencia pasa a ocupar un lugar dominante en la formula- ción de la normalidad. La invención de la categoría biotecnológica de género inaugura el régimen farma- copornográfico, donde no sólo el género, sino también el sexo, la sexualidad y la raza, funcionan como dispositivos tecnopolíticos de producción de subjetividad y de gestión corporal a través de dinámicas del capitalismo avanzado. Algunos hitos clave identifi- cados por Preciado (2008) en la tecnificación corporal de los dispositivos de regulación de la subjetividad son: la naturalización de la diferencia sexual mediante la patologiza- ción de la transexualidad y la intersexualidad; y la normalización de la heterosexuali- dad a través de la criminalización, persecución y patologización de las prácticas sexua- les consideradas perversas. A partir de este momento, en acuerdo con Preciado (2008), las diferencias corpo- rales y las prácticas sexuales generan identidad, es decir, la tecnificación corporal de los mecanismos que producen y regulan la subjetividad se articula a través del aparato identitario. Los dispositivos tecnopolíticos de subjetivación tienen sus raíces en la sociedad científica, industrial y colonial del siglo XIX y se harán visibles, en su configuración actual, tras la segunda guerra mundial con el desarrollo de la economía posfordista, donde el motor del capitalismo actual es el control farmacopornográfico de nuestra subjetividad (Preciado, 2008). Siguiendo los análisis de Preciado (2008) las tecnologías producidas en la segunda guerra mundial se van a transformar tanto en tecnologías de normalización y control del cuerpo, como en tecnologías de información, comunicación, mediatización y pro- ducción-consumo posfordista, como circuitos en donde nuestras corporalidades se re- suelven en el marco del capitalismo tardío. El funcionamiento de este conjunto de técnicas de guerra en el gobierno del cuer- po ya no es exclusivamente arquitectónico y jurídico, sino que, también es tecnológi- co, digital, cibernético, prostético, biomolecular, mercantil y mediático, pasando a transformar la misma materialidad y entrando dentro del propio cuerpo vivo (Precia- do, 2008). El cuerpo como ficción: de la biopolítica a la tecnopolítica En esta línea in- terpretativa, a través de la imaginería cyborg, la autora aborda los organismos, cuerpos 5 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías y subjetividades que habitan las actuales escenas posmodernas, regidos por tecnolo- gías del cuerpo de índole bien diferenciada a la modernidad. En este sentido, el cyborg se constituye como el paradigma que da cuenta de la codificación del cuerpo a través de las tecnologías de la información y la comunicación, y las biotecnologías. Para Haraway “el cyborg es nuestra ontología, nos otorga nuestra política” (1991/1995, p. 254), es decir, los cyborgs son los ciberorganismos resultantes de las con- temporáneas tecnologías de control y regulación, por lo que, también, son lugares es- tratégicos que desvelan, resisten y transgreden sistemas de dominación. Por tanto, la imaginería cyborg es subversiva, una potencia que Haraway (1990/1999) propone como política cyborg para otros inapropiados/bles. El cuerpo como ficción: de la biopolítica a la tecnopolítica 4 4 Lola Martínez Pozo En las sociedades disciplinarizadas teorizadas por Foucault (1976/1987) las tecno- logías biopolíticas de subjetivación regulaban el cuerpo desde el exterior, lo cuerpos habitaban las arquitecturas del poder. Sin embargo, en el régimen farmacopornográfi- co, las tecnologías de regulación corporal penetran y se transforman en cuerpo y sub- jetividad, la tecnopolítíca se incorpora, se hace carne. Por lo que el cuerpo del siglo XXI, “es una entidad tecnoviva multiconectada que incorpora tecnología” (Preciado, 2008, p. 39). Por tanto, siguiendo a Jaime del Val (2006b) las articulaciones del poder han expe- rimentado profundas transformaciones, donde las tecnologías son claves en los proce- sos de formalización, estandarización, asimilación y representación de los cuerpos en el mercado global. El advenimiento de dichas tecnologías ha desplazado al sistema biopolítico hacía nuevas formas de tecnopolítica, donde los cuerpos y lxs sujetos son discursivamente y tecnológicamente producidos, reproducidos y representados en base a categorías de estratificación y producción de subjetividad y sus márgenes, que ya no funcionan ex- clusivamente bajo lógicas esencialistas, sino, también, como instrumentos al servicio de la asimilación estandarizada (Del Val, 2006b). En este sentido, Del Val (2006a) denomina tecnopoder a esta derivación del biopo- der, donde el poder se conforma en torno a industrias, modelos e implantaciones tec- nológicas mediante estrategias de falso progreso y democratización, teniendo como consecuencia la estandarización de las sociedades y la aparición de nuevas formaliza- ciones del cuerpo. Así mismo, para abordar estas articulaciones entre cuerpos y tecnopoder en el marco de la tecnociencia recojo las aportaciones de Donna Haraway. En su trabajo Ciencia, cyborg y mujeres (1991/1995) no advertía de la urgencia feminista de interve- nir la tecnociencia. En este sentido, la autora desvela los discursos y prácticas tecno- científicas dirigidas a la construcción natural, estática y desigual del cuerpo, para desa- rrollar propuestas epistemológicas y políticas feministas, anticoloniales, antirracistas y anticlasistas orientadas a la reinvención de la naturaleza y a la reapropiación tecnoló- gica. Desde mi perspectiva, la deconstrucción crítica de la tecnociencia desarrollada por Haraway (1991/1995) continúa vigente y da cuenta de las contemporáneas tecnologías de regulación corporal. Sus análisis exceden las tecnologías del cuerpo contempladas por Foucault (1981/1990) como el conjunto de normas, mecanismos, técnicas, discipli- nas e instituciones que conforman y constituyen al sujeto moderno. Lecturas y prácticas posfeministas derivadas del cyborg Frente a la normalización tecnopolítica de los cuerpos desde las contemporáneas tec- nologías de subjetivación en las sociedades del capitalismo tardío emergen corrientes y prácticas de resistencia situadas en el marco de aquello ampliamente documentado como posfeminismos. Los posfeminismos implican un enfrentamiento y deconstrucción de los dualis- mos imperantes en los feminismos que les preceden. Asimismo, las perspectivas posfe- ministas suponen un giro político que se desplaza hacia la otredad, la diferencia, enfa- tizando ámbitos hasta el momento invisibilizados como son los cuerpos y las sexuali- dades no normativas (García, Moreno y Sánchez, 2004). Desde mi punto de vista, los posfeminismos, pero también lo queer y los feminis- mos decoloniales, han sido cruciales y el marco de referencia para la emergencia de enfoques y prácticas políticas que conectan cuerpos y tecnologías. Dichas corrientes han operado como una irrupción teórica, analítica y política desarrollando ejercicios críticos dirigidos a cuestionar y deconstruir los dualismos presentes en los feminismos occidentales, a articular las diversas categorizaciones sociales y ejes de poder que se inscriben en nuestros cuerpos y a conectar las diferentes opresiones y resistencias, te- niendo como consecuencia la desconstrucción del sujeto político feminista hegemóni- co y la apertura a otrxs subjetividades, ámbitos y prácticas políticas. En este contexto, desde la década de los setenta, la redefinición, reapropiación y politización del cuerpo ha adquirido una especial relevancia desde diversos ámbitos como las teorías y activismos queer, decoloniales, feministas y transfeministas. Las re- conceptualizaciones y posicionamientos en el cuerpo como ámbito político han su- puesto el cuestionamiento de los dispositivos implícitos del poder, así como la desarti- 6 Lola Martínez Pozo culación crítica de los procesos normativos de control, producción, naturalización y exclusión (Del Val, 2006a). Por otro lado, también en los años setenta, las teorizaciones y movimientos femi- nistas comenzaron a aproximarse a las tecnologías en el marco de las heterogéneas críticas feministas de la tecnociencia (Wajcman, 2009). No obstante, durante las déca- das de los setenta y ochenta, las perspectivas feministas de las tecnologías se caracteri- zaban por posturas pesimistas y de rechazo dado el carácter sexista, androcéntrico y patriarcal subyacente a las mismas (Romero, 2014; Vergés, 2013). Lecturas y prácticas posfeministas derivadas del cyborg Posteriormente, di- chas posturas, centradas en la relación género-tecnología y la exclusión de las mujeres, se vieron transformadas por las rupturas epistemológicas y políticas procedentes de las contribuciones de los feminismos queer, decoloniales y posfeministas, donde se produce una crítica a las tecnologías contemplando el potencial subversivo de las mis- mas (Wajcman, 2009). En este sentido, un punto de inflexión en las percepciones, narrativas y usos de las tecnologías por parte de los feminismos fue el Manifiesto para Cyborgs de Donna Haraway (1991/1995). La autora, a través de la figura del cyborg como parte humana y parte máquina, realiza, por un lado, una crítica a la tecnociencia y, por otro lado, pro- pone la política cyborg como una subjetividad potencialmente subversiva que no se construye desde las identificaciones tradicionales. Los ejercicios de deconstrucción, politización y articulación de los cuerpos y las tecnologías mediante la reapropiación de las mismas y a través de múltiples prácticas activistas han adquirido un peculiar protagonismo en las últimas décadas. Concreta- mente, desde los años noventa con el desarrollo de los ciberfeminismos, los tecnofemi- nismos y lo ciber-tecnoqueer, y más recientemente, desde las articulaciones transfemi- nistas con los ámbitos tecnológicos. Resituar el cuerpo en los ciberfeminismos En los años noventa las posibilidades creativas y políticas que ofrecía Internet consti- tuyeron el impulso para el desarrollo de los ciberfeminismos. Las visiones feministas de este ámbito tecnológico encontraron en las realidades on-line y en la red espacios e instrumentos para la articulación de nuevas formas de colectividad y activismo. Inter- net nos ofrecía a todxs la posibilidad de convertirnos en productorxs y de horizontali- zación. Todo ello constituyó una potencia política y revolucionaría que desde la hete- rogeneidad de las prácticas ciberfeministas se orientó a la deconstrucción de identida- des, y la desarticulación de las fronteras entre las esferas público-privada y produc- ción-reproducción (Zafra, 2011). 7 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías Las raíces teóricas de los ciberfeminismos las encontramos en la fusión entre Ha- raway, el feminismo francés de tercera ola y las corrientes posestructuralistas (Gallo- way, 1997). De igual modo, en las formas de abordar los cuerpos y las identidades “el ciberfeminismo entronca con el posfeminismo y las teorías posidentitarias y posgené- ricas” (Álvarez, 2014, p. 35). No obstante, desde el inicio de los ciberfeminismos el espacio cibernético fue des- corporeizado. Escapar de la carne, del cuerpo físico, del determinismo biológico asocia- do a las identidades, se constituían como frecuentes retóricas en las luchas feministas en la red (Romero, 2014). Tales narrativas las podemos encontrar en la autora ciberfeminista Sadie Plant (1997), quien conceptualizó Internet como un medio no jerarquizado, un espacio de li- beración para la construcción de identidades, para escapar del cuerpo, donde no existe el género. En su trabajo cuestiona la tecnofobia característica de los feminismos occidentales tradicionales, afirma que la tecnología es femenina y, a partir de la codificación binaria de Unos como masculinos y Ceros como femeninos y de las características de la red como la difusión, la flexibilidad, la descentralización y la horizontalidad, reivindica la creación de un mundo posgénero sin jerarquías (Plant, 1997). En el primer desarrollo de los ciberfeminismos, identificado por algunas investi- gadoras (Núñez, 2008; Vergés, 2013) como la primera ola ciberfeminista, la relación fe- minista con las tecnologías digitales se caracterizaba por la compresión neutral y utó- pica de las mismas, el dualismo entre cuerpo y tecnología y por una noción esencialis- ta y binaria de género. Por su parte, los tecnofeminismos emergieron en oposición a dichas posturas ci- berfeministas. Resituar el cuerpo en los ciberfeminismos Desde los tecnofeminismos se propone una relación de transformación crítica con las tecnologías, tomando como punto de partida los procesos de construc- ción mutuamente constitutivos entre género y tecnología, así como la reproducción de ideologías y prácticas desiguales de género en los ámbitos tecnológicos y en la red (Wajcman, 2004/2006). No obstante, a pesar de estas contribuciones críticas que permi- ten considerar los cuerpos generizados y la persistencia de jerarquías en relación con lo tecnológico, para Judith Wajcman (2004/2006), referente teórico de los tecnofemi- nismos, las tecnologías digitales y la red se conformaban como espacios descorporeiza- dos: “en el ciberespacio todas las señas físicas, corpóreas, se eliminan de la comunica- ción” (p. 104). 8 Lola Martínez Pozo Por el contrario, las aportaciones de Sandy Stone (1991), considerada referente en los activismos ciberfeministas y transgénero, han contribuido a enfatizar el cuerpo como un tema central para los ciberfeminismos. Su trabajo, lejos de considerar que el cuerpo y el género estén ausentes en las tecnologías, aborda cómo las tecnologías de la comunicación producen, representan y organizan cuerpos, identidades, comunidades y espacios mediante las interfaces. Han pasado casi tres décadas desde los inicios de los ciberfeminismos, desde el trabajo de las VNS Matrix y su Manifiesto ciberfeminista para el siglo XXI (1991)3, desde las primeras Internacionales Ciberfeministas4 y la publicación del Libro Ceros+Unos de Plant (1997). En el transcurso de este tiempo, los ciberfeminismos han sido protagonistas de múltiples debates fruto de las interpelaciones procedentes de los movimientos feminis- tas en relación a la utilidad política de los ciberfeminismos (Wilding, 2004), de las in- tersecciones ausentes en sus articulaciones con las tecnologías digitales (Salido, 2017), de las genealogías feministas decoloniales (Reyes, 2017), así como consecuencia de ejercicios críticos desarrollados dentro del propio ciberfeminismo (Núñez, 2008; Rever- ter, 2013). En este último sentido, en el marco de la segunda ola ciberfeminista, aparecen de- bates en torno al determinismo tecnológico y el esencialismo sexual activando postu- ras ciberfeministas que incluyen la multiplicidad del cuerpo en relación con lo tecnoló- gico. 4 La primera edición de la Internacional Ciberfeminista fue en 1997 con La X Documenta de Kassel, Alemania. 3 VNS Matrix fue un colectivo de artistas feministas que realizaban acciones políticas en la red y son un referente tanto para los ciberfeminismos como para la politización queer y transfeminista de las tecnologías.. Resituar el cuerpo en los ciberfeminismos De esta forma, las tecnologías digitales empiezan a ser abordadas desde una no- ción de género constructivista y no binaria (Vergés, 2013), se desarrollan narrativas y prácticas ciberfeministas que potencian la transformación crítica tanto del género como de la propia tecnología (Braidotti, 1996; Faulkner, 2001; Flanagan y Both, 2002), así mismo, emergen críticas que abordan la red como un espacio jerarquizado, sexista, racista, clasista y etnocéntrico (Reverter, 2001; Salido, 2017) e irrumpen debates que contemplan la articulación entre género, raza, tecnología y ciberfeminismos (Fernán- dez, Wilding y Writte, 2002). Precisamente las articulaciones que interrelacionaron raza, racismo, tecnologías y ciberfeminismos (Daniels, 2009) posibilitaron corporeizar el ciberespacio, la red y los ciberfeminismos. Las contribuciones procedentes de otras geografías políticas del co- nocimiento introdujeron el cuerpo desde la crítica decolonial yendo más allá de las crí- ticas ciberfeministas centradas en el género y abordando ámbitos excluidos como las 9 9 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías diferencias y desigualdades raciales (Galpin, 2002; Nakamura, 2002) y la crítica al ra- cismo y etnocentrismo característico de los feminismos occidentales de las tecnologías (Fernández, 2002; Gajjala, 1999; Kvasny, 2006; Wright, 2002). Así mismo, los enfoques decoloniales sobre los ciberfemismos (Reyes, 2017) abor- dan la producción y desarrollo de la tecnociencia, las tecnologías y las TIC en el marco de relaciones de poder entre el norte global y el sur global, las cuales tienen sus raíces en el colonialismo precedente y persisten debido a sus continuidades. En este contexto, las ciberfeministas del sur global atienden a las particularidades e intersecciones de las resistencias ciberfeministas de las áreas periféricas (Flores, 2012; Tapia, 2012) e impli- can revisiones epistemológicas y prácticas de ruptura con los imaginarios ciberfemi- nistas del norte global (Reyes, 2017). En este sentido, Roxana Reyes (2017) interpreta los ciberfeminismos de los ámbitos periféricos como la expresión contemporánea de la conciencia y política opositiva cyborg de Chela Sandoval. Resituar el cuerpo en los ciberfeminismos Por otro lado, en relación a las perspectivas ciberfeministas descorporeizadas so- bre la red como un espacio donde el género pueden suprimirse y superarse hacia un mundo posgénero (Zafra, 2014) y donde las identidades se escinden de los cuerpos, emergiendo el poscuerpo o no-cuerpo en Internet (Zafra, 2011), las reflexiones de Reme- dios Zafra (2008, 2011) acentúan cómo en la cultura-red se reproducen los modelos identitarios y corporales normativos, las representaciones y las desigualdades asocia- das a los mismos, a la par que se generan nuevas relaciones de poder limitando el ejer- cicio deconstructivo. Así mismo, en relación a las visiones ciberfeministas poscorporales y posidentita- rias en la red, se han sucedido re-elaboraciones que cuestionan aquellas perspectivas que sugieren la eliminación de las identidades y los cuerpos on-line, ya que implica su- primir las diferentes trayectorias y opresiones. Dicha deconstrucción conlleva la su- presión de la categoría social de visibilidad en los entornos on-line con una conse- cuente pérdida del potencial subversivo. Igualmente, desde ejercicios críticos ciberfe- ministas se ha constatado que prevalece el énfasis en las identidades como práctica subversiva para encontrar y construir comunidades, otras formas de comunicación y producción de conocimiento, como pueden ser colectivos y grupos feministas y lgtbiq (Daniels, 2009). Por otro lado, Diane Currier (2003) plantea otros enfoques en la intersección entre cuerpos y tecnologías, y aborda dicha articulación desde el concepto de ensamblajes de Deleuze y Guattari, ofreciendo posibilidades analíticas y políticas a los feminismos de las tecnologías para situar los cuerpos en la era tecnológica más allá de las lógicas identitarias. 10 Lola Martínez Pozo Las disidencias sexuales y de género: articulaciones tecnológicas queer Las disidencias sexuales y de género: articulaciones tecnológicas queer Junto con los feminismos, lo queer también se ha aproximado a los ámbitos tecnológi- cos. En este sentido, tecnoqueer es la orientación de las teorías y políticas queer hacía la tecnología mediante la relocalización y reconfiguración de los ámbitos abiertos por lo cyborg y lo ciberqueer, por lo que es la irrupción e intervención de las críticas, teo- rías y políticas queer en los estudios y ámbitos tecnoculturales (Chang, 2012). Desde mi punto de vista, las perspectivas tecnoqueer cuestionan, desvelan y resisten los sis- temas tecnológicos como estructuras de poder tecnonormativas, es decir, como estruc- turas y prácticas heterosexistas, cisnormativas y lgtbiqfóbicas, donde se conforman y reproducen construcciones interseccionales de sexo, género, sexualidad, cuerpo y tec- nología. En relación a las imbricaciones tecno-corpóreas, siguiendo a Edmond Chang (2012), las aproximaciones tecnoqueer nos permiten una inclusión e integración de los cuerpos y subjetividades en los circuitos tecnológicos: The technoqueer renders and articulates the ways technology mediates these identities, subjectivities, and embodiments, the ways ideas, selves, and bodies mediate technology, and our discourses and understandings of these forma- tions. By queering technology and moving queer theory toward technocul- ture, the tecnoqueer hopes to show how bodies, identities, and subjectivities are gendered, sexualized, raced, and technologized by how they are exten- ded, transformed, even contained by technologies. (p. 47) Así mismo, la orientación de las teorías y prácticas queer hacía la tecnocultura produce lecturas, imaginarios, representaciones y configuraciones corporales alternati- vas (Chang, 2012). Las raíces teóricas de lo tecnoqueer son identificadas con los planteamientos Cy- borg de Donna Haraway (1991/1995), en la medida en que dicha imaginería acoge la monstruosidad, la distorsión y la rareza. Así mismo, lo tecnoqueer encuentra su afini- dad en la figuración cyborg, dadas las potenciales acciones de lucha que sugiere la au- tora, las políticas cyborg. Katherine N. Hayles (1999) pone de manifiesto la conjunción tecnológica y discur- siva en la formación del cyborg y es, precisamente, en dicha articulación, donde lo tec- noqueer indaga, re-elaborando los límites y las posibilidades del cyborg inicialmente teorizado por Haraway. Las disidencias sexuales y de género: articulaciones tecnológicas queer Asimismo, Sandy Stone (1995) constituye un referente en la investigación sobre las huellas de lo tecnoqueer por sus análisis sobre las formas, a través de las cuales, las 11 11 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías tecnologías de la comunicación revelan a la par que desafían nuestras nociones tradi- cionales de identidad en relación al género, la sexualidad, la raza, etc. Y, por sus inda- gaciones, en torno a las posibilidades deconstructivas y alternativas de las tecnologías de la comunicación. En esta línea, el trabajo de la autora se convierte en un impulso y punto de partida para lo tecnoqueer a partir de sus planteamientos sobre lo cyborg como una figuración monstruosa en el límite de todas las fronteras. En este último sentido, en términos de desafío, transformación y subversión de las tecnologías como sistemas de producción y representación corporal-sexual normati- vos, es donde conectan el cyborg con lo queer, conformando aquello que se ha denomi- nado ciberqueer y, posteriormente, tecnoqueer, ampliando la vinculación de lo queer a todos los ámbitos tecnológicos y a la tecnocultura. En relación a lo ciberqueer, dicho concepto aglutina una amalgama de perspecti- vas, prácticas e indagaciones que emergen y se consolidan a lo largo de la década de los noventa. El cibersujeto apareció como una de las manifestaciones de lo queer en su paso de la realidad física a la realidad virtual (Tudor, 2012). Ciberqueer se presenta como un híbrido entre lo queer y lo ciber que se genera y vincula con prácticas diarias de personas y comunidades lgtbiq en la red, así como, con los estudios e investigacio- nes sobre tales realidades. En los ámbitos académicos, una de las primeras conceptualizaciones del término ciberqueer se identifica con el ensayo de Donald Morton (1995/2002). No obstante, el autor, desde perspectivas críticas hacia lo queer, el posmodernismo y desde posturas fóbicas hacia las tecnologías digitales, no reconoce el potencial político de subversión y transformación que la articulación queer y ciber pueden suponer. Por su parte, Nina Wakeford (2000) inaugura perspectivas más positivas de lo ci- berqueer, ya que reconoce el ciberespacio como un espacio queer —aunque no sola- mente queer— desmontando la convencional comprensión del ciberespacio como una realidad única y monolítica. Las disidencias sexuales y de género: articulaciones tecnológicas queer Desde el punto de vista de la autora, la conexión de lo queer con lo cyborg y el ciberespacio radica en las diferencias, la multiplicidad, en la dispersión, los solapamientos, superposiciones e inestabilidad. En este sentido, Wakeford (2000) plantea lo ciberqueer como espacios que, aun- que inmersos en ámbitos y prácticas tecnoculturales normativas, suponen lugares de resistencias y protestas mediante la construcción de comunidades, alianzas, políticas de representación y activismos conformados por gays, lesbianas, trans* y queer que desafían y confrontan la heteronormatividad. 12 12 Lola Martínez Pozo Los activismos ciberqueer, mediante la reapropiación de las tecnologías digitales como herramientas políticas, han desarrollado ejercicios y prácticas de visibilización, manifestación y representación de las sexualidades y géneros disidentes posibilitando el empoderamiento colectivo a través del uso de la red, ejercicios que también per- mean en la materialidad de las prácticas e identidades sexuales en las realidades off-li- ne (Landström, 2007). Así mismo, tales activismos han contribuido a superar y cuestionar la influencia de la heteronormatividad en perspectivas y análisis feministas sobre la relación entre género y tecnología, así como, en prácticas ciberfeministas (Landström, 2007). Por otro lado, en los estudios e investigaciones culturales en torno a las tecnolo- gías se han destacado ausencias en relación a temas y experiencias vinculadas con el ámbito de la sexualidad y lxs cibersujetos lgtbiq, de ahí la aparición y desarrollo de los denominados estudios ciberqueer orientados a visibilizar y resistir frente a dichas su- presiones (Wakeford, 2002). En esta línea, la compilación de The Cyberculture Reader, editada por David Bell y Bárbara Kennedy (2000) y la antología Queer online: Media Technology and Sexuality editada por Kate O´riordan y David Philips (2007), nos permiten recorrer los emergen- tes, y aún presentes, campos y temáticas de investigación en los estudios ciberqueer. Donde Wakeford (2002) destaca la identidad y auto-representación on-line; la produc- ción del espacio virtual queer; la construcción y gestión de comunidades on-line y re- des virtuales, y la transformación de las prácticas eróticas y sexuales a través de las nuevas tecnologías. Considero que la apropiación de las tecnologías digitales por parte de géneros y sexualidades no normativas, para visibilizar y conectar sus luchas, ha forzado a los ci- berfeminismos a incorporar las disidencias corporales en sus reflexiones. Las colectivos queer y trans se han nutrido del universo cyborg y se apropian del ciberespacio (Salido, 2017). No obstante, están inmersxs en una multiplicidad. Las disidencias sexuales y de género: articulaciones tecnológicas queer En la red se trasladan la pluralidad de prácticas feministas, los distintos colectivos lgtbiq, las di- ferentes posiciones de las movilizaciones y colectivos trans, las subjetividades y co- lectividades migrantes y racializadas, los activismos crip y por la diversidad funcional, lxs trabajdorxs sexuales, etc. Asimismo, la red propicia la expansión y continuidad de los feminismos “hegemónicos”, de las “tradicionales” controversias feministas en torno a lo trans, el trabajo sexual, las tecnologías reproductivas y gestación subrogada, el ra- cismo y el clasismo, entre muchas otras. 13 13 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías Desde mi punto de vista este panorama múltiple y conectado implica nuevos retos epistemológicos y políticos para la reflexión feminista y ciberfeminista actual en rela- ción a las fricciones, sinergias y articulaciones entre los diversos feminismos en red y las disidencias corporales. Abriendo cuerpos y tecnologías desde diálogos transfeministas El desarrollo de lo queer en el estado español, a lo largo de la década de los noventa, ha tenido su particularidades locales autoenunciándose como transfeminismos a partir de las Jornadas Estatales Feministas de Granada en 2009 y de los manifiestos transfemi- nistas de 2010 (Ideadestroyingmuros, 2009; Manifiesto Transfeminista- transfronterizo, s/f). Desde mi experiencia activista, los transfeminismos han posibilitado la articula- ción de narrativas y prácticas políticas entre comunidades, colectivos y luchas feminis- tas, bolleras, maricas, trans, de trabajadorxs sexuales, migrantes, racializadas, perso- nas con diversidad funcional, okupas, anticapitalistas, antiespecistas y personas con prácticas sexuales no normativas. La confluencia de todo ello ha dado lugar a toda una serie de activismos denomi- nados feminismos queer, feminismos posidentitarios, postporno, feminismos postpor- nopunk, transmaricabollo, transfeministas. Dichas denominaciones emplean nombres compuestos con una voluntad inclusiva, y prefijos como “trans” y “post” como una forma de cuestionar y ampliar estrategias de los propios discursos y políticas que los preceden (Sentamans, 2013). Toda una amalgama que encuentra su herencia en los fe- minismos radicales, lesbianos, queer, autónomos y autogestionados. A partir de las discusiones teóricas expuestas en lo referente a la articulación fe- minista de cuerpos y tecnologías, y tomando en consideración las contribuciones epis- temológicas y metodológicas de la etnografía feminista (Gregorio Gil, 2006, 2019), con- sideré adentrarme en los activismos transfeministas de Barcelona. Durante los años 2017 y 2018 participé en talleres, jornadas y realicé entrevistas. En este texto me centraré en las aportaciones de lxs activistas entrevistadxs, dado que sus discursos y experiencias me han posibilitado localizar una riqueza de enfoques y prácticas que, a mi modo de ver, suponen fructíferas y potentes conexiones tecno- corporales subversivas tanto en el plano digital como en el plano corporal de produc- ción subjetiva. En este marco, continuando con las críticas y cuestionamientos que los ciberfemi- nismos, lo queer y los feminismos decoloniales han realizado sobre las tecnologías, los 14 Lola Martínez Pozo transfeminismos también se han vinculado con los ámbitos tecnológicos y desarrollan reconceptualizaciones de las nociones tradicionales de las tecnologías entendiéndolas como herramientas que van más allá de lo tecno-maquínico y se extienden e incorpo- ran los cuerpos y sus intersecciones. En este sentido y en acuerdo con Lucia Egaña (2013) en la politización transfemi- nista de las tecnologías, los cuerpos, subjetividades y trayectorias se ubican en el cen- tro de los entramados de poder y resistencias. Abriendo cuerpos y tecnologías desde diálogos transfeministas Desde mi trabajo de investigación y a partir de las percepciones de algunas de las personas entrevistadas la vinculación entre transfeminismos y tecnologías no se co- rrespondería con un movimiento social en sí mismo sino, más bien, se trataría de una escena propiciada por determinados eventos, trayectorias, redes, colectivos y proyec- tos que abordan las conexiones entre disidencias sexuales, corporales y tecnológicas, así como entre feminismos y hacking (Lucía Egaña, entrevista personal, 16 de febrero de 2017; Alex Hache, entrevista personal, 14 de junio de 2017). Conectar la autonomía tecnológica con las disidencias corporales y feministas Durante los años setenta comienza la era del uso de las herramientas tecnológicas como una caja negra (Soria, 2016). La privatización y comercialización del software por parte de las industrias a través del código cerrado (Stallman, 2004) inaugura una etapa caracterizada por la generación de usuarixs dependientes y el incremento del po- der, del control y de la vigilancia (Soria, 2016). Por ello, en el marco de la cultura ha- cker, el movimiento open source (código abierto) y software libre se basan en abrir los códigos de programación y en la libertad de estudiarlos, modificarlos y distribuirlos (Stalllman, 2004). Las prácticas hacker se vinculan con la cultura libre y con la produc- ción de conocimiento de libre acceso y anticapitalista tanto en relación al software como al hardware. Es decir, los dispositivos que utilizamos también cuentan con un código cerrado de fabricación, por lo que el hardware libre consiste en que el código fuente de su diseño sea abierto, para poder modificarlo o replicarlo mediante compo- nentes y materiales disponibles, de recicle o bajo coste (Cabañez, 2016). A finales de la década de los noventa emergen espacios como hacklabs y hackerspaces para experi- mentar con software y hardware libre, potenciar el acceso libre a las tecnologías y a Internet, compartir conocimientos y generar proyectos colaborativos (Maxigas, 2012). Concretando en los hacklabs, estos son espacios de experimentación tecnológica que aparecen en Europa vinculados con movimientos anarquistas y autónomos frecuente- mente emplazados en espacios okupados (Maxigas, 2012). 15 15 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías Software libre, hardware libre y los espacios de experimentación potencian, por tanto, la autonomía tecnológica frente a la obsolescencia programada, la comercializa- ción y el control de las corporaciones industriales a menudo aliadas con las diferentes estructuras gubernamentales. En este sentido, como me compartía Klau Kinki, la vinculación de los transfemi- nismos con las tecnologías y las prácticas hacker se dirige a la apropiación y autono- mía tecnológica para generar y colectivizar nuestras propias herramientas, evitar la censura y la dependencia de estructuras tecnológicas hegemónicas y patriarcales (En- trevista personal, 1 de marzo de 2017). En esta línea, a continuación, recojo las aporta- ciones y proyectos de las personas entrevistadas en la medida que interpreto estable- cen conexiones entre la autonomía tecnológica y las disidencias feministas y corpora- les. Conectar la autonomía tecnológica con las disidencias corporales y feministas Durante el desarrollo de trabajo de campo me he aproximado a la autonomía tec- nológica feminista a partir de las aportaciones de Alex Hache y su trabajo con el colec- tivo Donestech (http://donestech.net/). Este proyecto desarrolla diversas actividades en el marco de la investigación, la formación y la creación ciberfeminista. Concretando en la formación, desde Donestech se realizan talleres en materia de producción audiovi- sual, software libre u otras herramientas alternativas dirigidos a mujeres, feministas y personas lgtbiq. Asimismo, trabajan herramientas y metodologías en torno a la priva- cidad y seguridad digital para contrarrestar las violencias on-line mediadas por las TIC (Alex Hache, entrevista personal, 14 de junio de 2017). Otro proyecto que refleja la articulación con la autonomía tecnológica fue Minipi- mer.tv (https://minipimer.tv/), del cual Lucía Egaña y Veronica Lahitte formaban parte. Este proyecto no se autodenomina transfeminista pero las trayectorias de algunas de sus integrantes se desarrollaban en vinculación con los transfeminismos (Lucía Egaña, entrevista personal, 16 de febrero de 2017). El colectivo empezó como un proyecto residente en Hangar trabajando a través de streaming con software libre y dando soporte para la transmisión en directo de dife- rentes eventos como La Muestra Marrana (Lucía Egaña, entrevista personal, 16 de fe- brero de 2017). Verónica Lahitte me compartió parte del proceso, ejes y metodologías de trabajo del colectivo Minipimer.tv. Donde además de streaming trabajaban la tecnología y el arte mediante la investigación, grupos de estudio, la realización de talleres, instalacio- nes y el desarrollo de hardware y software libre. Desde las experiencias compartidas por Verónica Lahitte , Minipimer se orientaba a hacer un cruce entre el espacio de la investigación y el espacio de las herramientas libres a través de grupos horizontales de 16 Lola Martínez Pozo trabajo, funcionando como un nexo donde se encontraban diversas trayectorias y pro- yectos. De manera más concreta, el diálogo con Verónica me amplió la reflexión sobre cuerpos y tecnologías libres, a partir de la experiencia del grupo de estudio, donde uno de los bloques abordados implicaba la investigación y producción de herramientas li- bres en relación a los cuerpos (Entrevista personal, 11 de marzo de 2017). Conectar la autonomía tecnológica con las disidencias corporales y feministas Por otro lado, Beka Iglesias me trasladaba su trabajo (https://geekshabeka.ho- tglue.me/) que explora la conexión cuerpo y tecnología en el marco de los transfemi- nismos y el hacking a través de la mediación tecnocultural, la activación y coordina- ción de eventos, las tecnologías autónomas, y el desarrollo de performances, instala- ciones interactivas y dispositivos que derivan de investigaciones colaborativas con tec- nologías libres y de código abierto (Entrevista personal, 30 de marzo de 2017). Centrándome en sus aportaciones en relación a la autonomía tecnológica y la me- diación tecnocultural, los diálogos con Beka Iglesias me han permitido compartir su trabajo donde la vinculación entre lo hacker, los feminismos y los transfeminismos se ha desarrollado de forma articulada desde el inicio de sus colaboraciones con diversos movimientos sociales, grupos políticos, redes o en los diversos eventos que ha coordi- nado. Tales diálogos me han trasladado a experiencias que conectan la autonomía tec- nológica y los feminismos mediante su trabajo de desarrollo de infraestructuras tecno- lógicas autónomas en diferentes grupos feministas. En este sentido de vincular los feminismos y la autonomía tecnológica, Beka Igle- sias me compartía su trabajo en relación a los servidores feministas, en tanto que im- pulsora y administradora de uno de ellos, Anarcha Server. Un servidor que dota de una infraestructura autónoma en Internet a proyectos políticos feministas y fue desarrolla- do en el primer encuentro Tranhackfeminista (Entrevista personal, 30 de marzo de 2017). Así mismo, a través de su trayectoria en Generatech o como coordinadora de di- versos eventos como Summerlabs, los encuentros con Beka Iglesias me hicieron parti- cipe de diferentes experiencias relacionadas con su trabajo en mediación tecnocultural. Ésta se dirige a generar espacios de intercambio, contaminación y contagio entre per- sonas y colectivos procedentes del ámbito de las tecnologías libres, por un lado, y per- sonas que trabajan otras temáticas o vinculadas a diferentes grupos feministas o movi- mientos sociales, por otro lado, con objeto de activar procesos de construcción colecti- va de conocimientos y conexiones entre proyectos políticos (Entrevista personal, 30 de marzo de 2017). 17 17 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías Además, este proceso de investigación me ha trasladado diversas experiencias de politización de espacios tecnológicos libres como hacklabs, debido a la predominante normatividad cisheteropatriarcal y a la ausencia de crítica a las jerarquías y relaciones de poder. Conectar la autonomía tecnológica con las disidencias corporales y feministas Politización de espacios que gira en torno a problematizar el género, la se- xualidad, el cuerpo, la raza, la clase, la capacidad, etc., más allá de la licencia y el códi- go libre. En este último sentido me he acercado a iniciativas que han activado otros espa- cios y laboratorios donde confluyen la experimentación tecnológica y corporal, la pro- ducción de conocimientos y el aprendizaje colectivo con los transfeminismos y con trayectorias de disidencia corporal, sexual y migrante. Al respecto, es un referente la trayectoria compartida por Klau Kinki (Entrevista personal, 1 de marzo de 2017) acerca de la activación y desarrollo de Mutanger. Laboratorio Tóxico de Alta Tensión (https:// mutangerlab.wordpress.com/) y Pechblenda, un hacklab transhackfeminista (http://pe- chblenda.hotglue.me/). Así mismo, Quimera Rosa es un laboratorio de investigación y experimentación sobre cuerpo y tecnología (http://quimerarosa.net/). Desde perspectivas posidentitarias y transfeministas, inspiradas en el cyborg de Haraway, me compartían su trabajo con el cuerpo y la producción de subjetividades mediante diferentes herramientas y pro- yectos transdisciplinares que conectan ciencia, arte, tecnología y hacking (Ce, entre- vista personal, 28 de marzo de 2017). Concretando en su trabajo con los talleres El cuerpo como instrumento sonoro post- género, este workshop es una propuesta para experimentar lo cyborg desde la mezcla de cuerpo y sonido. Para ello, a través de la construcción de un dispositivo electrónico (BodyNoise Amp) que genera sonido mediante el contacto y se emplaza en el cuerpo a modo de prótesis, experimentan y producen otras subjetividades resultado del aparato sonoro como extensión corporal. Desde estos talleres trabajan con metodologías DIY (Do it yourself), DIT (Do it together) y DIWO (Do it with others) para generar una re- lectura y experimentación con las tecnologías como producción de subjetividades no normativas (Ce, entrevista personal, 28 de marzo de 2017). Asimismo, su trabajo con electricidad y sonido deriva de los diálogos con el ámbito hacker y open source, de tal forma que el dispositivo BodyNoise AMP se desarrolla con tecnologías libres, recicla- bles y accesibles (Kina, entrevista personal, 28 de marzo de 2017). En esta línea de politización transfeminista de la autonomía tecnológica e inclu- sión de cuerpos no normativos, en agosto de 2014 se desarrolló el primer Transhackfe- mifest en la colonia ecoindustrial postcapitalista de Calafou (Barcelona). Conectar la autonomía tecnológica con las disidencias corporales y feministas Este encuen- tro Transhackfeminista (THF) fue resultado de las colaboraciones entre Pechblenda, 18 Lola Martínez Pozo Donestech, individualidades colectivas de Calafou y de las redes heredadas del Ecleptic Tech Carnival, donde se renombró y replanteó este evento sobre mujeres, género y tec- nologías para incluir otros cuerpos y subjetividades políticas (Klau Kinki, entrevista personal, 1 de marzo de 2017). A este primer THF le siguió el segundo encuentro Transhackfeminista en México bajo el nombre de “Llamado (Error 404) Tecnologías disidentes no encontradas” THF! Mx2015/pUEBLA, celebrado en julio de 2015 en el Centro Comunitario ADA de la ciu- dad de Puebla. Por su parte, la tercera edición del encuentro Transhackfeminista fue desarrollado del 8 al 22 de agosto del 2016 en Montreal. Desde la experiencia de Anamhoo, como participante en las dos primeras edicio- nes del transhackfemifest y organizadora de la segunda, tales encuentros transhackfe- ministas han contribuido a abordar la participación de mujeres, queer, trans*, migran- tes y racializadxs en los ámbitos tecnológicos y en el hacking. Así mismo, han poten- ciado la crítica a las narrativas hegemónicas y a las desigualdades presentes en las áreas tecnológicas en base al género, la sexualidad, la clase o la raza. Por otro lado, es- tos espacios han contribuido a generar infraestructuras tecnológicas autónomas, como los servidores feministas, y a la descolonización tecnológica (Anamhoo, entrevista per- sonal, 28 de febrero de 2017). Postpornografía: tecnologías del género y de la sexualidad En este proceso de investigación me he re-encontrado con narrativas, prácticas y re- presentaciones postpornográficas desde una nueva perspectiva que dialoga los cuerpos y las tecnologías, no exclusivamente porque las producciones postpornográficas están permeadas por diversos aparatos tecnológicos (audiovisuales, digitales, electro-mecá- nicos, prótesis, etc.), sino también porque los cuerpos, el género y las sexualidades son considerados, en sí mismos, como tecnológicos. En este último sentido que interpreta los cuerpos desde una perspectiva tecnoló- gica, en el marco de las prácticas postpornográficas, Lucía Egaña (entrevista personal, 16 de febrero de 2017) me compartía su trabajo como organizadora de La Muestra Ma- rrana, directora del documental Mi Sexualidad es una Creación Artística (2011) e inves- tigadora sobre las prácticas postpornográficas en Barcelona (Egaña, 2015. Trayectoria posteriormente reflejada en su trabajo Atrincheradas en la carne (Egaña, 2018). Desde su recorrido Lucía Egaña me trasladaba una lectura de las prácticas pos- tpornográficas como prácticas hacker que abren y desvelan los códigos normativos de género y sexualidad para generar múltiples y diversas reprogramaciones, constituyén- 19 19 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías dose el cuerpo como un ámbito de desprogramación (Entrevista personal, 16 de febre- ro de 2017). Precisamente, la trayectoria de Quimera Rosa comenzó como un trabajo sobre el propio cuerpo, la propia sexualidad y la subjetividad vinculándose con la emergencia del postporno y la escena transmaricabollo de Barcelona (Kina, entrevista personal, 28 de marzo de 2017). Parte del trabajo de Quimera Rosa se ha desarrollado desde la pos- tpornografía, en colaboración con otras redes, donde las prótesis, estética y prácticas BDSM y dispositivos electrónicos se hallaban presentes en numerosas performances y proyectos: como O-kaña (junto a Post-op, Mistress Liar y Dj Doroti,) Akelarre Cyborg (junto a Transnoise), SexosKeleton o Sexus 3 (Ce, entrevista personal, 28 de marzo de 2017). En el anterior contexto, emergería Post-op, un proyecto de investigación y experi- mentación con cuerpo y postpornografía (http://postop-postporno.tumblr.com/). Al respecto, Elena Urko me compartía el inició la trayectoria de Post-op como un grupo que mediante la performance llevaban al espacio público aquello que estaban experi- mentando en relación al género, la sexualidad, la identidad y el cuerpo. En este senti- do, el trabajo de Post-op ha estado orientado a desarrollar otros imaginarios, prácticas y representaciones en torno a sexualidades y cuerpos no normativos desde la propia experimentación corporal. Postpornografía: tecnologías del género y de la sexualidad Para ello han abordado la postpornografía desde la perfor- mance, la producción audiovisual y la realización de talleres (Entrevista personal, 18 de octubre de 2017). Biohacking: tecnologías corporales Otra fructífera articulación transfeminista entre cuerpos y tecnologías a la que me he aproximado a través mi proceso de investigación es el biohacking. El biohacking engloba numerosas prácticas que vinculan la biología con la ética hacker, desplazando la investigación y experimentación biotecnológica para reformu- lar la ciencia, la biología y la tecnología de forma colectiva. Los espacios de biohacking se organizan en torno a tecnologías y herramientas libres para abrir e intercambiar co- nocimientos, facilitar medios y recursos que permitan investigar colaborativamente y colectivamente a bajo coste, documentando, habitualmente, estos procesos y resulta- dos en plataformas web como wiki para que puedan ser replicadxs por otrxs (Pin, 2014). En este sentido, el proyecto de Transplant (http://quimerarosa.net/transplant/) im- pulsado por Quimera Rosa es un espacio de biohacking transdisciplinar donde, a partir 20 Lola Martínez Pozo de un trabajo de investigación y auto-experimentación corporal, se reflexiona y gene- ran transiciones entre lo humano y lo vegetal deconstruyendo las identidades y produ- ciendo otras subjetividades (Kina, entrevista personal, 28 de marzo de 2017). Trans- plant se materializa en varios tentáculos, uno de ellos es un proceso de modificación de subjetividad mediante la implantación intravenosa de clorofila en el cuerpo. Y otro, consiste en repensar el virus del papiloma humano, las sexualidades y cuerpos no nor- mativos mediante la apropiación y experimentación de conocimientos y técnicas bio- médicas, desvelando, replicando, difundiendo y abaratando la terapia fotodinámica para tratar los condilomas derivados del VPH (Ce, entrevista personal, 28 de marzo de 2017). Asimismo, este proyecto es colectivizado e implica la participación de otrxs per- sonas a través del trabajo en diferentes biolabs, los talleres y las performances. Por otro lado, Gynepunk (https://gynepunk.tumblr.com/) es un proyecto de bioha- cking derivado del laboratorio transhackfeminista Pechblenda, anteriormente mencio- nado, que desarrolla talleres nómadas dirigidos a descolonizar y hackear el cuerpo, la ginecología hegemónica y la biomedicina. Klau Kinki me compartía cómo llegó a Gy- nepunk a través de un trabajo de investigación de los orígenes de la ginecología mo- derna vertebrada sobre la mutilización, experimentación y explotación de tres esclavas negras, materializado en el proyecto AnarchaGland (https:// anarchagland.hotglue.me/). Ello, junto con la experiencia de violencia vivida en los es- cenarios de la ginecología biomédica, la impulsó, junto con Paula Pin, a activar este proyecto. Biohacking: tecnologías corporales Desde los talleres promovidos por Gynepunk se desarrollan conocimientos en relación a VPH, cáncer de útero, aborto, ETS, infecciones, etc., técnicas de docu- mentación, análisis y diagnóstico e instrumentos de laboratorio DIY y DIWO (Entre- vista personal, 1 de marzo de 2017). Para Klau Kinki, la metodología de estos talleres se basa en el conocimiento colec- tivo desde las experiencias corporales y en construir y hackear nuestras propias técni- cas. Asimismo, me compartía que la intencionalidad de estos talleres también es el contagio, es decir, que emerjan laboratorios DIT por todos los lugares para que estos conocimientos y técnicas sean accesibles y desarrollados por todxs, por migrantes, tra- bajadorxs sexuales, bolleras, mujeres, maricas, etc. (Entrevista personal, 1 de marzo de 2017). Conclusiones Las articulaciones entre cuerpos, tecnologías y poder en las ciber-tecno sociedades del capitalismo tardío derivan hacia nuevas formas de tecnopolítica. En este contexto, des- de la década de los setenta, los cuerpos y las tecnologías han sido protagonistas de 21 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías múltiples debates donde diversos feminismos han contribuido a su articulación crítica. Especialmente a partir de la imaginería y política cyborg, propuesta por Haraway, que se constituyó como un punto de inflexión para el desarrollo de posfeminismos orienta- dos a la lectura crítica y reapropiación tecnológica. Los emergentes ciberfeminismos desarrollaron articulaciones con las tecnologías desde una visión utópica, decorporeizada y centrada en el género. No obstante, los de- bates e interpelaciones han dado lugar a la incorporación de los cuerpos y sus inter- secciones sociales en la articulación política con la tecnología. Desde mi punto de vista, las críticas procedentes de las genealogías feministas de- coloniales y queer en relación a la exclusión e incorporación de otras subjetividades, corporalidades y sexualidades han supuesto retos para los ciberfeminismos llevándolos a explorar las intersecciones entre cuerpo y tecnología. A partir del anterior marco epistemológico y político he indagado en las conexio- nes transfeministas entre disidencias corporales y tecnológicas. En este sentido, inter- preto que éstas abren nuevos horizontes para las articulaciones entre cuerpos y tecno- logías y entre feminismos y hacking, las cuales identifico como retos a continuar in- vestigando desde los estudios sociales, tecnoculturales y feministas. Mi proceso de investigación, que en estas líneas he sintetizado destacando las aportaciones de activistas y proyectos, me ha activado cuestionamientos que me han supuesto revisar, repensar y ampliar qué entiendo por cuerpos y tecnologías, orientán- dome a percibir los cuerpos como códigos, como dispositivos, como tecnológicos, y a entender las tecnologías más allá de las máquinas. Tales aportaciones me han permiti- do aproximarme a enfoques y prácticas que generan nuevas articulaciones para ha- ckear las máquinas desde la autonomía tecnológica, para hackear la sexualidad desde la postpornografía y la subjetividad mediante el biohacking. Ello me ha llevado a desa- rrollar lecturas de los transfeminismos como prácticas hacker, en la medida en que in- terpreto que este tipo de experiencias, trayectorias y proyectos se orientan a abrir los cuerpos y las máquinas, desvelar cómo funcionan para generar otras formas de habi- tarlos, otras formas de subjetividad. Agradecimientos A Carmen Gregorio Gil, por su trabajo de revisión. A Alex Hache, Beka Iglesias, Klau Kinki, Anamhoo, Verónica Lahitte, Lucía Egaña, Elena Urko, Ce y Kina, por todas sus aportaciones. 22 Lola Martínez Pozo Referencias Referencias Álvarez, Marta (2014). Ciberfeminismo: prácticas postidentitarias. M Arte y cultura Visual, 11, 33-39. Recuperado de http://www.m-arteyculturavisual.com/2014/11/10/pdf-11-m- arteyculturavisual/ Bell, David & Kennedy, Barbara M (Eds.). (2000). The Cybercultures Reader. London and New York: Routledge. bell hooks; Brah, Avtar; Sandoval, Chela; Anzaldúa, Gloria; Levins Morales, Aurora; Kum-Kum, Bhavnani… Talpade Mohanty, Chandra (2004). Otras inapropiables. Feminismos desde las fronteras. Madrid: Traficantes de sueños. Braidotti, Rossi (1996). Cyberfeminism with a difference. Recuperado de https://rosibraidotti.com/publications/cyberfeminism-with-a-difference/ Butler, Judith (1993/2002). Cuerpos que importan. Sobre los límites materiales y discursivos del “sexo”. Buenos Aires: Paidós. Cabañez, Eurídice (2016). Hackers, software, hardware libre y trabajo colaborativo: la resistencia política del procomún frente al discurso capitalista. En Irene Soria (Coord.), Ética Hacker. Seguridad y Vigilancia (pp. 67-110). México D. F.: Universidad del Claustro de Sor Juana. Chang, Edmond (2012). Technoqueer: Re/Con/Figuring Posthuman Narratives. Tesis doctoral inedita, University of Washington, Washington. Recuperado de https://digital.lib.washington.edu/researchworks/handle/1773/22631? show=full Citro, Silvia (2010). La antropología del cuerpo y los cuerpos en-el-mundo. Indicios para una genealogía (in)disciplinar. En Silvia Citro (Coord.), Cuerpos plurales: antropología de y desde los cuerpos (pp. 17-58). Buenos Aires: Editorial Biblos. Córdoba, David; Sáez, Javier & Vidarte, Paco (Eds.). (2005/2007). Teoría Queer. Políticas Bolleras, maricas, Trans, Mestizas. Barcelona: Egales. Csordas, Thomas (1993). Somatic Modes of Attention. Cultural Anthropology, 8(2), 135- 156. https://doi.org/10.1525/can.1993.8.2.02a00010 Currier, Diane (2003). Feminist Technological Futures: Deleuze and Body/Technology Assemblages. Feminist Theory, 4(3), 321-338. Doi: t ttps://doi.org/10.1177/14647001030043005 Daniels, Jessie (2009). Rethinking ciberfeminism(s): Race, gender and embodiment. WSK's Women's Studies Quarterly, 37(1&2), 101-124. https://doi.org/10.1353/wsq.0.0158 De Lauretis, Teresa (1987/1989). Technologies of gender: essays on theory, film, and fiction. London: MacMillan. Del Val, Jaime (2006a). Cuerpos frontera. Imperios y resistencias en el pos- posmodernismo. Organicidades [nodo en línea]. Artnodes. Revista de Intersecciones entre artes, ciencias y tecnologías, 6, 31-43. Recuperado de https://dialnet.unirioja.es/descarga/articulo/2277316.pdf Del Val, Jaime (2006b) Situated Tékhne. Beyond the performative: metaformative bodies and the politics of technology in post-postmodernism. International 23 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías Journal of Performace Art and Digital Media, 2(2), 187-207. Recuperado de http://www.tandfonline.com/doi/pdf/10.1386/padm.2.2.187_1? needAccess=true Egaña, Lucía (2013). Tecnofeminismo. Apuntes para una tecnología transfeminista. En Miriam Solá & Elena Urko (Eds.), Transfeminismos. Epístemes, fricciones y flujos (pp. 313-323). Tafalla: Txalaparta. Egaña, Lucía (2015). Trincheras de carne. Una visión localizada de las prácticas postpornográficas en Barcelona. Tesis doctoral inédita, Universidad de Barcelona, Barcelona. Egaña, Lucía (2018). Atrincheradas en la carne. Lecturas en torno a las prácticas postpornográficas. Barcelona: Bellaterra. Esteban, Mari Luz (2004). Antropología del cuerpo. Género, Itinerarios Corporales, Identidad y Cambio. Barcelona: Bellaterra. Faulkner, Wendy (2001). The technology question in feminism: a view from feminist technology studies. Women’s studies International Forum, 24(1), 79-95. https:// doi.org/10.1016/S0277-5395(00)00166-7 Fausto-Sterling, Anne (2000/2006). Cuerpos sexuados. La política del género y construcción de la sexualidad. Barcelona: Melusina. Fernández, Maria (2002). Cyberfeminism, Racism, Embodiment. En María Fernández, Faith Wilding & Michelle Wrigth (Eds.), Domain Errors! Cyberfeminist Practices (pp. 29-44). New York: Autonomedia. Fernández, Maria; Wilding, Faith & Wrigth, Michelle (Eds.). (2002). Domain Errors! Cyberfeminist practices. New York: Autonomedia. Flanagan, Mary & Booth, Austin (2002). Reload. Rethinking women + cyberulture. Massachussetts: MIT press. Flores, Daisy (2012). Reflexiones sobre…Descubriendo mundos y compartiendo resistencias en línea. Boletín de GenderIT.org (Abril). Recuperado de http://www.genderit.org/es/node/3545/ Foucault, Michel (1979). Microfísica del poder. Madrid: la Piqueta. Foucault, Michel (1976/1987). Historia de la sexualidad. V. 1. La voluntad del saber. Madrid Siglo XXI. Foucault, Michel (1981/1990). Tecnologías del yo. Y otros textos afines. Barcelona: Paidos. García, Almudena; Moreno, Pilar & Sánchez, Jesús (2004). Las Nuevas Identidades de Género en el Marco del Siglo XXI: del Cyborg a las Identidades Queer. Revista de Antropología Experimental, 4, 1-15. Recuperado de http://revistaselectronicas.ujaen.es/index.php/rae/article/view/2086 Gajjala, Radhika (1999). “Third World” perspectives on cyberfeminism. Development in Practice, 9(5), 616-619. https://doi.org/10.1080/09614529952774 Galloway, Alex (1997). Un informe sobre ciberfeminismo: Sadie Plant y VNS Matrix. Un análisis comparativo. Mujeres en Red. El periódico feminista. ttps://doi.org/10.1177/14647001030043005 Recuperado de http://www.mujeresenred.net/spip.php?article1531Galpin, Vashti (2002). 24 24 Lola Martínez Pozo Women in Computing around the World. SIGSE Butlletin, 34(2), 94-100. https://doi.org/ 10.1145/543812.543839 Women in Computing around the World. SIGSE Butlletin, 34(2), 94-100. https://doi.org/ 10.1145/543812.543839 Women in Computing around the World. SIGSE Butlletin, 34(2), 94-100. https://doi.org/ 10.1145/543812.543839 Gregorio Gil, Carmen (2006). Contribuciones feministas a problemas epistemológicos de la disciplina antropológica: representación y relaciones de poder. AIBR Revista de Antropología Iberoamericana, 1(1), 22-39. https://doi.org/10.11156/aibr.010104 Gregorio Gil, Carmen (2019). Explorar posibilidades y potencialidades de una etnografía feminista. Disparidades, 74(1), e002a. https://doi.org/10.3989/dra.2019.01.002.01. Haraway, Donna (1990/1999). Las promesas de los monstruos: una política regeneradora para otros inapropiados/bles. Política y sociedad, 30, 121-163. Recuperado de https://dialnet.unirioja.es/servlet/articulo?codigo=154534 Haraway, Donna (1991/1995). Ciencia, cyborg y mujeres: la reinvención de la naturaleza. Madrid: Cátedra. Hayles, N. Katherine (1999). How We Become Posthuman. Virtual Bodies in Cybernetics, Literature and Informactics. Chicago: University of Chicago Press. Ideadestroyingmuros (2009, 10 de diciembre). Manifiesto para la insurrección transfeminista//manifesto per un'insurrezione transfemminista [Entrada de blog]. Ideadestroyingmuros. Recuperado de http://ideadestroyingmuros.blogspot.com/2009/12/manifiesto-para-la- insurreccion.html Kinki, Klau (2013). Ofensiva Transhackfeminista. Your Machine is a Battleground. En Miriam Solá & Elena Urko (Eds.), Transfeminismos. Epístemes, fricciones y flujos (pp. 305-307). Tafalla: Txalaparta. Kvasny, Lynete (2006). Let the Sisters Speak: Understanding InformationTechnology from the Standpoint of the ‘Other’. The DATA BASE for Advances in Information Systems, 37(4),13-25. https://doi.org/ 10.1145/1185335.1185342 Landström, Catharina (2007). Queering feminist technology studies. Feminist Theory, 8(1), 7-26. https://doi.org/10.1177/1464700107074193 Le Breton, David (2002). Antropología del cuerpo y modernidad. Buenos Aires: Nueva visión. Manifiesto Transfeminista- transfronterizo (s/f). Recuperado de https://sindominio.net/ karakola/IMG/pdf_Manifiestofinal2.pdf Maxigas (2012). Hacklabs and hackerspaces: tracing two genealogies. Journal of Peer Production, 2. Recuperado de http://peerproduction.net/issues/issue-2/peer- reviewed-papers/hacklabs-and-hackerspaces/ Morton, Donald (1995/2002). El nacimiento de lo ciberqueer. En Rafael Mérida Jiménez (Ed.), Sexualidades Transgresoras. Una Antología de Estudios Queer (pp. 11- 140). Barcelona: Icaria. Nakamura, Lisa (2002). Cybertypes: Race, Ethnicity, and Identity on the Internet. New York: Routledge. Núñez, Sonia (2008). From cyberfeminism to technofeminism: From an essentialist perspective to social cyberfeminism in certain feminist practices in Spain. 25 25 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías Women’s Studies International Forum, 31, 434-440. https://doi.org/10.1016/j.wsif.2008.09.005 O´riordan, Kate & Philips, David J. (2007). Queer online: Media Technology and Sexuality. New York: Peter Landg Publishers. Pin, Paula (2014). Biolab. Biohacking: Investigación científica como capacidad de performar la realidad. Una revisión transhackfeminista del hackeo de la ciencia. En Dossier Ritimo, Soberanía Tecnológica (pp. 91-95). Recuperado de: https://www.ritimo.org/IMG/pdf/dossier-st1-es.pdf Planella, Jordi (2006). Corpografías: dar la palabra al cuerpo. Artnodes. Revista de Intersecciones entre Artes, Ciencias y Tecnologías, 6, 13-23. https://doi.org/10.7238/a.v0i6.754 Plant Sadie (1997). Zeros + Ones. Women in Computing around the World. SIGSE Butlletin, 34(2), 94-100. https://doi.org/ 10.1145/543812.543839 Digital Wome + The New Techonoculture. London: Fourth Estate. Preciado, Paul (2008). Testo Yonqui. Madrid: Espasa. Preciado, Paul (2008). Testo Yonqui. Madrid: Espasa. Preciado, Paul (2014). Las subjetividades como ficciones políticas. En Hay Festival, Cartagena (Colombia), febrero 2014. Recuperado de https://www.youtube.com/watch?v=R4GnRZ7_-w4 Reverter, Sonia (2001). Reflexiones en torno al Ciberfeminismo. Asparkía: Investigación Feminista, 12, 35-49. Recuperado de http://www.raco.cat/index.php/Asparkia/ article/download/108578/154997 Reverter, Sonia (2013). Ciberfeminismo: de virtual a político. Teknokultura Revista de Cultura Digital y Movimientos Sociales, 10(2): 451-461. Recuperado de https://revistas.ucm.es/index.php/TEKN/article/view/51905 Reyes, Roxana (2017). Cyborgs en Centroamérica. Ciberespacio, políticas oposicionales y pensamiento feminista decolonial. Revista Interritórios, 3(4), 1-11. Recuperado de https://periodicos.ufpe.br/revistas/interritorios/article/view/22532 Romero, Amparo (2014). La utopía postfeminista: del ciberfeminismo al tecnofeminismo. Cuadernos del Ateneo, 32, 156-169. Recuperado de https://dialnet.unirioja.es/descarga/articulo/5241123.pdf Romero, Carmen; García, Silvia & Bargueiras, Carlos (Grupo de trabajo queer) (Eds.). (2005). El eje del mal es heterosexual. Figuraciones, movimientos y prácticas feministas queer. Madrid: Traficantes de sueños. Salido, Elena (2017). Ciberfeminismo. Disidencias corporales y género itinerante. REVELL, 3(17), 47-75. Recuperado de https://www.researchgate.net/publication/320111771 Sentamans, Tatiana (O.R.G.I.A.) (2013). Redes transfeministas y nuevas políticas de representación sexual (I) Diagramas de flujos. En Miriam Solá & Elena Urko (Eds.), Transfeminismos. Epístemes, fricciones y flujos (pp. 31-44). Tafalla: Txalaparta. Soria, Irene (2016). El software libre y la cultura hacker como vehículo para la emancipación tecnológica y su vínculo con la lucha feminista. En Irene Soria (Coord.), Ética Hacker. Seguridad y Vigilancia (pp. 183-216). México D. F.: Universidad del Claustro de Sor Juana. 26 Lola Martínez Pozo Stallman, Richard (2004). Software libre para una sociedad libre. Madrid: Traficantes de Sueños. Stone, Allucquère Rosanne “Sandy” (1991). Will the Real Body Please Stand Up?: Boundary Stories about Virtual Cultures. En Michael Benedikt (Ed.), Cyberspace: First Step, (pp. 81-118). Cambridge: MIT Press. Stone, Allucquère Rosanne “Sandy” (1995). The War of Desire and Technology at the Close of the Mechanical Age. Cambridge: MIT Press. Suárez, Liliana & Hernández, Rosalía (Eds). (2008). Descolonizando el feminismo: Teorías y prácticas desde los márgenes. Madrid: Cátedra Tapia, Danae (2012). Por un ciberfeminismo interseccional en América Latina. Revista Pillku, 18. Recuperado de https://pillku.org/article/por-un-ciberfeminismo- interseccional-en-america-la Tudor, Matilda (2010). Cyberqueer Techno-Practices. Digital Space-Making and Networking among Swedish Gay Men. Tesis doctoral inédita. The Department of Journalism, Media and Communication (JMK), Stockholm University. Stockholm, Swedish. Recuperado de: http://www.diva-portal.org/smash/get/diva2:532984/FULLTEXT01.pdf Turner, Bryan (1989). El cuerpo y la sociedad: exploraciones en la teoría social. México D. F.: Fondo de Cultura Económica. Turner, Bryan (1994). Avances recientes en la teoría del cuerpo. Women in Computing around the World. SIGSE Butlletin, 34(2), 94-100. https://doi.org/ 10.1145/543812.543839 Revista Española de Investigaciones Sociológicas, 68, 11-40. https://doi.org/10.2307/40183756 Vergés, Nuria (2013). Teorías principales de las tecnologías: Evolución y principales debates. Barcelona, Universidad de Barcelona. Recuperado de http://diposit.ub.edu/dspace/handle/2445/45624 Wajcman, Judith (2004/2006). El Tecnofeminismo. Madrid: Wajcman, Judith (2004/2006). El Tecnofeminismo. Madrid: Cátedra. Wajcman, Judith (2009). Feminist theories of technology. Cambridge Journal of Economics, 34(1), 143-152. https://doi.org/10.1093/cje/ben057 Wakeford, Nina (2000). Cyberqueer. En David Bell & Barbara Kennedy (Eds.), The Cybercultures Reader (pp. 403-415). London and New York: Routledge. Wakeford, Nina (2002). New technologies and cyber-queer research. En Diane Richardson & Steven Seidman (Eds.), Handbook of lesbian and gay studies (pp. 115-144). London: Sage Publications. Wilding, Faith (2004). ¿Dónde está el feminismo en el ciberfeminismo? Lectora: revista de dones y textualitat, 10, 141-151. Recuperado de http://www.raco.cat/index.php/Lectora/article/view/205482/284670 Wrigh, Michelle (2002). Racism, techcologies and the limit of the Western Knowledge. En Maria Fernández, Faith Wilding & Michelle Wrigth (Eds.), Domain Errors! Cyberfeminist practices (pp. 45-62). New York: Autonomedia. Zafra, Remedios (2008) Conectar-hacer-deshacer (los cuerpos). Zehar: Revista de Arteleku-ko aldizkaria, 64, 138-145. Recuperado de http://artxibo.arteleku.net/ es/islandora/object/arteleku%3A6023 Zafra, Remedios (2011). Un cuarto propio conectado. Feminismos y creación desde la esfera público-privada on-line. Asparkía Investigació Feminista, 22, 115-129. 27 27 Error en el sistema//Ctrl+Alt+Supr: Conectar cuerpos y tecnologías Recuperado de http://www.e-revistes.uji.es/index.php/asparkia/article/view/602/513 Zafra, Remedios (2014). Arte, Feminismo y Tecnología. Reflexiones sobre formas creativas y formas de domesticación. Quaderns de Psicologia, 16(1), 97-109. https://doi.org/10.5565/rev/qpsicologia.1212 Usted es libre para Compartir —copiar y redistribuir el material en cualquier medio o formato— y Adaptar el documen- to —remezclar, transformar y crear a partir del material— para cualquier propósito, incluso comercialmente, siempre que cumpla la condición de: Atribución: Usted debe reconocer el crédito de una obra de manera adecuada, proporcionar un enlace a la licencia, e in- di i h li d bi P d h l l i f bl d f l i i Usted es libre para Compartir —copiar y redistribuir el material en cualquier medio o formato— y Adaptar el documen- to —remezclar, transformar y crear a partir del material— para cualquier propósito, incluso comercialmente, siempre que cumpla la condición de: Atribución: Usted debe reconocer el crédito de una obra de manera adecuada, proporcionar un enlace a la licencia, e in- dicar si se han realizado cambios . Puede hacerlo en cualquier forma razonable, pero no de forma tal que sugiera que tie- ne el apoyo del licenciante o lo recibe por el uso que hace. Resumen de licencia - Texto completo de la licencia Resumen de licencia - Texto completo de la licencia 28
W2079702662.txt	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0083875&type=printable	en		The Personalized Advantage Index: Translating Research on Prediction into Individualized Treatment Recommendations. A Demonstration	PloS one	2,014	cc-by	6,338	The Personalized Advantage Index: Translating Research on Prediction into Individualized Treatment Recommendations. A Demonstration Robert J. DeRubeis1, Zachary D. Cohen1, Nicholas R. Forand2, Jay C. Fournier3, Lois A. Gelfand1, Lorenzo Lorenzo-Luaces1 1 Department of Psychology, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America, 2 Department of Psychiatry, The Ohio State University, Columbus, Ohio, United States of America, 3 Department of Psychiatry, University of Pittsburgh, Pittsburgh, Pennsylvania, United States of America Abstract Background: Advances in personalized medicine require the identification of variables that predict differential response to treatments as well as the development and refinement of methods to transform predictive information into actionable recommendations. Objective: To illustrate and test a new method for integrating predictive information to aid in treatment selection, using data from a randomized treatment comparison. Method: Data from a trial of antidepressant medications (N = 104) versus cognitive behavioral therapy (N = 50) for Major Depressive Disorder were used to produce predictions of post-treatment scores on the Hamilton Rating Scale for Depression (HRSD) in each of the two treatments for each of the 154 patients. The patient’s own data were not used in the models that yielded these predictions. Five pre-randomization variables that predicted differential response (marital status, employment status, life events, comorbid personality disorder, and prior medication trials) were included in regression models, permitting the calculation of each patient’s Personalized Advantage Index (PAI), in HRSD units. Results: For 60% of the sample a clinically meaningful advantage (PAI$3) was predicted for one of the treatments, relative to the other. When these patients were divided into those randomly assigned to their ‘‘Optimal’’ treatment versus those assigned to their ‘‘Non-optimal’’ treatment, outcomes in the former group were superior (d = 0.58, 95% CI .17—1.01). Conclusions: This approach to treatment selection, implemented in the context of two equally effective treatments, yielded effects that, if obtained prospectively, would rival those routinely observed in comparisons of active versus control treatments. Citation: DeRubeis RJ, Cohen ZD, Forand NR, Fournier JC, Gelfand LA, et al. (2014) The Personalized Advantage Index: Translating Research on Prediction into Individualized Treatment Recommendations. A Demonstration. PLoS ONE 9(1): e83875. doi:10.1371/journal.pone.0083875 Editor: William C. S. Cho, Queen Elizabeth Hospital, Hong Kong Received August 17, 2013; Accepted November 9, 2013; Published January 8, 2014 Copyright: ß 2014 DeRubeis et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Support provided by NIMH grant 2-R01-MH-060998-06, ‘‘Prevention of Recurrence in Depression with Drugs and CT’’ (http://www.nimh.nih.gov). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. E-mail: derubeis@psych.upenn.edu how to combine such predictive information, especially in cases in which the recommendations from multiple predictors conflict. As Meehl and colleagues have observed, actuarial approaches are preferred to clinical judgment in such cases [14], yet the potential for actuarial methods to inform personalized medicine by making prescriptive recommendations has not been realized. In 1996, Barber and Muenz introduced a ‘‘matching method’’ to mental health researchers, with data from a randomized comparison of two different psychotherapies, cognitive behavioral therapy and interpersonal therapy. Utilizing three pre-treatment variables (marital status, avoidant personality style, and obsessive personality style), they calculated for each patient a score on a ‘‘matching factor [15].’’ On average, patients with positive matching scores fared better in one of the two treatments, whereas those with negative scores fared better in the other. Based on these Introduction The call for an increased focus on ‘‘personalized medicine [1] ’’ is being met by efforts across medical fields to identify predictors of treatment response [2]. In mental health, this includes recent attempts to identify genetic [3–6] and neuroimaging [7–9] indices that predict differential response to pharmacological interventions. Variables from other domains (e.g., treatment history, course, comorbidities) that predict differential response to pharmacologic versus psychological treatments have also been identified [10–13]. Insofar as pretreatment patient characteristics predict differential response to the interventions, patient outcomes can be optimized by the systematic use of predictive information. Published reports of prescriptive relationships tend to be limited to examinations of single pre-treatment variables or of multiple variables that are each considered in isolation. Clinicians are left with little guidance as to PLOS ONE \| www.plosone.org 1 January 2014 \| Volume 9 \| Issue 1 \| e83875 Making Personalized Treatment Recommendations end-HRSD resulted in distributions of raw scores and residuals that did not differ from normality, allowing the use of the standard linear regression models [21]. The values we report from the models were squared so that they would be interpretable in terms of the original HRSD scale. findings, the authors recommended that clinicians consider these variables when deciding which of these treatments to recommend their patients. Their effort was a positive step towards personalizing treatment for depression, but neither their statistical approach nor the clinical recommendations it generated has been adopted by mental health researchers or practitioners. In this paper we illustrate an approach to the use of predictive information that builds upon Barber and Muenz’s efforts. The methods we describe produce point predictions of symptom severity at post-treatment for each individual in each of two interventions. The comparison of the two estimates yields an index, which we call the Personalized Advantage Index (PAI). The PAI identifies the treatment predicted to produce the better outcome for a given patient, and it provides the patient with a quantitative estimate of the magnitude by which that treatment is predicted to outperform the other. The utility of the approach is then tested by comparing the outcomes of those who had been randomly assigned to their indicated treatment versus those assigned to their non-indicated treatment. Selection of the variables to include in the models Nine variables were found to be either prognostic or prescriptive in our sample. The details concerning these findings can be found in three published works [10–12]. All nine variables were measured prior to randomization. Four of these were prognostic [22], in that they predicted end-HRSD scores irrespective of treatment. These were: 1) pre-treatment HRSD, where higher scores predicted higher end-HRSD scores; 2) Chronic versus Nonchronic course of major depressive disorder, where chronicity was associated with poorer outcome; 3) age, where older patients fared more poorly [12]; and 4) low (,100), middle (. = 100 and ,115), or high (. = 115) scores on the Shipley Institute of Living Scale, a brief measure of intellectual functioning [23], where higher scores predicted better outcomes. The other five variables were identified as prescriptive in that they predicted different outcomes depending on the treatment (ADM versus CBT) that was received. These variables were detected as a statistical interaction between that variable and treatment (ADM versus CBT): 1) presence (favoring ADM) versus absence (favoring CBT) of comorbid personality disorder [11]; 2) married or cohabiting (favoring CBT) versus single; 3) employed or not expected to work versus unemployed (favoring CBT); 4) number of stressful life events (more events favoring CBT) [12]; 5) number of prior antidepressant trials, capped at 2 trials (more trials favored CBT) [10]. Like any prescriptive variable, these characteristics also produced general effects on outcome, on average across treatments [24]. The direction of these effects was as follows: being married, employed, or having a higher number of life events predicted lower end-HRSD scores, whereas having a personality disorder or having had a larger number of prior medication attempts predicted higher end-HRSD scores. Descriptive statistics for the sample as a whole and for each treatment condition separately are provided in Table 1 for each of the nine predictive variables. There were no significant differences between ADM and CBT on any of the variables (t-test for continuous variables, chi-square for categorical variables; all p’s .0.1). Methods The approach we introduce and describe in this section of the paper can be used in any context in which patients have been randomized to two or more treatment conditions. For illustrative purposes, we use data drawn from a randomized comparative trial of cognitive behavioral therapy (CBT) versus the antidepressant medication (ADM) paroxetine in the treatment of outpatients with moderate to severe Major Depressive Disorder [16]. Each treatment was provided for 16 weeks. The trial was conducted at the University of Pennsylvania and Vanderbilt University during the period 1996 to 2002. The sampling method and outcomes have been described elsewhere [16,17]. The data are hosted at the University of Pennsylvania. The protocol for the study, titled ‘‘Cognitive Therapy and Pharmacotherapy in Major Depression,’’ was approved by the respective institutional review boards at the University of Pennsylvania, Philadelphia (Protocol #034900), and Vanderbilt University, Nashville, Tennessee (Protocol #7638). The data were de-identified before use in these analyses. Following the approval of an appropriate request, the data can be anonymized and provided to researchers. Written consent was given by the patients for their information to be stored in the university database and used for research. To simplify the presentation of our approach, we focus on data from the 154 patients for whom end-of-treatment scores were available, in either CBT (N = 50 of 60 assigned) or ADM (N = 104 of 120 assigned). End-of-treatment scores were calculated as the average of the final two scores (typically weeks 14 and 16) on the primary outcome measure, the 17-item version of the clinicianrated Hamilton Rating Scale for Depression (end-HRSD) [18]. The HRSD is the most commonly used assessment of depression symptom severity in depression treatment outcome research. In the present study, pre-treatment scores ranged from 20 to 36, where scores of 20 to 22 indicate ‘‘moderate’’ severity and higher scores indicate ‘‘severe’’ levels of depressive symptoms [19]. Differences of 3 or more points on the HRSD are considered to be ‘‘clinically significant [19].’’ In placebo-controlled randomized trials, medications tend to result in HRSD scores that are 2 to 3 points lower than placebo, on average, over the typical 4–8 week comparison period. This difference is associated with d-type effect size estimates of approximately 0.3 to 0.4 [20]. The end-HRSD scores in this sample were not normally distributed, which resulted in non-normal residuals when standard regression models were calculated. A square root transformation of PLOS ONE \| www.plosone.org Generation of the predicted end-HRSD scores We analyzed our data in MATLAB (The Mathworks Inc., Natick, MA). Using the GLMFIT procedure, we generated a prediction of the end-HRSD score for each participant in each of the two treatments. Hereafter we will refer to the prediction of the end-HRSD score for the treatment the participant actually received as the ‘‘factual prediction.’’ The ‘‘counterfactual prediction’’ was the estimate of the participant’s end-HRSD score in the treatment he or she did not receive. Both predictions were generated by the same model, in which end-HRSD was the dependent variable. To generate these predictions, we used techniques employed in leave-one-out cross-validation [25,26]. The leave-one-out procedure (also known as a jackknife [27]) required the creation of 154 models, each with a sample size of 153. Main effects for ‘‘Treatment’’ and the prognostic and prescriptive variables, as well as terms representing the interactions of Treatment and the prescriptive variables, served as independent variables. For each of the 154 patients, the factual prediction was calculated by entering the patient’s observed values on all of the independent variables into the prediction model. All values were centered using Kraemer et 2 January 2014 \| Volume 9 \| Issue 1 \| e83875 Making Personalized Treatment Recommendations Table 1. Descriptive statistics for baseline variables. Total (n = 154) ADM (n = 104) CBT (n = 50) Role Variable Mean or % SD Mean or % SD Mean or % SD Prognostic Intake-HRSD 23.8 3.2 23.8 3.2 23.7 3.4 Prognostic Chronic Subtype 55.2% — 58.7% — 48.0% — Prognostic Age 40.3 11.3 40.0 11.2 40.9 11.6 Prognostic IQ Lower IQ (IQ,100) 15.6% — 19.2% — 8.0% — Mid IQ (100, = IQ,115) 52.6% — 52.9% — 52.0% — Higher IQ (IQ. = 115) 31.8% — 27.9% — 40.0% — Prescriptive Married 37.7% — 39.4% — 34.0% — Prescriptive Employed 85.1% — 86.5% — 82.0% — Prescriptive Comorbid Personality Disorder 48.1% — 51.0% — 42.0% — Prescriptive Number Life Stressors Reported 6.6 4.8 6.7 5.1 6.3 4.3 Prescriptive Number Prior ADM Trialsa 0.7 0.8 0.7 0.8 0.8 0.8 ADM = Antidepressant Medication. CBT = Cognitive Behavioral Therapy. HRSD = Hamilton Rating Scale for Depression. a = Capped at 2; sample breakdown for number prior medications: 0 = 52% (55% in ADM, 46% in CBT), 1 = 24% (21% in ADM, 30% in CBT), 2 or more = 24% (24% in ADM, 24% in CBT). doi:10.1371/journal.pone.0083875.t001 al. ’s recommendations [24], whereby continuous measures were mean-centered, and dummy code values for dichotomous variables, including Treatment, were set at K and -K. We then computed each patient’s counterfactual prediction by substituting the value of the other treatment (either K or -K depending on the patient’s actual assignment) in the Treatment main effect term, as well as in all the terms representing the interactions of Treatment and the prescriptive variables. Because each model is estimated absent any information about the patient whose scores are to be predicted, the predictions are considered to contain little or no bias [25]. In essence, the accuracy of the set of predictions is what would be expected if the procedure had been used to predict outcomes in another set of patients who were drawn randomly from the same population of patients, assuming they would be assigned to the same treatments in the same way (i.e., randomly) [27]. A worked example of the approach Tables 2, 3, 4 illustrate how the procedure generated the predictions for CBT and ADM, using one of the 154 patients from the sample. This patient was selected because the PAI, the observed end-HRSD, and the prediction error were near the mean for the sample. Table 2 shows how this patient’s values on two of the four prognostic variables (low intake HRSD; high intellectual level) predicted better outcome (i.e., lower end-HRSD scores, as indicated by negative values of ab), whereas values on the other two prognostic variables (older; chronic course) predicted poorer outcome for this patient. As can be seen in the lower portion of Table 2, the patient’s values on three of the prescriptive variables (unmarried, unemployed, two prior ADM trials) predicted poorer outcome irrespective of treatment. On two others (three life stressors, no comorbid Personality Disorder), the values of ab are close to zero, indicating little influence on their own in the prediction of outcome. Table 3 shows how treatment affects the prediction of outcome, both as a main effect and in interactions with each of the five prescriptive variables. This patient’s values on three of the five prescriptive variables indicated CBT as the Optimal Treatment (unemployed, no comorbid Personality Disorder, two prior ADM trials) as reflected in the negative bc values. Values on the other two variables indicated ADM as the Optimal Treatment (unmarried, three life stressors), reflected in negative bm values. The model’s outputs (see Table 4) indicate that the patient’s predicted end-HRSD is 13.0 in CBT and 18.6 in ADM. The Personalized Advantage Index (PAI) for this individual is 5.6 in favor of CBT; it represents the difference between the endpoint scores predicted for each treatment. Properties of the predictions that will be examined Using the predicted scores, we estimated: (1) the ‘‘true error’’ of the factual predictions (i.e., the mean of the absolute value of the difference between the observed scores and factual predictions); (2) the standard error of the set of predictions; and (3) the magnitude of the predicted difference, for each patient, of receiving the treatment with the greater predicted benefit (Optimal) versus the other (Non-optimal) treatment. This last value is an index of ‘‘predicted advantage’’ which we call the Personalized Advantage Index (PAI). Because each individual is left out of the model from which their end point values are predicted, and because the Optimal treatment predicted for an individual is not tied to the treatment actually received, we can take advantage of the initial randomization of patients to treatments in order to test the utility of the PAI by comparing the mean observed difference, in endHRSD units, between the set of patients who had been randomly assigned to their Optimal treatment versus those who had been assigned to their Non-Optimal treatment. PLOS ONE \| www.plosone.org Results The true error of the end-HRSD score predictions (the average absolute difference between the predicted and actual scores, across the 154 patients) was 4.9. The standard error of prediction was 6.2. Figure 1 displays the distributions of the predicted end-HRSD scores for the Optimal and Non-Optimal treatments across the 154 patients. 3 January 2014 \| Volume 9 \| Issue 1 \| e83875 Making Personalized Treatment Recommendations Table 2. How the weights associated with prognostic and prescriptive variables combine with a patient’s values to contribute to the calculation of the patient’s Personalized Advantage Index. Variable Patient’s value Transformation Input value (a) Beta in LOO model (b) ab Intercept n/a n/a 1 3.15 3.15 Intake-HRSD (M = 23.8)a 20 Mean-centered 23.75 0.05 20.20 Age (M = 40.3)a 56 Mean-centered 15.67 0.01 0.20 IQ (Low, Middle, High)a High 21,0,1 1 20.18 20.18 Chronic Subtypea Yes 2.5, .5 0.5 0.39 0.19 Marital Statusb Unmarried 2.5, .5 20.5 20.45 0.22 Employment Statusb Unemployed 2.5, .5 20.5 20.50 0.25 Number Life Stressors Reported (M = 6.57)b 3 Mean-centered 20.65 20.07 0.04 Comorbid Personality Disorderb No 2.5, .5 20.5 0.17 20.08 Number Prior ADM Trials (capped at 2; M = 0.72)b 2 Mean-centered 1.28 0.28 0.35 Sum ab 3.96 Total, for use in end-HRSD predictionsc LOO = Leave One Out. ADM = Antidepressant Medication. HRSD = Hamilton Rating Scale for Depression. a = Prognostic variable. b = Prescriptive variable. c = See Table 4. doi:10.1371/journal.pone.0083875.t002 Optimal treatment. Given that in 40% of the sample the Optimal versus Non-Optimal difference was quite small, it is not surprising that the observed difference between the Optimal and NonOptimal means in the full sample was relatively small; they differed at the level of a nonsignificant trend (mean difference = 1.78; pooled SD = 6.38; t = 1.73, 152, p = .09; d = .28, 95% confidence interval 2.04 to .60). The right side of the figure gives the means for the 60% of the sample for whom the predicted advantage of the Optimal treatment was clinically significant. Here, the observed mean difference was both clinically and statistically significant (mean difference = 3.58; pooled SD = 6.12; t = 2.84, 90, p = .006; d = .58, 95% confidence interval .17 to 1.01). The distribution of PAI scores is shown in Figure 2. The average PAI was 4.2 (SD = 2.9), representing a 4.2 point difference in endHRSD scores between the Optimal treatment (predicted mean = 7.4, SD = 3.0) versus the Non-Optimal treatment (predicted mean = 11.6; SD = 3.9). Note that a patient’s PAI can be as low as 0, which would occur if the same outcome is predicted for both treatments, irrespective of whether high or low end-HRSD scores are predicted. As can be seen, whereas for some patients the predicted advantage of being assigned to their Optimal treatment was large, for others it was very small. For 62 (40%) of the patients, the PAI did not meet the National Institute for Health and Care Excellence (NICE) criterion (three points on the HRSD) for a ‘‘clinically significant’’ difference. For such patients, little weight would be given to the model’s predictions in a treatment selection decision; other factors (e.g., cost or patient preference) would likely be used to guide treatment. We test our approach, therefore, using the full sample of 154 patients as well as a reduced sample of those 92 patients (60%) whose PAI was ‘‘clinically significant.’’ The left side of Figure 3 shows, for the full sample, a comparison of the average end-HRSD score for those assigned randomly to their Optimal treatment versus those assigned to their Non- Discussion The method we have illustrated can be used to optimize treatment selection in any context in which: a) more than one intervention is under consideration, b) comparative outcome data are available, and c) pre-treatment factors can be identified that predict outcomes differentially across the interventions. In our example, a randomized comparison of cognitive behavioral Table 3. The treatment (Tx) main effect and interactions of Tx with the prescriptive variables. Tx = CBT Variable Beta in LOO model (b) Input (c) Tx = ADM bc Input (m) bm 0.21 CBT (0.5) or ADM (20.5) 20.42 0.5 20.21 20.5 TxMarital Status 21.10 20.25 0.27 0.25 20.27 TxEmployment Status 1.03 20.25 20.26 0.25 0.26 TxLife Stressors 20.35 20.32 0.11 0.32 20.11 TxPersonality Disorder 0.66 20.25 20.16 0.25 0.16 20.17 0.64 20.11 20.64 0.11 Sum bc 20.35 Sum bm 0.35 TxPrior ADMs Total, for use in end-HRSD predictions a LOO = Leave One Out. CBT = Cognitive Behavioral Therapy. ADM = Antidepressant Medication. Tx = Treatment. HRSD = Hamilton Rating Scale for Depression. a = See Table 4. doi:10.1371/journal.pone.0083875.t003 PLOS ONE \| www.plosone.org 4 January 2014 \| Volume 9 \| Issue 1 \| e83875 Making Personalized Treatment Recommendations treatments is likely to be large, as well as those for whom the predictions are similar and, thus, should not be given substantial weight in a choice between the two treatments. In applications of this approach, other factors, such as patient preference or treatment costs would likely weigh heavily in treatment selection decisions, when the PAI is small. It is important to emphasize that both ADM and CBT are evidence-based treatments for depression. Thus, all patients, including those identified as having received what for them was their Non-optimal treatment, received what is considered, absent any contraindications, a valid and appropriate treatment. Although we could not conduct a prospective test with our data, we approximated a critical feature of such a test by leaving each patient’s data out of the model that was used to make predictions for him or her. Thus, the benefits of treatment optimization we observed should provide a good estimate of the advantage that would have accrued to future patients from the same population had the prediction algorithm been used to assign them to the same treatments we studied. In a real world clinic, a consecutive series of patients would be randomized to one of two evidence-based treatments. Patient outcomes would be tracked, and baseline characteristics would be used to generate the predictive algorithm that would inform treatment decisions for future patients. The weight given to each new patient’s treatment recommendation would depend on the magnitude of the PAI generated by the algorithm. A true prospective test of our approach would begin with a randomized trial of two interventions. A predictive model, as we have described here, would be derived from the data obtained during the randomized trial. The model would then be tested in sample of patients who seek treatment in the same clinic in which the randomized trial was performed, using the same treatments. Outcomes of patients who are randomized to one of two conditions would then be compared: (a) those whose treatment is determined by random assignment, as in the first phase of the study; versus (b) those whose assignment is determined by the output of the predictive algorithm that was generated in the first phase. Table 4. How the estimates from Tables 2 and 3 combine to produce a patient’s estimated end-HRSD scores in CBT and ADM, respectively, and the PAI. Tx = CBT Value Source Value Tx = ADM 3.96 ,—Sum ab Table 2 Sum ab—. 3.96 20.35 ,—Sum bc Table 3 Sum b*m—. 0.35 3.6 Sum of sums 4.3 13.0 Predicted end-HRSDa 18.6 PAI = 5.6, favoring CBT Tx = Treatment. CBT = Cognitive Behavioral Therapy. ADM = Antidepressant Medication. HRSD = Hamilton Rating Scale for Depression. PAI = Personalized Advantage Index; the difference between the predictions for CBT and ADM, in end-HRSD units. a = The square of the model output. doi:10.1371/journal.pone.0083875.t004 therapy versus medications for depression, the treatments produced similar average levels of symptom reduction [16]. We used our approach to predict, for each patient, which treatment was more likely to lead to a better outcome. We then examined the results of the natural experiment that occurred whereby some patients had been randomized to their Optimal treatment and some to their Non-optimal treatment. In line with our hypothesis, patients randomized to their Optimal treatment tended to fare better than those who were randomized to their Non-optimal treatment. When we restricted our test of the method to those for whom the PAI was clinically significant, the advantage of assignment to the Optimal treatment was, in effect size terms, approximately twice the difference reported in a recent systematic review of antidepressant drug versus placebo comparisons [28], and larger than the average effect size observed between control and active treatments utilized in general medical contexts [29]. This result exemplifies an important feature of the approach: the ability to identify individuals for whom the difference in outcome between Figure 1. Frequency histogram showing predicted end-HRSD scores for each patient in their Optimal and their Non-Optimal treatment, as indicated by the treatment selection algorithm. doi:10.1371/journal.pone.0083875.g001 PLOS ONE \| www.plosone.org 5 January 2014 \| Volume 9 \| Issue 1 \| e83875 Making Personalized Treatment Recommendations Figure 2. Frequency histogram showing Personalized Advantage Index (PAI) scores for all patients in the sample. doi:10.1371/journal.pone.0083875.g002 It is often challenging to identify prescriptive variables that will replicate in a different population. Several features of the study from which the present data were drawn likely contributed to the strong prescriptive findings we obtained, and might also support a successful effort to replicate them. Cognitive behavioral therapy and antidepressant medications are both effective interventions for depression, but they are very different methods of treatment that likely work through different mechanisms [30]. We therefore expected to be able to identify prescriptive variables, especially given that several of the pre-treatment variables were included in Figure 3. Comparison of mean end-HRSD scores for patients randomly assigned to their Optimal treatment versus those assigned to their Non-Optimal treatment. The left side gives the results for the full sample. The right side includes only patients for whom the algorithm predicted a clinically significant advantage on the PAI of $3. doi:10.1371/journal.pone.0083875.g003 PLOS ONE \| www.plosone.org 6 January 2014 \| Volume 9 \| Issue 1 \| e83875 Making Personalized Treatment Recommendations the intake battery precisely because prior research had suggested that they predict differential response to these treatments. In comparisons of two treatments that work through similar mechanisms, such as might be true of two medications that operate on similar neurotransmitter systems, the power of this approach, or any approach that is contingent on the presence of significant treatment-by-patient-characteristic interaction effects, would likely be limited. The variables in our example comprised information from structured interviews, self-report questionnaires, and demographic forms, any of which can readily be obtained in a routine clinical setting. Other groups have begun to explore the potential of genetics or neuroimaging to inform treatment decisions in depressed patient populations [5,31–33]. In pharmacogenetic and pharmacogenomic studies, perhaps because the interventions included are mechanistically similar, the effects have thus far been small [6]. In principle, however, information from multiple different kinds of measures could be combined using the procedures we describe above in order to provide more accurate predictions than could be generated from any one predictor considered in isolation. The potential for neuroimaging-based treatment selection was evidenced recently in an investigation by Mayberg and colleagues, who explored the associations between pre-treatment brain activation and outcomes in a randomized comparison of CBT and ADM [32]. They reported that indexes of brain activity in six regions, as assessed with positron emission tomography, were associated with differential response to the two treatments. They focused on their strongest finding, which was obtained from the right anterior insula. Patients who remitted with CBT, as well as those who did not remit with ADM, exhibited relatively low activity in this region, whereas those who remitted in ADM, as well as those who did not remit in CBT, exhibited relatively high activity in that area. These findings represent a major contribution to prediction of treatment response. However, they examined each of the six indexes in isolation, and thus did not make maximal use of the predictive information provided from the multiple brain regions. Moreover, their approach does not allow for the quantification of benefit from treatment matching. As we have shown, some patients would be expected to derive comparable benefits from either treatment, whereas for others there would be little if any difference in outcomes expected between the two treatments. Considering these factors, it is not clear how their findings, or any set of findings in which multiple different predictors are identified, would be used in clinical decisionmaking on their own. Conversely, our approach produces a clinically interpretable index of the size of the expected difference in outcomes between the treatments. Future studies of neuroimaging or genetic markers as differential predictors of treatment response would do well to include a wide variety of variables and modalities in pre-treatment assessments and to take advantage of the multivariate nature of the set of potential predictors [34–37]. Biostatisticians have described analytic frameworks to identify prescriptive (moderator) variables [38,39], but less attention has been paid to the development of procedures to translate prescriptive findings into clear, actionable recommendations for individual patients. We were alerted to the points of contact between our approach and that of Barber and Muenz [15] while we were developing and testing our method, at which time a thorough review of the literature revealed no further developments along these lines in the mental health field. Only after an extensive review of the literature in other medical fields did we locate similar efforts, in oncological medicine [40–42]. To our knowledge, none of that prior work has been developed further or applied to the differential prediction of individual patient outcomes. The time is right for the revival, further development, and application of these methods, first introduced 35 years ago [40], as such approaches are suited perfectly to advance the goals of personalized medicine. With the present effort we hope to inspire renewed interest across medical fields in the development and application of prescriptive algorithms that combine multiple sources of information to yield estimates of patients’ outcomes in more than one treatment. This approach promises to enhance therapeutics by promoting the selection of the best treatment among available options, with the additional feature that it provides quantitative estimates of the benefits that can be expected when such an algorithm is implemented. Acknowledgments We would like to thank Steven Hollon for his encouragement throughout this project, and for his helpful and insightful feedback on several versions of our manuscript. Author Contributions Conceived and designed the experiments: RJD ZDC NRF JCF LG LLL. Analyzed the data: ZDC. Wrote the paper: RJD ZDC NRF JCF LG LLL. References 8. Gong Q, Wu Q, Scarpazza C, Lui S, Jia Z, et al. (2011) Prognostic prediction of therapeutic response in depression using high-field MR imaging. NeuroImage 55: 1497–1503. doi:10.1016/j.neuroimage.2010.11.079. 9. Baskaran A, Milev R, McIntyre RS (2012) The neurobiology of the EEG biomarker as a predictor of treatment response in depression. Neuropharmacology 63: 507–513. doi:10.1016/j.neuropharm.2012.04.021. 10. Leykin Y, Amsterdam JD, DeRubeis RJ, Gallop R, Shelton RC, et al. (2007) Progressive resistance to a selective serotonin reuptake inhibitor but not to cognitive therapy in the treatment of major depression. J Consult Clin Psychol 75: 267–276. doi:10.1037/0022-006X.75.2.267. 11. Fournier JC, DeRubeis RJ, Shelton RC, Gallop R, Amsterdam JD, et al. (2008) Antidepressant medications v. cognitive therapy in people with depression with or without personality disorder. The British Journal of Psychiatry 192: 124–129. doi:10.1192/bjp.bp.107.037234. 12. Fournier JC, DeRubeis RJ, Shelton RC, Hollon SD, Amsterdam JD, et al. (2009) Prediction of response to medication and cognitive therapy in the treatment of moderate to severe depression. J Consult Clin Psychol 77: 775–787. Available: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db = pubmed&cmd = Retrieve&dopt = AbstractPlus&list_uids = 19634969. 13. Jarrett RB, Schaffer M, McIntire D, Witt-Browder A, Kraft D, et al. (1999) Treatment of atypical depression with cognitive therapy or phenelzine: a doubleblind, placebo-controlled trial. Arch Gen Psychiatry 56: 431–437. doi:10.1001/ archpsyc.56.5.431. 1. Hamburg MA, Collins FS (2010) The Path to Personalized Medicine. New England Journal of Medicine 363: 301–304. doi:10.1056/NEJMp1006304. 2. Squassina A, Manchia M, Manolopoulos VG, Artac M, Lappa-Manakou C, et al. (2010) Realities and expectations of pharmacogenomics and personalized medicine: impact of translating genetic knowledge into clinical practice. Pharmacogenomics 11: 1149–1167. doi:10.2217/pgs.10.97. 3. Schosser A, Kasper S (2009) The role of pharmacogenetics in the treatment of depression and anxiety disorders. International Clinical Psychopharmacology 24: 277–288. doi:10.1097/YIC.0b013e3283306a2f. 4. Simon GE, Perlis RH (2010) Personalized medicine for depression: can we match patients with treatments? Am J Psychiatry 167: 1445–1455. Available: http:// eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom = pubmed&id = 20843873 &retmode = ref&cmd = prlinks. 5. McClay JL, Adkins DE, Åberg K, Stroup S, Perkins DO, et al. (2011) Genomewide pharmacogenomic analysis of response to treatment with antipsychotics. Mol Psychiatry 16: 76–85. doi:10.1038/mp.2009.89. 6. Malhotra AK, Zhang J-P, Lencz T (2012) Pharmacogenetics in psychiatry: translating research into clinical practice. Mol Psychiatry 17: 760–769. doi:10.1038/mp.2011.146. 7. Guo Y, DuBois Bowman F, Kilts C (2008) Predicting the brain response to treatment using a Bayesian hierarchical model with application to a study of schizophrenia. Hum Brain Mapp 29: 1092–1109. doi:10.1002/hbm.20450. PLOS ONE \| www.plosone.org 7 January 2014 \| Volume 9 \| Issue 1 \| e83875 Making Personalized Treatment Recommendations 29. Leucht S, Hierl S, Kissling W, Dold M, Davis JM (2012) Putting the efficacy of psychiatric and general medicine medication into perspective: review of metaanalyses. The British Journal of Psychiatry 200: 97–106. doi:10.1192/ bjp.bp.111.096594. 30. DeRubeis RJ, Siegle GJ, Hollon SD (2008) Cognitive therapy versus medication for depression: treatment outcomes and neural mechanisms. Nat Rev Neurosci 9: 788–796. doi:10.1038/nrn2345. 31. Uhr M, Tontsch A, Namendorf C, Ripke S, Lucae S, et al. (2008) Polymorphisms in the Drug Transporter Gene ABCB1 Predict Antidepressant Treatment Response in Depression. Neuron 57: 203–209. doi:10.1016/ j.neuron.2007.11.017. 32. McGrath CL, Kelley ME, Holtzheimer PE, Dunlop BW, Craighead WE, et al. (2013) Toward a Neuroimaging Treatment Selection Biomarker for Major Depressive Disorder. JAMA Psychiatry: 1–9. doi:10.1001/jamapsychiatry.2013.143. 33. Ising M, Lucae S, Binder EB, Bettecken T, Uhr M, et al. (2009) A genomewide association study points to multiple loci that predict antidepressant drug treatment outcome in depression. Arch Gen Psychiatry 66: 966–975. doi:10.1001/archgenpsychiatry.2009.95. 34. Arranz MJ, Kapur S (2008) Pharmacogenetics in Psychiatry: Are We Ready for Widespread Clinical Use? Schizophrenia Bulletin 34: 1130–1144. doi:10.1093/ schbul/sbn114. 35. Perlis RH (2011) Translating biomarkers to clinical practice. Mol Psychiatry 16: 1076–1087. doi:10.1038/mp.2011.63. 36. Uher R (2011) Genes, environment, and individual differences in responding to treatment for depression. Harvard review of psychiatry 19: 109–124. doi:10.3109/10673229.2011.586551. 37. Dunlop BW, Binder EB, Cubells JF, Goodman MG, Kelley ME, et al. (2012) Predictors of Remission in Depression to Individual and Combined Treatments (PReDICT): Study Protocol for a Randomized Controlled Trial. Trials 13. doi:10.1186/1745-6215-13-106. 38. Kraemer HC, Wilson GT, Fairburn CG, Agras WS (2002) Mediators and moderators of treatment effects in randomized clinical trials. Arch Gen Psychiatry 59: 877. 39. Kraemer HC, Frank E, Kupfer DJ (2006) Moderators of treatment outcomes: clinical, research, and policy importance. JAMA 296: 1286–1289. doi:10.1001/ jama.296.10.1286. 40. Byar DP, Corle DK (1977) Selecting optimal treatment in clinical trials using covariate information. J Chronic Dis 30: 445–459. 41. Byar DP (1985) Assessing apparent treatment—covariate interactions in randomized clinical trials. Stat Med 4: 255–263. doi:10.1002/sim.4780040304. 42. Yakovlev A, Goot RE, Osipova TT (1994) The choice of cancer treatment based on covariate information. Stat Med 13: 1575–1581. doi:10.1002/ sim.4780131508. 14. Dawes RM, Faust D, Meehl PE (1989) Clinical versus actuarial judgment. Science 243: 1668–1674. 15. Barber JP, Muenz LR (1996) The role of avoidance and obsessiveness in matching patients to cognitive and interpersonal psychotherapy: Empirical findings from the Treatment for Depression Collaborative Research Program. J Consult Clin Psychol 64: 951. 16. DeRubeis RJ, Hollon SD, Amsterdam JD, Shelton RC, Young PR, et al. (2005) Cognitive therapy vs medications in the treatment of moderate to severe depression. Arch Gen Psychiatry 62: 409–416. doi:10.1001/archpsyc.62.4.409. 17. Hollon SD, DeRubeis RJ, Shelton RC, Amsterdam JD, Salomon RMR, et al. (2005) Prevention of relapse following cognitive therapy vs medications in moderate to severe depression. Arch Gen Psychiatry 62: 417–422. Available: http://www.ncbi. nlm.nih.gov/entrez/query.fcgi?db = pubmed&cmd = Retrieve&dopt = AbstractPlus &list_uids = 15809409. 18. Hamilton M (1960) A rating scale for depression. J Neurol Neurosurg Psychiatry 23: 56–62. doi:10.1016/j.jad.2012.02.042. 19. National Institute for Clinical Excellence (2004) Depression: Management of Depression in Primary and Secondary Care. London, England: National Institute for Clinical Excellence. 20. Kirsch I, Deacon BJ, Huedo-Medina TB, Scoboria A, Moore TJ, et al. (2008) Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration. PLoS Med 5: e45–e45. doi:10.1371/ journal.pmed.0050045. 21. Draper NR, Smith H (2003) Applied regression analysis. Singapore: John Wiley and Sons. 22. Hollon SD, Beck AT (1986) Predicting outcome versus differential response: Matching clients to treatment. Rockville, MD. 23. Zachary RA, Western Psychological Services Firm (1991) Shipley Institute of Living Scale. Los Angeles, CA: WPS, Western Psychological Services. 24. Kraemer H, Blasey C (2004) Centring in regression analyses: a strategy to prevent errors in statistical inference. International Journal of Methods in Psychiatric Research 13: 141–151. 25. Efron B, Gong G (1983) A leisurely look at the bootstrap, the jackknife, and cross-validation. The American Statistician 37: 36–48. 26. Harrell FE, Lee KL, Mark DB (1996) Tutorial in biostatistics multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med 15: 361–387. Available: http://www.unt.edu/rss/class/Jon/MiscDocs/Harrell_1996.pdf. 27. Abdi H, Williams LJ (2010) Jackknife. In Salkind NJ, Dougherty DM, Frey B, editors. Encyclopedia of Research Design. Thousand Oaks, CA: Sage Publications, Incorporated. pp. 655–60. 28. Turner EH, Matthews AM, Linardatos E, Tell RA, Rosenthal R (2008) Selective Publication of Antidepressant Trials and Its Influence on Apparent Efficacy. N Engl J Med 358: 252–260. PLOS ONE \| www.plosone.org 8 January 2014 \| Volume 9 \| Issue 1 \| e83875
https://openalex.org/W4312102087	https://journal.unpar.ac.id/index.php/JurnalIlmiahHubunganInternasiona/article/download/5255/3736	English	null	UPHOLDING MULTILATERALISM: INDONESIA’S FOREIGN POLICY IN RESPONDING TO COVID-19 PANDEMIC	null	2,022	cc-by	18,189	https://doi.org/10.26593/jihi.v18i1.5255.29-54 azharisetiawan@umrah.ac.id azharisetiawan@umrah.ac.id ABSTRACT The 2019 Novel Coronavirus disease or Covid-19 has apparently become a new global challenge. Not only did the pandemic drive all actors to make response but it also affected the relations among them. That Indonesia raised multilateralism in the unprecedented situation while more unilateral or populist actions taken by a number of states encouraged this research. This paper attempts to explain Indonesia’s foreign policy in upholding multilateralism to respond to the COVID-19 during the first year of the pandemic. Such response was intended to mitigate the impacts caused by the pandemic. This research applied holistic constructivism in understanding the determinants of Indonesia’s foreign policy by investigating both domestic and international cause. This research utilized the qualitative method with an explanatory analysis. The findings show that such Indonesia’s foreign policy was driven by its identity constructed by both indigenous norm of ‘Gotong Royong’ and global norm of ‘International Health Regulation’. The norm-laden or identity-based foreign policy was leading it to uphold multilateralism which was considered appropriate in order to coordinate, collaborate and cooperate with international communities. In addition, Indonesia maintained its trust on and support to the World Health Organization as the most leading actor in health governance championing fight against the pandemic. This paper argues that the norm factors do matter in Indonesia’s foreign policy in facing uncertainties in the vulnerable and interconnected world. Through the case studied, this paper suggests that looking at the domestic actors as well as the state in international system help provide a better understanding on the state behavior in international relations. Keywords: Gotong royong; International Health Regulation; multilateralism; constructivist. Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic Luerdi1 & Azhari Setiawan2 1Doctoral Student in International Relations, Faculty of Law and International Relations, Universiti Sultan Zainal Abidin, Malaysia luerdi2202@gmail.com. 2 Department of International Relations, Faculty of Social and Political Science, Universitas Maritim Raja Ali Haji, Indonesia azharisetiawan@umrah.ac.id 1Doctoral Student in International Relations, Faculty of Law and International Relations, Universiti Sultan Zainal Abidin, Malaysia luerdi2202@gmail.com. 2 Department of International Relations, Faculty of Social and Political Science, Universitas Maritim Raja Ali Haji, Indonesia azharisetiawan@umrah.ac.id 1 Arry Bainus and Junita Budi Rachman, “Editorial: Pandemi Penyakit Menular (Covid-19) Hubungan Internasional,” Intermestic: Journal of International Studies; Vol 4 No 2 (2020)DO - 10.24198/Intermestic.V4n2.1 , May 31, 2020, 111, http://intermestic.unpad.ac.id/index.php/intermestic/article/view?path=. 2 Christine Riley, Bo Xie, and Anjum Khurshid, “Challenges Encountered in Comparing International Policy Responses to COVID-19 and Their Effects,” Health Research Policy and Systems 19, no. 1 (2021): 134, https://doi.org/10.1186/s12961- 021-00783-1. ( ) , p g 6 Jeff Tollefson, “How Trump Damaged Science — and Why It Could Take Decades to Recover,” Nature, 2020, https://www.nature.com/articles/d41586-020-02800-9. ABSTRAK Wabah penyakit Novel Coronavirus 2019 atau Covid-19 telah menjadi tantangan global baru. Pandemi Covid-19 tidak hanya mendorong semua aktor untuk merespon, tetapi juga mempengaruhi hubungan di antara mereka. Bahwa Indonesia mengangkat multilateralisme pada sebuah situasi yang belum pernah terjadi sebelumnya sementara tindakan yang lebih sepihak atau populis dilakukan oleh sejumlah negara lain mendorong penulis untuk melakukan penelitian ini. Tulisan ini menjelaskan kebijakan luar negeri Indonesia dalam mengedepankan multilateralisme untuk merespons COVID-19 selama tahun pertama pandemi tersebut. Respons tersebut dimaksudkan untuk memitigasi dampak yang ditimbulkan oleh pandemi. Penelitian ini menerapkan konstruktivisme holistik dalam memahami determinan politik luar negeri Indonesia dengan menyelidiki faktor domestik dan internasional. Penelitian ini menggunakan metode kualitatif dengan analisis eksplanatori. Temuan menunjukkan bahwa kebijakan luar negeri Indonesia tersebut didorong oleh identitasnya yang dikonstruksikan oleh norma domestik ‘gotong royong’ dan norma global ‘Peraturan Kesehatan Internasional’. Politik luar negeri yang sarat norma atau berbasis identitas mengarahkannya untuk mengedepankan multilateralisme yang dianggap tepat dalam rangka berkoordinasi, berkolaborasi dan bekerjasama dengan masyarakat internasional. Selain itu, Indonesia mempertahankan kepercayaan dan Wabah penyakit Novel Coronavirus 2019 atau Covid-19 telah menjadi tantangan global baru. Pandemi Covid-19 tidak hanya mendorong semua aktor untuk merespon, tetapi juga mempengaruhi hubungan di antara mereka. Bahwa Indonesia mengangkat multilateralisme pada sebuah situasi yang belum pernah terjadi sebelumnya sementara tindakan yang lebih sepihak atau populis dilakukan oleh sejumlah negara lain mendorong penulis untuk melakukan penelitian ini. Tulisan ini menjelaskan kebijakan luar negeri Indonesia dalam mengedepankan multilateralisme untuk merespons COVID-19 selama tahun pertama pandemi tersebut. 29 30 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic dukungannya kepada Organisasi Kesehatan Dunia (WHO) sebagai aktor utama dalam tata kelola kesehatan untuk menghadapi pandemi. Tulisan ini berargumen bahwa faktor norma penting dalam menganalisis kebijakan luar negeri Indonesia menghadapi ketidakpastian di dunia yang rentan dan saling terhubung. Penulis menyarankan bahwa dengan melihat aktor domestik dan juga negara dalam sistem internasional, akan membantu memberikan pemahaman yang lebih baik dan komprehensif tentang perilaku negara dalam hubungan internasional. Kata Kunci: Gotong royong, Regulasi Kesehatan Internasional, multilateralisme, konstruktivis 3 Riley, Xie, and Khurshid. Segun Oshewolo and Agaptus Nwozor, COVID 19: Projecting the National Security Dimensions of Pandemics, Strategic Analysis 44, no. 3 (May 3, 2020): 269, https://doi.org/10.1080/09700161.2020.1767911. 5 C M Weible et al., “COVID-19 and the Policy Sciences: Initial Reactions and Perspectives,” Policy Sciences 53, no. 2 (2020): 237, https://doi.org/10.1007/s11077-020-09381-4. y , ( y , ) , p g 5 C M Weible et al., “COVID-19 and the Policy Sciences: Initial Reactions and Perspectives,” Policy Sciences 53, n (2020): 237, https://doi.org/10.1007/s11077-020-09381-4. y, , egun Oshewolo and Agaptus Nwozor, “COVID-19: Projecting the National Security Dimensions of Pandemics,” Stra alysis 44, no. 3 (May 3, 2020): 269, https://doi.org/10.1080/09700161.2020.1767911. Arry Bainus and Junita Budi Rachman, “Editorial: Pandemi Penyakit Menular (Covid-19) Hubungan Internasional,” ermestic: Journal of International Studies; Vol 4 No 2 (2020)DO - 10.24198/Intermestic.V4n2.1 , May 31, 2020, 111 y, , 4 Segun Oshewolo and Agaptus Nwozor, “COVID-19: Projecting the National Security Dimensions of Pandemics,” Strategic Analysis 44, no. 3 (May 3, 2020): 269, https://doi.org/10.1080/09700161.2020.1767911. 5 C M Weible et al “COVID-19 and the Policy Sciences: Initial Reactions and Perspectives ” Policy Sciences 53 no 2 f ; ( ) http://intermestic.unpad.ac.id/index.php/intermestic/article/view?path=. p p p p p Christine Riley, Bo Xie, and Anjum Khurshid, “Challenges Encountered in Comparing International Policy Responses OVID-19 and Their Effects,” Health Research Policy and Systems 19, no. 1 (2021): 134, https://doi.org/10.1186/s1296 1-00783-1. Introduction Hardly do states neglect a disease outbreak somewhere in the globe inasmuch as it can spread beyond borders and turn to a global concern. Therefore, pandemic has been a rising issue either in international politics and security or in foreign policy since the end of Cold War. The international politics of transmissible disease emerges as a challenge in international relations. 1 The challenges are related to how states and other international actors reconfigure their policy and national interests’ articulation—in term of health, education, security, politics, and economy—in response to the Covid- 19. 2 According to Christine Riley et al the challenges consist of (1) no single reputable source of information and too much ‘noise’, (2) a lack of standards for how to measure and report data across countries, (3) variation in the content, implementation, and enforcement of policies, and (4) politics, instead of science, leading the efforts in pandemic management. 3 It is urged for all international actors and world leaders to collaborate in a unified mission to decrease the imperative mortality of Covid-19. While Segun Oshewolo and Agaptus Nwozor suggested that pandemic be raising scientific awareness about national security dimensions, 4 Christopher M. Weible et al. 5 argued that post-pandemic phenomena are still understudied in policy sciences. The study in this paper belongs to foreign policy with which state behaviors can be analyzed and foreign policy determinants can be explained. That while a number of states acted unilaterally in order to protect their citizens and save their economy in addition to blaming others and current global health system in the aftermath of Covid-19 outbreak even the United States (U.S.) for example. Donald Trump was mentioned as “damaged [the] science and it could take decades to recover” by Jeff Tollefson in an article from Nature because his actions which was perceived to have exacerbated the pandemic that killed more than 200,000 people in the U.S., rolled back environment and public-health regulations and undermined science and scientific institutions which some of them could be permanent. 6 Indonesia was one of the actors rejecting such populist trend. Instead, Indonesia complied with the existing international architecture emphasizing cooperation among states and international communities. Introduction 31 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic The 2019 Novel Coronavirus disease or so-called Covid-19 which formerly was just a local contagious virus-caused pneumonia was believed to have first emerged in the Chinese city of Wuhan in the end December 2019. It was officially reported by the World Health Organization (WHO) of China country office on 7 January 2020 7 and then characterized as a pandemic by the WHO on 11 March 2020. 8 By the end of April 2020, the virus was reported to have spread to over 200 states. 9 State governments in the world had to choose a difficult option between a large number of deaths and economic collapse by whether or not to impose lockdowns. 10 The Covid-19 affected domestic politics and relations among states in addition to disrupting economy globally The situation deaths and economic collapse by whether or not to impose lockdowns. 10 The Covid-19 affected domestic politics and relations among states in addition to disrupting economy globally. The situation could be best displayed by domestic pro-con protests and tight security measures in a number of states, distrust on international regimes and organizations, slowing economic growths and even recession threat either in developed or in developing states. So did the pandemic intensify the ongoing tension between two giants; the U.S. and China 11 and their competition for the post-Covid-19 global leadership. 12 Consequently, humanitarian aids during the pandemic were frequently politized regarding the competitions between the two blocks. 13 The Indonesian President Joko Widodo officially announced the first two confirmed COVID- 19 cases of citizens on 3 March 2020. 14 Criticisms and doubts rose as the Indonesian government could have announced the confirmed cases earlier while the virus was crippling China, Republic of Korea (ROK) and Iran in January and February 2020. 15 Indonesia has been one of the most affected in term of both confirmed cases and deaths in Southeast Asia since then. Indonesia was suffering from the economic loss due to restriction of domestic businesses. So were Indonesian overseas both migrants and students affected as foreign governments issued the temporary, strict policies targeting foreign nationals. As the Covid-19 was inevitable and all states were hit, Indonesia was actively responding to it. The Asian Development Bank (ADB) predicted that Indonesia, the world’s fourth most populous state, would suffer greatly and need longer time to recover. 7 Mayusef Sukmana, Muhammad Aminuddin, and Dwi Nopriyanto, “Indonesian Government Response In COVID-19 Disaster Prevention,” East African Scholars Journal of Medical Sciences 3, no. 3 (March 15, 2020): 81, https://doi.org/10.36349/EASMS.2020.v03i03.025. p g 10 Damien Bol et al., “The Effect of COVID-19 Lockdowns on Political Support: Some Good News for Democracy?,” European Journal of Political Research 60, no. 2 (May 1, 2021): 1, https://doi.org/https://doi.org/10.1111/1475-6765.12401. 11 Seth Schindler, Nicholas Jepson, and Wenxing Cui, “Covid-19, China and the Future of Global Development,” Research in Globalization 2 (2020): 1, https://doi.org/https://doi.org/10.1016/j.resglo.2020.100020; J.J. Zhang and V.R. Savage, “The Geopolitical Ramifications of COVID-19: The Taiwanese Exception,” Eurasian Geography and Economics, 2020, 465–66, https://doi.org/10.1080/15387216.2020.1779773. p g 8 Wasiu Abiodun Balogun and Oluwaseun Soile, “‘Pandemic Diplomacy’ and The Politics of Paradox: International Cooperation in The Age of Distancing,” Gaziantep University Journal of Social Sciences 19, no. COVID-19 Special Issue (October 31, 2020): 415, https://doi.org/10.21547/JSS.780792. 9 Javad Yoosefi Lebni et al., “How the COVID-19 Pandemic Effected Economic, Social, Political, and Cultural Factors: A Lesson from Iran,” International Journal of Social Psychiatry 67, no. 3 (July 2, 2020): 1, https://doi.org/10.1177/0020764020939984. p g 12 Md. Saifullah Akon and Mahfujur Rahman, “Reshaping the Global Order in the Post COVID-19 Era: A Critical Analysis,” Chinese Journal of International Review 02, no. 01 (June 1, 2020): 2, https://doi.org/10.1142/S2630531320500067. 13 Michael Dunford and Bing Qi, “Global Reset: COVID-19, Systemic Rivalry and the Global Order,” Research in Globalization 2 (2020): 6, https://doi.org/https://doi.org/10.1016/j.resglo.2020.100021. 7 Mayusef Sukmana, Muhammad Aminuddin, and Dwi Nopriyanto, “Indonesian Government Response In COVID-19 Disaster Prevention,” East African Scholars Journal of Medical Sciences 3, no. 3 (March 15, 2020): 81, https://doi.org/10.36349/EASMS.2020.v03i03.025. 8 Wasiu Abiodun Balogun and Oluwaseun Soile, “‘Pandemic Diplomacy’ and The Politics of Paradox: International Cooperation in The Age of Distancing,” Gaziantep University Journal of Social Sciences 19, no. COVID-19 Special Issue (October 31, 2020): 415, https://doi.org/10.21547/JSS.780792. 9 Javad Yoosefi Lebni et al., “How the COVID-19 Pandemic Effected Economic, Social, Political, and Cultural Factors: A L f I ” I i l J l f S i l P hi 67 3 (J l 2 2020) 1 7 Mayusef Sukmana, Muhammad Aminuddin, and Dwi Nopriyanto, “Indonesian Government Response In COVID-19 Disaster Prevention,” East African Scholars Journal of Medical Sciences 3, no. 3 (March 15, 2020): 81, https://doi.org/10.36349/EASMS.2020.v03i03.025. 8 Wasiu Abiodun Balogun and Oluwaseun Soile, “‘Pandemic Diplomacy’ and The Politics of Paradox: International Cooperation in The Age of Distancing,” Gaziantep University Journal of Social Sciences 19, no. COVID-19 Special Issue (October 31, 2020): 415, https://doi.org/10.21547/JSS.780792. 9 Javad Yoosefi Lebni et al., “How the COVID-19 Pandemic Effected Economic, Social, Political, and Cultural Factors: A Lesson from Iran,” International Journal of Social Psychiatry 67, no. 3 (July 2, 2020): 1, https://doi.org/10.1177/0020764020939984. 10 Damien Bol et al., “The Effect of COVID-19 Lockdowns on Political Support: Some Good News for Democracy?,” European Journal of Political Research 60, no. 2 (May 1, 2021): 1, https://doi.org/https://doi.org/10.1111/1475-6765.12401. 11 Seth Schindler, Nicholas Jepson, and Wenxing Cui, “Covid-19, China and the Future of Global Development,” Research in Globalization 2 (2020): 1, https://doi.org/https://doi.org/10.1016/j.resglo.2020.100020; J.J. Zhang and V.R. Savage, “The Geopolitical Ramifications of COVID-19: The Taiwanese Exception,” Eurasian Geography and Economics, 2020, 465–66, https://doi.org/10.1080/15387216.2020.1779773. 12 Md. Saifullah Akon and Mahfujur Rahman, “Reshaping the Global Order in the Post COVID-19 Era: A Critical Analysis,” Chinese Journal of International Review 02, no. 01 (June 1, 2020): 2, https://doi.org/10.1142/S2630531320500067. 13 Michael Dunford and Bing Qi, “Global Reset: COVID-19, Systemic Rivalry and the Global Order,” Research in Globalization 2 (2020): 6, https://doi.org/https://doi.org/10.1016/j.resglo.2020.100021. 14 Shindy Diah Ayu Lestari, “Pandemic Of Covid-19 in Indonesia: Social Politics Perspective (Pandemi Covid-19 Di Indonesia: Perspektif Sosial Politik),” ETNOREFLIKA: Jurnal Sosial Dan Budaya 10, no. 1 (February 27, 2021): 40, https://doi.org/10.33772/ETNOREFLIKA.V10I1.915. 15 Marcus Mietzner, “Populist Anti-Scientism, Religious Polarisation, and Institutionalised Corruption: How Indonesia’s Democratic Decline Shaped Its COVID-19 Response,” Journal of Current Southeast Asian Affairs 39, no. 2 (August 1, 2020): 228, https://doi.org/10.1177/1868103420935561. 16 Riyanti Djalante et al., “Review and Analysis of Current Responses to COVID-19 in Indonesia: Period of January to March 2020,” Progress in Disaster Science 6 (2020): 2, https://doi.org/https://doi.org/10.1016/j.pdisas.2020.100091. Introduction 16 Just like other natural disasters, Klaus Dodds et al. suggested that the pandemic unveil the uneven capacity among states to respond and allow others 12 Md. Saifullah Akon and Mahfujur Rahman, “Reshaping the Global Order in the Post COVID-19 Era: A Critical Analysis,” Chinese Journal of International Review 02, no. 01 (June 1, 2020): 2, https://doi.org/10.1142/S2630531320500067. 13 Michael Dunford and Bing Qi, “Global Reset: COVID-19, Systemic Rivalry and the Global Order,” Research in Globalization 2 (2020): 6, https://doi.org/https://doi.org/10.1016/j.resglo.2020.100021. p g p g j g 14 Shindy Diah Ayu Lestari, “Pandemic Of Covid-19 in Indonesia: Social Politics Perspective (Pandemi Covid-19 Di Indonesia: Perspektif Sosial Politik),” ETNOREFLIKA: Jurnal Sosial Dan Budaya 10, no. 1 (February 27, 2021): 40, https://doi.org/10.33772/ETNOREFLIKA.V10I1.915. 14 Shindy Diah Ayu Lestari, “Pandemic Of Covid-19 in Indonesia: Social Politics Perspective (Pandemi Covid-19 Di Indonesia: Perspektif Sosial Politik),” ETNOREFLIKA: Jurnal Sosial Dan Budaya 10, no. 1 (February 27, 2021): 40, https://doi.org/10.33772/ETNOREFLIKA.V10I1.915. 15 Marcus Mietzner, “Populist Anti-Scientism, Religious Polarisation, and Institutionalised Corruption: How Indonesia’s Democratic Decline Shaped Its COVID-19 Response,” Journal of Current Southeast Asian Affairs 39, no. 2 (August 1, 2020): 228, https://doi.org/10.1177/1868103420935561. 15 Marcus Mietzner, “Populist Anti-Scientism, Religious Polarisation, and Institutionalised Corruption: How Indonesia’s Democratic Decline Shaped Its COVID-19 Response,” Journal of Current Southeast Asian Affairs 39, no. 2 (August 1, 2020): 228, https://doi.org/10.1177/1868103420935561. 16 Riyanti Djalante et al., “Review and Analysis of Current Responses to COVID-19 in Indonesia: Period of January to March 2020,” Progress in Disaster Science 6 (2020): 2, https://doi.org/https://doi.org/10.1016/j.pdisas.2020.100091. erdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic to seek profits from such situation. 17 So did the pandemic expose deep sosio-economic inequalities within states especially migrants and lower income communities. 18 On the other hand, Indonesia was upholding multilateralism in its foreign policy which was not only determined by its interests alone. Indonesia commenced to carry out multilateralism as soon as it gained its independence. The involvement in the Non-Alliance Movement could be its best post- colonial experience. The practice of multilateralism was conducted through various leaderships despite differing foreign policy orientations. When the Covid-19 became a global issue, similar policy was still a preference under President Joko Widodo. At the 2019 annual speech of the Ministry of Foreign Affairs (MoFA), Minister Retno L.P. Introduction Marsudi explicitly said, “Indonesia will continue to work in multilateral setting in spite of the fact that the value of multilateralism is under pressure.” 19 The statement was made in the middle of the pessimistic trend on multilateralism shown by the U.S. in responding to the COVID-19. In addition, there was also increasing questions from rising actors in the Global South on the legitimacy of international architecture which benefitted the U.S. and the West in conserving their influence. 20 However, Indonesia was still committed to the track of multilateralism which was invested within its foreign policy. The foreign policy making was driven by ideational instead of material factors such as power and economic capability. 24 Davies, Kamradt-Scott, and Rushton, Disease Diplomacy: International Norms and Global Health Security; Kamradt- Scott, “The WHO Secretariat, Norm Entrepreneurship, and Global Disease Outbreak Control”; Ferhani and Rushton, “The International Health Regulations, COVID-19, and Bordering Practices: Who Gets in, What Gets out, and Who Gets Rescued?” 23 Davies, Kamradt-Scott, and Rushton, Disease Diplomacy: International Norms and Global Health Security; Kamradt- Scott, “The WHO Secretariat, Norm Entrepreneurship, and Global Disease Outbreak Control”; Ferhani and Rushton, “The International Health Regulations, COVID-19, and Bordering Practices: Who Gets in, What Gets out, and Who Gets Rescued?” 22 Adam Kamradt-Scott, “The WHO Secretariat, Norm Entrepreneurship, and Global Disease Outbreak Control,” Journal of International Organization Studies 1 (January 1, 2010); Sara E Davies, Adam Kamradt-Scott, and S Rushton, Disease Diplomacy: International Norms and Global Health Security (Baltimore: John Hopkins University Press, 2015); A Ferhani and S Rushton, “The International Health Regulations, COVID-19, and Bordering Practices: Who Gets in, What Gets out, and Who Gets Rescued?,” Contemporary Security Policy 41, no. 3 (2020): 458–77, https://doi.org/10.1080/13523260.2020.1771955. 17 Klaus Dodds et al., “The COVID-19 Pandemic: Territorial, Political and Governance Dimensions of the Crisis,” Territory, Politics, Governance 8, no. 3 (May 26, 2020): 290, https://doi.org/10.1080/21622671.2020.1771022. 18 R Djalante et al., “COVID-19 and ASEAN Responses: Comparative Policy Analysis,” Progress in Disaster Science 8 (2020): 11, https://doi.org/10.1016/j.pdisas.2020.100129. 19 Retno LP Marsudi, “Cooperation: The Catalyst for Fighting the COVID-19 Crisis - Opinion - The Jakarta Post,” The Jakarta Post, March 25, 2020, https://www.thejakartapost.com/academia/2020/03/24/cooperation-the-catalyst-for-fighting- the-Covid-19-crisis.html at Jakarta Post. 20 Jürgen Rüland, “The Rise of ‘Diminished Multilateralism’: East Asian and European Forum Shopping in Global Governance,” Asia Europe Journal 9, no. 2 (2012): 255–70, https://doi.org/10.1007/s10308-012-0311-9. 21 Sara E Davies, “Securitizing Infectious Disease,” International Affairs 84, no. 2 (March 1, 2008): 295–313, https://doi.org/10.1111/j.1468-2346.2008.00704.x; Melissa G Curley and Jonathan Herington, “The Securitisation of Avian Influenza: International Discourses and Domestic Politics in Asia,” Review of International Studies 37, no. 1 (2011): 141–66, https://doi.org/DOI: 10.1017/S0260210510000537. 22 Adam Kamradt-Scott, “The WHO Secretariat, Norm Entrepreneurship, and Global Disease Outbreak Control,” Journal of International Organization Studies 1 (January 1, 2010); Sara E Davies, Adam Kamradt-Scott, and S Rushton, Disease Diplomacy: International Norms and Global Health Security (Baltimore: John Hopkins University Press, 2015); A Ferhani and S Rushton, “The International Health Regulations, COVID-19, and Bordering Practices: Who Gets in, What Gets out, and Who Gets Rescued?,” Contemporary Security Policy 41, no. 3 (2020): 458–77, https://doi.org/10.1080/13523260.2020.1771955. 23 Davies, Kamradt-Scott, and Rushton, Disease Diplomacy: International Norms and Global Health Security; Kamradt- Scott, “The WHO Secretariat, Norm Entrepreneurship, and Global Disease Outbreak Control”; Ferhani and Rushton, “The International Health Regulations, COVID-19, and Bordering Practices: Who Gets in, What Gets out, and Who Gets Rescued?” 24 Davies, Kamradt-Scott, and Rushton, Disease Diplomacy: International Norms and Global Health Security; Kamradt- Scott, “The WHO Secretariat, Norm Entrepreneurship, and Global Disease Outbreak Control”; Ferhani and Rushton, “The I t ti l H lth R l ti COVID 19 d B d i P ti Wh G t i Wh t G t t d Wh G t 17 Klaus Dodds et al., “The COVID-19 Pandemic: Territorial, Political and Governance Dimensions of the Crisis,” Territory, Politics, Governance 8, no. 3 (May 26, 2020): 290, https://doi.org/10.1080/21622671.2020.1771022. 18 R Djalante et al., “COVID-19 and ASEAN Responses: Comparative Policy Analysis,” Progress in Disaster Science 8 (2020): 11, https://doi.org/10.1016/j.pdisas.2020.100129. 19 Retno LP Marsudi, “Cooperation: The Catalyst for Fighting the COVID-19 Crisis - Opinion - The Jakarta Post,” The Jakarta Post, March 25, 2020, https://www.thejakartapost.com/academia/2020/03/24/cooperation-the-catalyst-for-fighting- the-Covid-19-crisis.html at Jakarta Post. Jürgen Rüland, “The Rise of ‘Diminished Multilateralism’: East Asian and European Forum Shopping in Globa overnance,” Asia Europe Journal 9, no. 2 (2012): 255–70, https://doi.org/10.1007/s10308-012-0311-9. 21 Sara E Davies, “Securitizing Infectious Disease,” International Affairs 84, no. 2 (March 1, 2008): 295–313, https://doi.org/10.1111/j.1468-2346.2008.00704.x; Melissa G Curley and Jonathan Herington, “The Securitisati Influenza: International Discourses and Domestic Politics in Asia,” Review of International Studies 37, no. 1 (20 https://doi.org/DOI: 10.1017/S0260210510000537. Literature Review There exists some literature discussing health issues in both regional and international level. Securitizing infectious diseases by states and international organizations was studied by a number of scholars such as Sarah E. Davies, and Melissa G. Curley and Jonathan Herington. 21 Meanwhile, Sarah E. Davies, Adam Kamradt-Scott and S. Rushton studied the role of norm in international practices regarding health issues. 22 The emergence of new health threats led to the implementation of 2005 International Health Regulation (IHR) as a new international norm. 23 The study investigated the role of the WHO’s norm entrepreneurship in order to ensure its member states to comply with the new norm. 24 In another study, Kamradt-Scott revealed the role of smaller group within the WHO in successfully introducing and preserving the norm of utilizing non-government information sources in 23 Davies, Kamradt-Scott, and Rushton, Disease Diplomacy: International Norms and Global Health Security; Kamradt- Scott, “The WHO Secretariat, Norm Entrepreneurship, and Global Disease Outbreak Control”; Ferhani and Rushton, “The International Health Regulations, COVID-19, and Bordering Practices: Who Gets in, What Gets out, and Who Gets Rescued?” erdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic the revised IHR. 25 The global cooperation along with the influence of donors was studied by Devi Sridhar and Ngaire Woods. 26 The study showed that donors could redirect the international organizations to fight specific diseases. 27 The studies on Asian States regarding health issues were conducted by David P. Fidler 28, Sarah E. Davies 29, Shahar Hameiri (2014) and Clare Wenham30. Fidler suggested that Asian normative ideas encourage their participation and contribution to global health diplomacy and governance. 31 Davies argued that having hit by previous infectious diseases created a strong will among the Southeast Asia states for cooperation in responding to public health emergencies. 32 The studies on Asian States regarding health issues were conducted by David P. Fidler 28, Sarah E. Davies 29, Shahar Hameiri (2014) and Clare Wenham30. Fidler suggested that Asian normative ideas encourage their participation and contribution to global health diplomacy and governance. 31 Davies argued that having hit by previous infectious diseases created a strong will among the Southeast Asia states for cooperation in responding to public health emergencies. 32 Further, Davies found a number of East Asian states’ compliance with the reporting of the infectious avian influenza despite their sovereignty. Literature Review 33 Indonesia was included in the study and it was once the state with the highest human infections. However, Indonesia’ compliance with the reporting was contrast to its decision over the viral sharing dispute with the WHO. Hameiri argued that Indonesia asserted sovereignty to justify ceasing the virus-sharing was merely domestic actors’ struggles to gain their political goals. 34 On the other hand, Murray Hiebert noted that Indonesia attempted to play more serious role to step up global health diplomacy and specifically demanded a guarantee from the WHO that international cooperation would benefit lower income states especially in accessing affordable vaccine of the bird flu35. While, Wenham examined Thailand’s ambitions to dominate the normative process of disease control in mainland Southeast Asia. 36 Indonesia’s foreign policy and regional diplomacy by some states towards Indonesia during the Covid-19 pandemic were studied by a number of scholars such as Asep Setiawan, and Adiasri Putri Purbantina and Renitha Dwi Hapsari respectively. 37 While Indonesia’s inward looking foreign policy was raised in order to save its citizens and build cooperation regarding the health crisis situation and its effects on economic slowdown, 38 the ROK and Japan’s health diplomacy towards Indonesia was driven by health and economic recovery interests. 39 28 David P Fidler, “Asia’s Participation in Global Health Diplomacy and Global Health Governance,” Asian Journal of WTO & International Health Law and Policy 5, no. 2 (2010), p p p p p g 29 Sara E Davies, “The International Politics of Disease Reporting: Towards Post-Westphalianism?,” International Politics 49, no. 5 (2012): 591–613, https://doi.org/10.1057/ip.2012.19; Davies, Kamradt-Scott, and Rushton, Disease Diplomacy: International Norms and Global Health Security; Sara E Davies, Containing Contagion: The Politics of Disease Outbreaks in Southeast Asia (Baltimore: John Hopkins University Press, 2019). p y 30 C Wenham, “Regionalizing Health Security: Thailand’s Leadership Ambitions in Mainland Southeast Asian Disease Control,” Contemporary Southeast Asia: A Journal of International and Strategic Affairs 40, no. 1 (2018): 126–51. 31 idl i i i i i l b l l h i l d l b l 32 Davies, Containing Contagion: The Politics of Disease Outbreaks in Southeast Asia. 33 32 Davies, Containing Contagion: The Politics of Disease Outbreaks in Southeast Asia. 33 Davies, “The International Politics of Disease Reporting: Towards Post-Westphalianism?” Shahar Hameiri, “Avian Influenza, ‘Viral Sovereignty’, and the Politics of Health Security in Indonesia,” The Pacific view 27, no. 3 (May 27, 2014): 333–56, https://doi.org/10.1080/09512748.2014.909523. p p 26 Devi Sridhar and Ngaire Woods, “Trojan Multilateralism: Global Cooperation in Health,” Global Policy 4, no. 4 (November 1, 2013): 325–35, https://doi.org/https://doi.org/10.1111/1758-5899.12066. 27 Sridhar and Woods. 25 Kamradt-Scott, “The WHO Secretariat, Norm Entrepreneurship, and Global Disease Outbreak Control.” 26 Devi Sridhar and Ngaire Woods, “Trojan Multilateralism: Global Cooperation in Health,” Global Policy 4, no. 4 (November 1, 2013): 325–35, https://doi.org/https://doi.org/10.1111/1758-5899.12066. 27 Sridhar and Woods 25 Kamradt-Scott, “The WHO Secretariat, Norm Entrepreneurship, and Global Disease Outbreak Control.” 28 David P Fidler, “Asia’s Participation in Global Health Diplomacy and Global Health Governance,” Asian Journal of WTO & International Health Law and Policy 5, no. 2 (2010), 55 25 Kamradt-Scott, “The WHO Secretariat, Norm Entrepreneurship, and Global Disease Outbreak Control.” 26 Devi Sridhar and Ngaire Woods, “Trojan Multilateralism: Global Cooperation in Health,” Global Policy 4, no. 4 (November 1, 2013): 325–35, https://doi.org/https://doi.org/10.1111/1758-5899.12066. 27 Sridhar and Woods. 28 David P Fidler, “Asia’s Participation in Global Health Diplomacy and Global Health Governance,” Asian Journal of WTO & International Health Law and Policy 5, no. 2 (2010), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1729455#references-widget. 29 Sara E Davies, “The International Politics of Disease Reporting: Towards Post-Westphalianism?,” International Politics 49, no. 5 (2012): 591–613, https://doi.org/10.1057/ip.2012.19; Davies, Kamradt-Scott, and Rushton, Disease Diplomacy: International Norms and Global Health Security; Sara E Davies, Containing Contagion: The Politics of Disease Outbreaks in Southeast Asia (Baltimore: John Hopkins University Press, 2019). 30 C Wenham, “Regionalizing Health Security: Thailand’s Leadership Ambitions in Mainland Southeast Asian Disease Control,” Contemporary Southeast Asia: A Journal of International and Strategic Affairs 40, no. 1 (2018): 126–51. 31 Fidler, “Asia’s Participation in Global Health Diplomacy and Global Health Governance.” 32 Davies, Containing Contagion: The Politics of Disease Outbreaks in Southeast Asia. 33 Davies, “The International Politics of Disease Reporting: Towards Post-Westphalianism?” 34 Shahar Hameiri, “Avian Influenza, ‘Viral Sovereignty’, and the Politics of Health Security in Indonesia,” The Pacific Review 27, no. 3 (May 27, 2014): 333–56, https://doi.org/10.1080/09512748.2014.909523. 35 Murray Hiebert, “Indonesia Steps Up Global Health Diplomacy: Bolsters Role in Addressing International Medical Challenges,” Center for Strategic and International Studies, 2013, https://www.csis.org/analysis/indonesia-steps-global- health-diplomacy. 36 Wenham, “Regionalizing Health Security: Thailand’s Leadership Ambitions in Mainland Southeast Asian Disease Control.” 37 Asep Setiawan, “Politik Luar Negeri Indonesia Era Covid-19: Penyelamatan dan Kerjasama,” Independen: Jurnal Politik Indonesia Dan Global 1 (September 1, 2020): 65–74, https://doi.org/10.24853/independen.1.2.65-74; Adiasri Putri Purbantina and Renitha Dwi Hapsari, “Diplomasi Kesehatan Di Era Pandemik Global: Analisa Bantuan Penanganan Covid- 19 Dari Negara Jepang Dan Korea Selatan Ke Indonesia,” Global and Policy Journal of International Relations 8, no. 01 (July 9, 2020), https://doi.org/10.33005/JGP.V8I01.2167. 38 Setiawan, “Politik Luar Negeri Indonesia Era Covid-19: Penyelamatan dan Kerjasama.” 29 Sara E Davies, “The International Politics of Disease Reporting: Towards Post-Westphalianism?,” International Politics 49, no. 5 (2012): 591–613, https://doi.org/10.1057/ip.2012.19; Davies, Kamradt-Scott, and Rushton, Disease Diplomacy: International Norms and Global Health Security; Sara E Davies, Containing Contagion: The Politics of Disease Outbreaks in Southeast Asia (Baltimore: John Hopkins University Press, 2019). https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1729455#references-widget. 29 Sara E Davies, “The International Politics of Disease Reporting: Towards Post-Westphalianism?,” International Politics 49, no. 5 (2012): 591–613, https://doi.org/10.1057/ip.2012.19; Davies, Kamradt-Scott, and Rushton, Disease Diplomacy: International Norms and Global Health Security; Sara E Davies, Containing Contagion: The Politics of Disease Outbreaks in Southeast Asia (Baltimore: John Hopkins University Press, 2019). 30 34 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic 34 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic The literature confirms that norms have played significant roles in international relations regarding the infectious disease issues affecting actor behaviors in term of either political struggle or cooperation. However, neither of the studies describes nor explains Indonesia’s behavior in addressing the Covid-19. Furthermore, the study on Indonesia’s foreign policy in addressing the Covid-19 is still rare due to the fact that the pandemic issue is still new and most states are still struggling to deal with it. The study by Asep Setiawan slightly relates to this research describing Indonesia’s behavior and policy in responding to the pandemic40. However, the study only saw tangible national interest as the only driver of Indonesia’s foreign policy. Meanwhile, this paper suggests that such interest be constructed by values, norms, ideas, and identity. We propose that Indonesia’s multilateralism in regard to Covid-19 response is constructed by those values, norms, ideas, and identity—which we pack up as ‘gotong royong’ and the IHR. This paper would narrow the gap and focus on the role of both domestic and international norms in Indonesia’s foreign policy. This paper provides the answer to question “Why did Indonesia raise multilateralism in its foreign policy in responding to the Covid-19 pandemic?” This paper aims to explain Indonesia’s foreign policy in upholding multilateralism in responding to the pandemic in addition to the drivers of its multilateralism. This research was limited on the Indonesia’s foreign policy during the pandemic in 2020. This paper contributes to understanding Indonesia’s foreign policy and behavior during difficult or challenging situations not limited to the pandemic. 40 Setiawan, “Politik Luar Negeri Indonesia Era Covid-19: Penyelamatan Dan Kerjasama.” 41 Matthew B Miles and A. Michael Huberman, Qualitative Data Analysis: An Expanded Sourcebook, 2nd Ed. (London: SAGE Publications Inc, 1994). 42 “The Ethics of Constructivism,” in The Oxford Handbook of International Relations (New York: Oxford University Press, Inc, 2008), 300–305. Setiawan, “Politik Luar Negeri Indonesia Era Covid-19: Penyelamatan Dan Kerjasama.” Literature Review 35 Murray Hiebert, “Indonesia Steps Up Global Health Diplomacy: Bolsters Role in Addressing International Medical Challenges,” Center for Strategic and International Studies, 2013, https://www.csis.org/analysis/indonesia-steps-global- health-diplomacy. Wenham, “Regionalizing Health Security: Thailand’s Leadership Ambitions in Mainland Southeast Asian Disease ontrol.” 37 Asep Setiawan, “Politik Luar Negeri Indonesia Era Covid-19: Penyelamatan dan Kerjasama,” Independen: Jurnal Politik Indonesia Dan Global 1 (September 1, 2020): 65–74, https://doi.org/10.24853/independen.1.2.65-74; Adiasri Putri Purbantina and Renitha Dwi Hapsari, “Diplomasi Kesehatan Di Era Pandemik Global: Analisa Bantuan Penanganan Covid- 19 Dari Negara Jepang Dan Korea Selatan Ke Indonesia,” Global and Policy Journal of International Relations 8, no. 01 (July 9, 2020), https://doi.org/10.33005/JGP.V8I01.2167. 38 Setiawan, “Politik Luar Negeri Indonesia Era Covid-19: Penyelamatan dan Kerjasama.” 39 38 Setiawan, “Politik Luar Negeri Indonesia Era Covid-19: Penyelamatan dan Kerjasama.” 39 Purbantina and Hapsari, “Diplomasi Kesehatan Di Era Pandemik Global: Analisa Bantuan Penanganan Covid-19 Dari Negara Jepang Dan Korea Selatan Ke Indonesia,” 10. 39 Purbantina and Hapsari, “Diplomasi Kesehatan Di Era Pandemik Global: Analisa Bantuan Penanganan Covid-19 Dari Negara Jepang Dan Korea Selatan Ke Indonesia,” 10. Methodology This research applied the qualitative method with an explanatory analysis. It was a library research in which a range of recorded data such as journal articles, research papers, scholarly commentaries and other relevant online sources were collected. So was this research benefitted from information provided in the institutional website of MoFA as well as officials’ statements and writings in explaining Indonesia’s foreign policy to respond to the COVID-19. The data were analysed through the Miles and Hubberman’s interactive model consisting of four steps; collecting data, reducing data, displaying data and drawing conclusion41. The application of method can be seen as following: Firstly, collecting data. An amount of information on the COVID-19 and Indonesia’s response was collected. Secondly, reducing data. Data were sorted out in order to raise critical questions so that the issue became specific. Meanwhile, the unnecessary data were taken aside. The critical questions led to Indonesia’s foreign policy in responding to the pandemic. Thirdly, displaying data. At this stage, pieces of information were organized so that an early conclusion could be drawn. However, collecting and reducing data still continued simultaneously. The last, verification. The conclusion appeared clear and strong after the researchers found meanings in addition to recording regularities, explanatory patterns, configurations, causalities and propositions. This research found that the factor of identity constructed by domestic and global norm drove Indonesia to raise multilateralism in its foreign policy in responding to the Covid-19. This research applied the holistic constructivist approach of international relations. Ian Hurd 42 includes several features of constrcutivism; first, it is an alternative to materialism. Unlike the materialist approach such as neorealism and neoliberalism which believe in military power, strategic resources or other causes of direct outcomes in world politics, the constructivist approach embraces that the world is socially constructed depending on the web of meanings and practices among actors. Second, state interests are constructed. Social relations play important role in shaping actor 5 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandem perceptions and identities which contribute to interest construction. Third, structure and agents are mutually constituted. While state actions affect norm and institution making, they affect to defining, socializing and influencing states and both can be redefined. Fourth, multiple logics of anarchy are present. 43 “The Power of (Emotion) Words: On the Importance of Emotions for Social Constructivist Discourse Analysis in IR,” Journal of International Relations and Development 21, no. 3 (2018): 495–522, https://doi.org/10.1057/s41268-017-0086-0. 44 Martin Weber, “Between ‘Isses’ and ‘Oughts’: IR Constructivism, Critical Theory, and the Challenge of Political Philosophy,” European Journal of International Relations 20, no. 2 (January 4, 2013): 516–43, https://doi.org/10.1177/1354066112466573. 45 “Introduction: Alternative Perspective on National Security,” in The Culture of National Security: Norms and Identity in World Politics, ed. P.J Katzenstein (New York: Columbia University Press, 1996). 46 R.I Jepperson, Alexander Wend, and P.J Katzenstein, “Norms, Identity, and Culture in National Security,” in The Culture of National Security: Norms and Identity in World Politics (New York: Columbia University Press, 1996). 43 “The Power of (Emotion) Words: On the Importance of Emotions for Social Constructivist Discourse Analysis in IR,” Journal of International Relations and Development 21, no. 3 (2018): 495–522, https://doi.org/10.1057/s41268-017-0086-0. 44 Martin Weber, “Between ‘Isses’ and ‘Oughts’: IR Constructivism, Critical Theory, and the Challenge of Political Philosophy,” European Journal of International Relations 20, no. 2 (January 4, 2013): 516–43, https://doi.org/10.1177/1354066112466573. 45 “Introduction: Alternative Perspective on National Security,” in The Culture of National Security: Norms and Identity in World Politics, ed. P.J Katzenstein (New York: Columbia University Press, 1996). 46 R.I Jepperson, Alexander Wend, and P.J Katzenstein, “Norms, Identity, and Culture in National Security,” in The Culture World Politics, ed. P.J Katzenstein (New York: Columbia University Press, 1996). 46 R.I Jepperson, Alexander Wend, and P.J Katzenstein, “Norms, Identity, and Culture in National Security,” in The Culture of National Security: Norms and Identity in World Politics (New York: Columbia University Press, 1996). Methodology Source: Author’s own visualization, adapted from M M Nia, ‘A Holistic Constructivist Approach to Iran’s Foreign Policy’, International Journal of Business and Society, 2.4 (2011); M Alexandrov, ‘The Concept of State Identity in International Relations: A Theoretical Analysis’, 2003. Source: Author’s own visualization, adapted from M M Nia, ‘A Holistic Constructivist Approach to Iran’s Foreign Policy’, International Journal of Business and Society, 2.4 (2011); M Alexandrov, ‘The Concept of State Identity in International Relations: A Theoretical Analysis’, 2003. Despite the fact that each constructivist approach portrays different ideas on norms and actor identities, they arrive at the same principle of norm and identity theory; norms refer to the environment’s cultural and regulative contents; on the other hand, identities refer to the actor cultural representation shaped by such environment. 47 In other words as Maxym Alexandrov 48 proposed, state identity is “a set of broadly accepted representations of state, in particular in its relation to other states, together with the corresponding beliefs about the appropriate behavior, rights or responsibilities.” This research emphasized the role of domestic norm and actor as well as global norm in creating Indonesia’s identity. A domestic norm contributes to the construction of domestic actor’s identity which then represents state identity. Gotong royong (mutual cooperation) was the domestic norm and the MoFA was the main actor in constructing Indonesia’s identity. As a collective expectation, indigenous value of gotong royong guided the MoFA to issue foreign policy decision; multilateralism which was proper in addressing the Covid-19. Its comitment to such practice could be seen through its reoccurring practices and the information stated in its website as well as its officials’ statements and writings. Despite our argument that the MoFA was the most apparent actor in Indonesia’s foreign policy in responding to the Covid-19, other domestic branches of government such as the President, the Ministry of Health (MoH) and Parliament also played important role in boldening the MoFA’s commitment to the indigenous norm and multilateralism. The perspective is in line with the characteristic of foreign policy in democratic states in which foreign policy making is an arena of involvement for diverse actors. The idea and practice of multilateralism were not solely constructed by the elements within domestic level. In addition, the international normative structure resulted in by interactions among states played a significant role in constructing Indonesia’s identity. p y y 49 John Gerard Ruggie, “Multilateralism: The Anatomy of an Institution,” International Organization 46, no. 3 (1992): 561, https://www.jstor.org/stable/2706989. 47 Jepperson, Wend, and Katzenstein. person, Wend, and Katzenstein. e Concept of State Identity in International Relations: A Theoretical Analysis,” 2003, 38. 47 Jepperson, Wend, and Katzenstein. 48 “The Concept of State Identity in International Relations: A Theoretical Analysis,” 2003, 38. 49 John Gerard Ruggie, “Multilateralism: The Anatomy of an Institution,” International Organization 46, no. 3 (1992): 561, https://www.jstor.org/stable/2706989. Methodology The anarchic structure is not natural and permanent, it depends on the ideas of actors whether they see others as enemies and rivals or friends instead. The structure changes provided that ideas change. The constructivist approach emphasizes the role of non-material drivers such as whether they are ideas, norms, cultures and historical legacies which contribute to shaping actor identities and behaviors. In addition, Simon Koschut 43 suggests the importance of shared emotion allowing political actors to participate in international politics. The rise of constructivism mainstreams norms as efficacious explanatory variables in analyzing world politics. 44 Consequently, the foreign policy study recognizes the importance of norms in policy making determining either cooperations or conflicts. Peter J. Katzenstein 45 defines norms as “collective expectations for the proper behavior of actors with a given identity.” Norms have both constitutive and regulative function. While the former means that norms act as the rules which determine actions causing ‘relevant others’ recognize their peculiar identities, the latter means that norms act as the standards which prescribe proper behaviors according to such defined identities. 46 Identities are constructed through the norms coming from either internal or external environment or both. While the systemic constructivism believes in the role of international norms thanks to the interactions between states and the unit-level constructivism believes in domestic norms affecting domestic actors’ identities, the holistic constructivism bridges the two in which both sorts of norms contribute to such identity construction. The holistic constructivist framework of construction between normative structure and actor can be seen as following: 36 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic Figure 1. Holistic Constructivist Framework of Normative Structure and Actor Construction Source: Author’s own visualization, adapted from M M Nia, ‘A Holistic Constructivist Approach to Iran’s Foreign Policy’, International Journal of Business and Society, 2.4 (2011); M Alexandrov, ‘The Concept of State Identity in International Relations: A Theoretical Analysis’, 2003 ure 1. Holistic Constructivist Framework of Normative Structure and Actor Construction Source: Author’s own visualization, adapted from M M Nia, ‘A Holistic Constructivist Approach to Iran’s Foreign Policy’, International Journal of Business and Society, 2.4 (2011); M Alexandrov, ‘The Concept of State Identity in International Relations: A Theoretical Analysis’, 2003. pp , , 48 “The Concept of State Identity in International Relations: A Theoretical Analysis,” 2003, 38. a. Drivers of Indonesia’s Multilateralism Indonesia’s multilateralism was initially expressed in 1945 Constitution Preamble and implemented in the 1955 Bandung Asia-Africa Conference. Indonesia encouraged the spirit of non- alliance among the post-colonial nations. These are significant precedents within Indonesia’s foreign policy which emphasized on multilateralism and international solidarity. John Gerard Ruggie refers multilateralism as “coordinating relations among three or more states in accordance with certain principles.” 50 Similarly, Fen Osler Hampson and Paul Heinbecker defines multilateralism as “collectively agreed norms, rules and principles guiding and governing interstate behavior,” while the generalized reciprocity principles are the basis of multilateral institutions. 51 In other words, multilateralism means cooperation among states which are encouraged by mutual understanding and shared values frequently facilitated by international organizations and institutions. This paper suggests that Indonesia’s multilateralism in responding to the COVID-19 be driven by a couple of norms, namely gotong royong and the IHR. Not only was it a political decision but it also became another institution (a set of norm and practice) in Indonesia’s foreign policy. Gotong royong is an indigenous institution of communality among Indonesian valued even before the republic modern time. Koentjaraningrat 52 categorizes gotong royong into ‘spontaneous help’ and ‘mutual assistance’; while the former occures generally in collective activities in a variety of social aspects, the latter is based on the reciprocal principle either initiated by citizens or expressed as mutual cooperation. It is believed that both various common problems and community development activities could be solved easily and cheaply thanks to the existence of mutual assistance. 53 Gotong royong was constructed as a means of nation building by Indonesian founding fathers in times of the republic early inception as well. The spirit of gotong royong was believed to be incorporated in national ideology of Pancasila and to constitute the national political system. However, gotong royong played different roles during the Indonesian leadership in either domestic politics and foreign policy. I.A.A. Gde Agung 54 noted that the Indonesian first president Soekarno once utilized gotong royong to justify his guided democracy and criticized those rejecting the idea of Nasakom (Nationalist-Religion-Communist). So did President Soekarno try to get advantage of gotong royong slogan in order to gain domestic support for his anti-imperialism and colonialism rhetoric within his active militant foreign policy. Methodology As an active member of the WHO, Indonesia already showed its committment to the IHR as moral principles and a guidance when facing the international public health concern. Embracing the IHR meant that Indonesia was supposed to share responsibility and work together with other actors against any threat of globally transmitted infectious disease. Multilateralism was taken by Indonesia as it was encouraged by the IHR. Furthermore, multilateralism had a number of success experiences since the end of Cold War. Multilateral norms and institutions helped bring about stabilization and management of regional and global changes in today’s world system. 49 37 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic As strongly embraced by the constructivist approach, only if an actor realizes its identity can it determine its interest and role or behavior. In order to respond to the Covid-19, Indonesia was keeping up with international collaboration and cooperation driven by its norm-based identity. Meanwhile, to uphold multilateralism, Indonesia played a number of roles such as maintaining trust on and support to the WHO and engaging within global and regional frameworks. That Indonesia had right and responsibility protecting its citizens and mitigating economic impacts as well as cooperating with other actors either regionally or internationally was deemed appropriate in accordance with its identity. 50 Ruggie, “Multilateralism: The Anatomy of an Institution,” 568. 51 “The ‘New’ Multilateralism of the Twenty-First Century,” Global Governance: A Review of Multilateralism and International Organizations 17, no. 3 (2011): 300, https://doi.org/https://doi.org/10.1163/19426720-01703003. 52 as cited in L.J Slikkerveer, “Gotong Royong: An Indigenous Institutions of Community and Mutual Assistance in Indonesia,” in Integrated Community Managed Development: Strategizing Indigenous Knowledge and Institutions for Poverty Reduction and Sustainable Community Develpment, ed. L.J. Slikkerveer (Cham: Springer, 2019), 308. 53 Slikkerveer, 308. 50 Ruggie, “Multilateralism: The Anatomy of an Institution,” 568. 51 “The ‘New’ Multilateralism of the Twenty-First Century,” Global Governance: A Review of Multilateralism and International Organizations 17, no. 3 (2011): 300, https://doi.org/https://doi.org/10.1163/19426720-01703003. 52 as cited in L.J Slikkerveer, “Gotong Royong: An Indigenous Institutions of Community and Mutual Assistance in Indonesia,” in Integrated Community Managed Development: Strategizing Indigenous Knowledge and Institutions for Poverty Reduction and Sustainable Community Develpment, ed. L.J. Slikkerveer (Cham: Springer, 2019), 308. 53 Slikkerveer, 308. 54 Twenty Years Indonesian Foreign Policy 1945–1965 (De Gruyter Mouton, 1973), 283, https://doi.org/doi:10.1515/9783111558226. 55 A 283 84 54 Twenty Years Indonesian Foreign Policy 1945–1965 (De Gruyter Mouton, 1973), 283, https://doi.org/doi:10.1515/9783111558226. 55 Agung, 283–84. 50 Ruggie, “Multilateralism: The Anatomy of an Institution,” 568. a. Drivers of Indonesia’s Multilateralism 55 During President Soeharto’s regime, despite less sounding, gotong royong was valued to strengthen the concept of integralist state in which rulers and the ruled should be in harmony instead of conflict. The top-down governance style was deemed a typical authoritarian erdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic state in which people should support the government’s domestic and foreign policy. I.G. Wahyu Wicaksana argues that gotong royong is a fundamental force of Indonesian societal cohesion in which conflicts and disputes should be solved through conciliations constituted Indonesia as a family state. 56 The family state foreign policy has led it to cooperate with others and avoid conflicts post-Soeharto regime. Before elected as a President of Indonesia, Joko Widodo once reiterated the spirit of gotong royong in Indonesia’s foreign policy within his vision “creating a sovereign, independent Indonesia.” 57 The spirit of gotong royong would reflect Indonesia’s foreign policy favoring the mutual assistance and win-win solution. Hence, multilateralism could be the most suitable decision when dealing with other international actors in a number of issues. Multilateralism is a belief fitting the Indonesian indigenous value living for centuries among the people as the Indonesian Permanent Representative Mr. Dian Triansyah Djani stated at the April 2019 High-Level Meeting of the United Nations’ General Assembly (HLM-UNGA) recorded in the A’ b i “Indonesia believes in multilateralism because for centuries we believe on the principle of gotong royong, loosely translated as working together, shoulder to shoulder for the greater good. Multilateralism is working together for the greater good, to face global challenges together.” 58 “Indonesia believes in multilateralism because for centuries we believe on the principle of gotong royong, loosely translated as working together, shoulder to shoulder for the greater good. Multilateralism is working together for the greater good, to face global challenges together.” 58 Multilateralism was infused in one of the MoFA’s strategic purposes emphasizing the Indonesia’s leadership and role in influential international cooperation. 59 In addition, the MoFA formed a directorate of multilateral affairs in order to be able to engage Indonesia’s diplomacy more actively and effectively in multilateral affairs. The step was deemed necessary that Indonesia was taking into account multilateralism as an international practice as well as a belief. The officials’ statements and strategic purpose could be evidences that multilateralism was a preference which was embraced by the MoFA as a state institution. 56 “The Family State: A Non-Realist Approach to Understanding Indonesia’s Foreign Policy,” Asian Journal of Political Science 27, no. 3 (September 2, 2019): 6–7, https://doi.org/10.1080/02185377.2019.1686997. 57 R.L Pattiradjawane and N. Soebagio, “Global Maritime Axis: Indonesia, China, and a New Approach to Southeast Asian Regional Resillience,” International Journal of China Studies 6, no. 2 (2015): 179. 58 MoFA-RI, “International Day of Multilateralism and Diplomacy for Peace High-Level Meeting of the General Assembly,” Ministry of Foreign Affairs of The Republic of Indonesia, April 24, 2019, https://kemlu.go.id/newyork-un/en/read/statement-at-the-international-day-of-multilateralism-and-diplomacy- for-peace-high-level-meeting-of-the-general-assembly/2610/etc-menu. 59 MoFA-RI, “Strategic Purposes Of The Indonesian Foreign Policy (Accessed 1 August 2020) \| Portal Kementerian Luar Negeri Republik Indonesia,” Ministry of Foreign Affairs of The Republic of Indonesia, March 25, 2019, https://kemlu.go.id/portal/en/read/19/halaman_list_lainnya/strategic-purposes-of-the-indonesian- foreign-policy (accessed 1 August 2020). 60 Ted Piccone “Do New Democracies Support Democracy? The Multilateral Dimension ” Journal of 61 “The Impact of Democratisation on Indonesia’s Foreign Policy: Regional Cooperation” (The London School of Economics and Political Science, 2014), 77–78. 60 Ted Piccone, “Do New Democracies Support Democracy? The Multilateral Dimension,” Journal of Democracy 22, no. 4 (2011): 140, https://doi.org/10.1353/JOD.2011.0063. 62 Desca LN and Rahmat Nasution, “Indonesia Needs Spirit of ‘gotong Royong’ in COVID-19 Fight: Jokowi,” Antara New, 2020, https://en.antaranews.com/news/146268/indonesia-needs-spirit-of-gotong-royong-in-covid- 19-fight-jokowi. 63 Office of Assistant to Deputy Cabinet Secretary for State Documents and Translation, “President Jokowi Calls for Close Cooperation among Countries to Tackle COVID-19,” Office of Assistant to Deputy Cabinet Secretary for State Documents and Translation, 2021, https://setkab.go.id/en/president-jokowi-calls-for-close-cooperation- among-countries-to-handle-covid-19/. 64 Hiebert, “Indonesia Steps Up Global Health Diplomacy: Bolsters Role in Addressing International Medical Challenges.” 65 Kiswondari, “DPR Minta Kemlu Maksimalkan Diplomasi Dalam Penanggulangan Covid-19,” Sindonews.com, 2020, https://nasional.sindonews.com/read/78374/12/dpr-minta-kemlu-maksimalkan-diplomasi- dalam-penanggulangan-covid-19-1592878004. 66 Lawrence O Gostin and Rebecca Katz, “The International Health Regulations: The Governing Framework for Global Health Security,” The Milbank Quarterly 94, no. 2 (June 1, 2016): 266–67, https://doi.org/https://doi.org/10.1111/1468-0009.12186. 67 Uluslararası İlişkiler Konseyi and Dilaver Açar, Global Governance and the COVID-19 Pandemic, 2021, 10, https://doi.org/10.13140/RG.2.2.35706.31687. 68 Gostin and Katz, “The International Health Regulations: The Governing Framework for Global Health Security,” 267–68. a. Drivers of Indonesia’s Multilateralism As the constructivist approach suggests, an actor’s behavior is dependent on the norm-based identity. In the domestic level, Indonesia’s multilateralism reflecting gotong royong was embraced and practiced by the MoFA. The norm was a basis of the MoFA’s identity construction through which represented the state identity, interest and behavior. Ted Piccone noted that Indonesia already experienced a significant reorientation in its foreign policy due to the remarkable transformation to open and pluralist democracy following the regime change. 60 Despite the fact that democratization has also occured to the foreign policy making institution during reformation era, Muhammad Hadianto Wirajuda in his dissertation “The Impact of Democratization on Indonesia’s Foreign Policy” argues that the MoFA is still the most central body in Indonesian foreign policy making. Such situation enabled what the MoFA believed to be tranformed into Indonesia’s policy. 61 By this explanation, we erdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic assumed and believed that the idea of ‘gotong royong’ fits in to our explanation of Indonesia’s multilateralism as its approach to response and overcome the pandemic of Covid-19. In addition to the MoFA’s capability of constructing the state identity, this paper attempts not to ignore other executive bodies such as the President and the MoH in the Indonesian foreign policy. President Jokowi has repeatedly utilized the speech on gotong royong in fighting the Covid-19 in domestic level62 and called on close cooperation among states to tackle the pandemic63. The MoFA’s actions would not be independent from the president’s approval as he acted as a head of the republic’s government. The MoH also recorded a number of international engagements regarding global health issues64. Thus, the MoH has been the important part of foreign policy making when it comes to health issues. In addition to executive body, the Indonesian parliament also demanded the MoFA to maximize Indonesia’s diplomacy to defend multilateralism practice65. sThose actors’ role played a part as the push factor strengthening the MoFA’s commitment to the norm of gotong royong which then influenced the state preference on multilateralism. The IHR was officially established in 1969 as a revision of the previously known the 1951 International Sanitary Regulation (ISR). The regulation was fundamentally revised in 2005 following the spread of new endemic diseases as well as viral outbreak in some parts of the word. 66 The regulation aims to: 62 Desca LN and Rahmat Nasution, “Indonesia Needs Spirit of ‘gotong Royong’ in COVID-19 Fight: Jokowi,” Antara New, 2020, https://en.antaranews.com/news/146268/indonesia-needs-spirit-of-gotong-royong-in-covid- 19-fight-jokowi. 66 Lawrence O Gostin and Rebecca Katz, “The International Health Regulations: The Governing Framework for Global Health Security,” The Milbank Quarterly 94, no. 2 (June 1, 2016): 266–67, https://doi.org/https://doi.org/10.1111/1468-0009.12186. 63 Office of Assistant to Deputy Cabinet Secretary for State Documents and Translation, “President Jokowi Calls for Close Cooperation among Countries to Tackle COVID-19,” Office of Assistant to Deputy Cabinet Secretary for State Documents and Translation, 2021, https://setkab.go.id/en/president-jokowi-calls-for-close-cooperation- among-countries-to-handle-covid-19/. 67 Uluslararası İlişkiler Konseyi and Dilaver Açar, Global Governance and the COVID-19 Pandemic, 2021, 10, https://doi.org/10.13140/RG.2.2.35706.31687. 68 Gostin and Katz, “The International Health Regulations: The Governing Framework for Global Health Security,” 267–68. 65 Kiswondari, “DPR Minta Kemlu Maksimalkan Diplomasi Dalam Penanggulangan Covid-19,” Sindonews.com, 2020, https://nasional.sindonews.com/read/78374/12/dpr-minta-kemlu-maksimalkan-di dalam-penanggulangan-covid-19-1592878004. among-countries-to-handle-covid-19/. 64 Hiebert, “Indonesia Steps Up Global Health Diplomacy: Bolsters Role in Addressing International Medical Challenges.” 65 Kiswondari, “DPR Minta Kemlu Maksimalkan Diplomasi Dalam Penanggulangan Covid-19,” “Prevent, protect against, control and provide a public health response to the international spread of disease in ways that are commensurate with and restricted to public health risks, and which avoid unnecessary interference with international traffic and trade.” 67 “Prevent, protect against, control and provide a public health response to the international spread of disease in ways that are commensurate with and restricted to public health risks, and which avoid unnecessary interference with international traffic and trade.” 67 The IHR is an international treaty upholding global cooperation to make the world more secured which is also often called a governing framework for global health security and the WHO is the one responsible to ensure its implementation among the member states. The regulation suggests that all states cooperate when facing a global threat of infectious disease. The WHO requires the member states to develop and meet core capacities in the IHR such as to detect, assess, report and respond to potential public emergencies of international concern. In spite of aiming to prevent, protect against, control and provide a public health response to international spread of disease, the IHR embraces a balancing dynamic of three things; public health, commerce and human rights. 68 It means that state domestic measures to tackle any infectious disease and other health 62 Desca LN and Rahmat Nasution, “Indonesia Needs Spirit of ‘gotong Royong’ in COVID-19 Fight: Jokowi,” Antara New, 2020, https://en.antaranews.com/news/146268/indonesia-needs-spirit-of-gotong-royong-in-covid- 19-fight-jokowi. 63 Office of Assistant to Deputy Cabinet Secretary for State Documents and Translation, “President Jokowi Calls for Close Cooperation among Countries to Tackle COVID-19,” Office of Assistant to Deputy Cabinet Secretary for State Documents and Translation, 2021, https://setkab.go.id/en/president-jokowi-calls-for-close-cooperation- among-countries-to-handle-covid-19/. 64 Hiebert, “Indonesia Steps Up Global Health Diplomacy: Bolsters Role in Addressing International Medical Ch ll ” 67 Uluslararası İlişkiler Konseyi and Dilaver Açar, Global Governance and the COVID-19 Pandemic, 2021, 10, https://doi.org/10.13140/RG.2.2.35706.31687. erdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic problem risks should be neither restrictive to international travel and trade nor harmful to human rights as written in a number of IHR’s articles. Indonesia has been complying with the revised IHR since 2007 and established its national commission for the implementation in 2011. 69 Indonesia announced its achievement to meet the IHR core capacity requirements in 2014. 70 The committment to the preparedness from detecting to responding to the previous avian influenza, SARS and MERS-CoV could be examples of Indonesia’s favor of the IHR. “Prevent, protect against, control and provide a public health response to the international spread of disease in ways that are commensurate with and restricted to public health risks, and which avoid unnecessary interference with international traffic and trade.” 67 This research not only viewed gotong royong as an Indonesian domestic norm but it also viewed the IHR as a global norm in relations between states and other international communities as suggested by the holistic constructivism. The WHO was the most important actor which encouraged the member states to adopt the IHR in their domestic efforts to handle any threat of infectious disease. As a norm entrepreneur, the health organization is at the forefront of international parnership including with Indonesia. In addition, Indonesia’s experience in handling the previous infectious diseases strengthened its belief to be with the WHO and IHR. Handling the current Covid-19 could be a dilemma for Indonesia; saving domestic public health without harming international travel and trade. Nevertheless, the dynamic balance prescribed in the IHR was a value to which Indonesia was committed. The global norm contributed to the construction of Indonesia’s identity through interactions with other international actors as well as its interpretation of their behaviors. The identity led Indonesia to prefer multilateralism in responding to the Covid-19. Multilateralism which is a common good has been adopted by the United Nations (UN) as written in its charter. As the representation of states, the UN could be the most active body promoting multilateralism and the United Nations’ General Assembly (UNGA) already adopted a resolution on the International Day of Multilateralism and Diplomacy for Peace. Multilateralism is also a common value adopted by the European Union and emerging powers. 71 So do small and developing states convey their concern on multilateral efforts in many challenging global issues. Considering the trend of multilateralism, Indonesia trusted that working with international institutions would more effectively solve global issues rather than with unilateral measures. According to the constructivist approach, normative structure and state mutually constitute. This paper argues that Indonesia’s identity was constructed by the domestic norm of gotong royong and international norm of IHR leading its foreign policy to embrace multilateralism. In other words, multilateralism was a collective expectation constructed by the state (and domestic actors). Upon realizing its norm-based identity, Indonesia had to consider what action was proper to respond to the Covid-19. In this case, taking multilateralism-laden foreign policy to mitigate the pandemic was Indonesia’s preference. 69 Nyoman Kumara Rai et al., “Strengthening Emergency Preparedness and Response Systems: Experience from Indonesia,” WHO South-East Asia Journal of Public Health 9, no. 1 (2020): 27, https://apps.who.int/iris/handle/10665/331827. 70 i l 2 71 Hans W Maull, “Multilateralisml: Variants, Potential, Constraints and Conditions for Success,” March (Berlin, 2020), 1. W Maull, “Multilateralisml: Variants, Potential, Constraints and Conditions for Success,” March (Berlin, b. Upholding Multilateralism to Mitigate the Covid-19 The Covid-19 is a global problem which has been affecting all states very quickly regardless their geographical distance, economic capability and preparedness. The actions of a few states 1 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandem ironically did not reflect constructive diplomacy since they tended to be uncoordinated and encouraged by the nationalistic sentiment instead of response as part of international society. 72 One of the phenomena is state isolationism. In classical realism thoughts, it is rational for state to put forward its own national interest regarding to its consideration on anarchic international system and its struggle for power among other nations. This isolationism apparently has influenced and shaped—some of but not all—major powers’ behaviors and policies toward Covid-19. United Kingdom is one of countries which applied some isolationist policy in response to Covid-19 regarding to the travel restrictions and visas suspension for immigrant workers. 73 Ralf Roloff also noted that a few European states took similar measures such as France, Italy and Spain at the beginning of the pandemic74. Another example is related to ‘vaccine nationalism’ where states regulate agreements with its pharmaceutical sectors to only supply their own populations with vaccines ahead of them becoming available for others due to strong public and political pressures and fear of waning immunity. 75 While some countries exhibit characteristics of isolationism, endangering the health and food supply chains that are currently required to combat the effect of the Covid-19, it shows that many other states are doing the exact opposite, reaching out through international cooperation, and seeing the benefits of internationalism rather than isolationism. No exception for Indonesia, which realized that each state would hardly be able to avoid it. Instead, Indonesia believed that only with cooperation and collaboration would all states be able to mitigate the pandemic. Indonesia believed that multilateralism opened the pathway for all sides to tackle all issues including the Covid-19 through coordinations to cooperations and from which the interests of states could find middle grounds. Indonesia was putting efforts to engage with both global and regional organizations in order to mitigate the pandemic in addition to maintainig support to the WHO. 75 Caroline E Wagner et al., “Vaccine Nationalism and the Dynamics and Control of SARS-CoV-2.,” Science (New York, N.Y.) 373, no. 6562 (September 2021): eabj7364, https://doi.org/10.1126/science.abj7364. 76 Rebecca Katz et al., “Global Health Security Agenda and the International Health Regulations: Moving Forward,” Biosecurity and Bioterrorism-Biodefense Strategy Practice and Science 12 (2014): 231–38. , p y , g , 73 James Melville, “Isolationism Is Not the Way Forward during Coronavirus \| Coronavirus Pandemic \| Al Jazeera,” Aljazeera, April 10, 2020, https://www.aljazeera.com/opinions/2020/4/10/isolationism-is-not-the-way- forward-during-coronavirus. A Triwibowo, “Diplomacy and Covid-19: A Reflection,” Jurnal Ilmiah Hubungan Internasional, 2020. g 74 Ralf Roloff, “COVID-19 and No One’s World,” Connections 19, no. 2 (2020): 25–37. 72 A Triwibowo, “Diplomacy and Covid-19: A Reflection,” Jurnal Ilmiah Hubungan Internasional, 2020. 73 James Melville, “Isolationism Is Not the Way Forward during Coronavirus \| Coronavirus Pandemic \| Al Jazeera,” Aljazeera, April 10, 2020, https://www.aljazeera.com/opinions/2020/4/10/isolationism-is-not-the-way- forward-during-coronavirus. 74 Ralf Roloff, “COVID-19 and No One’s World,” Connections 19, no. 2 (2020): 25–37. 75 Caroline E Wagner et al., “Vaccine Nationalism and the Dynamics and Control of SARS-CoV-2.,” Science (New York, N.Y.) 373, no. 6562 (September 2021): eabj7364, https://doi.org/10.1126/science.abj7364. 76 Rebecca Katz et al., “Global Health Security Agenda and the International Health Regulations: Moving Forward,” Biosecurity and Bioterrorism-Biodefense Strategy Practice and Science 12 (2014): 231–38. ew York, N.Y.) 373, no. 6562 (September 2021): eabj7364, https://doi.org/10.1126/science.abj7364. Rebecca Katz et al., “Global Health Security Agenda and the International Health Regulations: Moving rward,” Biosecurity and Bioterrorism-Biodefense Strategy Practice and Science 12 (2014): 231–38. better deal with future outbreaks.” 77 better deal with future outbreaks.” 77 The WHO was facing some critisisms over its lack of transparency and China-centric response accusation. 78 In addition, A. Chapelan 79 argued that the termination of the U.S - WHO relationships could worsen existing global health governanance crisis. Brett Meyer noted that populist leaders in a few states would either downplay or take the Covid-19 seriously80. Though they took different initial actions to respond to the pandemic, they had similar policy putting aside the WHO’s views on the new disease. Indonesia, however, still retained trust on the health organization which would coordinate the international cooperation against the Covid-19. Instead, as a member of the WHO, Indonesia expected that the global health organization would be able to facilitate cooperation for effective mitigation on the pandemic. Furthermore, it would suggest that the WHO evaluate how much the current health governance system could work against the pandemic and some future similar issues. The stand indicated that Indonesia admitted the importance to continue standing behind the organization. That the Indonesian government such as the most apparent the Indonesian Minister of Health, Lt Gen (Ret.) Terawan Agus Putranto downplayed the Covid-19 at the beginning of the Covid-19 spread in January and February 2020 did not mean Indonesia rejected multilateralism. We assume it was merely a domestic effort to prevent panic among citizens and maintain domestic stability. Furthermore, Indonesia and other states in the world were still blind about the new disease during the first two months of its spreading and were still waiting for more official guidances from the WHO. In fact, the MoH was one of the participants in the Foreign Policy and Global Health (FPGH) which officially stated to back the WHO and international cooperation against the Covid-1981. The commitment was increasing following the WHO’s declaration on the pandemic status of the Covid-19. The WHO issued priority strategies in March 2020 in responding to the Covid-19 ranging from expanding, training and deploying health-care workers; implementing systems to locate suspected cases; raising production of tests and availability; identifying facilities which could be transformed into health centres; developing plans to quarantine cases; and refocussing measures on suppressing the virus. 82 The strategies were to be undertaken by all states including Indonesia which counted on the organization as its commitment to the multilateralism. Maintaining Trust on and Support to the WHO The The IHR urges the WHO and global communities’ responsibility to share resources, information and expertise to help states to prepare for and respond to the health emergencies76. The regulation entrusts the WHO to be the most leading global organization to consolidate efforts and encourage states to comply with it. Embracing the global norm, Indonesia suggested the global health organization be not ignored during the Covid-19 as Minister Marsudi said during her remark at the video conference of Ministerial Meeting for the Alliance for Multilateralism (MM-AM) on 16 April 2020 recorded by the MoFA: “No country is able to overcome this crisis alone. In this connection, we have no other option but to count on the WHO to be a platform of cooperation among countries. This crisis provides a momentum to review the global health governance system so that it will be able to 42 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic 77 MoFA-RI, “Joint Statement Of The Special Asean Plus Three Summit On Coronavirus Disease 2019 Covid 19 \| Portal Kementerian Luar Negeri Republik Indonesia,” Ministry of Foreign Affairs of the Republic of Indonesia, April 14, 2020, https://kemlu.go.id/portal/en/read/1206/pidato/joint-statement-of-the-special-asean-plus-three- summit-on-coronavirus-disease-2019-Covid-19. 78 Balogun and Soile, “‘Pandemic Diplomacy’ and The Politics of Paradox: International Cooperation in The Age of Distancing,” 423. 79 “The World Health Organization and the Political Embroilment of Wounded Multilateralism,” Revista de Stiinte Politice 8, no. March 29 (2020): 137, https://library.laredo.edu/eds/detail?db=edsdoj&an=edsdoj.383b84c377194dd7a91923e34c7e954a. 80 Brett Meyer, “Pandemic Populism: An Analysis of Populist Leaders’ Responses to Covid-19.,” The Tony Blair Institute for Global Change., 2020, https://institute.global/policy/pandemic-populism-analysis-populist- leaders-responses-covid-19. 81 Ministry of Health Republic of Indonesia, “Komitmen Kemenkes Mendukung WHO Dan Kerjasama Internasional Menghadapi COVID 19.,” Ministry of Health Republic of Indonesia, 2020, https://sehatnegeriku.kemkes.go.id/baca/umum/20200518/3033922/indonesia-pimpin-foreign-policy-and-global- health-dukung-upaya-global-akhiri-pandemi-covid-19/. 82 Djalante et al., “Review and Analysis of Current Responses to COVID-19 in Indonesia: Period of January to March 2020,” 2. 77 MoFA-RI, “Joint Statement Of The Special Asean Plus Three Summit On Coronavirus Disease 2019 Covid 19 \| Portal Kementerian Luar Negeri Republik Indonesia,” Ministry of Foreign Affairs of the Republic of Indonesia, April 14, 2020, https://kemlu.go.id/portal/en/read/1206/pidato/joint-statement-of-the-special-asean-plus-three- summit-on-coronavirus-disease-2019-Covid-19. 81 Ministry of Health Republic of Indonesia, “Komitmen Kemenkes Mendukung WHO Dan Kerja Internasional Menghadapi COVID 19.,” Ministry of Health Republic of Indonesia, 2020, https://sehatnegeriku.kemkes.go.id/baca/umum/20200518/3033922/indonesia-pimpin-foreign-pol health-dukung-upaya-global-akhiri-pandemi-covid-19/ 81 Ministry of Health Republic of Indonesia, “Komitmen Kemenkes Mendukung WHO Dan Kerjasama Internasional Menghadapi COVID 19.,” Ministry of Health Republic of Indonesia, 2020, h // h ik k k id/b / /20200518/3033922/i d i i i f i li Balogun and Soile, “‘Pandemic Diplomacy’ and The Politics of Paradox: International Cooperation in Th ge of Distancing,” 423. g p , y p , , https://sehatnegeriku.kemkes.go.id/baca/umum/20200518/3033922/indonesia-pimpin-foreign-policy-and health-dukung-upaya-global-akhiri-pandemi-covid-19/. p 81 Ministry of Health Republic of Indonesia, “Komitmen Kemenkes Mendukung WHO Dan Kerjasama Internasional Menghadapi COVID 19.,” Ministry of Health Republic of Indonesia, 2020, https://sehatnegeriku.kemkes.go.id/baca/umum/20200518/3033922/indonesia-pimpin-foreign-policy-and-global- health-dukung-upaya-global-akhiri-pandemi-covid-19/. 82 Djalante et al “Review and Analysis of Current Responses to COVID 19 in Indonesia: Period of January to p y j j Brett Meyer, “Pandemic Populism: An Analysis of Populist Leaders’ Responses to Covid-19.,” The Tony air Institute for Global Change., 2020, https://institute.global/policy/pandemic-populism-analysis-populist aders-responses-covid-19. Ministry of Health Republic of Indonesia, “Komitmen Kemenkes Mendukung WHO Dan Kerjasama ternasional Menghadapi COVID 19 ” Ministry of Health Republic of Indonesia 2020 82 Djalante et al., “Review and Analysis of Current Responses to COVID-19 in Indonesia: Period of January to March 2020,” 2. 83 Hameiri, “Avian Influenza, ‘Viral Sovereignty’, and the Politics of Health Security in Indonesia,” 334. 84 David P Fidler, “Influenza Virus Samples, International Law, and Global Health Diplomacy,” Emerging Infectious Diseases 14, no. 1 (2008): 88. 85 Shawn Smallman, “Biopiracy and Vaccines: Indonesia and the World Health Organization’s New Pandemic Influenza Plan,” Journal of International & Global Studies, 2013. 86 Cornelia Clara Tifany, “Securitization on the Spread of COVID-19 in Indonesia [Sekuritasasi Dalam Penyebaran Covid-19 Di Indonesia],” Jurnal Politica Dinamika Masalah Politik Dalam Negeri Dan Hubungan Internasional 11, no. 2 (December 5, 2020): 139, https://doi.org/10.22212/JP.V11I2.1749. better deal with future outbreaks.” 77 Indonesia also stepped up its support to the Global Alliance for Vaccine and Immunization (GAVI) as the WHO’s partner in 82 Djalante et al., “Review and Analysis of Current Responses to COVID-19 in Indonesia: Period of January to March 2020,” 2. erdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic researching, developing, producing and distributing the Covid-19 vaccines to all states. Such support was aimed to prevent any state of finding itself difficult to access the vaccines. Indonesia kept announcing the vital role of the WHO in a number of international fora and kindly demanded international communities to keep cooperating with the organization. By doing so, Indonesia and others could support the health organization as a global platform to address the Covid- 19 collectively. Indonesia believed in the organization for having experience and expertise when facing many challenges on past global health issues. Indonesia cooperated in a number of health issues together with it as well. Even when the viral dispute of avian influenza between the two once occured, the health cooperations between the officials preserved. 83 Regarding Indonesia’s decision to withhold the avian influenza virus samples to the WHO by the end of 2006, we argue that the action did not mean that Indonesia withdrew itself from multilateralism practice. Indonesia’s demand to the WHO was the effort to call on major changes in global influenza vaccine production and distribution mechanism towards an equitable, fair, and transparent mechanism. Fidler noted Indonesia was discontent with the Global Influenza Surveillance Network led by the WHO which shared virus samples to the vaccine manufacturers based in the developed states without the consent of source states84. While the developed states could produce the vaccines and patent them, the developing states would have a little access to them. Indonesia’s stand to dispute the viral sharing was supported by other developing states and the Third World Network (TWN); an international non-governmental advocacy group promoting the reform on the global health governance. Indonesia demanded multilateralism in fighting avian influenza providing benefits for all. In addition, Indonesia was willing to resolve the dispute with the WHO through multilateral processes mediated by the World Health Assembly (WHA). The WHO accommodated Indonesia and developing world states’ aspiration by issuing the 2011 Pandemic Influenza Plan in order to provide better access to the vaccines85. Engaging within Global and Regional Frameworks In order to ensure all international entities to engage in international collaboration and cooperation, Indonesia believed that solidifying international solidarity was prerequisite. The cooperation emboldened by solidarity without stigma and discrimination should be advanced as all states shared responsibility to fight the Covid-19. Indonesia together with other five states, namely Ghana, Liechtenstein, Norway, Singapore and Switzerland pushed the global solidarity against the Covid-19 pandemic to be a resolution of the UNGA on 2 April 2020 which was supported by 188 member states. 87 It was the effort within Indonesia’s diplomacy mainly to solidify the global solidarity facing adversity amidst the rising trend of unilateralism among a number of states. Due to the fact that the Covid-19 caused numerous impacts, Indonesia expected that through multilateral institutions all states would not face difficulties in accessing all needs such as essential protective equipment as well as other medical supplies. Meanwhile, following the outbreak the trade should not be disrupted in order to minimize the economic loss. In this regard, at the Video Conference Ministerial Meeting for the Alliance for Multilateralism as recorded by the MoFA, Minister Marsudi said, “We must ensure that our multilateral system delivers to address the immediate need of our people.” 88 Indonesia conveyed a message that all states regardless their status should be able to take advantage of international cooperation and collaboration which were valued in multilateralism especially during the deadly pandemic. In times of early Covid-19, as delivered by Minister Marsudi at the CSIS Indonesia’s discussion forum, Indonesia was engaging with several regional or global groupings to respond to the pandemic such as the ASEAN Plus Three Special Summit (APT-SS), Organization of Islamic Cooperation Steering Committee on Health (OIC-SCH), Group of Twenty Extraordinary Leaders Summit (G20-ELS), and International Coordination Group on Covid-19 (ICGC). 89 ASEAN was the first regional institution maintaining commitment to colaborative culture when self-help mechanism was increasing among many states in response to the Covid-19. 90 On 14 April 2020, ASEAN member states along with China, Japan and ROK held a joint statement on the Covid-19 following a few meetings among their leaders. 87 Harun Umar and Irma Indrayani, “Foreign And Domestic Policy, Diplomacy and Cooperation, International Political Dynamics),” Journal of Social Political Sciences 1, no. 2 (May 31, 2020): 113, http://e- journal.unas.ac.id/index.php/jsps/article/view/14. 88 MoFA-RI, “Remarks Minister For Foreign Affairs Of The Republic Of Indonesia Video Conference Ministerial Meeting For The Alliance For Multilateralism \| Portal Kementerian Luar Negeri Republik Indonesia,” Ministry of Foreign Affairs of The Republic of Indonesia, April 16, 2020, https://kemlu.go.id/portal/id/read/1226/pidato/remarks-minister-for-foreign- affairs-of-the-republic-of-indonesia-video-conference-ministerial-meeting-for-the-alliance-for-multilateralism. 89 Retno LP Marsudi, “International Cooperation to Mitigate #COVID19: Indonesia, Southeast Asia, and Beyond \| CSIS,” CSIS Indonesia, April 17, 2020, https://csis.or.id/events/international-cooperation-to-mitigate-covid19-indonesia-southeast- asia-and-beyond. 90 Riska Putri Hariyadi, “Singapura Dan Asean: Analisis Relasi Negara dan Institusi Kawasan di Tengah Pandemi Covid-19,” Jurnal Dinamika Global 5, no. 02 (December 10, 2020): 312, https://doi.org/10.36859/JDG.V5I02.219. 91 MoFA-RI, “Joint Statement Of The Special Asean Plus Three Summit On Coronavirus Disease 2019 Covid 19 \| Portal Kementerian Luar Negeri Republik Indonesia.” 90 Riska Putri Hariyadi, “Singapura Dan Asean: Analisis Relasi Negara dan Institusi Kawasan di Tengah Pandemi Covid-19,” Jurnal Dinamika Global 5, no. 02 (December 10, 2020): 312, https://doi.org/10.36859/JDG.V5I02.219. 91 MoFA-RI, “Joint Statement Of The Special Asean Plus Three Summit On Coronavirus Disease 2019 Covid 19 \| Portal Kementerian Luar Negeri Republik Indonesia.” 92 MoFA RI y Riska Putri Hariyadi, “Singapura Dan Asean: Analisis Relasi Negara dan Institusi Kawasan di Tengah Pandemi Covi rnal Dinamika Global 5, no. 02 (December 10, 2020): 312, https://doi.org/10.36859/JDG.V5I02.219. better deal with future outbreaks.” 77 During the dispute, Indonesia did not cease its relations with the WHO and other aspects of cooperation beyond the viral sharing continued between the two. We propose an argument that apart from the its experience and expertise, the WHO’s decision to finally accommodate Indonesia’s demand has helped the state regain trust on the organization in fighting the pandemic. The latter event contributed to the Indonesia’s position towards the WHO in tackling the Covid-19 as well. Furthermore, Indonesia’s willingness to finally securitize the Covid-19 pandemic at home was influenced by the WHO. 86 Indonesia saw the organization having a role as a global health governance in disease control as mandated by the IHR despite its past unpleasant experience. Maintaining trust on and support to the WHO indicated that Indonesia upholded multilateralism in its foreign policy in responding to the Covid-19. As suggested by the constructivism, Indonesia’s behavior towards the organization was driven by its identity constructed by both domestic and global norm. While the norm of gotong royong prescribed Indonesia to cooperate with the WHO as frequently argued by Minister Marsudi and the MoFA, the norm of IHR suggested that the organization receive mandate to facilitate and coordinate the health emergency response among states. In other words, the norms regulated 4 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandem Indonesia to stand together with the health organization to fight against the pandemic. 45 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic 45 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic The OIC played its role in curbing the Covid-19 in the Moslem world as its charter paved the cooperation and coordination among member states in humanitarian emergencies. 93 The OIC-SCH facilitated the sharing of knowledge and information on the pandemic and discussion of ways and means to coordinate individual and collective response alike among the member states. Indonesia kept inviting the member states to strengthen coordination and cooperation against the pandemic. In March 2020, G20 leaders held an extraordinary summit on Covid-19 and delivered their commitment to overcoming the pandemic together with its global partners. 94 During the summit, the G20 leaders was raising concern on fighting the pandemic, safeguarding global economy, addressing international trade disruptions and enhancing global cooperation. The summit recognized the pandemic as a common threat that solidarity along with a transparent, robust, coordinated, large-scale and science-based global response was necessary in interconnected and vulnerable world. Minister Marsudi reiterated such virtual meetings were the concrete forms of partnership which had to be continuously carried out to tackle global problem especially in the unprecedented situation caused by the Covid-19. 95 Moreover, Indonesia realized that its stand could only be heard in international fora in which it actively participated in addressing the pandemic. Therefore, Indonesia used all channels available to coordinate, collaborate and cooperate with international communities in both regional and global level. Not only could such channels facilitate cooperation in responding to the pandemic, they could be suitable media to promote Indonesia’s call for sustaining multilateralism. The ICGC was aimed to resolve challenges regarding the Covid-19 especially the vaccines. The vaccine invention was the key global effort that would only stop the virus and it was, unfortunately, like a race among states and their companies to be the first producers. Indonesia’s efforts to ensure the vaccine affordability could be the challenge in its diplomacy since the production and distribution would be concentrated in developed states. However, Indonesia kept sounding that the vaccine allocation should be equitable for all including the vulnerable groups96. In addition, at the ICGC forum, Indonesia also proposed a platform with which companies from different states could cooperate in increasing capacity and production of health equipment and medicines within joint production schemes97. The above mentioned roles were played by Indonesia to show its commitment to multilateralism practice. 93 M Hussain, Socio-Economic Impacts of Covid-19 Pandemic in OIC Member Countries: Prospects and Challenges (Ankara: Statistical, Economic, and Social Research and Training Centre for Islamic Countries, 2020). 94 Wiwiek Rukmi Dwi Astuti, “Kerja Sama G20 Dalam Pemulihan Ekonomi Global Dari COVID-19,” Andalas Journal of International Studies (AJIS) 9, no. 2 (November 30, 2020): 138, https://doi.org/10.25077/AJIS.9.2.131-148.2020. 95 Jessica Martha, “Pemanfaatan Diplomasi Publik Oleh Indonesia Dalam Krisis Covid-19,” Jurnal Ilmiah Hubungan Internasional, 2020. 96 Abdul Kadir Jailani, “How Indonesian Diplomacy Can Ensure Equitable Access to Covid-19 Vaccine,” Jakarta Globe, 2020, https://jakartaglobe.id/opinion/how-indonesian-diplomacy-can-ensure-equitable-access-to-covid19-vaccine. 97 Anadolu Agency, “Indonesia Usulkan Negara ICGC Kerja Sama Alat Kesehatan Covid-19,” Anadolu Agency, 2020, https://www.aa.com.tr/id/nasional/indonesia-usulkan-negara-icgc-kerja-sama-alat-kesehatan-Covid-19/1822241. 98 Yulius P. Hermawan and Ahmad D. Habir, “Indonesia and International Institutions: Treading New Territory,” in Indonesia’s Ascent: Power, Leadership, and the Regional Order (London: Palgrave Macmillan UK, 2015), 188, https://doi.org/10.1057/9781137397416_9. 99 “Th R l f Middl P i A i M ltil t li ” A i P li 25 4 (2018) 42 43 p g 99 “The Role of Middle Powers in Asian Multilateralism,” Asia Policy 25, no. 4 (2018): 42–43, https://doi.org/10.1353/ASP.2018.0060. Engaging within Global and Regional Frameworks 91 The MoFA announced that the summit acknowledged people’s participation and welcomed international financial institutions to support states in need using and enhancing their instruments in addition to stressing the importance of the WHO’s role in leading campaign to control and contain the viral spread and to implement health measures under the IHR. 92 Since the APT had role in creating peace, security and prosperity in East Asia, Indonesia expected the role of APT to build solidarity and cooperation in responding to the pandemic in the region. 46 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic 46 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic Not only was Indonesia promoting the spirit of multilateralism but it was also encouraging collective leadership and win-win solutions in responding to the Covid-19. Such action was in accordance with the characteristic of today global problems. They could be more complex, multidimensional and wide-ranging so that international actors need global institutional cooperation more than ever. 100 Engaging within global and regional frameworks reflected Indonesia (and the MoFA)’s preference of multilateralism practice. In addition to the domestic norm of gotong royong, Indonesia realized that the IHR demanded states not only to cooperate but also to maintain the economy without harming people’s rights in the middle of national and international efforts to mitigate the health emergency. None of state policies should risk other states including people during the the Covid-19 pandemic. Therefore, cooperation was the best option appropriate to its identity which was constituted by the norms as suggested by the constructivist approach. 100 Konseyi and Açar, Global Governance and the COVID-19 Pandemic, 1. 101 Marsudi, “Cooperation: The Catalyst for Fighting the COVID-19 Crisis - Opinion - The Jakarta Post.” 102 Marsudi, “International Cooperation to Mitigate #COVID19: Indonesia, Southeast Asia, and Beyond \| CSIS.” 103 Marsudi. 104 Marsudi. 105 Dyah Gandasari and Diena Dwidienawati, “Content Analysis of Social and Economic Issues in Indonesia during the COVID-19 Pandemic,” Heliyon 6, no. 11 (November 1, 2020), https://doi.org/10.1016/J.HELIYON.2020.E05599. 105 Dyah Gandasari and Diena Dwidienawati, “Content Analysis of Social and Economic Issues in Indonesia during the COVID-19 Pandemic,” Heliyon 6, no. 11 (November 1, 2020), https://doi.org/10.1016/J.HELIYON.2020.E05599. 103 Marsudi. 104 Marsudi. 105 Dyah Gandasari and Diena Dwidienawati, “Content Analysis of Social and Economic Issues in Indonesia during the COVID-19 Pandemic,” Heliyon 6, no. 11 (November 1, 2020), https://doi.org/10.1016/J.HELIYON.2020.E05599. y ç , , 101 Marsudi, “Cooperation: The Catalyst for Fighting the COVID-19 Crisis - Opinion - The Jakarta Post.” 102 Marsudi, “International Cooperation to Mitigate #COVID19: Indonesia, Southeast Asia, and Beyond \| CSIS.” 103 M di 45 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic Indonesia could undertake such roles thanks to its active engagement in various international institutions long attached to its foreign policy under a traditional independent and active guiding narrative. In addition, Indonesia believed that multilateral institutions could work better to solve global problems as it had repeatedly expressed idealistic views on the role of international institutions as agencies sharing benefits to all sides. 98 Indonesia’s role was aligned with the common practice among the Asian states considering norms into their foreign policy as well. Ralf Emmers 99 suggested that the Asian states show its involvement with global institutions through normative strategy in order to manage negative impacts of competitions. 106 Prawita Meidi Handayani, “Nationalism Over Globalization amidst COVID-19 Pandemic and Its Impact on Indonesia’s Food Security,” Jurnal Politica Dinamika Masalah Politik Dalam Negeri Dan Hubungan Internasional 11, no. 2 (December 5, 2020): 151, https://doi.org/10.22212/JP.V11I2.1751. 107 Gostin and Katz, “The International Health Regulations: The Governing Framework for Global Health Security,” 268. 108 Muhyiddin Muhyiddin and Hanan Nugroho, “A Year of Covid-19: A Long Road to Recovery and Acceleration of Indonesia’s Development,” Jurnal Perencanaan Pembangunan: The Indonesian Journal of Development Planning 5, no. 1 (April 30, 2021): 3, https://doi.org/10.36574/JPP.V5I1.181. 109 Luerdi Luerdi, “POLITIK LUAR NEGERI INDONESIA DI AWAL PANDEMI NOVEL CORONAVIRUS DISEASE: JANUARI–AGUSTUS 2020,” 2022. 110 Marsudi, “Cooperation: The Catalyst for Fighting the COVID-19 Crisis - Opinion - The Jakarta Post.” 111 Emmers, “The Role of Middle Powers in Asian Multilateralism,” 43. 112 Davies, Kamradt-Scott, and Rushton, Disease Diplomacy: International Norms and Global Health Security; Kamradt- Scott, “The WHO Secretariat, Norm Entrepreneurship, and Global Disease Outbreak Control”; Ferhani and Rushton, “The c. Priorities of Indonesia’s Foreign Policy 105 In order to mitigate the economic impact, Indonesia urged global and regional economic institutions to be reliable platforms to prevent global economy from deteriorating and help weaker states to minimize damages to their national economy amidst the global efforts to contain the further disease spread and 47 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic 47 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic flatten curves of the infected. flatten curves of the infected. Indonesia rejected export restriction and import suspension during the Covid-19. Those two trade policies were commonly taken by a number of states to tackle the pandemic at home, but they potentially harmed the weak states. 106 Indonesia believed that protectionism would not work and none of states would survive on its own during the pandemic. The flow of goods should be maintained despite the pandemic threat as one of the measures to mitigate the economic impact. Similar stand was also shown by Indonesia at other international fora. Indonesia’s demands were aligned with a number of IHR articles regulating and ensuring trade and travel during public health emergencies. 107 As the consequence, Indonesia was struggling to combine countermeasures against the pandemic without sacrificing the economy. 108 The Covid-19 forced many governments to enact firm measures on foreign nationals from quarantine requirement to expulsion. As the disease spread very quickly, many foreign governments imposed a strict measure like either partial or full lockdown in order to prevent further spread which impacted Indonesian nationals residing abroad such as becoming unemployed or being unable to continue their study. Realizing this situation, Indonesia’s foreign policy was directed to assure its nationals safe. The MoFA coordinated with foreign governments as well as Indonesian embassies or consulates so that they could either leave the states voluntarily in orderly manner or receive humanitarian aids unless they returned to Indonesia. For an exception, Indonesia conducted an evacuation mission of citizens trapped in isolation-imposed Wuhan after gaining permit from the Chinese government in early February 2020109. For foreign nationals at home, Indonesia maintained communication with their respective embassies regarding updates such as the people-under-surveilance, infected, death and recovered cases. In June 2020, Minister Marsudi announced that as many as 311 foreigners tested positive for the Covid-19, 26 of whom passed away and 204 recovered and another 485 classified as people under surveilance. c. Priorities of Indonesia’s Foreign Policy Despite multilateralism rhetoric, there existed several national interests which Indonesia was pursuing. Following the Covid-19 outbreak, Minister Marsudi announced that Indonesia’s foreign policy had three priorities, namely addressing the pandemic itself, mitigating economic impacts and protecting Indonesian citizens abroad and foreign nationals at home. 101 In term of addressing the pandemic, at many international fora, Indonesia kept inviting all states to work together to tackle the pandemic instead of blaming each other. During the pandemic, Indonesia suggested that protectionism and isolationism be avoided. The flows of medical equipment and other relevant needs across border should not meet any barrier so every government would be able to access them in order to save people’s lives or minimize victims. Minister Marsudi uttered that Indonesia already sent protection equipment to the states infected earlier before the virus was announced as a pandemic. 102 So should other states take the similar decision as more and more states would suffer due to lack of protection equipment. In her writing entitiled “Cooperation: The Catalyst for Fighting the Covid-19 Crisis” published in the Jakarta Post, Minister Marsudi argued that information sharing, technical assistance and joint production of basic medical equipment which had to be virtues in global friendship during difficult time were Indonesia’s diplomacy messages. 103 In addition, Indonesia believed that coordination with the WHO was imperative as ignoring its health protocol and recommendations would only cause the situation worse. Once the vaccines were found in the future, they should be affordable for all states including the third world instead of monopolizing them. The Covid-19 already impacted the global economy as the results of global supply chains disruption and restriction of people movements and border closures. The ADB projected that the Southeast Asia’s economic growth would drop by 1 percent in 2020, while the World Bank Group (WBG) predicted there would be another 11 million people falling into poverty in East Asia and the Pacific. 104 Indonesia’s Rupiah and many industries were weakening in addition to capital flight. Conclusion Indonesia’s active foreign policy during the unprecedented Covid-19 could be an oppositional option compared to what to be taken by more established states and other parts of the world. This paper argues that non-material factors do matter in defining Indonesia’s foreign policy and behavior especially when confronting a new challenging global issue. Upholding multilateralism and pursuing national interests would not contradict as the Indonesia’s foreign policy was affected by the indigenous and global norm complementary to each other. International fora were the main media for Indonesia to promote the practice of multilateralism. In addition, this paper reveals that looking at the domestic actors is as important as looking at state in international system in order to reach a better kowledge on the issue. There has been a trend showing that a number of developed states with better health infrastructure and stronger economic foundation seemed to act uncoordinatedly to take serious initial action to respond to the pandemic or downplay the severity of the disease at the beginning of the pandemic. Meanwhile, Indonesia as one of developing states raised multilateralism suggesting a few actions such as coordinating, collaborating and cooperating with international communities. The policy was not the result of applying rationalist principle of foreign policy decision makers who attempted to utilize the existing international norm and overcome domestic shortages for maximizing national interests as suggested by the liberal approach. Neither would Indonesia have taken the former action in order to protect its people and save economy ahead before others if it had been better equipped with health infrastructure as the neorealist approach suggests. Otherwise, we propose that Indonesia’s foreign policy would remain to uphold multilateralism when facing other similar crisis thanks to its identity as suggested by the constructivist approach. Only when the identity changes would Indonesia redirect its foreign policy and interests. It is hardly seen in the near future. This paper has stated a novel theoretical argument in which Indonesia’s multilateralism in the pandemic era represents a confluence of domestic norms (gotong royong) and international norms (IHR). The gotong royong along with other international norms is potential to be expanded into other areas of foreign policy including—but not limited to—security, climate change, human rights, developmentalism and socio-economic inequality. c. Priorities of Indonesia’s Foreign Policy 110 The Indonesian government only restricted foreign nationals to enter the state by suspending the visa exemption policy and visa on arrival facilities along with the firm WHO health protocol instead of totally barring them. In addition, such measures were in accordance with the principles of respecting human rights as stipulated in the IHR. Indonesia’s foreign policy during the Covid-19 was not only aimed to protect its citizens and economy but also others. Communication and coordination with other foreign governments were necessary to pursue the three priorities. Unilateral and uncoordinated actions could risk Indonesia’s interests as well as others’ during the time of uncertainties. It is in line with the Indonesian foreign policy practice as noted by Emmers that Indonesia – just like other Asian states – preserved its national interests by encouraging the adoption of norms and standards of good international behavior. 111 Even though a few states ignored the IHR, the regulation’s role as a new norm sustained during the pandemic. Furthermore, the emergence of new transmissible diseases has made health an important element in state diplomacy guided by the regulation. 112 48 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic Considering the constructivist approach, the roles and interests within Indonesia’s foreign policy were determined by its norm-based identity. Gotong royong emphasizing mutual assistance and the IHR suggesting the balance of trade, travel and health led Indonesia to uphold multilateralism. In domestic context, these priorities refer to Indonesia’s national interests which were not only determined by the MoFA alone but also other domestic actors. Such national interests reflect the identity which belongs to Indonesia influenced by these actors through the MoFA’s embracing of multilateralism. Meanwhile, in the systemic level the interactions between Indonesia and other international communities brought about an understanding that the IHR was an international norm which was infused to its identity. Only after Indonesia realized its identity could it determine its interests deemed appropriate. Indonesia would have responsibility to pursue such interests as a consequence of the identity awareness. International Health Regulations, COVID-19, and Bordering Practices: Who Gets in, What Gets out, and Who Gets Resc ed?” Conclusion Though this paper provides a view of Indonesia’s foreign policy in responding to the Covid-19 through the lense of holistic constructivism, it leaves questions – as of the time of writing – if such foreign policy effectively contributes to tackle the pandemic at home and save its economy as the number of confirmed cases keeps surging and recession threat alarms in addition to if Indonesia is able to leverage its leadership and role in regional and 9 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandem global politics in addressing the pandemic as many see multilateral institutions underperforming without the role of more established states. A few future researches can be expected to fill the gaps. Above all, the study benefits scholars or students who have interest in Indonesia’s foreign policy in order to understand how the state behaves during difficult times. References “COVID-19 and ASEAN Responses: Comparative Policy Analysis.” Progress in Disaster Science 8 (2020). https://doi.org/10.1016/j.pdisas.2020.100129. Djalante, R, L Nurhidayah, H Van Minh, N T N Phuong, Y Mahendradhata, A Trias, J Lassa, and M A Miller. “COVID-19 and ASEAN Responses: Comparative Policy Analysis.” Progress in Disaster Science 8 (2020). https://doi.org/10.1016/j.pdisas.2020.100129. Djalante, Riyanti, Jonatan Lassa, Davin Setiamarga, Aruminingsih Sudjatma, Mochamad Indrawan, Budi Haryanto, Choirul Mahfud, et al. “Review and Analysis of Current Responses to COVID- 19 in Indonesia: Period of January to March 2020.” Progress in Disaster Science 6 (2020): 100091. https://doi.org/https://doi.org/10.1016/j.pdisas.2020.100091. Dodds, Klaus, Vanesa Castan Broto, Klaus Detterbeck, Martin Jones, Virginie Mamadouh, Maano Ramutsindela, Monica Varsanyi, David Wachsmuth, and Chih Yuan Woon. “The COVID-19 Pandemic: Territorial, Political and Governance Dimensions of the Crisis.” Territory, Politics, Governance 8, no. 3 (May 26, 2020): 289–98. https://doi.org/10.1080/21622671.2020.1771022. Dunford, Michael, and Bing Qi. “Global Reset: COVID-19, Systemic Rivalry and the Global Order.” Research in Globalization 2 (2020): 100021. https://doi.org/https://doi.org/10.1016/j.resglo.2020.100021. Emmers, Ralf. “The Role of Middle Powers in Asian Multilateralism.” Asia Policy 25, no. 4 (2018): 42–47. https://doi.org/10.1353/ASP.2018.0060. Ferhani, A, and S Rushton. “The International Health Regulations, COVID-19, and Bordering Practices: Who Gets in, What Gets out, and Who Gets Rescued?” Contemporary Security Policy 41, no. 3 (2020): 458–77. https://doi.org/10.1080/13523260.2020.1771955. Fidler, David P. “Asia’s Participation in Global Health Diplomacy and Global Health Governance.” Asian Journal of WTO & International Health Law and Policy 5, no. 2 (2010). https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1729455#references-widget. ———. “Influenza Virus Samples, International Law, and Global Health Diplomacy.” Emerging Infectious Diseases 14, no. 1 (2008): 88. Gandasari, Dyah, and Diena Dwidienawati. “Content Analysis of Social and Economic Issues in Indonesia during the COVID-19 Pandemic.” Heliyon 6, no. 11 (November 1, 2020). https://doi.org/10.1016/J.HELIYON.2020.E05599. Gostin, Lawrence O, and Rebecca Katz. “The International Health Regulations: The Governing Framework for Global Health Security.” The Milbank Quarterly 94, no. 2 (June 1, 2016): 264– 313. https://doi.org/https://doi.org/10.1111/1468-0009.12186. Hameiri, Shahar. “Avian Influenza, ‘Viral Sovereignty’, and the Politics of Health Security in Indonesia.” The Pacific Review 27, no. 3 (May 27, 2014): 333–56. https://doi.org/10.1080/09512748.2014.909523. Hampson, Fen Osler, and Paul Heinbecker. “The ‘New’ Multilateralism of the Twenty-First Century.” Global Governance: A Review of Multilateralism and International Organizations 17, no. 3 (2011): 299–310. https://doi.org/https://doi.org/10.1163/19426720-01703003. Handayani, Prawita Meidi. “Nationalism Over Globalization amidst COVID-19 Pandemic and Its Impact on Indonesia’s Food Security.” Jurnal Politica Dinamika Masalah Politik Dalam Negeri Dan Hubungan Internasional 11, no. 2 (December 5, 2020): 143–61. https://doi.org/10.22212/JP.V11I2.1751. Hariyadi, Riska Putri. References Agung, Ide Anak Agung Gde. Twenty Years Indonesian Foreign Policy 1945–1965. De Gruyter Mouton, 1973. https://doi.org/doi:10.1515/9783111558226. Akon, Md. Saifullah, and Mahfujur Rahman. “Reshaping the Global Order in the Post COVID-19 Era: A Critical Analysis.” Chinese Journal of International Review 02, no. 01 (June 1, 2020): 2050006. https://doi.org/10.1142/S2630531320500067. Alexandrov, M. “The Concept of State Identity in International Relations: A Theoretical Analysis,” 2003. Anadolu Agency. “Indonesia Usulkan Negara ICGC Kerja Sama Alat Kesehatan Covid-19.” Anadolu Agency, 2020. https://www.aa.com.tr/id/nasional/indonesia-usulkan-negara-icgc-kerja-sama-alat- kesehatan-Covid-19/1822241. Astuti, Wiwiek Rukmi Dwi. “Kerja Sama G20 Dalam Pemulihan Ekonomi Global Dari COVID-19.” Andalas Journal of International Studies (AJIS) 9, no. 2 (November 30, 2020): 131–48. https://doi.org/10.25077/AJIS.9.2.131-148.2020. Bainus, Arry, and Junita Budi Rachman. “Editorial: Pandemi Penyakit Menular (Covid-19) Hubungan Internasional.” Intermestic: Journal of International Studies; Vol 4 No 2 (2020)DO - 10.24198/Intermestic.V4n2.1 , May 31, 2020. http://intermestic.unpad.ac.id/index.php/intermestic/article/view?path=. Balogun, Wasiu Abiodun, and Oluwaseun Soile. “‘Pandemic Diplomacy’ and The Politics of Paradox: International Cooperation in The Age of Distancing.” Gaziantep University Journal of Social Sciences 19, no. COVID-19 Special Issue (October 31, 2020): 413–28. https://doi.org/10.21547/JSS.780792. Bol, Damien, Marco Giani, Andre Blais, and Peter John Loewen. “The Effect of COVID-19 Lockdowns on Political Support: Some Good News for Democracy?” European Journal of Political Research 60, no. 2 (May 1, 2021): 497–505. https://doi.org/https://doi.org/10.1111/1475-6765.12401. Chapelan, Alexis. “The World Health Organization and the Political Embroilment of Wounded Multilateralism.” Revista de Stiinte Politice 8, no. March 29 (2020). https://library.laredo.edu/eds/detail?db=edsdoj&an=edsdoj.383b84c377194dd7a91923e34c7e954 a. Curley, Melissa G, and Jonathan Herington. “The Securitisation of Avian Influenza: International Discourses and Domestic Politics in Asia.” Review of International Studies 37, no. 1 (2011): 141–66. https://doi.org/DOI: 10.1017/S0260210510000537. Davies, Sara E. Containing Contagion: The Politics of Disease Outbreaks in Southeast Asia. Baltimore: John Hopkins University Press, 2019. ———. “Securitizing Infectious Disease.” International Affairs 84, no. 2 (March 1, 2008): 295–313. https://doi.org/10.1111/j.1468-2346.2008.00704.x. ———. “The International Politics of Disease Reporting: Towards Post-Westphalianism?” ———. “Securitizing Infectious Disease.” International Affairs 84, no. 2 (March 1, 2008): 295–313. https://doi.org/10.1111/j.1468-2346.2008.00704.x. ———. “The International Politics of Disease Reporting: Towards Post-Westphalianism?” International Politics 49, no. 5 (2012): 591–613. https://doi.org/10.1057/ip.2012.19. Davies, Sara E, Adam Kamradt-Scott, and S Rushton. Disease Diplomacy: International Norms and Global Health Security. Baltimore: John Hopkins University Press, 2015. 0 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandem Djalante, R, L Nurhidayah, H Van Minh, N T N Phuong, Y Mahendradhata, A Trias, J Lassa, and M A Miller. References “Singapura Dan ASEAN: Analisis Relasi Negara Dan Institusi Kawasan Di Tengah Pandemi Covid-19.” Jurnal Dinamika Global 5, no. 02 (December 10, 2020): 299–318. https://doi.org/10.36859/JDG.V5I02.219. Hermawan, Yulius P., and Ahmad D. Habir. “Indonesia and International Institutions: Treading New Territory.” In Indonesia’s Ascent: Power, Leadership, and the Regional Order, 177–94. London: Palgrave Macmillan UK, 2015. https://doi.org/10.1057/9781137397416_9. Hiebert, Murray. “Indonesia Steps Up Global Health Diplomacy: Bolsters Role in Addressing International Medical Challenges.” Center for Strategic and International Studies, 2013. 51 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic https://www.csis.org/analysis/indonesia-steps-global-health-diplomacy. https://www.csis.org/analysis/indonesia-steps-global-health-diplomacy. Hurd, Ian. “The Ethics of Constructivism.” In The Oxford Handbook of International Relations. New York: Oxford University Press, Inc, 2008. Hussain, M. Socio-Economic Impacts of Covid-19 Pandemic in OIC Member Countries: Prospects and Challenges. Ankara: Statistical, Economic, and Social Research and Training Centre for Islamic Countries, 2020. Jailani, Abdul Kadir. “How Indonesian Diplomacy Can Ensure Equitable Access to Covid-19 Vaccine.” Jakarta Globe, 2020. https://jakartaglobe.id/opinion/how-indonesian-diplomacy-can- ensure-equitable-access-to-covid19-vaccine. Jepperson, R.I, Alexander Wend, and P.J Katzenstein. “Norms, Identity, and Culture in National Security.” In The Culture of National Security: Norms and Identity in World Politics. New York: Columbia University Press, 1996. Kamradt-Scott, Adam. “The WHO Secretariat, Norm Entrepreneurship, and Global Disease Outbreak Control.” Journal of International Organization Studies 1 (January 1, 2010). Katz, Rebecca, M Sorrell, Erin, A Kornblet, Sarah, and E Fischer, Julie. “Global Health Security Agenda and the International Health Regulations: Moving Forward.” Biosecurity and Bioterrorism-Biodefense Strategy Practice and Science 12 (2014): 231–38. Katzenstein, P.J. “Introduction: Alternative Perspective on National Security.” In The Culture of National Security: Norms and Identity in World Politics, edited by P.J Katzenstein. New York: Columbia University Press, 1996. Kiswondari. “DPR Minta Kemlu Maksimalkan Diplomasi Dalam Penanggulangan Covid-19.” Sindonews.com, 2020. https://nasional.sindonews.com/read/78374/12/dpr-minta-kemlu- maksimalkan-diplomasi-dalam-penanggulangan-covid-19-1592878004. Konseyi, Uluslararası İlişkiler, and Dilaver Açar. Global Governance and the COVID-19 Pandemic, 2021. https://doi.org/10.13140/RG.2.2.35706.31687. Koschut, Simon. “The Power of (Emotion) Words: On the Importance of Emotions for Social Constructivist Discourse Analysis in IR.” Journal of International Relations and Development 21, no. 3 (2018): 495–522. https://doi.org/10.1057/s41268-017-0086-0. Lestari, Shindy Diah Ayu. “PANDEMIC OF COVID-19 IN INDONESIA: SOCIAL POLITICS PERSPECTIVE (Pandemi Covid-19 Di Indonesia: Perspektif Sosial Politik).” ETNOREFLIKA: Jurnal Sosial Dan Budaya 10, no. 1 (February 27, 2021): 31–45. https://doi.org/10.33772/ETNOREFLIKA.V10I1.915. LN, Desca, and Rahmat Nasution. “Indonesia Needs Spirit of ‘gotong Royong’ in COVID-19 Fight: Jokowi.” Antara New, 2020. https://en.antaranews.com/news/146268/indonesia-needs-spirit-of- gotong-royong-in-covid-19-fight-jokowi. Luerdi, Luerdi. “POLITIK LUAR NEGERI INDONESIA DI AWAL PANDEMI NOVEL CORONAVIRUS DISEASE: JANUARI–AGUSTUS 2020,” 2022. References Marsudi, Retno LP. “Cooperation: The Catalyst for Fighting the COVID-19 Crisis - Opinion - The Jakarta Post.” The Jakarta Post, March 25, 2020. https://www.thejakartapost.com/academia/2020/03/24/cooperation-the-catalyst-for-fighting-the- covid-19-crisis.html. “I i l C i Mi i #COVID19 I d i S h A i d B d \| Marsudi, Retno LP. “Cooperation: The Catalyst for Fighting the COVID-19 Crisis - Opinion - The Jakarta Post.” The Jakarta Post, March 25, 2020. https://www.thejakartapost.com/academia/2020/03/24/cooperation-the-catalyst-for-fighting-the- covid-19-crisis.html. ———. “International Cooperation to Mitigate #COVID19: Indonesia, Southeast Asia, and Beyond \| CSIS.” CSIS Indonesia, April 17, 2020. https://csis.or.id/events/international-cooperation-to- ———. “International Cooperation to Mitigate #COVID19: Indonesia, Southeast Asia, and Beyond \| CSIS.” CSIS Indonesia, April 17, 2020. https://csis.or.id/events/international-cooperation-to- mitigate-covid19-indonesia-southeast-asia-and-beyond. Martha, Jessica. “Pemanfaatan Diplomasi Publik Oleh Indonesia Dalam Krisis Covid-19.” Jurnal Ilmiah Hubungan Internasional, 2020. Maull, Hans W. “Multilateralisml: Variants, Potential, Constraints and Conditions for Success.” March. Berlin, 2020. 52 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic Melville, James. “Isolationism Is Not the Way Forward during Coronavirus \| Coronavirus Pandemic \| Al Jazeera.” Aljazeera, April 10, 2020. https://www.aljazeera.com/opinions/2020/4/10/isolationism-is-not-the-way-forward-during- coronavirus. Melville, James. “Isolationism Is Not the Way Forward during Coronavirus \| Coronavirus Pandemic \| Al Jazeera.” Aljazeera, April 10, 2020. https://www.aljazeera.com/opinions/2020/4/10/isolationism-is-not-the-way-forward-during- coronavirus. Meyer, Brett. “Pandemic Populism: An Analysis of Populist Leaders’ Responses to Covid-19.” The Tony Blair Institute for Global Change., 2020. https://institute.global/policy/pandemic-populism- analysis-populist-leaders-responses-covid-19. Mietzner, Marcus. “Populist Anti-Scientism, Religious Polarisation, and Institutionalised Corruption: How Indonesia’s Democratic Decline Shaped Its COVID-19 Response.” Journal of Current Southeast Asian Affairs 39, no. 2 (August 1, 2020): 227–49. https://doi.org/10.1177/1868103420935561. Miles, Matthew B, and A. Michael Huberman. Qualitative Data Analysis: An Expanded Sourcebook, 2nd Ed. London: SAGE Publications Inc, 1994. Ministry of Health Republic of Indonesia. “Komitmen Kemenkes Mendukung WHO Dan Kerjasama Internasional Menghadapi COVID 19.” Ministry of Health Republic of Indonesia, 2020. https://sehatnegeriku.kemkes.go.id/baca/umum/20200518/3033922/indonesia-pimpin-foreign- policy-and-global-health-dukung-upaya-global-akhiri-pandemi-covid-19/. MoFA-RI. “International Day of Multilateralism and Diplomacy for Peace High-Level Meeting of the General Assembly.” Ministry of Foreign Affairs of The Republic of Indonesia, April 24, 2019. https://kemlu.go.id/newyork-un/en/read/statement-at-the-international-day-of-multilateralism- and-diplomacy-for-peace-high-level-meeting-of-the-general-assembly/2610/etc-menu. ———. “Joint Statement Of The Special Asean Plus Three Summit On Coronavirus Disease 2019 Covid 19 \| Portal Kementerian Luar Negeri Republik Indonesia.” Ministry of Foreign Affairs of the Republic of Indonesia, April 14, 2020. https://kemlu.go.id/portal/en/read/1206/pidato/joint- statement-of-the-special-asean-plus-three-summit-on-coronavirus-disease-2019-covid-19. ———. “Remarks Minister For Foreign Affairs Of The Republic Of Indonesia Video Conference Ministerial Meeting For The Alliance For Multilateralism \| Portal Kementerian Luar Negeri Republik Indonesia.” Ministry of Foreign Affairs of The Republic of Indonesia, April 16, 2020. References https://kemlu.go.id/portal/id/read/1226/pidato/remarks-minister-for-foreign-affairs-of-the- republic-of-indonesia-video-conference-ministerial-meeting-for-the-alliance-for-multilateralism. ———. “Strategic Purposes Of The Indonesian Foreign Policy (Accessed 1 August 2020) \| Portal Kementerian Luar Negeri Republik Indonesia.” Ministry of Foreign Affairs of The Republic of Indonesia, March 25, 2019. https://kemlu.go.id/portal/en/read/19/halaman_list_lainnya/strategic- purposes-of-the-indonesian-foreign-policy (accessed 1 August 2020). Muhyiddin, Muhyiddin, and Hanan Nugroho. “A Year of Covid-19: A Long Road to Recovery and Acceleration of Indonesia’s Development.” Jurnal Perencanaan Pembangunan: The Indonesian Journal of Development Planning 5, no. 1 (April 30, 2021): 1–19. https://doi.org/10.36574/JPP.V5I1.181. Office of Assistant to Deputy Cabinet Secretary for State Documents and Translation. “President Jokowi Calls for Close Cooperation among Countries to Tackle COVID-19.” Office of Assistant to Deputy Cabinet Secretary for State Documents and Translation, 2021. https://setkab.go.id/en/president-jokowi-calls-for-close-cooperation-among-countries-to-handle- covid-19/. Oshewolo, Segun, and Agaptus Nwozor. “COVID-19: Projecting the National Security Dimensions of Pandemics.” Strategic Analysis 44, no. 3 (May 3, 2020): 269–75. https://doi.org/10.1080/09700161.2020.1767911. Pattiradjawane, R.L, and N. Soebagio. “Global Maritime Axis: Indonesia, China, and a New Approach to Southeast Asian Regional Resillience.” International Journal of China Studies 6, no. 2 (2015). erdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic Piccone, Ted. “Do New Democracies Support Democracy? The Multilateral Dimension.” Journal of Democracy 22, no. 4 (2011): 139–52. https://doi.org/10.1353/JOD.2011.0063. Piccone, Ted. “Do New Democracies Support Democracy? The Multilateral Dimension.” Journal of Democracy 22, no. 4 (2011): 139–52. https://doi.org/10.1353/JOD.2011.0063. Purbantina, Adiasri Putri, and Renitha Dwi Hapsari. “Diplomasi Kesehatan Di Era Pandemik Global: Analisa Bantuan Penanganan Covid-19 Dari Negara Jepang Dan Korea Selatan Ke Indonesia.” Global and Policy Journal of International Relations 8, no. 01 (July 9, 2020). https://doi.org/10.33005/JGP.V8I01.2167. Rai, Nyoman Kumara, Rim Kwang Il, Endang Widuri Wulandari, Frieda Subrata, Anung Sugihantono, and Vensya Sitohang. “Strengthening Emergency Preparedness and Response Systems: Experience from Indonesia.” WHO South-East Asia Journal of Public Health 9, no. 1 (2020): 26–31. https://apps.who.int/iris/handle/10665/331827. Riley, Christine, Bo Xie, and Anjum Khurshid. “Challenges Encountered in Comparing International Policy Responses to COVID-19 and Their Effects.” Health Research Policy and Systems 19, no. 1 (2021): 134. https://doi.org/10.1186/s12961-021-00783-1. Roloff, Ralf. “COVID-19 and No One’s World.” Connections 19, no. 2 (2020): 25–37. Ruggie, John Gerard. “Multilateralism: The Anatomy of an Institution.” International Organization 46, no. 3 (1992). https://www.jstor.org/stable/2706989. Rüland, Jürgen. “The Rise of ‘Diminished Multilateralism’: East Asian and European Forum Shopping in Global Governance.” Asia Europe Journal 9, no. 2 (2012): 255–70. https://doi.org/10.1007/s10308-012-0311-9. Schindler, Seth, Nicholas Jepson, and Wenxing Cui. “Covid-19, China and the Future of Global Development.” Research in Globalization 2 (2020): 100020. https://doi.org/https://doi.org/10.1016/j.resglo.2020.100020. References Setiawan, Asep. “Politik Luar Negeri Indonesia Era Covid-19: Penyelamatan Dan Kerjasama.” Independen: Jurnal Politik Indonesia Dan Global 1 (September 1, 2020): 65–74. https://doi.org/10.24853/independen.1.2.65-74. Slikkerveer, L.J. “Gotong Royong: An Indigenous Institutions of Community and Mutual Assistance in Indonesia.” In Integrated Community Managed Development: Strategizing Indigenous Knowledge and Institutions for Poverty Reduction and Sustainable Community Develpment, edited by L.J. Slikkerveer. Cham: Springer, 2019. Smallman, Shawn. “Biopiracy and Vaccines: Indonesia and the World Health Organization’s New Pandemic Influenza Plan.” Journal of International & Global Studies, 2013. Sridhar, Devi, and Ngaire Woods. “Trojan Multilateralism: Global Cooperation in Health.” Global Policy 4, no. 4 (November 1, 2013): 325–35. https://doi.org/https://doi.org/10.1111/1758- 5899.12066. Sukmana, Mayusef, Muhammad Aminuddin, and Dwi Nopriyanto. “Indonesian Government Response In COVID-19 Disaster Prevention.” East African Scholars Journal of Medical Sciences 3, no. 3 (March 15, 2020). https://doi.org/10.36349/EASMS.2020.v03i03.025. Tifany, Cornelia Clara. “Securitization on the Spread of COVID-19 in Indonesia [Sekuritasasi Dalam Penyebaran Covid-19 Di Indonesia].” Jurnal Politica Dinamika Masalah Politik Dalam Negeri Dan Hubungan Internasional 11, no. 2 (December 5, 2020): 127–42. https://doi.org/10.22212/JP.V11I2.1749. Tollefson, Jeff. “How Trump Damaged Science — and Why It Could Take Decades to Recover.” Nature, 2020. https://www.nature.com/articles/d41586-020-02800-9. Triwibowo, A. “Diplomacy and Covid-19: A Reflection.” Jurnal Ilmiah Hubungan Internasional, 2020. Umar, Harun, and Irma Indrayani. “Foreign and Domestic Policy, Diplomacy and Cooperation, International Political Dynamics.” Journal of Social Political Sciences 1, no. 2 (May 31, 2020): 102–16. http://e-journal.unas.ac.id/index.php/jsps/article/view/14. 54 Luerdi, Azhari Setiawan \| Upholding Multilateralism: Indonesia’s Foreign Policy in Responding to Covid-19 Pandemic Wagner, Caroline E, Chadi M Saad-Roy, Sinead E Morris, Rachel E Baker, Michael J Mina, Jeremy Farrar, Edward C Holmes, et al. “Vaccine Nationalism and the Dynamics and Control of SARS- CoV-2.” Science (New York, N.Y.) 373, no. 6562 (September 2021): eabj7364. https://doi.org/10.1126/science.abj7364. Weber, Martin. “Between ‘Isses’ and ‘Oughts’: IR Constructivism, Critical Theory, and the Challenge of Political Philosophy.” European Journal of International Relations 20, no. 2 (January 4, 2013): 516–43. https://doi.org/10.1177/1354066112466573. Weible, C M, D Nohrstedt, P Cairney, D P Carter, D A Crow, A P Durnová, T Heikkila, K Ingold, A McConnell, and D Stone. “COVID-19 and the Policy Sciences: Initial Reactions and Perspectives.” Policy Sciences 53, no. 2 (2020): 225–41. https://doi.org/10.1007/s11077-020- 09381-4. Wenham, C. “Regionalizing Health Security: Thailand’s Leadership Ambitions in Mainland Southeast Asian Disease Control.” Contemporary Southeast Asia: A Journal of International and Strategic Affairs 40, no. 1 (2018): 126–51. Wicaksana, I Gede Wahyu. References “The Family State: A Non-Realist Approach to Understanding Indonesia’s Foreign Policy.” Asian Journal of Political Science 27, no. 3 (September 2, 2019): 308–28. https://doi.org/10.1080/02185377.2019.1686997. Wirajuda, Muhammad Hasan. “The Impact of Democratisation on Indonesia’s Foreign Policy: Regional Cooperation.” The London School of Economics and Political Science, 2014. Yoosefi Lebni, Javad, Jaffar Abbas, Farideh Moradi, Mohammad Reza Salahshoor, Fakhreddin Chaboksavar, Seyed Fahim Irandoost, Nazila Nezhaddadgar, and Arash Ziapour. “How the COVID-19 Pandemic Effected Economic, Social, Political, and Cultural Factors: A Lesson from Iran.” International Journal of Social Psychiatry 67, no. 3 (July 2, 2020): 298–300. https://doi.org/10.1177/0020764020939984. Zhang, J.J., and V.R. Savage. “The Geopolitical Ramifications of COVID-19: The Taiwanese Exception.” Eurasian Geography and Economics, 2020, 1–18. https://doi.org/10.1080/15387216.2020.1779773.
https://openalex.org/W2560215312	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0167452&type=printable	English	null	Cost Effectiveness of Screening Colonoscopy Depends on Adequate Bowel Preparation Rates – A Modeling Study	PloS one	2,016	cc-by	9,410	RESEARCH ARTICLE Results At the current assumed inadequate bowel preparation rate of 25%, the cost of colonoscopy as a screening strategy is above society’s willingness to pay (<$50,000/QALY). Threshold analysis demonstrated that an inadequate bowel preparation rate of 13% or less is neces- sary before colonoscopy is considered more cost effective than FIT. At inadequate bowel preparation rates of 25%, colonoscopy is still more cost effective compared to sigmoidos- copy and stool DNA test. Sensitivity analysis of all inputs adjusted by ±10% showed incre- mental cost effectiveness ratio values were influenced most by the specificity, adherence, and sensitivity of FIT and colonoscopy. Data Availability Statement: All relevant data are within the paper. Editor: John Green, University Hospital Llandough, UNITED KINGDOM Editor: John Green, University Hospital Llandough, UNITED KINGDOM A microsimulation model of CRC screening strategies for the general population at average risk for CRC. The strategies include fecal immunochemistry test (FIT) every year, colonos- copy every ten years, sigmoidoscopy every five years, or stool DNA test every 3 years. The screening could be performed at private practice offices, outpatient hospitals, and ambula- tory surgical centers. Received: August 22, 2016 Accepted: November 14, 2016 Published: December 9, 2016 Copyright: © 2016 Kingsley et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. James Kingsley1, Siddharth Karanth2, Frances Lee Revere2, Deepak Agrawal3* 1 Department of Internal Medicine, Texas Health Presbyterian Hospital, Dallas, Texas, United States of America, 2 School of Public Health, University of Texas Health Science Center, Houston, Texas, United States of America, 3 Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, Texas, United States of America * Deepak.Agrawal@UTSouthwestern.edu * Deepak.Agrawal@UTSouthwestern.edu OPEN ACCESS OPEN ACCESS Citation: Kingsley J, Karanth S, Revere FL, Agrawal D (2016) Cost Effectiveness of Screening Colonoscopy Depends on Adequate Bowel Preparation Rates – A Modeling Study. PLoS ONE 11(12): e0167452. doi:10.1371/journal. pone.0167452 Aim Citation: Kingsley J, Karanth S, Revere FL, Agrawal D (2016) Cost Effectiveness of Screening Colonoscopy Depends on Adequate Bowel Preparation Rates – A Modeling Study. PLoS ONE 11(12): e0167452. doi:10.1371/journal. pone.0167452 Citation: Kingsley J, Karanth S, Revere FL, Agrawal D (2016) Cost Effectiveness of Screening Colonoscopy Depends on Adequate Bowel Preparation Rates – A Modeling Study. PLoS ONE 11(12): e0167452. doi:10.1371/journal. pone.0167452 The aim of this study is to determine the impact of inadequate bowel preparation rate on the cost effectiveness of colonoscopy compared to other screening strategies for colorectal cancer (CRC). Abstract a1111 Cost Effectiveness of Screening Colonoscopy Depends on Adequate Bowel Preparation Rates – A Modeling Study James Kingsley1, Siddharth Karanth2, Frances Lee Revere2, Deepak Agrawal3* Background Inadequate bowel preparation during screening colonoscopy necessitates repeating colo- noscopy. Studies suggest inadequate bowel preparation rates of 20–60%. This increases the cost of colonoscopy for our society. OPEN ACCESS Citation: Kingsley J, Karanth S, Revere FL, Agrawal D (2016) Cost Effectiveness of Screening Colonoscopy Depends on Adequate Bowel Preparation Rates – A Modeling Study. PLoS ONE 11(12): e0167452. doi:10.1371/journal. pone.0167452 Data Availability Statement: All relevant data are within the paper. Funding: The author(s) received no specific funding for this work. Funding: The author(s) received no specific funding for this work. Competing Interests: The authors have declared that no competing interests exist. 1 / 17 PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 Colonoscopy and Bowel Preparation Rates Screening colonoscopy is not a cost effective strategy when compared with fecal immuno- chemical test, as long as the inadequate bowel preparation rate is greater than 13%. Screening colonoscopy is not a cost effective strategy when compared with fecal immuno- chemical test, as long as the inadequate bowel preparation rate is greater than 13%. Introduction Colonoscopy, sigmoidoscopy and stool tests for occult blood are recommended cost effective strategies for colorectal cancer (CRC) screening in the United States (US) [1]. The preferred method of CRC screening varies among different countries. Many countries have nationally organized screening programs that are centrally funded and controlled. These countries tend to favor stool tests. Other countries such as the US, have a more decentralized approach with multiple sources of funding and hence multiple modes of delivery, assessment and follow-up. In the US, screening colonoscopies are promoted as a preferred test. Despite significantly higher costs, colonoscopies are considered more cost-effective since they allow detection and removal of smaller precancerous polyps rather than simply early detection of cancer or in some cases advanced adenomas. The assumption in these analyses is that colonoscopies are performed and repeated per established guidelines i.e. if no polyp is detected (negative colo- noscopy) screening colonoscopy is repeated in ten years, if adenomatous polyps are detected (positive colonoscopy) the surveillance colonoscopy is performed in three to five years [2]. However, in reality, colonoscopies are often repeated earlier than recommended intervals. Studies suggest colonoscopy overuse rates up to 25%[3–15], which is a significant problem, since it exposes patients to procedure risks without added benefit and is a poor use of resources. Reasons for overuse of colonoscopies include endoscopists’ preferences [9–12] patients’ preferences [9–12] and inadequate bowel preparation [12–15]. Guidelines recommend repeat- ing colonoscopy within a year, when bowel prep is inadequate due to concern about missing polyps. The reported rate of inadequate bowel preparations ranges from 5%-60%, with most studies citing 25%[3–15]. This wide range indicates that it is possible to address the variables leading to inadequate bowel preparations. However, this is often not prioritized due to logisti- cal difficulties, resource investment and reimbursement pressures. It is estimated that population-wide inadequate bowel preparation results in colonoscopy being repeated every 7.8 years in average risk patients instead of every ten years [16]. This increases the cost of colonoscopy as a population screening strategy and may even make screening colonoscopy a less cost-effective strategy when compared to less efficacious strate- gies. We performed a cost analysis of colorectal cancer (CRC) screening strategies incorporat- ing inadequate bowel preparation rates to determine if colonoscopy is the most cost effective strategy. Decision Model We constructed a Markov state transition model to examine the cost-effectiveness of four CRC screening strategies. The model is based on the natural history of CRC. The model tracks indi- vidual patients from age 50 until death or age 100. All screening strategies stop at age 80 because screening beyond this age is not routinely recommended. Each cycle in the model is one year long. Each patient has an inherent risk of developing an adenoma depending on age, PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 2 / 17 Colonoscopy and Bowel Preparation Rates Fig 1. Markov Model. (Ω) varies by age according to US Actuarial Tables from 2014. (Φ) varies by risk factors: Age (0.011–0.019), History of High Grade Dysplasia (0.08), History of CRC (0.25). (Π) Diagnosis achieved through screening strategy (p value dependent on screening modality) & symptom presentation per year for total of 2 years (Local 0.22; Regional 0.40; Distant 1.00). (I) Costs of initial CRC care. (C) Costs of continual CRC care each year. (T) Costs of Terminal CRC care (End of Life). (LG) Low Grade. (HG) High Grade. doi:10.1371/journal.pone.0167452.g001 Fig 1. Markov Model. (Ω) varies by age according to US Actuarial Tables from 2014. (Φ) varies by risk factors: Age (0.011–0.019), History of High Grade Dysplasia (0.08), History of CRC (0.25). (Π) Diagnosis achieved through screening strategy (p value dependent on screening modality) & symptom presentation per year for total of 2 years (Local 0.22; Regional 0.40; Distant 1.00). (I) Costs of initial CRC care. (C) Costs of continual CRC care each year. (T) Costs of Terminal CRC care (End of Life). (LG) Low Grade. (HG) High Grade. doi:10.1371/journal.pone.0167452.g001 doi:10.1371/journal.pone.0167452.g001 sex, and history of high-grade polyp/CRC. The model tracks type of adenoma (low grade or high grade). High-grade adenomas are >1cm, three or more in number or with villous dys- plastic features on histology. All other adenomas are considered low-grade. These characteris- tics determine the probability of adenoma being found by a screening strategy and the risk of developing CRC, which progresses through stages: local, regional and distant colorectal cancer. Probability of survival after diagnosis of CRC depends on the stage of disease. Fig 1 depicts our model and possible transitions from one state to another. Screening strategies are superim- posed on the model. The effectiveness of a screening strategy is determined by the test’s ability to detect adenomas or CRC. Decision Model All key assumptions used in the model are listed in Table 1. The incidence of adenomatous polyps and colorectal cancer was determined from age-spe- cific (per year [prevalence rates at autopsy from years 1970 to 1990 reported in the Surveil- lance, Epidemiology, and End Results (SEER) program data [17]. These decades were chosen because population screening or surveillance was not widespread enough to skew the natural history model. If a person develops CRC he/she carries a new yearly probability rate of mortal- ity based on five-year survival rates and average length of survival by stage of cancer and age according to data from SEER [17,18]. To estimate the rates for death from causes other than CRC we used data from the USA Official Social Security Office’s Actuarial Life Tables of 2011. Initial, regional and distant CRC were assigned health utility weights for 0.9, 0.8 and 0.76 respectively [19]. Colonoscopy and Bowel Preparation Rates Table 1. Model Inputs: Case Base Estimates Used in the Model. Input Names Base Case Estimate Reference Used [Additional Support] FIT Col SDNA Sig Sensitivity Low Grade Polyp 0.1 0.85 0.172 0.85 [53] [2,19,22,23,54–64] Sensitivity High Grade Polyp 0.24 0.9 0.692 0.9 [53] [2,19,22,23,54–64] Sensitivity Colorectal Cancer 0.7 0.95 0.923 0.95 [53] [2,19,22,23,54–64] Speciﬁcity 0.95 0.95 0.866 0.92 [53] [2,19,22,23,54–64] Percent of polyps within reach of Endoscopy [regardless of whether strategy successfully detects polyp or not] - 100% - 60% [21] [2,22,23,35,55–63] Inadequate Bowel Preparation Rate 0.25 [33] [3–15] Probability that Inadequate Bowel Preparation is Poor Prep 0.29 [3] [57–64] Probability that Inadequate Bowel Preparation is Fair Prep 0.71 [3] [57–64] CRC Screening Adherence: % of population who are completely Up to Date 0.651 [29] [20,22] CRC Screening Adherence: % of population who have never been screened. Screening Strategies The Markov model assigns patients to undergo one of the following recommended screening strategies: annual fecal immunochemical testing (FIT), flexible sigmoidoscopy (FS) every five PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 3 / 17 PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 0.277 [29] [20,22] CRC Screening Adherence: % of population who have been screened but are not up to date 0.072 [29] [20,22] Surveillance Adherence for those Completely Up to Date 0.95 [19] [3,20,22] Surveillance Adherence for those Incompletely Up to Date 0.85 [19] [3,20,22] Surveillance Adherence for those Never Screened 0.75 [19] [3,20,22] Adherence to Instructions of Colonoscopy 0.53 [20] [2,21] Adherence to Instructions of FIT or SDNA 0.8 - 0.8 - [20] [2,21] Follow Up Adherence with Colonoscopy After + FIT or SDNA 0.83 - 0.83 0.83 [20] [2,3,22] Health Utility Weight for Initial CRC 0.9 [19] [21] Health Utility Weight for Regional CRC 0.8 [19] [21] Health Utility Weight for Distant CRC 0.76 [19][21] Probability of Local Cancer Diagnosis 0.4 [57–64] Probability of Regional Cancer Diagnosis 0.37 [57–64] Probability of Distant Cancer Diagnosis 0.23 [57–64] Probability of Local Cancer Cure 0.65 [57–64] Probability of Regional Cancer Cure 0.45 [57–64] Probability of Distant Cancer Cure 0 [57–64] CRC Treatment Mortality 0.02 [21][57–64] Complication Rate per Screening Event - 0.0016 - 0.00016 [65][21,26,33,57–64] Mortality Rate per Screening Event - 7.4x10-5 - - [65] [21,26,33,57–64] Prevalence of Polyp at First Screen 0.25 [66] [21,57–64] Incidence of Low Grade Polyp without history of polyps or CRC .011−.019* [66] [21,57–64] Incidence of High Grade Polyp without history of polyps or CRC 7.5x10-4 [66] [21,57–64] Incidence of Low & High Grade Polyp with history of polyps but not CRC 0.08 [66] [21,57–64] Incidence of Low & High Grade Polyp with history of CRC 0.25 [66] [21,57–64] Ratio of Polyp Incidence that are LG vs. HG < 65yrs/old 0.95 [66] [21,57–64] Ratio of Polyp Incidence that are LG vs. HG > 65yrs/old 0.679 [66] [21,57–64] Annual Transition Probability from: LG Polyp—> HG Polyp 0.02 [21] [57–64] Annual Transition Probability from: LG Polyp—> Local CRC Polyp 0.00697 [21] [57–64] Annual Transition Probability from: HG Polyp—> Local CRC 0.05 [21] [57–64] Annual Transition Probability from: Local CRC—> Regional CRC 0.28 [21] [57–64] Annual Transition Probability from: Regional CRC—> Distant CRC 0.63 [21] [57–64] Probability of Local Cancer Presentation each year for 2 years 0.22 [21] [57–64] Probability of Regional Cancer Presentation each year for 2 years 0.4 [21] [57–64] Probability of Distant Cancer Presentation over the course of 2 years 1 [21] [57–64] Probability of Colorectal Cancer Occurring without Polyp each year 6.0x10-5 [57–64] (Continued) PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 4 / 17 Colonoscopy and Bowel Preparation Rates Table 1. (Continued) ( ) Input Names Base Case Estimate Reference Used [Additional Support] FIT Col SDNA Sig Probability of Local CRC from Normal Epithelium 0.68 0.56 0.68 0.53 [21] [57–64] Probability of Regional from Normal Epithelium 0.22 0.32 0.22 0.33 [21] [57–64] Probability of Distant CRC from Normal Epithelium 0.1 0.12 0.1 0.14 [21] [57–64] Mortality Rate for Local Cancer each year for 5 years 0.0174 [20, 58–64] Mortality Rate for Regional Cancer each year for 5 years 0.086 [20, 58–64] Societal Willingness to Pay for 1 Additional QALY $50,000.00 [15] [25] Discount Rate Used for Costs & Effectiveness 3% [15] [20,25,58–64] CRC, colorectal cancer; LGD, low-grade dysplasia; HGD, high grade dysplasia; QALY, quality adjusted life years * varies with age years, colonoscopy every ten years and stool DNA test every three years. The characteristics of different screening tests were determined from independent review of literature (Table 1). Each screening strategy was assigned a different adherence rate based on published literature [20–22]. Patients undergoing colonoscopy could have an adequate or an inadequate prep. If the preparation was inadequate, colonoscopy was repeated at varying intervals based on aver- aged real world data [3]. Simulated patients with a positive stool test or polyps on sigmoidos- copy received a colonoscopy that same year. For the probability that the stool test was false positive, the patient underwent a negative colonoscopy, and then resumed with the primary screening strategy. If polyps were found, they were removed and patients were switched to sur- veillance colonoscopy in 3 years (high-grade adenomas) or 5 years (low-grade adenomas). We assumed that 95% of all colonoscopies reached the cecum [23]. PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 Cost of Screening Strategies and CRC Treatment The calculation of costs in this study is notable for a number of reasons. First, a weighted aver- age of Medicare and private insurance reimbursements was used for colonoscopies performed in an office setting, ambulatory surgical centers and hospital based endoscopy units. This is different from most prior studies on cost effectiveness, which calculated colonoscopy costs based only on Medicare payment rates. Medicare payment rate calculations result in underes- timation of total costs since Medicare accounts for only 31% of screening colonoscopies and private insurance pays 1.5 to 2 times more than Medicare [33]. Furthermore, the reimburse- ment for screening colonoscopies performed at hospital based endoscopy units can be up to 30% higher [34]. The second notable cost calculation is the incorporation of the cost of seda- tion provided by anesthesiologists. The study assumed that the sedation in 30% of colonosco- pies was administered by an anesthesiologist, although much higher numbers have been reported. Third, costs of lost wages and benefits due to the procedure were added, based on Bureau of Labor Statistics. Additionally, the cost for a follow-up colonoscopy was included for the fecal exams and the sigmoidoscopy that were positive (both true positives and false posi- tives). The baseline costs used for our analysis are given in the Fig 1. Net costs of CRC treatment from year 2009 were determined for each local, regional and advanced stage of cancer. These costs differed for the initial, continuing and terminal phases of care. All cost data associated with screening and treatment were adjusted to the real value of US dollar in 2014 using the Consumer Price Index for Medical Care. The costs of lost wages and benefits were derived from the Bureau of Labor Statistics and adjusted to 2014 US dollars using the Consumer Price Index. Colonoscopy and Bowel Preparation Rates entered a surveillance regime, with a 75% adherence rate [19,20]. Alternative adherence assumptions, including 100% adherence, were explored in a sensitivity analysis. Inadequate Bowel Preparation The definition of inadequate bowel preparation and the timing for repeat colonoscopy varies in both the theoretical realm and in practice. For this analysis, inadequate preparation was defined as those that result in a repeat colonoscopy earlier than the recommended interval for an inadequately prepared colon. Studies suggest that most gastroenterologists would repeat colonoscopies for intermediate (fair) and poor quality preps due to the possibility of missing polyps greater than five millimeter in size [20–32]. In our analysis we varied the inadequate bowel preparation rate using a threshold analysis, beginning at an ideal rate of 0% and ending at 30%. This range was based on the assumed cur- rent inadequate bowel preparation rate of 25% [3–15]. It was assumed that all patients with inadequate bowel preparations underwent repeat colonoscopy at the rates of 9% within one year, 5% in one to two years, 3% in two to three years and 83% in three to 10 years [14]. Nota- bly, the inadequate bowel preparation rate in the model had no influence on the effectiveness portion of the cost-effectiveness calculation as all patients were rescheduled for colonoscopy. Adherence Rates Most prior cost-effective studies assumed a 100% adherence rate for different screening strate- gies [24–27]. In reality, the adherence rates are much lower and differ among strategies. Since screening strategies vary considerably in costs, adherence rates can have large influence cost- effectiveness ratios. Non-adherence to subsequent screenings even within the same strategy can significantly affect the cost-effectiveness. However, even with a relatively a small degree of adherence with subsequent screenings, cost-effectiveness can be comparable to full participa- tion [24–28]. We used reported adherence rates for US population and accordingly divided the popula- tion into three groups. The adherence rate was broken down into two gates of adherence. The first gate was whether the patient received a referral/order for a screening test. The second gate was whether or not the patient actually adhered to the recommendations. 27% of patients were never recommended CRC screening, 8% received recommendation once in their lifetime and 65% received recommendations consistently according to guidelines [27]. The second gate evaluated whether or not patients adhered to the recommendations (of note, this doesn’t include inadequate bowel preparations). Adherence rates varied according to type of screening strategy with less adherence for endoscopic procedures (52%) versus stool tests (92%) [20]. Thus, the relative end of life adherence rate was 38% for colonoscopy and sigmoidoscopy and 67% for fecal exams. Patients with positive stool tests were assumed to be 83% adherent with follow-up colonoscopy [20]. In addition, while non-adherers did not receive screening, if they were diagnosed with CRC by symptom presentation, and received curative treatment, they PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 5 / 17 Comparison of Screening Strategies vs. No Screening The cost and effectiveness of different screening strategies are given in Table 2. Compared to no screening, FIT is the most cost effective (in terms of quality adjusted life year, QALY) screening strategy followed by colonoscopy, stool DNA and sigmoidoscopy. The cost of FIT ($1,579) is the least and stool DNA ($3,526) the most expensive. Only FIT and sigmoidoscopy are both more effective and less costly (dominant) compared with no screening. Colonoscopy cost $3,257 more per quality adjusted life year and stool DNA test cost $15,762 more per qual- ity adjusted life year when compared to no screening strategy. Results The base case scenario in our model was designed to reflect real-life CRC screening patterns in the US, which show an inadequate preparation rate of 25% and variation in adherence rates for different screening strategies. Colonoscopy and Bowel Preparation Rates utilize ratios, effectiveness calculations are not in the denominator and thus NMB is a better comparative measure when differences in effectiveness between strategies are small [34]. NMB was calculated by assuming a willingness to pay (WTP) of $50,000 per QALY gained. Costs and QALYs were discounted at an annual rate of 3%. The model was analyzed using TreeAge Pro 2015 (TreeAge Software Inc., Williamston, MA). The analysis included 1-way and probabilistic sensitivity analysis (PSA). One-way sensitiv- ity analysis was performed on key variables identified by a tornado diagram. For probabilistic sensitivity analysis- 100,000 individuals walked through a second order simulation and each individual’s probability (in second order) was the composite mean of 10,000 individuals walk- ing through a first order simulation. The PSA assumed a uniform distribution for the probabil- ities, and a beta distribution for the QALYs in the model. Model Validation External validation was performed by comparing our model’s outputs with those of well-estab- lished CEA models. For example, CEA performed by Sharaf et al [21] comparing FOBT, sig- moidoscopy and colonoscopy. The lifetime cost of colonoscopy per person in their model was $2564 compared to $2430 in our model. The Incremental Cost effective ratio (ICER) values for colonoscopy versus FIT reported by them were also similar to our model. We also compared intermediate outputs within our model with the National Polyp Study and other well-established epidemiological data produced by the American Cancer Society and the National Cancer Institute. In our model the current utilization rates for each strategy showed a 0.0593 lifetime CRC incidence rate compared to 0.05 published by the American Cancer Society [35]. Our natural history model, on which the screening strategies were super- imposed, had a standardized incidence based CRC mortality ratio of 0.40, which is comparable to the rate of 0.47 (95% CI .26–.80) published by the National Polyp Study [22]. Our standard- ized incidence based lifetime CRC incident ratio was 0.3989 compared to the National Polyp Study rate of 0.24 (95% CI .08–.56) [22]. Our model’s output for CRC mortality reduction with colonoscopies was 60%, which is comparable to the National Cancer Institute’s estimate of 60– 70% mortality reduction [35]. Methods of Model Analysis The cost-effectiveness of different CRC screening strategies was estimated using a societal perspective. Given the inherent limitations of ratios in cost-effectiveness (C/E), Net Monetary Benefit (NMB) was determined to allow for a more transparent comparison of multiple strategies. The NMB uses society’s willingness to pay for a quality adjusted life year (QALY) to convert the comparisons between strategies from ratios to arithmetic differences. Because NMB doesn’t 6 / 17 PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 Colonoscopy and Bowel Preparation Rates Table 2. CEA & ICER Results at 25% Inadequate Bowel Prep Rate for a cohort of 100,000 average-risk US citizens age 50 to 100yrs. FIT Sigmoidoscopy No Screening Colonoscopy 25% Stool DNA CEA Cost $1578.73 $1914.09 $2103.12 $2429.81 $3526.45 Effect 19.5057 19.4787 19.4168 19.5171 19.5071 NMB* $973706.27 $972020.91 $968736.88 $973425.19 $971246.00 ICER FIT - Sig is Dominated NS is Dominated $74656.14 $1391228.57 Sigmoidoscopy - - NS is Dominated $13430.21 $56773.24 No Screening - - - $3257.13 $15762.24 Colonoscopy 25% - - - - SDNA is Dominated Stool DNA - - - - - owel Prep Rate for a cohort of 100,000 average-risk US citizens age 50 to 100yrs. sults at 25% Inadequate Bowel Prep Rate for a cohort of 100,000 average-risk US citizens age 50 to 100yrs. le 2. CEA & ICER Results at 25% Inadequate Bowel Prep Rate for a cohort of 100,000 average-risk US citizen Abbreviations: CEA, Cost Effectiveness Analysis; ICER, incremental cost effectiveness ratio; FIT, fecal immunoassay test; NS, No Screening; NMB, Net Monetary Beneﬁt; SDNA, Stool DNA. * At a Willingness to Pay for 1 Quality Adjusted Life Year of $50,000. effective according to societies’ established willingness to pay (WTP) threshold for a single quality adjusted life year of $50,000. However, colonoscopy was more cost-effective compared to sigmoidoscopy and stool DNA. Stool DNA test was dominated as it was more costly and less effective than colonoscopy. Similarly the incremental cost effectiveness ratios of stool DNA compared with FIT and sigmoidoscopy were both above the WTP of $50,000. The net monetary benefit at WTP of $50,000 showed FIT to have more benefit than colonoscopy fol- lowed by sigmoidoscopy and stool DNA test. Comparison of Screening Strategies with One Another In incremental comparisons between each strategy (Table 2), colonoscopy at 25% inadequate preparation rate cost $74,656 per QALY more compared to FIT and hence was not cost- PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 7 / 17 Comparison of Different Inadequate Preparation Rates The cost-effectiveness and NMB of screening strategies were determined at different inade- quate bowel preparation rates using a threshold analysis from 0% to 30% (Table 3). Colonos- copy was more cost effective compared to FIT (i.e. ICER < $50,000) when inadequate bowel preparation rates were 13% or lower. At inadequate preparation rate of 13%, incremental cost effectiveness ratio of colonoscopy compared with FIT was at societies willingness to pay threshold of $50,000. This suggests that 13% inadequate preparation rate is the threshold above which FIT becomes the preferred screening strategy. Colonoscopy remained a preferred screening strategy over sigmoidoscopy even at 30% inadequate preparation rate. Colonoscopy dominated stool DNA test at all inadequate preparation rates. Sensitivity Analysis Colonoscopy and Bowel Preparation Rates Table 4. Sensitivity Analysis. Variables* Variable Range Input Low Value High Value Risk Percent Speciﬁcity of FIT 0.855 to 0.997 16,006.3790 96,141.9481 45.14% Adherence to FIT 0.72 to 0.88 59,286.6552 110,129.2013 18.17% Colonoscopy Sensitivity for LG Polyp 0.765 to 0.935 60,225.2048 99,345.4563 10.76% Adherence of Colonoscopy 0.586 to 0.716 66,029.7035 104,194.2306 10.24% Colonoscopy Sensitivity for HG Polyp 0.81 to 0.99 56,133.4962 89,003.9174 7.60% FIT Sensitivity for LG Polyp 0.09 to 0.11 65,414.5932 84,185.2703 2.48% FIT Sensitivity for HG Polyp 0.216 to 0.264 63,230.9017 81,665.0094 2.39% FIT Sensitivity for Colorectal Cancer 0.63 to 0.77 66,830.9051 83,756.4146 2.01% Colonoscopy Sensitivity for Colorectal Cancer 0.855 to 0.997 71,987.1885 85,105.9302 1.21% Cost of FIT $18 to $26 74,043.6500 75,342.3300 < 1% Cost of Colonoscopy (with Intervention) $900 to $1,200 74,309.7600 74,963.5435 < 1% Cost of Colonoscopy (without Intervention) $1,100 to $1,500 74,604.7600 74,724.9900 < 1% Cost of Initial Colon Cancer Treatment at a Stage of Diagnosis Speciﬁc Rate + 25% - 69,333.9211 - FIT, Fecal Immunochemical Test; LG, Low Grade; HG, High Grade * Variables were chosen based on the results of a tornado diagram in order of highest ICER variance. literature. The base results remained robust to the majority of the sensitivity analyses (Table 4). The incremental values for FIT and colonoscopy (25% inadequate preparation rate) are reported because they had the highest NMB compared to other tests. Specificity of FIT was the most influential variable; however, this is of limited practical significance as most studies have confirmed specificity of FIT for colon cancer to be 91–93%. The sensitivity of FIT had much less influence on the results. Provider reimbursement for a colonoscopy had a negligible impact on the results. Stool DNA is a newly approved test, which is generally considered to be more effective than FIT but is significantly more expensive. Threshold analysis (at WTP of $50,000) show stool DNA test reimbursement would have to be $75 for it to be the preferred screening strategy. The cost-effectiveness acceptability curves (CEAC) from probabilistic sensitivity analysis were determined. Fig 2 shows the probability, based on the proportion of simulations in which a given CRC screening strategy is most cost effective for a range of maximum WTP thresholds. FIT is the most cost-effective strategy at WTP of $50,000 according to 36% of 10,000 model iterations. Sensitivity Analysis One-way sensitivity analysis was performed on all inputs by increasing and decreasing the var- iables within 10% of the base case values to ensure inclusion of all the values reported in tios comparing colonoscopy at various bowel prep rates with other screening strategies. Table 3. Incremental Cost Effectiveness Ratios comparing colonoscopy at various bowel prep rates with other screening strategies. Inadequate Preparation Rate Cost FIT Sigmoidoscopy No Screening Stool DNA (SDNA) 0% $1,982.50 $35,418.42 $1,781.51 NS is Dominated SDNA is Dominated 5% $2,043.70 $40,786.84 $3,375.26 NS is Dominated SDNA is Dominated 10% $2,102.86 $45,976.32 $4,915.89 NS is Dominated SDNA is Dominated 15% $2,181.05 $52,835.09 $6,952.08 $776.97 SDNA is Dominated 20% $2,241.03 $58,096.49 $8,514.06 $1,374.98 SDNA is Dominated 25% $2,429.81 $74,656.14 $13,430.21 $3,257.13 SDNA is Dominated 30% $2,515.11 $82,138.60 $15,651.56 $4,107.58 SDNA is Dominated doi:10 1371/journal pone 0167452 t003 Table 3. Incremental Cost Effectiveness Ratios comparing colonoscopy at various bowel prep rates wit PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 8 / 17 Colonoscopy (25% inadequate bowel preparation rate) becomes cost effective as WTP increases beyond $55,000 (in 35% iterations). Sigmoidoscopy is cost effective in less than 19% iterations and decreases slightly as the WTP increases beyond $55,000. Stool DNA test is not cost effective compared to any screening strategy except at WTP of $85,000 at which point it becomes more cost effective than only sigmoidoscopy. Colonoscopy and Bowel Preparation Rates Discussion Inadequate bowel preparation is defined as the inability to identify lesions larger than five mil- limeters [36,37]. However, judging bowel preparation as inadequate remains inconsistent and subjective. Guidelines suggest repeating colonoscopy for inadequate preparations, but they do not make a distinction between grades of inadequacy (intermediate or poor) [22]. The timing of repeat colonoscopy is left to the endoscopists’ discretion and thus highly variable, with most recommending shorter intervals as preparation worsens [20]. While colonoscopy has been shown to be cost-effective for colon cancer screening, the research does not take into account the cost of repeat colonoscopies due to inadequate preps. This analysis shows that screening PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 9 / 17 Colonoscopy and Bowel Preparation Rates Fig 2. Cost Effectiveness Acceptability Curve. DNA Stool Test (blue square). Colonoscopy (red triangle). FIT (yellow circle). Sigmoidoscopy (green triangle). Natural History (turquoise diamond). doi:10.1371/journal.pone.0167452.g002 Fig 2. Cost Effectiveness Acceptability Curve. DNA Stool Test (blue square). Colonoscopy (red triangle). FIT (yellow circle). Sigmoidoscopy (green triangle). Natural History (turquoise diamond). Fig 2 C colonoscopy is not a cost-effective strategy if more than 13% of colonoscopies are repeated because of inadequate bowel preps. These findings are more pronounced given the model assumed the current high costs of CRC treatment, which yields a cost-effectiveness advantage for colonoscopy because it can prevent, not simply detect, cancer. Our analysis has several strengths. First, to calculate the cost of colonoscopy we used weighted averages of reimbursement by different third party payers. This best reflects, settings where colonoscopies are performed and the costs of anesthesia. Most cost-effective analyses use only Medicare reimbursement rates and thus underestimate the true cost burden of colo- noscopies. Second, our model accounts for population adherence to screening across all strate- gies. This has been an important limitation of prior analyses because studies have shown higher adherence rates for stool tests compared to endoscopies. In fact, this can influence the Fig 2. Cost Effectiveness Acceptability Curve. DNA Stool Test (blue square). Colonoscopy (red triangle). FIT (yellow circle). Sigmoidoscopy (green triangle). Natural History (turquoise diamond). doi:10.1371/journal.pone.0167452.g002 PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 10 / 17 Fig 2. Cost Effectiveness Acceptability Curve. DNA Stool Test (blue square). Colonoscopy (red triangle). FIT (yellow circle). Sigmoidoscopy (green triangle). Natural History (turquoise diamond). doi:10.1371/journal.pone.0167452.g002 Fig 2. Cost Effectiveness Acceptability Curve. DNA Stool Test (blue square). Colonoscopy (red triangle). FIT (yellow circle). Sigmoidoscopy (green triangle). PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 Colonoscopy and Bowel Preparation Rates outcomes of incremental cost effectiveness ratios. For example, in the cost-effective analysis of CRC screening strategies by Sharraf et al, FIT dominated colonoscopy since it was less expen- sive and more effective [21]. In our model, colonoscopy was more effective because FIT’s effec- tiveness is largely dependent on yearly adherence to avoid the emergence of CRC given its lower negative predictive rate. FIT has a poor sensitivity for dysplastic polyps, but this poor performance is made up by yearly exams. In a model, such as ours, where adherence to FIT is lower than 100%, the effectiveness of FIT would expectedly be lower. In fact, Sharaf et al dem- onstrated this point in a sensitivity analysis, which showed that as adherence rates for FIT were decreased toward real world rates, colonoscopy became increasingly cost effective [21]. Third, the recently approved stool DNA test was included as a screening strategy, using newly published data. The model and analyses are based on a few significant assumptions. First, it was assumed that colonoscopies are repeated in all patients with inadequate bowel preparations. This assumption is based on both guidelines that recommend a repeat colonoscopy and the stan- dard practice among gastroenterologists [30,31]. However, not all patients who are recom- mended repeat colonoscopy may get it. A study analyzing commercial and Medicare claims data found inadequate bowel preparation in 15%- 25% patients, but lower rates of repeat colo- noscopies. The second assumption was colonoscopies with inadequate preparations were reimbursed at the same rate as those with adequate preparations. Medicare pays for a screening colonoscopy (without polyps) only once every ten years unless the physician informs Medicare that the colonoscopy was incomplete, by using billing modifier 53 [38]. Medicare then pays for the incomplete colonoscopy at half the rate of a regular colonoscopy. However, there is ambi- guity around this billing practice and the billing modifier is rarely used. Medicare Part B claims data from 2009–2013 shows that only 2–3% of all colonoscopies were billed with modifier 53 [38]. A study of sample Medicare screening colonoscopies from 2000–2008 showed that of all the claims for colonoscopies without clear indications, only 2% were denied reimbursement by Medicare [39]. Hence, we feel that using full reimbursement rates for inadequate colonos- copies is justified. Third, it was assumed that all colonoscopies where a polyp was removed were adequate. PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 Discussion Natural History (turquoise diamond). doi:10.1371/journal.pone.0167452.g002 colonoscopy is not a cost-effective strategy if more than 13% of colonoscopies are repeated because of inadequate bowel preps. These findings are more pronounced given the model assumed the current high costs of CRC treatment, which yields a cost-effectiveness advantage for colonoscopy because it can prevent, not simply detect, cancer. Our analysis has several strengths. First, to calculate the cost of colonoscopy we used weighted averages of reimbursement by different third party payers. This best reflects, settings where colonoscopies are performed and the costs of anesthesia. Most cost-effective analyses use only Medicare reimbursement rates and thus underestimate the true cost burden of colo- noscopies. Second, our model accounts for population adherence to screening across all strate- gies. This has been an important limitation of prior analyses because studies have shown higher adherence rates for stool tests compared to endoscopies. In fact, this can influence the PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 10 / 17 Colonoscopy and Bowel Preparation Rates outpatients. Furthermore, the treatment arms of many randomized control trials had inade- quate bowel preparation rates of 4–10% [42–48]. There are many variables that determine the quality of colon preparation. These include: patient related variables (medical comorbidities, demographics, socio-economic, literacy); preparation related factors (type, dose, timing); system related constraints (quality of prepara- tion instructions, scheduling) and endoscopist related patterns of care (subjective rating of quality of colon preparation and time spent to clean the colon). Most of the factors that deter- mine colon preparation quality can be controlled. For example, Hassan et al proposed a pre- dictive model based on patients’ demographics and comorbidities that could theoretically decrease the inadequate preparation rate from 33% to 13% [48]. In practice, these proven interventions are infrequently adopted and most endoscopy units still prescribe a standard bowel preparation to all patients. This is likely because individualizing bowel preps is resource intensive and many institutions may not feel necessary to invest in these resources. Based on these findings, perhaps inadequate bowel preparation rate should be a reportable quality indicator. After all, better quality of bowel preparation is linked to higher adenoma detection rate (ADR), which is now the most recognized and promoted quality indicator of screening colonoscopy. Proper bowel preparation provides significant additive value to ADR. For example, a polyp may be removed from a colon with an inadequate preparation counting towards ADR, but there may be other undetected polyps. Thus, total adenoma detection may be lower with intermediate versus high-quality bowel preparation, even as ADRs remain equivalent [49,50]. Lower inadequate bowel preparation rates would result in fewer repeat colonoscopies. This data could be used by third party payers, accountable care organizations and primary care physicians to guide their patients to endoscopy units that consistently have better preparation rates. It is likely that inadequate bowel preparation rates as a reportable met- ric would be an incentive for institutions to improve. It may also be an impetus for the endos- copists to try harder to wash some of the intermediate preps converting them to adequate. In a study comparing a public and private hospital affiliated to the same university, patients at the public hospital were more likely to have a repeat colonoscopy for imperfect preps (20% vs. 12.5%) and the endoscopists spent less time in washing the colon (7.5% vs. 10.3% of the total procedure time) [14]. PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 In reality, it is not uncommon for an endoscopist to remove a small polyp when the preparation is intermediate and have the patient follow-up earlier than the recom- mended three to five years. Studies suggest that in 25–50% of patients with small adenomas, surveillance colonoscopies are recommended earlier than recommended in guidelines. These patients were more likely to have less than excellent preps. Thus, this analysis underestimates the burden of overuse of surveillance colonoscopy. The final assumption presumed patients follow only one screening strategy. In reality, patients may change between different strategies, but no data exists to quantify this potential and the purpose of the study was to compare one strategy to the other. Our study has important clinical, financial and policy implications. Costs of screening colo- noscopy have been under scrutiny in the last few years by third party payers and lay media. The increasing costs have prompted Medicare to propose more cuts in reimbursement rates for colonoscopy. This analysis shows that decreasing reimbursements would have minimal impact on overall costs. The professional societies fear that the cuts may even negatively impact the colonoscopy screening rates [40]. The approach instead should be to promote more effective use of screening colonoscopies, such as improving the inadequate bowel prepa- ration rates and decreasing the costs associated with repeat colonoscopies. Screening colonos- copies are cost-effective if the inadequate bowel preparation rate is less than 13%; a rate that is certainly an achievable target with concerted efforts. The European Society of Gastroenterol- ogy [8] has a recommendation of inadequate bowel preparation rate of less than 10% while the American College of Gastroenterology [41] recently set this target as less than 15% for PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 11 / 17 Software: JK. Supervision: DA. Supervision: DA. Validation: JK SK FR. Visualization: JK DA. Writing – original draft: DA. Writing – review & editing: JK DA. Colonoscopy and Bowel Preparation Rates valid. Data on stool DNA tests was derived mainly from one study and this is subjected to bias. Second, while the analysis looked at differential adherence among screening tests, the possibil- ity of decreasing adherence rates over time was not factored in. FIT is generally repeated every year, whereas the recommended interval for colonoscopy is ten years. It is likely that with fre- quent testing, adherence of FIT may decline over time. Third, the risk of CRC was not varied by socio-demographic status. This may be important because African-Americans have higher incidence of CRC and are more likely to have inadequate preps. Fourth, the model applies to a US population and setting, and may not be generalizable. The way screening programs are organized, funded and controlled are different in European countries and the US. The treat- ment protocols and costs can be significantly different in the US compared to single payor sys- tem in many European countries.” In conclusion, this research shows that for colonoscopy to be a cost effective screening strat- egy for CRC, the inadequate bowel preparation rate should be less than 13%. For others, FIT may be the preferred screening strategy. Inadequate bowel preparation rates should be a reportable quality indicator. Author Contributions Conceptualization: DA JK. Data curation: JK SK FR. Formal analysis: JK SK FR. Funding acquisition: DA. Investigation: JK DA. Methodology: JK SK FR. 3. Lebwohl B, Kastrinos F, Glick M, Rosenbaum AJ, Wang T, Neugut AI. The Impact of Suboptimal Prepa- ration on Adenoma Miss Rates and the Factors Associated with Early Repeat Colonoscopy: Gastroint- est Endosc. 2011; 73(6):1207–1214. doi: 10.1016/j.gie.2011.01.051 PMID: 21481857 Finally, one could argue that endoscopy units or institutions that consis- tently fail to improve their inadequate bowel preparation rates to less than 13% should offer FIT instead of colonoscopy as the preferred CRC screening strategy. There has been much emphasis and resources spent on increasing CRC screening rates to “80% by 2018” [51,52]. This is certainly a worthy goal and it is estimated that increasing CRC rates from 60% to 80% could decrease CRC mortality by almost 33%. If this increase in CRC screening comes from increasing colonoscopy screening rates, it will cost significantly more with an inadequate bowel preparation rate of 25% compared to 13%. In fact, if the current inadequate preparation rates cannot be improved, then it is more cost effective to increase rates of FIT since 100% adherence rate with FIT offers more QALYs at a lesser cost compared to even 80% colonoscopy adherence rates. Stool DNA test was compared with other screening strategies and found to be more cost effective than no screening but not compared to other screening strategies. This analysis deter- mined $78 to be the threshold cost for stool DNA test to be cost effective. These results are sim- ilar to the cost effectiveness study performed by the Agency of Healthcare Research and Quality that calculated a cost of $34 -$60 for stool test to be a non-dominated option [40]. The slight difference in results is likely due to new data on cost and efficacy. There are some limitations of the study, apart from those due to the aforementioned model assumptions. First, while the best available data were used for estimates, the source of data ran- ged from randomized control trials to observational studies and such comparisons may not be PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 12 / 17 Project administration: DA. Project administration: DA. Conceptualization: DA JK. Methodology: JK SK FR. References 1. Levin B, Lieberman DA, McFarland B, Smith RA, Brooks D, Andrews KS, et al. Screening and Surveil- lance of the Early Detection of Colorectal Cancer And Adenomatous Polyps. CA Cancer J Clin. 2008; 58:130–160. 1. Levin B, Lieberman DA, McFarland B, Smith RA, Brooks D, Andrews KS, et al. Screening and Surveil- lance of the Early Detection of Colorectal Cancer And Adenomatous Polyps. CA Cancer J Clin. 2008; 58:130–160. 2. Lieberman DA, Rex DK, Winawer SJ, Giardiello FM, Johnson DA, Levin TR. Guidelines for colonoscopy surveillance after screening and polypectomy: a consensus update by the us multi-society task force on colorectal cancer. Gastroenterology. 2012; 143:844–857. doi: 10.1053/j.gastro.2012.06.001 PMID: 22763141 3. Lebwohl B, Kastrinos F, Glick M, Rosenbaum AJ, Wang T, Neugut AI. The Impact of Suboptimal Prepa- ration on Adenoma Miss Rates and the Factors Associated with Early Repeat Colonoscopy: Gastroint- est Endosc. 2011; 73(6):1207–1214. doi: 10.1016/j.gie.2011.01.051 PMID: 21481857 PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 13 / 17 Colonoscopy and Bowel Preparation Rates 4. Serper M, Gawron AJ, Smith SG, Pandit AA, Dahlke AR, Bojarski EA, et al. Patient factors that affect quality of colonoscopy preparation. Clin Gastroenterol Hepatol. 2014; 12:451–457. doi: 10.1016/j.cgh. 2013.07.036 PMID: 23954647 5. Ness RM, Manam R, Hoen H, Chalasani N. Predictors of inadequate bowel preparation for colonos- copy. Am J Gastroenterol. 2001; 96:1797–1802. doi: 10.1111/j.1572-0241.2001.03874.x PMID: 11419832 6. Harewood GC, Sharma VK, de Garmo P. Impact of colonoscopy preparation quality on detection of sus- pected colonic neoplasia. Gastrointest Endosc. 2003; 58:76–79. doi: 10.1067/mge.2003.294 PMID: 12838225 7. Froehlich F, Wietlisbach V, Gonvers JJ, Burnand B, Vader JP. Impact of colonic cleansing on quality and diagnostic yield of colonoscopy: the European Panel of Appropriateness of Gastrointestinal Endos- copy European multicenter study. Gastrointest Endosc. 2005; 61:378–384. PMID: 15758907 8. Hendry PO, Jenkins JT, Diament RH. The impact of poor bowel preparation on colonoscopy: a prospec- tive single centre study of 10,571 colonoscopies. Colorectal Dis. 2007; 9:745–748. doi: 10.1111/j.1463- 1318.2007.01220.x PMID: 17477852 9. Chung YW, Han DS, Park KH, Kim KO, Park CH, Hahn T, et al. Patient factors predictive of inadequate bowel preparation using polyethylene glycol: a prospective study in Korea. J Clin Gastroenterol. 2009; 43:448–452. doi: 10.1097/MCG.0b013e3181662442 PMID: 18978506 10. Chan WK, Saravanan A, Manikam J, Goh KL, Mahadeva S. Appointment waiting times and education level influence the quality of bowel preparation in adult patients undergoing colonoscopy. BMC Gastro- enterol. 2011; 11:86. References doi: 10.1186/1471-230X-11-86 PMID: 21798022 11. Fatima H, Johnson CS, Rex DK. Patients’ description of rectal effluent and quality of bowel preparation at colonoscopy. Gastrointest Endosc. 2010; 71:1244–1252. doi: 10.1016/j.gie.2009.11.053 PMID: 20362286 12. Borg BB, Gupta NK, Zuckerman GR, Banerjee B, Gyawali CP. Impact of obesity on bowel preparation for colonoscopy. Clin Gastroenterol Hepatol. 2009; 7:670–675. doi: 10.1016/j.cgh.2009.02.014 PMID: 19245852 13. Rex DK, Imperiale TF, Latinovich DR, Bratcher LL. Impact of bowel preparation on efficiency and cost of colonoscopy. Am J Gastroenterol. 2002; 97:1696–1700. doi: 10.1111/j.1572-0241.2002.05827.x PMID: 12135020 14. Lebwohl B, Kastrinos F, Glick M, Rosenbaum AJ, Wang T, Neugut AI. The impact of suboptimal bowel preparation on adenoma miss rates and the factors associated with early repeat colonoscopy. Gastro- intest Endosc. 2011; 73:1207–1214. doi: 10.1016/j.gie.2011.01.051 PMID: 21481857 15. Tan-Torres Edejer T, Baltussen R, Adam T, Hutubessy R, Acharya A, Evans D B., et al. Making Choices In Health: WHO Guide To Cost-Effectiveness Analysis. Geneva, Switzerland: World Health Organization; 2003. 16. Kruse GR, Khan SM, Zaslavsky AM, Ayanian JZ, Sequist TD. Overuse of colonoscopy for colorectal cancer screening and surveillance. J Gen Intern Med. 2015; 30(3):277–83. doi: 10.1007/s11606-014- 3015-6 PMID: 25266407 17. Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER Stat Database: Incidence—SEER 9 Regs Research Data, Nov 2014 Sub (1973–2012) <Katrina/Rita Popu- lation Adjustment>—Linked To County Attributes—Total U.S., 1969–2013 Counties, National Cancer Institute, DCCPS, Surveillance Research Program, Surveillance Systems Branch, released April 2015, based on the November 2014 submission. 18. Ries LAG, Kosary CL, Hankey BF, Miller BA, Harras A, Edwards BK, et al. SEER Cancer Statistics Review, 1973–1994, NIH Pub. No. 97–2789, National Cancer Institute: Bethesda, MD, 1997. 19. Djalalov S, Rabeneck L, Tomlinson G, Bremner KE, Hilsden R, Hoch JS. A Review and Meta-analysis of Colorectal Cancer Utilities. Med Decis Making. 2014; 34:809. doi: 10.1177/0272989X14536779 PMID: 24903121 20. Wong MCS, Ching JYL, Chan VCW, Lam TYT, Luk AKC, Ng SC, et al. Informed Choice vs. No Choice in Colorectal Cancer Screening Tests: A Prospective Cohort Study in Real-Life Screening Practice. Am J Gastroenterol. 2014; 109:1072–1079. doi: 10.1038/ajg.2014.136 PMID: 24935273 21. Sharaf RN, Ladabaum U. Comparative Effectiveness and Cost-Effectiveness of Screening Colonos- copy vs. Sigmoidoscopy and Alternative Strategies. Am J Gastroenterol. 2013; 108:120–132. doi: 10. 1038/ajg.2012.380 PMID: 23247579 22. Winawer SJ, Zauber AG, Ho MN, O’Brien MJ, Gottlieb LS, Sternberg SS, et al. PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 References Prevention of colorectal cancer by colonoscopic polypectomy: The National Polyp Study Workgroup, N Engl J Med. 1993; 329:1977–81. doi: 10.1056/NEJM199312303292701 PMID: 8247072 14 / 17 PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 Colonoscopy and Bowel Preparation Rates 23. Zauber AG, Lansdorp-Vogelaar I, Wilschut J, Knudsen AB, van Ballegooi- jen M, Kuntz KM. Cost-effec- tiveness of DNA stool testing to screen for colorectal cancer: report to AHRQ and CMS from the Cancer Intervention and Surveillance Modeling Network (CISNET) for MISCAN and SimCRC Models. 24. Ginsberg GM, Lim SS, Lauer JA, Johns BP, Sepulveda CR. Prevention, screening and treatment of colorectal cancer: a global and regional generalized cost effectiveness analysis. Cost Effectiveness and Resource Allocation. 2010; 8:2. doi: 10.1186/1478-7547-8-2 PMID: 20236531 25. Maciosek MV, Solberg LI, Coffield AB, Edwards NM, Goodman MJ. Colorectal Cancer Screening: Health Impact and Cost Effectiveness: Am J Prev Med. 2006; 31(1):80–89. doi: 10.1016/j.amepre. 2006.03.009 PMID: 16777546 26. Ladabaum U, Phillips KA. Colorectal Cancer Screening Differential Costs for Younger Versus Older Americans: Am J Prev Med. 2006; 30(5). 27. Song K, Fendrick AM, Ladabaum U. Fecal Dna Testing Compared With Conventional Colorectal Can- cer Screening Methods: A Decision Analysis. Gastroenterology 2004; 126:1270–1279. PMID: 15131787 28. Howard K, Salkeld G, Irwig L, Adelstein BA. High participation rates are not necessary for cost-effective colorectal cancer screening. J Med Screen. 2005; 12(2):96–102. doi: 10.1258/0969141053908276 PMID: 15949121 29. Centers for Disease Control and Prevention (CDC).Vital signs: colorectal cancer screening test use— United States, 2012: MMWR Morb Mortal Wkly Rep. 2013 Nov 8; 62(44):881–8. PMID: 24196665 30. Larsen M, Hills N, Terdiman J. The impact of the quality of colon preparation on follow-up colonoscopy recommendations. Am J Gastroenterol. 2011; 106(12):2058–62. doi: 10.1038/ajg.2011.238 PMID: 22138933 31. Menees SB, Elliott E, Govani S, Anastassiades C, Judd S, Urganus A, et al. The impact of bowel cleansing on follow-up recommendations in average-risk patients with a normal colonoscopy. Am J Gastroenterol, 2014; 109(2):148–54. doi: 10.1038/ajg.2013.243 PMID: 24496417 32. Ben-Horin S, Bar-Meir S, Avidan B. The impact of colon cleanliness assessment on endoscopists’ rec- ommendations for follow-up colonoscopy. Am J Gastroenterol. 2007; 102(12):2680–5. doi: 10.1111/j. 1572-0241.2007.01486.x PMID: 17714555 33. Ladabaum U, Levin Z, Mannalithara A, Brill JV, Bundorf MK. Colorectal Testing Utilization and Pay- ments in a Large Cohort of Commercially Insured US Adults: Am J Gastroenterol. 2014. 34. Siebert U, Alagoz O, Bayoumi AM, Jahn B, Owens DK, Cohen DJ, et al. PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 References State-Transition Modeling: A Report Of The ISPOR-SMDM Modeling Good Research Practices Task Force-3. Value In Health. 2012; 15:812–820. doi: 10.1016/j.jval.2012.06.014 PMID: 22999130 35. American Cancer Society. Colorectal Cancer Facts & Figures 2011–2013: American Cancer Society, 2011. 36. Ladabaum U, Song K. Projected National Impact of Colorectal Cancer Screening on Clinical and Eco- nomic Outcomes and Health Services Demand. Presented at Annual Meeting of the American Gastro- enterological Association 2004. 37. Rex DK. Bowel Preparation for Colonoscopy: Entering an Era of Increased Expectations for Efficacy. Clinical Gastroenterology and Hepatology 2014; 12:458–462. doi: 10.1016/j.cgh.2013.11.003 PMID: 24239858 38. Sheffield KM, Han Y, Kuo YF, Taylor S. Riall TS, Goodwin JS. Potentially Inappropriate Screening Colo- noscopy in Medicare Patients: Variation by Provider and Geographic Region: JAMA Intern Med. 2013; 173(7): there are no pages. 39. Goodwin JS, Singh A, Reddy N, Riall TS, Kuo YF. Overuse of Screening Colonoscopy in the Medicare Population: Arch Intern Med. 2011; 171(15) 40. Zauber AG, Lansdorp-Vogelaar I, Wilschut J, Knudsen AB, Ballegooijen MV, Kuntz KM. Cost-Effective- ness of DNA Stool Testing to Screen for Colorectal Cancer. Agency for Healthcare Research and Qual- ity. Published December 20, 2007. 41. Johnson DA, Barkun AN, Cohen LB, Dominitz JA, Kaltenbach T, Martel M, et al. Optimizing Adequacy of Bowel Cleansing for Colonoscopy: Recommendations From the US Multi-Society Task Force on Colorectal Cancer. Am J Gastroenterol. 2014 42. Pyenson B, Scammell C, Broulette J. Costs and repeat rates associated with colonoscopy observed in medical claims for commercial and Medicare populations. BMC Health Serv Res. 2014; 14: 92. doi: 10. 1186/1472-6963-14-92 PMID: 24572047 43. Henry SG, Ness RM, Stiles RA, Shintani AK, Dittus RS. A Cost Analysis Of Colonoscopy Using Micro- costing And Time-And-Motion Techniques. J Gen Intern Med. 2002; 22(10):1415–21. 44. Matro R, Shnitser A, Spodik M, Daskalakis C, Katz L, Murtha A, et al. Efficacy of morning-only com- pared with split-dose polyethylene glycol electrolyte solution for afternoon colonoscopy: a randomized PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 15 / 17 Colonoscopy and Bowel Preparation Rates controlled single-blind study. Am J Gastroenterol. 2010; 105(9): 1954–61. doi: 10.1038/ajg.2010.160 PMID: 20407434 controlled single-blind study. Am J Gastroenterol. 2010; 105(9): 1954–61. doi: 10.1038/ajg.2010.160 PMID: 20407434 45. Manno M, Pigò F, Manta R, Barbera C, Bertani H, Mirante VG, et al. Bowel preparation with polyethyl- ene glycol electrolyte solution: optimizing the splitting regimen. Dig Liver Dis. 2012; 44(7):576–9. doi: 10.1016/j.dld.2012.02.012 PMID: 22456624 46. References De Leone A, Tamayo D, Fiori G, Ravizza D, Trovato C, De Roberto G, et al. Same-day 2-L PEG-citrate- simethicone plus bisacodyl vs split 4-L PEG: Bowel cleansing for late-morning colonoscopy. World J Gastrointest Endosc. 2013; 5(9):433–9. doi: 10.4253/wjge.v5.i9.433 PMID: 24044042 47. Chiu HM, Lin JT, Wang HP, Lee YC, Wu MS. The impact of colon preparation timing on colonoscopic detection of colorectal neoplasms—a prospective endoscopist-blinded randomized trial. Am J Gastro- enterol, 2006; 101(12):2719–2725. doi: 10.1111/j.1572-0241.2006.00868.x PMID: 17026559 48. Hassan C, Fuccio L, Bruno M, Pagano N, Spada C, Carrara S, et al. A predictive model identifies patients most likely to have inadequate bowel preparation for colonoscopy. Clin Gastroenterol Hepatol, 2012; 10(5):501–6. doi: 10.1016/j.cgh.2011.12.037 PMID: 22239959 49. Liang PS, Dominitz JA. Editorial: Bowel preparation: is fair good enough? Am J Gastroenterol, 2014; 109(11):1725–7. doi: 10.1038/ajg.2014.328 PMID: 25373582 50. Wang HS, Pisegna J, Modi R, Liang LJ, Atia M, Nguyen M, et al. Adenoma detection rate is necessary but insufficient for distinguishing high versus low endoscopist performance. Gastrointest Endosc. 2013; 77(1):71–8. doi: 10.1016/j.gie.2012.08.038 PMID: 23261096 51. Simon S. Organizations Commit to Goal of 80% Colon Cancer Testing Rate by 2018. American Cancer Society. http://www.cancer.org/cancer/news/organizations-commit-to-goal-of-80-percent-coloncancer- testing-rate-by-2018. Published March 17, 2014. 52. National Colorectal Cancer Round Table. 80% by 2018. http://nccrt.org/tools/80-percent-by-2018/ 53. Zauber AG, Lansdorp-Vogelaar I, Knudsen AB, Wilschut J, van Ballegooijen M, Kuntz KM. Evaluating Test Strategies for Colorectal Cancer Screening: A Decision Analysis for the U.S. Preventive Services Task Force. Ann Intern Med. 2008; 149:659–669. PMID: 18838717 54. A Stool DNA Test (Cologuard) for Colorectal Cancer Screening. Med Lett Drugs Ther. 2014 Oct 13; 56 (1453):100–1. I’m not sure if this is a journal or website? PMID: 25296259 55. Bourikas LA, Tsiamoulos ZP, Haycock A, Thomas-Gibson S, Saunders BP. How we can measure qual- ity in colonoscopy?. World J Gastrointest Endosc. 2013; 5(10): 468–475. doi: 10.4253/wjge.v5.i10.468 PMID: 24147190 56. Brenner H, Chang-Claude J, Seiler CM, Rickert A, Hoffmeister M. Protection from colorectal cancer after colonoscopy: a population-based, case-control study. Ann Intern Med. 2011; 154:22–30. doi: 10. 7326/0003-4819-154-1-201101040-00004 PMID: 21200035 57. Zauber AG, Lansdorp-Vogelaar I, Knudsen AB, Wilschut J, van Ballegooijen M, Kuntz KM. Evaluating Test Strategies for Colorectal Cancer Screening: A Decision Analysis for the U.S. Preventive Services Task Force. Ann Intern Med. 2008; 149:659–669. PMID: 18838717 58. Song K, Fendrick AM, Ladabaum U. Fecal Dna Testing Compared With Conventional Colorectal Can- cer Screening Methods: A Decision Analysis. Gastroenterology 2004; 126:1270–1279. PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 65. Rabeneck L, Paszat LF, Hilsden RJ, Saskin R, Leddin D, Grunfeld E, et al. Bleeding And Perforation After Outpatient Colonoscopy And Their Risk Factors In Usual Clinical Practice. Gastroenterology. 2008; 135:1899–1906. doi: 10.1053/j.gastro.2008.08.058 PMID: 18938166 66. Tran AH, Ngor EWM, Wu BU. Surveillance Colonoscopy in Elderly Patients. A Retrospective Cohort Study. JAMA Intern Med. 2014; 174(10):1675–1682. doi: 10.1001/jamainternmed.2014.3746 PMID: 25111954 PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 Colonoscopy and Bowel Preparation Rates References PMID: 15131787 59. Pignone M, Russell L, Wagner J, eds. Institute of Medicine (US); National Research Council (US); Eco- nomic Models of Colorectal Cancer Screening in Average-Risk Adults: Workshop Summary.Washing- ton (DC): National Academies Press (US); 2005. The National Academies Collection: Reports funded by National Institutes of Health. 60. Frazier AL, Colditz GA, Fuchs CS, Kuntz KM. Cost-effectiveness of Screening for Colorectal Cancer in the General Population. JAMA. 2000; 284(15):1954–1961. PMID: 11035892 61. Khandker RK, Dulski JD, Kilpatrick JB, Ellis RP, Mitchell JB. A decision model and cost-effectiveness analysis of colorectal cancer screening and surveillance guidelines for average-risk adults. Int J Technol Assess Health Care. 2000; 16(3):799–810. PMID: 11028135 62. Vijan S, Hwang EW, Hofer TP, Hayward RA. Which Colon Cancer Screening Test? A Comparison of Costs, Effectiveness, and Compliance. Am J Med. 2001; 111:593–601. PMID: 11755501 63. Loeve F, Brown ML, Boer R, van Ballegooijen M, van Oortmarssen GJ, Habbema JDF. Endoscopic Colorectal Cancer Screening: a Cost-Saving Analysis. J Natl Cancer Inst. 2000; 92:557–63 PMID: 10749911 64. Ness RM, Holmes AM, Klein R, Dittus R. Cost-Utility of One-Time Colonoscopic Screening for Colorec- tal Cancer at Various Ages. Am J Gastroenterol. 2000; 95(7):1800–11 doi: 10.1111/j.1572-0241.2000. 02172.x PMID: 10925988 16 / 17 PLOS ONE \| DOI:10.1371/journal.pone.0167452 December 9, 2016 Colonoscopy and Bowel Preparation Rates 17 / 17
https://openalex.org/W2205158231	https://dash.harvard.edu/bitstream/1/26859863/1/4809575.pdf	English	null	Combined dabrafenib and trametinib therapy in metastatic melanoma and organ transplantation: Case report and review of the literature	JAAD case reports	2,015	cc-by	2,106	Permanent link http://nrs.harvard.edu/urn-3:HUL.InstRepos:26859863 Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#LAA Published Version doi:10.1016/j.jdcr.2015.09.020 doi:10.1016/j.jdcr.2015.09.020 Citation Garrett, Giorgia L., Steven Y. He, Nica Sabouni, Adil Daud, and Sarah T. Arron. 2015. “Combined dabrafenib and trametinib therapy in metastatic melanoma and organ transplantation: Case report and review of the literature.” JAAD Case Reports 1 (6): S23-S25. doi:10.1016/ j.jdcr.2015.09.020. http://dx.doi.org/10.1016/j.jdcr.2015.09.020. Published Version doi:10.1016/j.jdcr.2015.09.020 From the Departments of Dermatologya and Medicine,b University of California, San Francisco; the Department of Medicine, Harvard/Beth Israel Deaconess Medical Center, Bostonc; and the University of California, Berkeley.d Funding sources: None. Conflicts of interest: None declared. Correspondence to: Sarah T. Arron, MD, PhD, Department of Dermatology, 1701 Divisadero St, San Francisco, CA 94115. E-mail: arrons@derm.ucsf.edu. Combined dabrafenib and trametinib therapy in metastatic melanoma and organ transplantation: Case report and review of the literature Giorgia L. Garrett, MD,a Steven Y. He, MD,c Nica Sabouni,d Adil Daud, MD,b and Sarah T. Arron, MD, PhD San Francisco and Berkeley, California; and Boston, Massachusetts Key words: iatrogenic immunosuppression; metastatic melanoma; nonmelanoma skin cancer; solid organ transplantation. Abbreviations used: AJCC: American Joint Committee on Cancer CT: computed tomography OTRs: organ transplant recipients PET: positron emission tomography Abbreviations used: AJCC: American Joint Committee on Cancer CT: computed tomography OTRs: organ transplant recipients PET: positron emission tomography Share Your Story The Harvard community has made this article openly available. Please share how this access benefits you. Submit a story . Accessibility INTRODUCTION Improved surgical techniques and superior post- transplantcarehaveincreasedgraftsurvivalaftersolid organ transplantation, and complications of lifelong immunosuppression are increasingly prevalent in clinical practice. The excess relative risk of melanoma attributable to transplantation is between 2 and 12 times that of the general population,1 and given the high tumor burden, this has important implications for the surveillance and management of organ trans- plant recipients (OTRs). Several risk factors for melanoma are known, including immunosuppres- sion, the presence of atypical nevi, light skin photo- type, ultraviolet light exposure, and a family or personal history of skin cancer.1,2 Here we report a case of cutaneous melanoma arising posttransplant, highlightingthesubstantialmorbidityandmortality of this disease in OTR. In view of the increased risk of melanoma and high potential for metastasis, derma- tologists should monitor OTRs closely and institute multidisciplinary care without delay. excision was performed, and sentinel lymph node biopsy found 1 of 1 positive node in the left side of the groin and 0 of 1 positive nodes in the right side of the groin. In 2008, surveillance positron emission tomog- raphy (PET)/computed tomography (CT) detected left groin node recurrence, and a subsequent radical left inguinal dissection found nodal positivity in 18 of 21 nodes. At that time, the patient underwent deﬁn- itive adjuvant radiation treatment to the left groin. In 2011, a surveillance PET/CT found multifocal retroperitoneal lymphadenopathy, and a diagnosis of metastatic melanoma was conﬁrmed by CT- guided biopsy. A BRAF V600E mutation was de- tected, and the patient was enrolled in a phase 3 clinical trial evaluating combined dabrafenib and trametinib therapy (selective BRAF and mitogen- activated protein kinases inhibitors, respectively). She tolerated the regimen well, showing no evidence of disease recurrence for 15 months. Unfortunately, a single metastasis was found in the right side of the skull in 2013, which was managed with stereotactic radiotherapy (15000 Gy). Because of tumor progres- sion, the patient was withdrawn from the study and unfortunately died from complications of metastatic disease later in the year. JAAD Case Reports 2015;1:S23-5. 2352-5126 2015 by the American Academy of Dermatology, Inc. Published by Elsevier, Inc. This is an open access article under the CC BY- NC-ND license (http://creativecommons.org/licenses/by-nc-nd/ 4.0/). http://dx.doi.org/10.1016/j.jdcr.2015.09.020 DISCUSSION Preclinical and early development studies of combination dabrafe- nib/trametinib have shown promising results, with fewer side effects and less tumor resistance that either agent used alone.16 Our patient had BRAF V600 mutant metastatic melanoma and was therefore deemed a good candidate for enrollment in a phase 3 clinical trial testing combined BRAF and MEK inhibitor treatment. She had an excellent initial response to therapy, with progression-free survival of 15 months or 5.5 months higher than the median (9.4 months vs 6 to 7 months for monotherapy).16,17 The initial approach to posttransplant primary melanoma follows the same principles as the non- transplanted general population. Surgical excision with wide margins is the ﬁrst step after diagnostic biopsy and can cure tumors that have not spread. The size of the surgical margin is guided by the Breslow thickness of the primary lesion. Tumors of more than 4.0 mm require 2-cm margins, and, because of the high risk of nodal and distant metastasis, in this clinical setting more extensive surgical resection is unlikely to yield better outcomes. The AJCC pub- lished an international melanoma staging system that guides melanoma prognosis and treatment in immu- nocompetent patients.7 The AJCC predicted that 5- year survival for patients with early, regional nodal, and metastatic melanoma disease is 85%, 24% to 69.5%, and 6%, respectively, where the 24% to 69.5% range varies according to the number of positive lymph nodes at diagnosis.7 Sentinel lymph node biopsy is indicated for tumors greater than 1-mm Breslow thickness or in the presence of increased mitotic activity and ulceration on histology. Sentinel lymph node positivity is found to be the strongest predictor of mortality in melanoma,4,8 and in nodal disease, PET/CT or direct CT yield information about tumor staging and disease progression. Surgical Although these numbers are promising, at this stage we cannot generalize her response to the transplanted population as a whole. Further research on melanoma outcomes in OTRs is likely to inform effective treatment strategies in relation to both immune therapy and reduction of immunosuppression. Melanoma is a potentially fatal complication of transplantation, and severe disease can dramatically reduce both the quality of life and the chance of survival in OTRs. The optimal setting for treating OTRs who have melanoma is a multidisciplinary clinic led by the primary transplant team. DISCUSSION control was deemed unlikely in our patient, as she presented with a primary lesion of 17.5-mm Breslow thickness and positive lymph node status. Completion lymphadenectomy was not performed at diagnosis, as this procedure is not found to increase survival in early nodal disease.9 Three distinct clinical entities are of interest in relation to melanoma in the OTR: donor-derived melanoma, melanoma preceding organ transplanta- tion, and de novo melanoma arising after transplant. In this report, we focus on de novo posttransplant melanoma, the scenario most frequently encoun- tered in clinical practice. Melanoma is an immune responsive tumor, and the regression of benign and malignant melanocytic lesions in response to im- mune mechanisms is well documented.1,2 Iatrogenic immunosuppression, therefore, is expected to worsen melanoma outcomes after transplantation, although to date no population-based study with adequate power has addressed this issue in OTRs. Small, retrospective studies have found that Breslow thickness is inversely correlated with melanoma survival in this population.1,3-5 In a case series of 100 transplant patients, Matin et al6 found that OTRs with melanomas of greater than 2 mm thickness (American Joint Committee on Cancer [AJCC] TNM stage T3 or T4) had a significantly worse prognosis than did immunocompetent patients with similar tumors, whereas the prognosis did not differ in tumors of # 2 mm thickness (TNM stage T1 and T2). These data indicate that iatrogenic immunosup- pression may be associated with worse prognosis in patients with thicker tumors and highlight the need for close surveillance and early diagnosis in the OTR. y There is no cure for metastatic melanoma, which has a median survival time of 7.5 months from diagnosis.10 Conventional treatment strategies, such as surgery, chemotherapy, and radiotherapy are not effective in limiting the progression of tumor metas- tasis and are often associated with adverse effects and poor quality of life. Reassuringly, immunotherapy and molecular targeted therapy have achieved partial success in a subset of patients with unresectable metastatic melanoma. High-dose interleukin-211 and, more recently, antieCTLA-4 (ipilimumab)12 and antiePD-1 (nivolumab)13 used as single agents or in combination therapy, have achieved a long-term survival of 5 years or longer. Dabrafenib and trame- tinib, targeted BRAF and MEK1 inhibitors, respec- tively,haveachieved a progression-free survival of up to 3 years when used as monotherapy,14,15 although as with most treatments based on oncogene-targeted small molecules, tumor resistance has limited the therapeutic response to these agents. CASE 66 A 66-year-old white woman with Fitzpatrick skin type II received a heart transplant in 2001 for familial hypertrophic cardiomyopathy and was maintained on cyclosporine. She presented to the dermatology department in 2005 for the evaluation of a rapidly growing pigmented lesion on the umbilicus. A biopsied found this lesion was a primary malignant melanoma of 17.5 mm Breslow thickness. Wide local JAAD Case Reports 2015;1:S23-5. 2352-5126 2015 by the American Academy of Dermatology, Inc. Published by Elsevier, Inc. This is an open access article under the CC BY- NC-ND license (http://creativecommons.org/licenses/by-nc-nd/ 4.0/). http://dx.doi.org/10.1016/j.jdcr.2015.09.020 S23 JAAD CASE REPORTS NOVEMBER 2015 S24 Garrett et al REFERENCES 1. Zwald FO, Christenson LJ, Billingsley EM, et al. Melanoma in solid organtransplant recipients. Am J Transplant. 2010;10:1297-1304. 1. Zwald FO, Christenson LJ, Billingsley EM, et al. Melanoma in solid organtransplant recipients. Am J Transplant. 2010;10:1297-1304. 12. Hodi FS, O’Day SJ, McDermott DF, et al. Improved survival with ipilimumab in patients with metastatic melanoma. N Engl J Med [Internet]. 2010 Aug 19 [cited 2014 Jul 9];363(8): 711-23. Available from: http://www.pubmedcentral.nih.gov/ articlerender.fcgi?artid=3549297&tool=pmcentrez&rendertype =abstract. 2. Hollenbeak CS, Todd MM, Billingsley EM, Harper G, Dyer AM, Lengerich EJ. Increased incidence of melanoma in renal transplantation recipients. Cancer. 2005;104:1962-1967. 3. Le Mire L, Hollowood K, Gray D, Bordea C, Wojnarowska F. Melanomas in renal transplant recipients. Br J Dermatol. 2006; 154:472-477. 13. Mamalis A, Garcha M, Jagdeo J. Targeting the PD-1 pathway: a promising future for the treatment of melanoma. Arch Dermatol Res [Internet]. 2014 Aug [cited 2015 Jan 10]; 306(6):511-19. Available from: http://www.ncbi.nlm.nih.gov/ pubmed/24615548. 4. Brown VL, Matin RN, Cerio R, Leedham-Green ME, Proby CM, Harwood CA. Melanomas in renal transplant recipients: the London experience, and invitation to participate in a European study. Br J Dermatol [Internet]. 2007 Jan [cited 2014 Oct 13]; 156(1):165-7; author reply 167-9. Available from: http://www. pubmedcentral.nih.gov/articlerender.fcgi?artid=2423223&tool =pmcentrez&rendertype=abstract. 14. Falchook GS, Long GV, Kurzrock R, et al. Dabrafenib in patients with melanoma, untreated brain metastases, and other solid tumours: A phase 1 dose-escalation trial. Lancet. 2012;379: 1893-1901. 5. Penn I. Malignant melanoma in organ allograft recipients. Transplantation. 1996;61:274-278. 15. Flaherty KT, Infante JR, Daud A, et al. Combined BRAF and MEK inhibition in melanoma with BRAF V600 mutations. N Engl J Med. 2012;367:1694-1703. 6. Matin RN, Mesher D, Proby CM, et al. Melanoma in organ transplant recipients: clinicopathological features and outcome in 100 cases. Am J Transplant. 2008;8:1891-1900. 16. Johnson DB, Flaherty KT, Weber JS, et al. Combined BRAF (Dabrafenib) and MEK inhibition (Trametinib) in patients with BRAFV600-mutant melanoma experiencing progression with single-agent BRAF inhibitor. J Clin Oncol [Internet]. 2014 Oct 6 [cited 2014 Oct 7]; 32:3697-704. Available from: http://jco. ascopubs.org/cgi/doi/10.1200/JCO.2014.57.3535. 7. Balch CM, Buzaid AC, Soong SJ, et al. Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol. 2001;19:3635-3648. 8. Debarbieux S, Duru G, Dalle S, Beatrix O, Balme B, Thomas L. Sentinel lymph node biopsy in melanoma: A micromorpho- metric study relating to prognosis and completion lymph node dissection. Br J Dermatol. 2007;157:58-67. 17. DISCUSSION In the presence of life-threatening tumors, a reduction or modiﬁcation of immunosuppression should be considered, after carefully balancing the risk of allograft rejection against the risk of melanoma mortality. Dermatologists play a key role in this JAAD CASE REPORTS VOLUME 1, NUMBER 6 Garrett et al S25 Garrett et al S25 10. Barth A, Wanek LA, Morton DL. Prognostic factors in 1,521 melanoma patients with distant metastases. J Am Coll Surg. 1995;181:193-201. assessment, as they may be the only team members able to quantify the tumor burden. A decision was made against a reduction of immunosuppression in our patient, as her cyclosporine regime was deemed necessary to preserve allograft function. 11. Rosenberg SA, Yang JC, Topalian SL, et al. Treatment of 283 consecutive patients with metastatic melanoma or renal cell cancer using high-dose bolus interleukin 2. JAMA [Internet]. 1994 Jan [cited 2014 Oct 24];271(12):907-13. Available from: http://www.ncbi.nlm.nih.gov/pubmed/8120958. REFERENCES Sosman JA, Kim KB, Schuchter L, et al. Survival in BRAF V600-mutant advanced melanoma treated with vemurafenib. N Engl J Med [Internet]. 2012 Feb 23 [cited 2014 Oct 12];366(8): 707-14. Available from: http://www.pubmedcentral.nih.gov/ articlerender.fcgi?artid=3724515&tool=pmcentrez&rendertype =abstract. 9. Veronesi U, Adamus J, Bandiera DC, et al. Inefficacy of immediate node dissection in stage 1 melanoma of the limbs. N Engl J Med. 1977;297:627-630.
https://openalex.org/W2099747853	https://europepmc.org/articles/pmc4694916?pdf=render	English	null	Genomic amplification of 9p24.1 targeting <i>JAK2</i>, <i>PD-L1</i>, and <i>PD-L2</i> is enriched in high-risk triple negative breast cancer	Oncotarget	2,015	cc-by	7,004	ABSTRACT We used DNA content flow cytometry followed by oligonucleotide array based comparative genomic hybridization to survey the genomes of 326 tumors, including 41 untreated surgically resected triple negative breast cancers (TNBC). A high level (log2ratio ≥1) 9p24 amplicon was found in TNBC (12/41), glioblastomas (2/44), and colon carcinomas (2/68). The shortest region of overlap for the amplicon targets 9p24.1 and includes the loci for PD-L1, PD-L2, and JAK2 (PDJ amplicon). In contrast this amplicon was absent in ER+ (0/8) and HER2+ (0/15) breast tumors, and in pancreatic ductal adenocarcinomas (0/150). The PDJ amplicon in TNBCs was correlated with clinical outcomes in group comparisons by two-sample t-tests for continuous variables and chi-squared tests for categorical variables. TNBC patients with the PDJ amplicon had a worse outcome with worse disease-free and overall survival. Quantitative RT-PCR confirmed that the PDJ amplicon in TNBC is associated with elevated expression of JAK2 and of the PD-1 ligands. These initial findings demonstrate that the PDJ amplicon is enriched in TNBC, targets signaling pathways that activate the PD-1 mediated immune checkpoint, and identifies patients with a poor prognosis. Michael T. Barrett1, Karen S. Anderson2, Elizabeth Lenkiewicz1, Mariacarla Andreozzi1, Heather E. Cunliffe3, Christine L. Klassen4, Amylou C. Dueck5, Ann E. McCullough6, Srikanth K. Reddy7, Ramesh K. Ramanathan8, Donald W. Northfelt8, Barbara A. Pockaj4 1Department of Research, Mayo Clinic in Arizona, Scottsdale, Arizona, United States of America 2Biodesign Institute, Arizona State University, Tempe, Arizona, United States of America 3Department of Pathology, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand 4Division of General Surgery, Section of Surgical Oncology, Mayo Clinic in Arizona, Phoenix, Arizona, United States of America 5Section of Biostatistics, Mayo Clinic in Arizona, Scottsdale, Arizona, United States of America 6Department of Pathology and Laboratory Medicine, Mayo Clinic in Arizona, Scottsdale, Arizona, United States of America 7Vanderbilt University, Nashville, Tennessee, United States of America 8 n of Hematology-Oncology, Mayo Clinic in Arizona, Scottsdale, Arizona, United States of America pondence to: Oncotarget, Vol. 6, No. 28 Oncotarget, Vol. 6, No. 28 www.impactjournals.com/oncotarget/ www.impactjournals.com/oncotarget INTRODUCTION [2–4]. PD-1 is a CD28 and CTLA-4 homologue that is normally induced on activated T cells, but the chronic antigenic exposure in cancer may lead to high levels of PD-1 and T cell exhaustion [5]. Consequently PD-1 and its ligands are being investigated as candidate biomarkers for response to targeted immune checkpoint blockade in clinical trials of an increasing variety of tumors. Two PD-1 inhibitors, nivolumab and pembrolizumab, have been approved for clinical use in melanoma and are in clinical trials in other solid tumors [6]. In melanoma, the There is an emerging recognition of the role of immune checkpoints in the pathogenesis of solid tumors [1]. Distinct inhibitory pathways serve to regulate T cell activation and function. These include the immune regulatory molecules PD-L1 and PD-L2 that limit the duration and the level of the T-cell response. Increased expression of these ligands of the immune checkpoint receptor PD-1 has been reported in human tumors www.impactjournals.com/oncotarget Oncotarget 26483 and provide a candidate biomarker for patients who may benefit from immune checkpoint targeted therapies. In order to address tissue and clonal heterogeneity in clinical samples we used DNA content based flow sorting to identify and purify distinct tumor populations in each tissue of interest [17–19]. There are well-established DNA staining based methods for isolating nuclei of aneuploid, tetraploid, and diploid neoplastic populations from solid tumor samples [20–23]. Individual populations of cells or nuclei can be objectively and quantitatively purified to greater than 95% purity for molecular analyses even in heavily admixed and sub-optimal clinical samples. We have developed and extensively validated flow cytometry based methodologies to study a wide variety of clinical samples including small needle biopsies, post treatment tissues, and formalin fixed paraffin embedded (FFPE) tissues with high definition genome assays such as oligonucleotide based array comparative genomic hybridization (aCGH) and next generation sequencing NGS [18, 24]. combination of nivolumab plus ipilimumab which targets the PD-1 homologue CTLA-4 was associated with a > 80% decline in tumor burden at 12 weeks in respondents and a 53% overall response rate (ORR) [7]. More recent studies with a larger cohort of patients with untreated metastatic melanoma reported an ORR of 61% for the combination therapy with complete responses in 22% of patients [8]. However clinical studies have been limited by the lack of predictive biomarkers of disease response, the complexity of tumor genomes, and the degeneracy of the receptor/ligand interactions. INTRODUCTION The Janus kinase 2 gene (JAK2) is one of four members of the JAK family (which includes JAK1, JAK2, JAK3 and non-receptor protein-tyrosine kinase 2 (TYK2) [9]. JAKs associate with the cytoplasmic portion of a variety of transmembrane cytokine and growth factor receptors important for signal transduction in hematopoietic cells. Receptor binding by extracellular ligand causes receptor multimerization and brings JAK proteins together to allow activation by transphosphorylation. Activated JAK2 has emerged as an important target in myeloproliferative disorders, and increasingly, in solid tumors [10]. Significantly JAK2 has been implicated in interleukin (IL)-6-dependent breast cancer stem cell self-renewal [11], and in both IL-6- and IL-8-dependent growth of triple-negative breast cancers (TNBCs) [12]. Pre-clinical studies have implicated JAK2 signaling as a mechanism of escape from targeted therapies in TNBC and as a promoter for the emergence of more invasive tumor cells [13]. Thus, JAK2 inhibitors are being evaluated in patients with breast and other solid tumors. However a case control study of 223 breast tumors reported an association between increased JAK2 mRNA levels and favorable prognosis [14], and a strong correlation between JAK2 mRNA levels with the presence of tumor-infiltrating lymphocytes (TILs). The protective effect of elevated JAK2 expression was not associated with increased JAK2 protein levels in the tumor epithelial cells. These observations suggest that therapeutic targeting of JAK2 expression may abrogate the potential benefits arising from an antitumor immune response. Thus the clinical significance of selective amplification and overexpression of JAK2 arising from the often highly aberrant genomic landscapes of solid tumors remains to be elucidated. To test our hypothesis we surveyed the genomes of 326 clinical samples representing pancreatic ductal adenocarcinomas (PDA) (n = 150), glioblastomas (n = 44), colorectal carcinomas (n = 68), and breast carcinomas (n = 64). Each tumor sample in this study was flow sorted prior to copy number analysis with oligonucleotide based aCGH. These data were then used to determine the prevalence of amplification of PD-L1, PD-L2, and JAK2 at chromosome 9p24.1 within each tumor type. The presence of this “PDJ” amplicon was then correlated with expression of JAK2 and the PD-1 ligands, and to clinical outcomes in a subset of patients. RESULTS We detected and sorted an aneuploid and/or a proliferating tumor fraction from each clinical sample in this study. These fractions were gated during sorting providing highly purified and objectively defined tumor populations from each sample for analysis. The tissues included triple negative breast cancer n = 41, HER2+ breast cancer n = 15, ER+HER2- breast cancer n = 8, pancreatic adenocarcinoma n = 150 (including 30 liver metastases), colorectal carcinoma n = 68, and glioblastoma n = 44. These included both fresh frozen and formalin fixed paraffin embedded (FFPE) clinical samples. The tumor cellularity prior to sorting varied extensively from less than 10% to greater than 70% across tissue types. The genomes of each sorted tumor cell population were interrogated with whole genome oligonucleotide based aCGH. Copy number aberrant intervals were identified and their genomic boundaries mapped using a step gram algorithm [25]. Amplicons were then ranked within each sample based on their fold change and their overall prevalence in tumor genomes. A recurring top ranked and high level (log2ratio ≥1) amplicon that targeted 9p24.1 The PD-L1 and PD-L2 genes localize to 9p24.1 adjacent to JAK2 and there is emerging data that an amplicon targeting this 9p24.1 locus is present in lymphomas and a subset of EBV-positive gastric cancers [15, 16]. Furthermore JAK2 has been shown to up-regulate the transcription of both PD-1 ligands and increase sensitivity to JAK2 inhibitors in a dose dependent manner [16]. We hypothesized that targeted amplification of 9p24.1 in tumor genomes would result in co-amplification of PD-L1, PD-L2, and JAK2, identify a distinct molecular subtype arising in multiple cancers, www.impactjournals.com/oncotarget Oncotarget 26484 was detected in 12/41 TNBCs, 2/68 colon carcinomas, and 2/44 glioblastomas (Figs. 1 and 2). In contrast this amplicon was absent in ER+ (n = 8) and HER2+ (n = 15) breast tumors, and in pancreatic ductal adenocarcinomas (n = 150). The shortest region of overlap (SRO) spanned 777 kb and included the PD-1 ligands PD-L1, PD-L2, and the Janus kinase 2 (JAK2) loci (Fig. S1). The height of this recurring amplicon included mean log2ratios >4 consistent with amplification of genomic drivers such as HER2 and MYC described in breast cancer and other solid tumor genomes. genes compared to those without the amplicon (Fig. 2). The latter included samples with low level copy number gains (log2ratio >0 and <1) at 9p24.1 including increases of whole 9p arm and polysomy of chromosome 9. RESULTS In addition we identified another TNBC in the subset of 15 tumors without aCGH data with concurrent elevated expression of JAK2 and PD-L1 (Fig. S2). PD-L2 expression was also elevated in the presence of the PDJ amplicon however it did not reach statistical significance ( p < 0.0645) in this preliminary study. These observations are consistent with studies showing that genomic amplification of 9p24.1 leads to coordinated overexpression of these genes in human tumors [15, 16]. In order to determine the effect of the PDJ amplicon on JAK2, PD-L1, and PD-L2 expression we selected 31 TNBC samples, 16 of which were profiled in our copy number analysis, for qRT-PCR analysis. We used a pooled sample comprised of an unrelated normal breast, and individual TNBC, ER+, and HER2+ tumor tissues to generate a standard curve for assaying JAK2, PD-L1 and PD-L2 expression in our TNBC cohort. Tumors with a high level amplicon (4/16 TNBCs surveyed by qRT-PCR) had significantly higher expression of JAK2 and PD-L1 Clinical data was available on 36 of 41 (88%) of the TNBC patients that were flow sorted then profiled for copy number (Table 1). Patients with the high level PDJ amplicon (n = 8) were noted to have larger tumors (mean 3.9 cm vs. 1.9 cm, p = 0.04) and a higher incidence of lymph node metastases (75% vs. 26%, p = 0.01). Lymphocytic infiltration was noted in 4 of the 36 patients, none of whom had the PDJ amplicon in their tumor Figure 1: Whole genome and chromosome 9 aCGH plots of flow sorted tumor populations. A. Colorectal (CRC) and B–C. triple negative breast cancers (TNBC) with high level 9p24.1 amplicon. Amplicons were scored according to log2ratios >1. Blue arrows denote JAK2 locus. Figure 1: Whole genome and chromosome 9 aCGH plots of flow sorted tumor populations. A. Colorectal (CRC) and B–C. triple negative breast cancers (TNBC) with high level 9p24.1 amplicon. Amplicons were scored according to log2ratios >1. Blue arrows denote JAK2 locus. www.impactjournals.com/oncotarget Oncotarget 26485 Figure 2: The 9p24 amplicon in a triple negative breast cancer genome. A. Flow histogram of sorted 3.2N TNBC population from FFPE tissue. B. Chromosome 9 CGH plot and detection of 9p24 amplicon. C. Gene specific view of amplicon. Red shaded area denotes ADM2 defined copy number aberrant region. D. Gene expression of JAK2, PD-L1, and PD-L2 in TNBC. RESULTS Comparisons and correlations between the expression levels of PD-L1, PD-L2, and JAK2 genes and copy number status of chromosome 9p24.1 were performed using an unpaired t test and variation among and between groups were calculated using an ANOVA test (GraphPad Prism 6). Figure 2: The 9p24 amplicon in a triple negative breast cancer genome. A. Flow histogram of sorted 3.2N TNBC population from FFPE tissue. B. Chromosome 9 CGH plot and detection of 9p24 amplicon. C. Gene specific view of amplicon. Red shaded area denotes ADM2 defined copy number aberrant region. D. Gene expression of JAK2, PD-L1, and PD-L2 in TNBC. Comparisons and correlations between the expression levels of PD-L1, PD-L2, and JAK2 genes and copy number status of chromosome 9p24.1 were performed using an unpaired t test and variation among and between groups were calculated using an ANOVA test (GraphPad Prism 6) Figure 2: The 9p24 amplicon in a triple negative breast cancer genome. A. Flow histogram of sorted 3.2N TNBC population from FFPE tissue. B. Chromosome 9 CGH plot and detection of 9p24 amplicon. C. Gene specific view of amplicon. Red shaded area denotes ADM2 defined copy number aberrant region. D. Gene expression of JAK2, PD-L1, and PD-L2 in TNBC. Comparisons and correlations between the expression levels of PD-L1, PD-L2, and JAK2 genes and copy number status of chromosome 9p24.1 were performed using an unpaired t test and variation among and between groups were calculated using an ANOVA test (GraphPad Prism 6). inhibition occurs in a dose-dependent manner [26]. Thus genomic amplification of 9p24.1 may provide a selective tumor cell dependent increase in both JAK2 and immune checkpoint signaling. In addition, copy number gains on chromosome 9 that include 9p24.1 and elevated expression of PD-L1 within tumor cells have recently been associated with substantial therapeutic activity in patients with advanced or refractory Hodgkin’s lymphoma treated with the PD-1 inhibitor nivolumab [27]. Notably, two early phase I trials of PD-1/PD-L1 blockade (pembrolizumab or MPDL3280A) in TNBC have demonstrated overall response rates of 15–20% [28, 29]. genome. Twenty nine of these 36 TNBC patients received chemotherapy after definitive surgical therapy. The disease-free survival rate at 5 years was 25% in the PDJ amplified patients, and 66% in the unamplified patients ( p = 0.005) (Fig. 3). Overall survival (OS) at 5 years was 25% in the PDJ amplified patients, compared with 69% in the unamplified patients ( p = 0.004). RESULTS Thus our preliminary results suggest that the presence of the PDJ amplicon defines a clinically significant subset of high- risk TNBC patients. DISCUSSION Progression free survival. Lower disease- free survival at 5 years 25.0% vs. 66.0%, p = 0.005. B. Overall survival. Lower overall survival at 5 years 25.0% vs. 69.0%, p = 0.004. Median follow up is 4.7 years (range 0.9–12.0 years). Not Amplified (N = 28) Amplified (N = 8) Total (N = 36) p value Radiation [Mastectomy only] 0.882 Missing 17 (.%) 4 (.%) 21 No 5 (45.5%) 2 (50%) 7 (46.7%) Yes 6 (54.5%) 2 (50%) 8 (53.3%) 1Two-Sample T-Test Figure 3: Clinical outcomes for TNBC patients with or without PDJ amplicon. A. Progression free survival. Lower disease- free survival at 5 years 25.0% vs. 66.0%, p = 0.005. B. Overall survival. Lower overall survival at 5 years 25.0% vs. 69.0%, p = 0.004. Median follow up is 4.7 years (range 0.9–12.0 years). has been associated with a better prognosis and decreased risk of recurrence [14]. Thus an important clinical question is the role of a tumor cell driven elevation in JAK2 and PD-1 ligands versus increases arising from immune cells in solid tissue cancers [6, 27]. correlative studies with FISH based markers in lymphoma have proposed that even tumors with low level 9p gains and chromosome 9 polysomy may benefit from PD-1 targeted therapies [27].il Our preliminary results with highly purified flow sorted clinical samples and an array platform designed for whole genome copy number measures suggest that the PDJ amplicon has a SRO that targets JAK2 and the PD-1 ligands, is enriched in a subset of TNBCs with poor outcomes prior to therapy, and is distinct from background genomic aberrations affecting chromosome 9. Given the coordinated overexpression of PD-L1, PD-L2, and JAK2, antagonists targeting PD-1 signaling and JAK2 inhibitors should be evaluated in the context of PDJ amplification. Thus we propose that the PDJ amplicon provides a candidate biomarker for identifying high-risk patients and for advancing emerging immunotherapies in TNBC. The absence of this amplicon in pancreatic adenocarcinoma samples and its relatively low frequencies in colorectal cancers and glioblastomas are consistent with relatively low response rates reported in these and other solid tumors when compared to TNBC [33, 34]. DISCUSSION PD-L1 expression across multiple solid tumors varies significantly by tumor type and appears common in both tumor-infiltrating immune cells and in tumor cells [4, 30]. Elevated PD-L1 expression from tumor infiltrating immune cells had a stronger association with clinical response than that observed with expression from the tumor cells [4]. Significantly elevated expression of JAK2 in the non-epithelial compartment of breast tumors Studies in gastric cancer and in lymphomas have described a recurring 9p24.1 amplicon that includes PD-L1, PD-L2, and JAK2 [15, 16]. The presence of the amplicon was associated in a subset of each tumor type with distinct pathological and clinical features. Functional studies revealed that JAK2 is a transcriptional activator of both PD-1 ligands [16] and that sensitivity to JAK2 www.impactjournals.com/oncotarget Oncotarget 26486 Table 1: Comparison between patients with and without PDJ amplification Not Amplified (N = 28) Amplified (N = 8) Total (N = 36) p value Age [years] 0.791 Mean (SD) 53.4 (12.88) 54.8 (8.71) 53.7 (11.98) Median 54.0 53.5 54.0 Range (29.0–78.0) (45.0–72.0) (29.0–78.0) Tumor Size [cm] 0.041 Mean (SD) 1.9 (0.86) 3.9 (3.88) 2.4 (2.11) Median 1.9 2.5 2.0 Range (0.5–4.0) (0.4–11.0) (0.4–11.0) Grade 0.632 1 1 (3.6%) 0 (0%) 1 (2.8%) 2 7 (25%) 1 (12.5%) 8 (22.2%) 3 20 (71.4%) 7 (87.5%) 27 (75%) Lymphocytic Infiltrates 0.212 Missing 10 (.%) 2 (.%) 12 No 14 (77.8%) 6 (100%) 20 (83.3%) Yes 4 (22.2%) 0 (0%) 4 (16.7%) Lymph Nodes 0.012 Missing 1 (%) 0 (.%) 1 Negative 20 (74.1%) 2 (25%) 22 (62.9%) Positive 7 (25.9%) 6 (75%) 13 (37.1%) Cancer Stage 0.042 Missing 1 (.%) 0 (.%) 1 Stage I 11 (40.7%) 1 (12.5%) 12 (34.3%) Stage II 13 (48.1%) 3 (37.5%) 16 (45.7%) Stage III 3 (11.1%) 4 (50%) 7 (20%) Surgical Treatment 0.702 Missing 9 (.%) 2 (.%) 11 Breast Conservation Therapy (BCT) 8 (42.1%) 2 (33.3%) 10 (40%) Mastectomy 11 (57.9%) 4 (66.7%) 15 (60%) Chemotherapy 0.102 Neoadjuvant 1 (3.6%) 2 (25%) 3 (8.3%) No 4 (14.3%) 0 (0%) 4 (11.1%) Yes 23 (82.1%) 6 (75%) 29 (80.6%) (Continued) www.impactjournals.com/oncotarget Oncotarget 26487 Not Amplified (N = 28) Amplified (N = 8) Total (N = 36) p value Radiation [Mastectomy only] 0.882 Missing 17 (.%) 4 (.%) 21 No 5 (45.5%) 2 (50%) 7 (46.7%) Yes 6 (54.5%) 2 (50%) 8 (53.3%) 1Two-Sample T-Test 2Chi-Square Figure 3: Clinical outcomes for TNBC patients with or without PDJ amplicon. A. DISCUSSION However the presence and role of the PDJ amplicon in a smaller subset (2–5%) of colorectal cancers, including primary and matching Previous in vitro studies of a 9p amplicon in a small panel of established breast cancer cell lines excluded JAK2, PD-L1, and PD-L2 loci from the SRO reported [31]. However JAK2 copy number levels assessed in bulk TNBC biopsies by targeted resequencing of a panel of 196 cancer associated genes have been reported to be elevated in residual disease post neoadjuvant chemotherapy. The JAK2 amplification was associated with poor regression free survival (RFS) and predicted poor OS in a subset (~10%) of TNBCs [32]. The prevalence of JAK2 amplification was noted to be higher in these post treatment TNBCs than reported for treatment naïve TNBCs in TCGA. The latter are based on single nucleotide polymorphism (SNP) array data of bulk tumor samples. Consequently it remained to be resolved whether JAK2 amplifications and the potential co-amplifications of PD-L1 and PD-L2 are present in untreated TNBCs, and whether genomic amplification at 9p24.1 may provide a biomarker for clinical outcome. Furthermore recent www.impactjournals.com/oncotarget Oncotarget 26488 lymph node biopsies (Fig. S3), and glioblastomas warrants further investigation. relevant assays (e.g. targeted resequencing, FISH, IHC) to efficiently identify those patients with the PDJ amplicon. We propose that the PDJ amplicon in TNBC and other solid tumors represents a novel candidate biomarker with broad application for cancer research and for advancing personalized therapies for cancer patients. g Given the highly aberrant genomic landscapes of TNBCs and other solid tumors, alterations that disrupt immune regulatory signaling pathways upstream or downstream of PDJ amplicon may provide additional biomarkers for this emerging class of therapies [35]. Studies in cell line models suggest that deletions in PTEN and disruption of AKT signaling can increase PD-L1 levels [36]. Our use of flow sorted clinical samples provides a robust objective method to detect somatic lesions targeting clinically actionable pathways [18, 24]. For example we detected focal deletions of PTEN in the tumor genomes of 5/12 PDJ positive TNBCs (Fig. 4 and Fig. S4). Future studies will incorporate flow sorted tissue samples and both copy number and next generation sequencing analyses to provide a more comprehensive and unbiased profile of the genomic landscapes of tumors with the PDJ amplicon. Of significant interest will be biopsies from patients enrolled in clinical trials with agents that target the PD-1 immune checkpoint. DISCUSSION The ability to discriminate mutations and genomic lesions including the PDJ amplicon arising in the epithelial component of solid tumors and to comprehensively interrogate genomes in the pre and post adjuvant setting, regardless of tumor cell content, will be essential to understand clinical responses and to advance effective immunotherapies. These data will also be used to develop robust clinically Clinical samples TNBC, ER+, and HER2+ samples were obtained under approval from the Mayo Clinic Institutional Review Board prior to undertaking this study (IRB protocol 08-006579). Tumor specimens were obtained from formalin fixed paraffin embedded (FFPE) archived breast cancer samples obtained at the time of definitive surgical resection. All breast cancers underwent central pathologic review, and were evaluated by IHC for estrogen receptor (ER), progesterone receptor (PR), and by IHC +/− FISH for Her2/neu under CLIA/CAP guidelines. TNBC samples were also obtained from The Inflammatory Breast Cancer Research Foundation (IBCRF) Biobank. Samples were collected following informed consent by the IBCRF under approval of their IRB and their Medical Advisory Board. Additional flow sorted TNBC data was from a previous study of breast cancer genomes that was performed with informed consent and ethics committee approvals [37]. Figure 4: PTEN homozygous deletion in PDJ+ triple negative breast cancer genome. A. Flow histogram of sorted 5.0N TNBC population from FFPE tissue. B–C. Chromosome 10 and chromosome 9 CGH plots. D. Gene specific view of PTEN homozygous deletion. Blue shaded area denotes ADM2 defined copy number aberrant region. Figure 4: PTEN homozygous deletion in PDJ+ triple negative breast cancer genome. A. Flow histogram of sorted 5.0N TNBC population from FFPE tissue. B–C. Chromosome 10 and chromosome 9 CGH plots. D. Gene specific view of PTEN homozygous deletion. Blue shaded area denotes ADM2 defined copy number aberrant region. www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget Oncotarget 26489 FFPE sample preparation Prior to sorting, excess paraffin was removed with a scalpel from either side of 40–60 um scrolls then processed according to our published methods [24]. Briefly each sectioned piece was collected into individual microcentrifuge tubes then washed three times with 1 ml Xylene for 5 minutes to remove remaining paraffin. Each sample was rehydrated in sequential ethanol washes (100% 5 minutes × 2, then 95%, 70%, 50% and 30% ethanol) and washed 2 times in 1 ml 1 mM EDTA pH 8.0. A 1 ml aliquot of 1 mM EDTA pH 8.0 was added to the samples and incubated at 95°C for 80 minutes to facilitate the removal of protein cross-links present in FFPE tissue. Samples were then cooled to room temperature for ≥5 minutes, followed by addition of 300 ul PBS pH 7.4 and gentle centrifugation for 2 minutes at 3.6 × g. The supernatant was carefully removed and the pellet washed three times with 1 ml PBS pH 7.4/0.5 mM CaCl2 to remove EDTA. Each sample was digested overnight (6–17 hours) in 1 ml of a freshly prepared enzymatic cocktail containing 50 units/ml of collagenase type 3, 80 units/ml of purified collagenase, and 100 units/ml of hyaluronidase in PBS pH 7.4/0.5 mM CaCl2 buffer. Following overnight digestion 500 ul NST buffer (146 mM NaCl, 10 mM Tris-HCl, pH 7.5, 1 mM CaCl2, 0.5 mM MgSO4, 21 mM MgCl2, 0.05% bovine serum albumin, 0.2% Nonidet P40 (Sigma)) with 4, 6-diamindino-2-phenylindole (DAPI; 10 μg/ml) was added to each sample to facilitate pelleting. Samples were centrifuged for 5 minutes at 3000 × g, after which pellets were resuspended in 750 ul of NST/10% fetal bovine serum and then passed through a 25 G needle 10–20 times. We used a single 50 μm scroll from each FFPE tissue block to obtain sufficient numbers of intact nuclei for subsequent sorting and molecular assays. DNA extraction DNA from sorted nuclei was extracted using an amended protocol from QIAamp® DNA Micro Kit from Qiagen (Valencia, CA). Briefly each sorted sample was resuspended in 180 ul buffer ATL and 20 ul proteinase K (20 mg/ml) then incubated for 3 hours at 56°C for complete lysis. Samples were bound and washed according to QIAamp® DNA Micro Kit instructions, eluted into 50 ul of H20, then precipitated overnight with 5 ul 3 M sodium acetate and 180 ul 100% EtOH. Each sample was then centrifuged for 30 minutes at 20, 000 × g, washed in 1 ml of 70% EtOH for 30 minutes at 20, 000 × g. The samples were carefully decanted and the DNA pellet was dried by speed vacuum then resuspended in a small volume (e.g. 10–50 ul) of H20 for final concentrations suitable for accurate quantitation. Flow cytometry PDA samples were from previous studies obtained under a WIRB protocol (20040832) for an NIH funded biospecimen repository (NCI P01 Grant CA109552), Stand up To Cancer clinical trials 2026001 and 2026003, and with approved consent of the Ethics Committee of Basel (252/08, 302/09).The glioblastoma samples were collected for another previous study with informed consent from patients at the Neurosurgery Department of the Centre Hospitalier in Luxembourg (CHL) [38]. Samples were approved for study by the National Ethics Committee for Research (CNER) of Luxembourg. In all cases written consent was obtained. De-identified colon samples did not qualify as human subjects according to guidelines administered by the Translational Genomics Research Institute (TGen) Office of Research Compliance and were collected under an institutional blanket exemption as part of different research initiatives and from the Cooperative Human Tissue Network (CHTN). Biopsies were minced in the presence of NST buffer and DAPI according to published protocols [23, 39, 40]. Prior to sorting each sample was filtered through a 35 um mesh and collected into a 5 ml Polypropylene round bottom tube. The mesh was rinsed with an additional 750 ul of NST/10% fetal bovine serum and placed on ice while processing remaining samples. The total volume in the tube for each sample was approximately 1.5 ml. An equal volume of 20 ug/ml DAPI was added to each tube to achieve a final concentration of 10 ug/ml DAPI for flow sorting with a BD Influx cytometer with ultraviolet excitation (Becton-Dickinson, San Jose, CA). The optimal settings for sorting FFPE samples with the Influx sorter were as follows: Drop formation was achieved with piezzo amplitude of 6–10 volts and a drop frequency of 30 khertz. The sort mode was set to purity yield with a drop delay of 31.5–32. Sheath fluid pressure was typically 17–18 psi with a 100 um nozzle. For single parameter DNA content assays DAPI emission was collected at >450 nm. DNA content and cell cycle were then analyzed using the software program MultiCycle (Phoenix Flow Systems, San Diego, CA). REFERENCES Total RNA was extracted from one whole-tissue 50 μm thick section using RNeasy FFPE RNA Isolation Kit (Qiagen). RNA quantification was performed using Qubit 2.0 fluorometer (Life Technologies) and Qubit RNA HS assay kit (molecular probes). Reverse transcription was carried out using SuperScript® VILO™ cDNA Synthesis Kit (Life Technologies) and 200 ng of RNA per reaction, with triplicate reactions performed for each sample [41]. Each of the 31 samples produced sufficient RNA yield and quality. Quantitative Real-Time PCR was performed for CD274, PDCD1LG2, and JAK2 (Hs01125301_m1, Hs01057777_m1 Hs00234567_m1, respectively; Life Technologies) with TaqMan® chemistry on the ABI Prism 7900HT (Applied Biosystems), using the standard curve method. Furthermore, two reference genes were used TFRC and MRPL19 (Hs00174609_m1, Hs00608519_m1, respectively; Life Technologies), previously selected as effectively normalizing for degradation of the FFPE- RNA. Target gene expression levels were normalized to the geometric mean of the two reference genes and normalized to a pool of RNAs prepared from a normal and from 3 FFPE breast tumors (TNBC, ER+, and HER2+). All statistical comparisons and correlations between the expression levels of PD-L1, PD-L2, and JAK2 genes and copy number status of chromosome 9p24.1 were performed using an unpaired t test and variation among and between groups were calculated using an ANOVA test (GraphPad Prism 6). The p-values < 0.05 were considered significant. 1. Ribas A, Tumeh PC. The future of cancer therapy: selecting patients likely to respond to PD1/L1 blockade. Clinical can cer research: an official journal of the American Association for Cancer Research. 2014; 20:4982–4984. 2. Gao Q, Wang XY, Qiu SJ, Yamato I, Sho M, Nakajima Y, Zhou J, Li BZ, Shi YH, Xiao YS, Xu Y, Fan J. Overexpression of PD-L1 significantly associates with tumor aggressiveness and postoperative recurrence in human hepatocellular carcinoma. Clinical cancer research: an official journal of the American Association for Cancer Research. 2009; 15:971–979. 2. Gao Q, Wang XY, Qiu SJ, Yamato I, Sho M, Nakajima Y, Zhou J, Li BZ, Shi YH, Xiao YS, Xu Y, Fan J. Overexpression of PD-L1 significantly associates with tumor aggressiveness and postoperative recurrence in human hepatocellular carcinoma. Clinical cancer research: an official journal of the American Association for Cancer Research. 2009; 15:971–979. 3. Taube JM, Klein A, Brahmer JR, Xu H, Pan X, Kim JH, Chen L, Pardoll DM, Topalian SL, Anders RA. ACKNOWLEDGMENTS This study was supported by funding from the non-profit Desert Mountain Member’s C.A.R.E. (Cancer Awareness through Research and Education), Carefree AZ, and from The Inflammatory Breast Cancer Research Foundation (IBCRF) RFA-IBC-90-10-2011. REFERENCES Association of PD-1, PD-1 ligands, and other features of the tumor immune microenvironment with response to anti-PD-1 therapy. Clinical cancer research: an official journal of the American Association for Cancer Research. 2014; 20:5064–5074. 4. Herbst RS, Soria JC, Kowanetz M, Fine GD, Hamid O, Gordon MS, Sosman JA, McDermott DF, Powderly JD, Gettinger SN, Kohrt HE, Horn L, Lawrence DP, Rost S, Leabman M, Xiao Y, et al. Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients. Nature. 2014; 515:563–567. 5. Melero I, Gaudernack G, Gerritsen W, Huber C, Parmiani G, Scholl S, Thatcher N, Wagstaff J, Zielinski C, Faulkner I, Mellstedt H. Therapeutic vaccines for cancer: an overview of clinical trials. Nat Rev Clin Oncol. 2014; 11:509–24. aCGH analysis DNAs were treated with DNAse 1 prior to Klenow- based labeling. High molecular weight reference templates were digested for 30 minutes while the smaller fragmented FFPE-derived DNA samples were digested for only 1 minute. In each case 1 ul of 10× DNase 1 reaction buffer and 2 ul of DNase 1 dilution buffer were added to 7 ul of DNA sample and incubated at room temperature then transferred to 70°C for 30 minutes to deactivate DNase 1. Sample and reference templates were then labeled with Cy-5 dUTP and Cy-3 dUTP respectively using a BioPrime labeling kit (Invitrogen, Carlsbad, CA) according to our published protocols [18]. All labeling reactions were www.impactjournals.com/oncotarget Oncotarget 26490 assessed using a Nanodrop assay (Nanodrop, Wilmington, DE) prior to mixing and hybridization to CGH arrays (Agilent Technologies, Santa Clara, CA) for 40 hours in a rotating 65°C oven. All microarray slides were scanned using an Agilent 2565C DNA scanner and the images were analyzed with Agilent Feature Extraction version 10.7 using default settings. The aCGH data was assessed with a series of QC metrics then analyzed using an aberration detection algorithm (ADM2) [25]. The latter identifies all aberrant intervals in a given sample with consistently high or low log ratios based on the statistical score derived from the average normalized log ratios of all probes in the genomic interval multiplied by the square root of the number of these probes. This score represents the deviation of the average of the normalized log ratios from its expected value of zero and is proportional to the height h (absolute average log ratio) of the genomic interval, and to the square root of the number of probes in the interval. All aCGH data in this paper have been deposited at the National Center for Biotechnology Information Gene Expression Omnibus (accession number in progress). variables. Disease-free survival time was defined as the time from primary surgery to first local, regional, or distant recurrence or death regardless of cause. Overall survival analysis included all deaths as events regardless of cause. Disease-free and overall survival rates at 5 years were estimated using the method of Kaplan and Meier. Disease-free survival and overall survival were compared between groups using a log-rank test. Median follow-up was 4.7 years (range 0.9–12.0 years). P-values < 0.05 were considered statistical significant throughout. CONFLICTS OF INTEREST None. Statistical analysis Cytokeratin labeling of breast cancer cells extracted from paraffin-embedded tissue for bivariate flow cytometric anal ysis. Cytometry. 1996; 24:260–267. 10. Quintas-Cardama A, Kantarjian H, Cortes J, Verstovsek S. Janus kinase inhibitors for the treatment of myeloprolifera tive neoplasias and beyond. Nature reviews Drug discovery. 2011; 10:127–140. 21. Rabinovitch PS. DNA content histogram and cell-cycle analysis. Methods Cell Biol. 1994; 41:263–296. 22. Barrett MT, Sanchez CA, Prevo LJ, Wong DJ, Galipeau PC, Paulson TG, Rabinovitch PS, Reid BJ. Evolution of neoplastic cell lineages in Barrett oesophagus. Nat Genet. 1999; 22:106–109. 11. Marotta LL, Almendro V, Marusyk A, Shipitsin M, Schemme J, Walker SR, Bloushtain-Qimron N, Kim JJ, Choudhury SA, Maruyama R, Wu Z, Gonen M, Mulvey LA, Bessarabova MO, Huh SJ, Silver SJ, et al. The JAK2/STAT3 signaling pathway is required for growth of CD44(+)CD24(−) stem cell-like breast cancer cells in human tumors. J Clin Invest. 2011; 121:2723–2735. 23. Maley CC, Galipeau PC, Finley JC, Wongsurawat VJ, Li X, Sanchez CA, Paulson TG, Blount PL, Risques RA, Rabinovitch PS, Reid BJ. Genetic clonal diversity predicts progression to esophageal adenocarcinoma. Nat Genet. 2006; 38:468–473. 12. Hartman ZC, Poage GM, den Hollander P, Tsimelzon A, Hill J, Panupinthu N, Zhang Y, Mazumdar A, Hilsenbeck SG, Mills GB, Brown PH. Growth of triple- negative breast cancer cells relies upon coordinate autocrine expression of the proinflammatory cytokines IL-6 and IL-8. Cancer Res. 2013; 73:3470–3480. 24. Holley T, Lenkiewicz E, Evers L, Tembe W, Ruiz C, Gsponer JR, Rentsch CA, Bubendorf L, Stapleton M, Amorese D, Legendre C, Cunliffe HE, McCullough AE, Pockaj B, Craig D, Carpten J, et al. Deep clonal profiling of formalin fixed paraffin embedded clinical samples. PLoS One. 2012; 7:e50586. 13. Britschgi A, Andraos R, Brinkhaus H, Klebba I, Romanet V, Muller U, Murakami M, Radimerski T, Bentires-Alj M. JAK2/STAT5 inhibition circumvents resistance to PI3K/ mTOR blockade: a rationale for cotargeting these pathways in metastatic breast cancer. Cancer Cell. 2012; 22:796–811. 25. Lipson D, Aumann Y, Ben-Dor A, Linial N, Yakhini Z. Efficient calculation of interval scores for DNA copy num ber data analysis. J Comput Biol. 2006; 13:215–228. 26. Hao Y, Chapuy B, Monti S, Sun HH, Rodig SJ, Shipp MA. Selective JAK2 inhibition specifically decreases Hodgkin lymphoma and mediastinal large B-cell lymphoma growth in vitro and in vivo. Clin Cancer Res. 2014; 20:2674–2683. 14. Statistical analysis 6. Topalian SL, Sznol M, McDermott DF, Kluger HM, Carvajal RD, Sharfman WH, Brahmer JR, Lawrence DP, Atkins MB, Powderly JD, Leming PD, Lipson EJ, 6. Topalian SL, Sznol M, McDermott DF, Kluger HM, Carvajal RD, Sharfman WH, Brahmer JR, Lawrence DP, Group comparisons used two-sample t-tests for continuous variables and chi-squared tests for categorical Atkins MB, Powderly JD, Leming PD, Lipson EJ, www.impactjournals.com/oncotarget Oncotarget 26491 Puzanov I, Smith DC, Taube JM, Wigginton JM, et al. Survival, durable tumor remission, and long-term safety in patients with advanced melanoma receiving nivolumab. J Clin Oncol. 2014; 32:1020–1030. reveals selective 9p24.1 amplification, increased PD-1 ligand expression, and further induction via JAK2 in nodu lar sclerosing Hodgkin lymphoma and primary mediastinal large B-cell lymphoma. Blood. 2010; 116:3268–3277. 17. 17. Evers L, Perez-Mancera PA, Lenkiewicz E, Tang N, Aust D, Knosel T, Rummele P, Holley T, Kassner M, Aziz M, Ramanathan RK, Von Hoff DD, Yin H, Pilarsky C, Barrett MT. STAG2 is a clinically relevant tumor suppressor in pancreatic ductal adenocarcinoma. Genome medicine. 2014; 6:9. 7. Wolchok JD, Kluger H, Callahan MK, Postow MA, Rizvi NA, Lesokhin AM, Segal NH, Ariyan CE, Gordon RA, Reed K, Burke MM, Caldwell A, Kronenberg SA, Agunwamba BU, Zhang X, Lowy I, et al. Nivolumab plus ipilimumab in advanced melanoma. N Engl J Med. 2013; 369:122–133. 18. Ruiz C, Lenkiewicz E, Evers L, Holley T, Robeson A, Kiefer J, Demeure MJ, Hollingsworth MA, Shen M, Prunkard D, Rabinovitch PS, Zellweger T, Mousses S, Trent JM, Carpten JD, Bubendorf L, et al. Advancing a clinically relevant perspective of the clonal nature of cancer. Proc Natl Acad Sci U S A. 2011; 108:12054–12059. 8. Postow MA, Chesney J, Pavlick AC, Robert C, Grossmann K, McDermott D, Linette GP, Meyer N, Giguere JK, Agarwala SS, Shaheen M, Ernstoff MS, Minor D, Salama AK, Taylor M, Ott PA, et al. Nivolumab and Ipilimumab versus Ipilimumab in Untreated Melanoma. The New England journal of medicine. 2015; 372:2006–17. 19. Ibrahim SF, van den Engh G. Flow cytometry and cell sorting. Adv Biochem Eng Biotechnol. 2007; 106:19–39. 9. Verstovsek S. Janus-activated kinase 2 inhibitors: a new era of targeted therapies providing significant clinical benefit for Philadelphia chromosome-negative myeloproliferative neoplasms. Journal of clinical oncology: official jour nal of the American Society of Clinical Oncology. 2011; 29:781–783. 20. Glogovac JK, Porter PL, Banker DE, Rabinovitch PS. Statistical analysis Miller CP, Thorpe JD, Kortum AN, Coy CM, Cheng WY, Ou Yang TH, Anastassiou D, Beatty JD, Urban ND, Blau CA. JAK expression is associated with tumor- infiltrating lymphocytes and improved breast cancer out comes: implications for evaluating JAK inhibitors. Cancer immunology research. 2014; 2:301–306. 27. Ansell SM, Lesokhin AM, Borrello I, Halwani A, Scott EC, Gutierrez M, Schuster SJ, Millenson MM, Cattry D, Freeman GJ, Rodig SJ, Chapuy B, Ligon AH, Zhu L, Grosso JF, Kim SY, et al. PD-1 Blockade with Nivolumab in Relapsed or Refractory Hodgkin’s Lymphoma. N Engl J Med. 2015; 372:311–9. 15. Comprehensive molecular characterization of gastric adeno carcinoma. Nature. 2014; 513:202–209. 16. Green MR, Monti S, Rodig SJ, Juszczynski P, Currie T, O’Donnell E, Chapuy B, Takeyama K, Neuberg D, Golub TR, Kutok JL, Shipp MA. Integrative analysis 28. Emens LA, Braiteh FS, Cassier P, DeLord J-P, Eder JP, Shen X, Xiao Y, Wang Y, Hedge PS, Chen DS, Krop I. Inhibition of PD-L1 by MPDL3280A leads to clinical www.impactjournals.com/oncotarget Oncotarget 26492 in cancer. The New England journal of medicine. 2012; 366:2443–2454. activity in patients with metastatic triple-negative breast cancer. San Antonio Breast Cancer Symposium. (San Antonio TX) 2014. 35. Mittendorf EA, Philips AV, Meric-Bernstam F, Qiao N, Wu Y, Harrington S, Su X, Wang Y, Gonzalez-Angulo AM, Akcakanat A, Chawla A, Curran M, Hwu P, Sharma P, Litton JK, Molldrem JJ, et al. PD-L1 expression in triple- negative breast cancer. Cancer immunology research. 2014; 2:361–370. 29. Nanda R, Chow LQ, Dees EC, Berger R, Gupta S, Geva R, Pusztai L, Dolled-Filhart M, Emancipator K, Gonzalez E, Houp J, Pathiraja K, Karantza V, Iannone R, Gause C, Cheng J, et al. A phase Ib study of pembrolizumab (MK-3475) in patients with advanced triple-negative breast cancer. San Antonio Breast Cancer Symposium. (San Antonio TX) 2014. 36. Lehmann BD, Pietenpol JA. Identification and use of biomarkers in treatment strategies for triple-negative breast cancer subtypes. The Journal of pathology. 2014; 232:142–150. 30. Taube JM, Klein A, Brahmer JR, Xu H, Pan X, Kim JH, Chen L, Pardoll DM, Topalian SL, Anders RA. Association of PD-1, PD-1 ligands, and other features of the tumor immune microenvironment with response to anti-PD-1 therapy. Clin Cancer Res. 2014; 20:5064–5074. 37. Przybytkowski E, Lenkiewicz E, Barrett MT, Klein K, Nabavi S, Greenwood CM, Basik M. Chromosome- breakage genomic instability and chromothripsis in breast cancer. BMC Genomics. 2014; 15:579. 31. Statistical analysis Wu J, Liu S, Liu G, Dombkowski A, Abrams J, Martin- Trevino R, Wicha MS, Ethier SP, Yang ZQ. Identification and functional analysis of 9p24 amplified genes in human breast cancer. Oncogene. 2012; 31:333–341. 38. Stieber D, Golebiewska A, Evers L, Lenkiewicz E, Brons NH, Nicot N, Oudin A, Bougnaud S, Hertel F, Bjerkvig R, Vallar L, Barrett MT, Niclou SP. Glioblastomas are composed of genetically divergent clones with distinct tumourigenic potential and variable stem cell-associated phenotypes. Acta neuropathologica. 2014; 127:203–219. 32. Balko JM, Giltnane JM, Wang K, Schwarz LJ, Young CD, Cook RS, Owens P, Sanders ME, Kuba MG, Sanchez V, Kurupi R, Moore PD, Pinto JA, Doimi FD, Gomez H, Horiuchi D, et al. Molecular profiling of the residual dis ease of triple-negative breast cancers after neoadjuvant chemotherapy identifies actionable therapeutic targets. Cancer Discov. 2014; 4:232–245. 39. Galipeau PC, Li X, Blount PL, Maley CC, Sanchez CA, Odze RD, Ayub K, Rabinovitch PS, Vaughan TL, Reid BJ. NSAIDs modulate CDKN2A, TP53, and DNA content risk for progression to esophageal adenocarcinoma. PLoS Med. 2007; 4:e67. 33. Brahmer JR, Tykodi SS, Chow LQ, Hwu WJ, Topalian SL, Hwu P, Drake CG, Camacho LH, Kauh J, Odunsi K, Pitot HC, Hamid O, Bhatia S, Martins R, Eaton K, Chen S, et al. Safety and activity of anti-PD-L1 antibody in patients with advanced cancer. The New England journal of medicine. 2012; 366:2455–2465. 40. Rabinovitch PS, Longton G, Blount PL, Levine DS, Reid BJ. Predictors of progression in Barrett’s esophagus III: baseline flow cytometric variables. Am J Gastroenterol. 2001; 96:3071–3083. 41. Piscuoglio S, Ng CK, Martelotto LG, Eberle CA, Cowell CF, Natrajan R, Bidard FC, De Mattos-Arruda L, Wilkerson PM, Mariani O, Vincent-Salomon A, Weigelt B, Reis-Filho JS. Integrative genomic and transcriptomic characterization of papillary carcinomas of the breast. Mol Oncol. 2014; 8:1588–1602. 34. Topalian SL, Hodi FS, Brahmer JR, Gettinger SN, Smith DC, McDermott DF, Powderly JD, Carvajal RD, Sosman JA, Atkins MB, Leming PD, Spigel DR, Antonia SJ, Horn L, Drake CG, Pardoll DM, et al. Safety, activity, and immune correlates of anti-PD-1 antibody www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget Oncotarget 26493
https://openalex.org/W3004513325	https://europepmc.org/articles/pmc7007972?pdf=render	English	null	Assessment of the vulnerability of alpine grasslands on the Qinghai-Tibetan Plateau	PeerJ	2,020	cc-by	9,130	Subjects Biogeography, Ecosystem Science, Climate Change Biology, Environmental Impacts Keywords Alpine grasslands, Exposure, Sensitivity, Resilience, Vulnerability How to cite this article Li M, Zhang X, He Y, Niu B, Wu J. 2020. Assessment of the vulnerability of alpine grasslands on the Qinghai- Tibetan Plateau. PeerJ 8:e8513 http://doi.org/10.7717/peerj.8513 Assessment of the vulnerability of alpine grasslands on the Qinghai-Tibetan Meng Li1,2, Xianzhou Zhang1,3, Yongtao He1,3, Ben Niu1 and Jianshuang Wu4 1 Lhasa National Ecological Research Station, Key Laboratory of Ecosystem Network Observation and Modelling, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing, China 2 University of Chinese Academy of Science, Beijing, China 3 College of Resources and Environment, University of Chinese Academy of Sciences, Beijing, China 4 Institute of Environment and Sustainable Development in Agriculture, Chinese Academy of Agricultural Sciences, Beijing, China 3 College of Resources and Environment, University of Chinese Academy of Sciences, INTRODUCTION Assessment of the vulnerability of ecosystems which focuses on the assessment of the potential effects of perturbations on a specific ecosystem has become a major topic in the field of global change ecology and sustainability research (Urruty, Tailliez-Lefebvre & Huyghe, 2016; Xu et al., 2016). It is well recognized that climate change is driving structural and functional changes in natural ecosystems (Tilman & Lehman, 2001; Venter et al., 2016), especially in the sensitive and vulnerable alpine and montane ecosystems (Crossman, Bryan Assessment of the vulnerability of ecosystems which focuses on the assessment of the potential effects of perturbations on a specific ecosystem has become a major topic in the field of global change ecology and sustainability research (Urruty, Tailliez-Lefebvre & Huyghe, 2016; Xu et al., 2016). It is well recognized that climate change is driving structural and functional changes in natural ecosystems (Tilman & Lehman, 2001; Venter et al., 2016), especially in the sensitive and vulnerable alpine and montane ecosystems (Crossman, Bryan & Summers, 2012). In the future, increasing climate variability and more frequent and intense extreme events are likely to increase risks to natural ecosystems (Kharin et al., 2013; Li et al., 2018). Some studies have confirmed that some species or ecosystems have a high ability to adapt to climate change, whereas others are suffering negative consequences (Kozlov, 2008; Neilson et al., 2005). Therefore, identification and prioritization of the vulnerable areas are crucial to mitigate the threat of climate change to ecosystems and to achieve sustainable management and adaptive conservation of natural ecosystems. Assessment of vulnerability aims to measure the ability of an ecosystem to resist and cope with environmental perturbations (Crossman, Bryan & Summers, 2012). A more effective and accurate vulnerability assessment framework was presented by Intergovernmental Panel on Climate Change (IPCC), which described the vulnerability as a function of ecosystem sensitivity and adaptivity to climate change with a different character, magnitude, and rate (IPCC, 2001). This definition integrates the multiple properties and processes of ecosystems, including sensitivity, resilience, and exposure, into vulnerability assessment (Xu et al., 2016). Sensitivity measures the ability of ecosystems to withstand environmental disturbances, and is quantified by the magnitude of vegetation response at the moment of the climate anomaly (De Keersmaecker et al., 2015). ABSTRACT Assessing ecosystem vulnerability to climate change is critical for sustainable and adaptive ecosystem management. Alpine grasslands on the Qinghai-Tibetan Plateau are considered to be vulnerable to climate change, yet the ecosystem tends to maintain stability by increasing resilience and decreasing sensitivity. To date, the spatial pattern of grassland vulnerability to climate change and the mechanisms that vegetation applies to mitigate the impacts of climate change on grasslands by altering relevant ecosystem characteristics, especially sensitivity and resilience, remain unknown. In this study, we first assessed the spatial pattern of grassland vulnerability to climate change by integrating exposure, sensitivity, and resilience simultaneously, and then identified its driving forces. The results show that grasslands with high vulnerability were mainly located on the edges of the plateau, whereas alpine grasslands in the hinterlands of the plateau showed a low vulnerability. This spatial pattern of alpine grassland vulnerability was controlled by climatic exposure, and grassland sensitivity and resilience to climate change might also exacerbate or alleviate the degree of vulnerability. Climate change had variable impacts on different grassland types. Desert steppes were more vulnerable to climate change than alpine meadows and alpine steppes because of the high variability in environmental factors and their low ability to recover from perturbations. Our findings also confirm that grazing intensity, a quantitative index of the most important human disturbance on alpine grasslands in this plateau, was significantly correlated with ecosystem vulnerability. Moderate grazing intensity was of benefit for increasing grassland resilience and then subsequently reducing grassland vulnerability. Thus, this study suggests that future assessments of ecosystem vulnerability should not ignore anthropogenic disturbances, which might benefit environmental protection and sustainable management of grasslands on the Qinghai-Tibetan Plateau. Submitted 27 September 2019 Accepted 4 January 2020 Published 6 February 2020 Corresponding author Xianzhou Zhang, zhangxz@igsnrr.ac.cn Academic editor Jianhua Xu Additional Information and Declarations can be found on page 15 DOI 10.7717/peerj.8513 Copyright 2020 Li et al. Distributed under Distributed under Creative Commons CC-BY 4.0 Distributed under Creative Commons CC-BY 4.0 OPEN ACCESS How to cite this article Li M, Zhang X, He Y, Niu B, Wu J. 2020. Assessment of the vulnerability of alpine grasslands on the Qinghai- Tibetan Plateau. PeerJ 8:e8513 http://doi.org/10.7717/peerj.8513 Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 INTRODUCTION Resilience refers to the ability of ecosystems to recover to its original state after the disturbance, or the magnitude of absorbed disturbance by the ecosystem before the ecosystem’s structure begins to change (Holling, 1996; Scheffer et al., 2001). Exposure refers to the degree of climate disturbance experienced by a species or ecosystem and represents the rate of migration that species need to follow climate change (Loarie et al., 2009). More recently, various methods have been developed to quantify the vulnerability of ecosystems, including the comprehensive index method (Nguyen et al., 2016), the quantitative evaluation model method (Zhao & Wu, 2014), and the scenario analysis method (Woznicki et al., 2016). However, some shortcomings remain in those methods. For example, although many studies have used the comprehensive index method to assess the vulnerability, the indicator selection and weight determination of the indices remain controversial. The quantitative evaluation model method is limited by the complexity of the dynamic vegetation. Although scenario-based assessment provides an advance in predicting future ecological vulnerability, it requires large quantities of historical data to drive the model (Jiang et al., 2018). As a consequence, the assessment of vulnerability is labor-intensive and may be less accurate. With the development of remote sensing, some studies have quantified the components of vulnerability, such as sensitivity and resilience, based on the short-term dynamics of vegetation and climatic factors at regional and Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 2/20 global scales (De Keersmaecker et al., 2015; Li et al., 2018; Li et al., 2019a), which provides a feasible and efficient method to assess ecosystem vulnerability. Alpine grasslands dominate the Qinghai-Tibetan Plateau, covering 1.28 × 108 hm2 and accounting for more than 65% of its total land area (Long et al., 1999). These grasslands are critical for livestock husbandry and environmental security (Yang et al., 2017; Yao et al., 2012). As a result of the harsh characteristics of high altitude, drought, and cold, alpine grasslands on the Qinghai-Tibetan Plateau are sensitive and vulnerable to climate change and human activities (Harris, 2010; Shang et al., 2014). In the past decades, most grasslands have experienced accelerated warming (Chen et al., 2013; Li et al., 2010). Moreover, the effects of human disturbances (mainly referring to livestock grazing) on alpine grasslands have also intensified. INTRODUCTION Currently, under the influences of overgrazing and climate change, nearly 40% of the grassland has been degraded, resulting in decreases in plant diversity and productivity of alpine grasslands (Zhang et al., 2015b). Grassland degradation and even desertification are likely to increase ecosystems vulnerability and seriously threaten ecological security and regional sustainable development. Assessment of the vulnerability of alpine grasslands on the Qinghai-Tibetan Plateau is a necessary first step to explore target measures to eliminate or alleviate negative influences. Although several studies have explored the effects of climate change on grassland ecosystems (Fu, Shen & Zhang, 2018; Shen et al., 2015), to our knowledge, no studies have illustrated the patterns of the vulnerability of alpine grasslands on the Qinghai-Tibetan Plateau based on the IPCC definition of vulnerability. In this study, we first integrated sensitivity, resilience, and exposure to quantify the vulnerability of alpine grasslands on the Qinghai-Tibetan Plateau according to the IPCC framework. We then compared the differences in vulnerability among different eco- geographical regions, grassland types, and grazing intensities. We aimed to (1) identify where grasslands on the Qinghai-Tibetan Plateau are most vulnerable, and (2) detect what aspects determine their vulnerability. Overall, we hope this study can assist policymakers and stakeholders in achieving sustainable and adaptive management of alpine grasslands in the context of climate change on the Qinghai-Tibetan Plateau. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 Study area The Qinghai-Tibetan Plateau, known as ‘‘the water tower of Asia’’, is vital to safeguard ecological security for both China and South Asia (Yao et al., 2012). The plateau is located in the arid alpine climate zone with an average elevation over 4,000 m above sea level. The annual average temperatures range from −15 ◦C to 10 ◦C (You et al., 2013). The precipitation distribution in the Qinghai-Tibetan Plateau shows much spatial heterogeneity, with mean annual precipitation being about 50–150 mm in the northwest and 300–450 mm in the southeast. From southeast to northwest, grassland types alter from humid alpine meadow, semi-arid alpine steppe, to arid alpine desert-steppe (Fig. 1A). The Qinghai- Tibetan Plateau is separated by the Tanggula Mountains into two regions, Qinghai province and the Tibet Autonomous Region (Fig. 1B). The northern Qinghai-Tibetan Plateau, also Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 3/20 Figure 1 Eco-geographical regions (A) and alpine grassland types (B) on the Qinghai-Tibetan Plateau. IB1, Golog-Nagqu high-cold shrub-meadow zone; ICI, Southern Qinghai high-cold meadow steppe zone; IC2, Qiangtang high-cold steppe zone; ID1, Kunlun high-cold desert zone; IIAB1, Western Sichuan- eastern Tibet montane coniferous forest zone; IIC2, Eastern Qinghai-Qilian montane steppe zone; IIC1, Southern Tibet montane shrub-steppe zone; IID2, Qaidam montane desert zone; OA1, Southern slopes of Himalaya montane evergreen broad-leaved forest zone; IID1, Nagri montane desert-steppe and desert zone; IID3, Northern slopes of Kunlun montane desert zone. Full-size DOI: 10.7717/peerj.8513/fig-1 Figure 1 Eco-geographical regions (A) and alpine grassland types (B) on the Qinghai-Tibetan Plateau. IB1, Golog-Nagqu high-cold shrub-meadow zone; ICI, Southern Qinghai high-cold meadow steppe zone; IC2, Qiangtang high-cold steppe zone; ID1, Kunlun high-cold desert zone; IIAB1, Western Sichuan- eastern Tibet montane coniferous forest zone; IIC2, Eastern Qinghai-Qilian montane steppe zone; IIC1, Southern Tibet montane shrub-steppe zone; IID2, Qaidam montane desert zone; OA1, Southern slopes of Himalaya montane evergreen broad-leaved forest zone; IID1, Nagri montane desert-steppe and desert zone; IID3, Northern slopes of Kunlun montane desert zone. Full size DOI: 10 7717/peerj 8513/fig 1 named ‘‘Changtang’’, contains the largest national nature reserve in China. The Three Rivers Headwater region contains high biodiversity and a fragile and sensitive ecological environment, and is the birthplace of China’s three major rivers, the Yangtze River, the Yellow River and the Lancang River (Shao et al., 2017). The vulnerability of the grassland ecosystem In this study, we quantified the alpine grassland ecosystem vulnerability to climate change at the regional scale by simultaneously considering sensitivity, resilience, and exposure (Li et al., 2018). The formula is as follows: VI = ((EI×SI)/(1+RI))1/2 (1) where VI is the vulnerability index, SI is the sensitivity index, RI is the resilience index, and EI is the exposure index. where VI is the vulnerability index, SI is the sensitivity index, RI is the resilience index, and EI is the exposure index. In recent years, a novel method was presented by De Keersmaecker et al. (2015) to simultaneously quantify vegetation sensitivity and resilience under short-time climate anomalies, which has been successfully used at the regional and global scale (Geng et al., 2019; Seddon et al., 2016). In this study, we accepted this empirical methodology to quantify the sensitivity and resilience of alpine grasslands productivity to short-term climate variability for each pixel on the Qinghai-Tibetan Plateau. This method assumes that the change of vegetation index is a linear combination of climatic factors and the changes in vegetation index in the early stage. We took the grassland NDVI anomaly as the dependent variable and the temperature, precipitation (or water availability), incoming radiation anomaly and NDVI anomaly history as independent variables to model the vegetation response to short-term climate anomalies, as the following equation (autoregressive model, AR1): (2) NDVIt = α×Tt +β×Pt +γ×Rt +δ×NDVIt−1 +ε (2) NDVIt = α×Tt +β×Pt +γ×Rt +δ×NDVIt−1 +ε where NDVIt is the standardized NDVI anomaly at time t, Tt is the standardized temperature anomaly at time t, Pt is the standardized precipitation index at time t, Rt is the standardized radiation index at time t, NDVIt−1 is the standardized NDVI anomaly at time t−1 and ε is the residual error. α, β, γ, and δ are model coefficients. To ensure the comparability among the three coefficients, all-time series of NDVI and the three climate variables were de-trended and then transformed to z-score anomalies using variables’ means and standard deviations. The sensitivity index was quantified with the combination of α, β, and γ. The coefficient α is the temperature sensitivity metric denoting the response of grassland NDVI to instantaneous variation in temperature, where a higher α value indicates a higher sensitivity of grasslands to temperature, and vice versa. Similarly, β and γ represent drought sensitivity metric and radiation sensitivity metric, respectively. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 Data collection The normalized difference vegetation index (NDVI) has been widely applied in regional ecosystem monitoring and evaluation. In this study, we downloaded the Version 6 NDVI from the monthly MOD13A3 product of the Moderate Resolution Imaging Spectroradiometer (MODIS), which is at 1 km spatial resolution (https: //lpdaac.usgs.gov/get_data/data_pool). The monthly NDVI data were developed using the maximum value composition method, and have been calibrated for geometric effects, atmospheric effects, and cloud contamination. The NDVI dataset was smoothed and reconstructed with the Savitzky–Golay method to exclude the effects of cloud, snow, and ice contamination (Chen et al., 2004). Monthly mean temperature and total precipitation data between 2000 and 2017 were collected from the National Meteorological Information Center (NMIC) of the China Meteorological Administration (CMA) (http://www.cma.gov.cn/). The origin meteorological data were interpolated into raster surfaces at the same spatial resolution as the NDVI data using ANUSPLIN 4.3 (Hutchinson, 2004). ANUSPLIN is a professional interpolation software in which one or more influence factors can be included as covariates to improve the interpolation accuracy, especially for time series of meteorological data (Hutchinson, 2004). Our previous studies have shown the accuracy of interpolation data (Chen et al., 2014; Li et al., 2019b). Information on the distribution of grasslands was obtained from the vegetation atlas of China with a scale of 1:1,000,000 (Chinese Academy of Sciences, 2001). Eco-geographical Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 4/20 zones adopt the framework scheme of the China eco-geographic regional system drawn up by Zheng (1996). zones adopt the framework scheme of the China eco-geographic regional system drawn up by Zheng (1996). The vulnerability of the grassland ecosystem To calculate the sensitivity metric, we first quantified the relative contributions of temperature, precipitation, and radiation to the grassland sensitivity index for each pixel based on a principal component regression (PCR). Second, the sensitivity scores for each climatic variable were calculated by the ratio of NDVI variance and each climate variance in time series. Then, the ratios were weighted using the contribution of each climatic variable to NDVI variation. Finally, the sensitivity index was calculated by summing the sensitivity scores. More detailed methods and the R script for calculating the sensitivity index can be found in Seddon et al. (2016). Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 5/20 The coefficient δ is an indication of the similarity between the current state and the previous state. If δ is large, ecosystem anomalies at time t are strongly dependent on the anomaly at time t−1, and the ecosystem recovers slowly from any disturbance. Conversely, ecosystems with smaller δ values tend to recover quickly from any disturbance. As such, δ can be considered as a resilience metric, with higher absolute δ values representing lower resilience (De Keersmaecker et al., 2015; Li et al., 2018). Thus, in this study, the resilience index is equal to 1−δ. The exposure index was defined as the ratio of the temporal climate gradient to the spatial climate gradient (Loarie et al., 2009). The formula is as follows: Etem = (◦C×year−1)/(◦C×km−1) (3) Epre = (mm×year−1)/(mm×km−1) (4) Erad = (MJ×year−1)/(MJ×km−1) (5) EI = α×Etem +β×Epre +γ×Erad (6) (6) where α, β, γ, and δ are the coefficients from Eq. (2). The temporal trend from 2000 to 2017 was calculated using least squares regression, and the spatial gradient was calculated using the average maximum technique based on a 3 × 3 grid cell neighborhood (Burrows et al., 2011). where α, β, γ, and δ are the coefficients from Eq. (2). The temporal trend from 2000 to 2017 was calculated using least squares regression, and the spatial gradient was calculated using the average maximum technique based on a 3 × 3 grid cell neighborhood (Burrows et al., 2011). Sensitivity, resilience, exposure, and vulnerability of grassland ecosystems were normalized between 0 and 100 using the minimum and maximum values. The vulnerability of the grassland ecosystem To measure the relative vulnerability and ensure its spatial comparison, we binned grassland areas into five levels of ecosystem sensitivity, resilience, exposure, and vulnerability, such that each bin had an equal interval, and labeled these bins as slight (0–20), low (20–40), moderate (40–60), high (60–80), and extreme (80–100). Grazing intensity The main human activity on the Qinghai-Tibetan Plateau is livestock grazing, the extent of which was calculated based on the statistical data with the following equation: GI = S/A (7) (7) GI = S/A where GI represents the index of grazing intensity (GI), S denotes the numbers of livestock in each county, and A represents the area of available natural grassland in each county (ha). The numbers of livestock were acquired from the yearly ‘‘Statistical Yearbook of Tibet’’ and ‘‘Agricultural Statistics Manual of Qinghai Province’’ from 2000 to 2017. Sheep and large livestock were included in the original livestock data. In this study, different animals were standardized into sheep units (SHU) with the criterion that one sheep is equal to one SHU and one large livestock is equal to four SHU (Fan et al., 2010). The average grazing intensity from 2000 to 2017 for 78 counties (16 counties in Qinghai province and 62 counties in the Tibet autonomous region) was used to explore the effects of human activities on the vulnerability index as well as the three components, including sensitivity, resilience, and exposure. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 6/20 Figure 2 Spatial patterns of standardized grassland (A) sensitivity index (SI), (B) resilience index (RI), (C) exposure index (EI), and (D) vulnerability index (VI) on the Qinghai-Tibet Plateau. Full-size DOI: 10.7717/peerj.8513/fig-2 Figure 2 Spatial patterns of standardized grassland (A) sensitivity index (SI), (B) resilience index (RI), (C) exposure index (EI), and (D) vulnerability index (VI) on the Qinghai-Tibet Plateau. Full-size DOI: 10.7717/peerj.8513/fig-2 Statistical analysis We employed the analysis of variance (ANOVA) to analyze the difference of grassland vulnerability and its components among different eco-geographical regions and also among different vegetation types. General linear models (GLMs) were conducted to explore the relative effect strengths of grassland types and grazing intensity to the spatial patterns of grassland vulnerability and its components. Spatial patterns of grassland vulnerability and its components ulnerable grasslands (60 < VI < 80) and extremely vulnerable grasslands (80 < VI < 100) vulnerable grasslands (60 < VI < 80) and extremely vulnerable grasslands (80 < VI < 100) together accounted for 32.76% of the total grassland areas, which were mainly distributed in the eastern and western parts of this plateau. The grasslands with a slight (VI < 20) and low vulnerability index (20 < VI < 40) accounted for 16.82% and 24.51% of the entire grassland area on the plateau, respectively, and were mainly distributed in the center of the plateau. g ( ) y g ( ) together accounted for 32.76% of the total grassland areas, which were mainly distributed in the eastern and western parts of this plateau. The grasslands with a slight (VI < 20) and low vulnerability index (20 < VI < 40) accounted for 16.82% and 24.51% of the entire grassland area on the plateau, respectively, and were mainly distributed in the center of the plateau. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 Spatial patterns of grassland vulnerability and its components Spatial patterns of grassland vulnerability and its components The sensitivity, resilience, exposure, and vulnerability index of alpine grasslands showed distinct patterns on the Qinghai-Tibetan Plateau (Fig. 2). In terms of the sensitivity index, most of the alpine grasslands were moderately sensitive to climate change (40 < SI < 60), which accounted for 67.43% of grassland pixels (Fig. 2A). The extreme sensitivity level (SI > 80), accounting for only 2.72% of grassland pixels, was mainly distributed in the south of the plateau. The areas of slight and low sensitivity level were primarily located in the east of Qinghai province (SI < 20), accounting for 1.45% of the grasslands on the plateau (Table 1). For the resilience index (Fig. 2B), grasslands with slight resilience (RI < 20) accounted for 3.63% of grasslands on the whole plateau, and were mainly situated in the north of the plateau. Moderate (40 < RI < 60) and high resilience index (60 < RI < 80) areas were scattered across the whole plateau, accounting for 39.92% and 29.29% of the total grassland area, respectively. The spatial pattern and the proportion among five different levels of the exposure index were fairly consistent with the vulnerability index (Figs. 2C and 2D, Table 1). Highly Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 7/20 Table 1 Pixel percentage (%) of the standardized sensitivity index (SI), resilience index (RI), exposure index (EI), and vulnerability index (VI) on the Qinghai-Tibet Plateau. Index Pixel percentage (%) Slight (0–20) Low (20–40) Moderate (40–60) High (60–80) Extreme (80–100) SI 1.45 9.04 67.43 19.37 2.72 RI 3.63 18.94 39.92 29.29 8.21 EI 20.20 28.77 25.64 12.92 12.48 VI 16.82 24.51 25.92 17.05 15.71 vulnerable grasslands (60 < VI < 80) and extremely vulnerable grasslands (80 < VI < 100) together accounted for 32.76% of the total grassland areas, which were mainly distributed in the eastern and western parts of this plateau. The grasslands with a slight (VI < 20) and low vulnerability index (20 < VI < 40) accounted for 16.82% and 24.51% of the entire grassland area on the plateau, respectively, and were mainly distributed in the center of the plateau. Table 1 Pixel percentage (%) of the standardized sensitivity index (SI), resilience index (RI), exposure index (EI), and vulnerability index (VI) on the Qinghai-Tibet Plateau. Grassland vulnerability of diﬀerent eco-geographical regions and grassland types Grassland vulnerability of diﬀerent eco-geographical regions and grassland types The division of eco-geographical regions took both the characteristics of climatic factors and vegetation cover information into consideration at the same time. Except for the OA1 region, the remaining ten eco-geographic regions showed an alpine grassland distribution. The sensitivity, resilience, exposure, and vulnerability index within the ten eco-graphical regions were compared. As shown in Fig. 3, IIAB1 and IIC1 in the humid area of the eastern Qinghai-Tibetan plateau had the highest sensitivity index, whereas IC1 and IC2 in the center of the plateau had a relatively low sensitivity index. The resilience index in IIAB1 and IIC1 was higher than that in other eco-geographical regions. The vulnerability index among the ten eco-geographical regions was similar with the exposure index, but with some differences. For example, although the lowest exposure index was found in IID3, the vulnerability index in IID3 was not the lowest because of the higher sensitivity index. This indicates that the degree of exposure determined the vulnerability of alpine grasslands on the plateau, but it was also affected by grassland sensitivity and resilience simultaneously. Figure 4 shows the sensitivity, resilience, exposure, and vulnerability index for different grassland types on the Qinghai-Tibetan plateau. Compared with the other three indexes, the sensitivity index was less variable among alpine meadow (53.1), alpine steppe (51.8), and desert steppe (56.3), but with a significant difference (P < 0.05, ANOVA test). The average resilience index of the alpine meadow (57.7) was relatively higher than desert steppe (40.7) and alpine steppe (50.6). For the exposure index, desert steppe (49.3) was significantly higher than that in the alpine steppe (42.4) and alpine meadow (44.7). Desert steppe also had the highest vulnerability to climate change (63.7), whereas alpine steppe (47.9) and alpine meadow (47.4) were subject to less vulnerability. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 8/20 Figure 3 Grassland standardized (A) sensitivity index (SI), (B) resilience index (RI), (C) exposure in- dex (EI) and (D) vulnerability index (VI) for each eco-geographical region on the Qinghai-Tibetan Plateau. Bars in the figure represent standard error. Full-size DOI: 10.7717/peerj.8513/fig-3 Figure 4 Grassland sensitivity index (SI), resilience index (RI), exposure index (EI), and vulnerabil- ity index (VI) for each grassland type on the Qinghai-Tibetan Plateau. AM, alpine meadow; AS, alpine steppe; DS, desert steppe. Bars in the figure represent standard error. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 The eﬀects of grazing intensity on grassland vulnerability g g y g y Grazing intensity showed a distinct spatial pattern among counties on the Qinghai- Tibetan Plateau. Eastern and central counties in Tibet and Qinghai province had higher grazing intensity (above 1.5 SHU/ha) than western counties (below 0.3 SHU/ha) (Fig. 5). Quantitative analysis was conducted on the grazing intensity and the corresponding sensitivity index, resilience index, exposure index, and vulnerability index. As shown in Fig. 6, the index of grazing intensity had a significant impact on the vulnerability of alpine grasslands as well as the three components (P < 0.05). Specially, we found unimodal responses of sensitivity index and resilience index to grazing intensity across the counties (Figs. 6A and 6B). This indicates that grasslands with moderate grazing intensity might result in high sensitivity and resilience. The response of exposure index to the changes in grazing intensity was similar to that of the vulnerability index, which shown as a significant U-shaped pattern in Figs. 6C and 6D. Thus, moderate grazing intensity might play a crucial role in preventing alpine grasslands from becoming vulnerable. Grassland vulnerability of diﬀerent eco-geographical regions and grassland types Full-size DOI: 10.7717/peerj.8513/fig-4 Figure 3 Grassland standardized (A) sensitivity index (SI), (B) resilience index (RI), (C) exposure in- dex (EI) and (D) vulnerability index (VI) for each eco-geographical region on the Qinghai-Tibetan Plateau. Bars in the figure represent standard error. Full-size DOI: 10.7717/peerj.8513/fig-3 Figure 3 Grassland standardized (A) sensitivity index (SI), (B) resilience index (RI), (C) exposure in- dex (EI) and (D) vulnerability index (VI) for each eco-geographical region on the Qinghai-Tibetan Plateau. Bars in the figure represent standard error. Full-size DOI: 10.7717/peerj.8513/fig-3 Figure 3 Grassland standardized (A) sensitivity index (SI), (B) resilience index (RI), (C) exposure in- dex (EI) and (D) vulnerability index (VI) for each eco-geographical region on the Qinghai-Tibetan Plateau. Bars in the figure represent standard error. Full-size DOI: 10.7717/peerj.8513/fig-3 dex (EI) and (D) vulnerability index (VI) for each eco-geographical region on the Qinghai-Tibetan Plateau. Bars in the figure represent standard error. Full-size DOI: 10.7717/peerj.8513/fig-3 Figure 4 Grassland sensitivity index (SI), resilience index (RI), exposure index (EI), and vulnerabil- ity index (VI) for each grassland type on the Qinghai-Tibetan Plateau. AM, alpine meadow; AS, alpine steppe; DS, desert steppe. Bars in the figure represent standard error. Full-size DOI: 10.7717/peerj.8513/fig-4 Figure 4 Grassland sensitivity index (SI), resilience index (RI), exposure index (EI), and vulnerabil- ity index (VI) for each grassland type on the Qinghai-Tibetan Plateau. AM, alpine meadow; AS, alpine steppe; DS, desert steppe. Bars in the figure represent standard error. Full-size DOI: 10.7717/peerj.8513/fig-4 Figure 4 Grassland sensitivity index (SI), resilience index (RI), exposure index (EI), and vulnerabil- ity index (VI) for each grassland type on the Qinghai-Tibetan Plateau. AM, alpine meadow; AS, alpine steppe; DS, desert steppe. Bars in the figure represent standard error. Full-size DOI: 10.7717/peerj.8513/fig-4 Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 9/20 Figure 5 Mean grazing intensity (GI) from 2000 to 2017 for the 78 counties on the Qinghai-Tibet Plateau. White color counties in Tibetan indicate grassland areas less than 50 km2, and white color counties in Qinghai indicates null-value due to unavailable data. SHU represent sheep unit. Full-size DOI: 10.7717/peerj.8513/fig-5 Figure 5 Mean grazing intensity (GI) from 2000 to 2017 for the 78 counties on the Qinghai-Tibet Plateau. White color counties in Tibetan indicate grassland areas less than 50 km2, and white color counties in Qinghai indicates null-value due to unavailable data. SHU represent sheep unit. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 Grassland vulnerability of diﬀerent eco-geographical regions and grassland types Full-size DOI: 10.7717/peerj.8513/fig-5 Figure 5 Mean grazing intensity (GI) from 2000 to 2017 for the 78 counties on the Qinghai-Tibet Plateau. White color counties in Tibetan indicate grassland areas less than 50 km2, and white color counties in Qinghai indicates null-value due to unavailable data. SHU represent sheep unit. Full-size DOI: 10.7717/peerj.8513/fig Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 Notes. d.f., degree of freedom; SS, sum squares; MS, mean squares; F, variance ratio; P, significance; %SS, percentage of the total sum of squares explained. DISCUSSION With the increased variability of climate change, it is essential to understand the vulnerability of ecosystems, especially for high-altitude vegetation which is believed to be highly sensitive to climate change. Sensitivity, resilience, and exposure are essential components of vulnerability reflecting the degree of an ecosystem exposure and response to perturbations (Williams et al., 2008). In this study, the vulnerability of alpine grasslands on the Qinghai-Tibetan Plateau was calculated based on the combination of external (exposure) and intrinsic (sensitivity and resilience) components. In the following sections, we will discuss the spatial pattern of vulnerability and the driving factors for this pattern in detail. The relative contribution of grassland type and grazing intensity to the spatial pattern of vulnerability In the GLM analysis, grazing intensity, grassland type, and the interactions between them could explain 1.46%, 64.20%, 19.42%, and 28.24% of the total variation in sensitivity index, resilience index, exposure index, and vulnerability index, respectively (Table 2). For grassland vulnerability, we found that grassland type and grazing intensity accounted for 10.34% and 15.36% of its spatial variation, respectively. This indicates that grazing intensity had a relatively higher explanatory power than grassland type for the spatial variation in the vulnerability index. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 10/20 Figure 6 The relationship between grazing intensity and (A) sensitivity index (SI), (B) resilience index (RI), (C) exposure index (EI), and (D) vulnerability index (VI). AM, alpine meadow; AS, alpine steppe; DS, desert steppe. Full-size DOI: 10.7717/peerj.8513/fig-6 Figure 6 The relationship between grazing intensity and (A) sensitivity index (SI), (B) resilience index (RI), (C) exposure index (EI), and (D) vulnerability index (VI). AM, alpine meadow; AS, alpine steppe; DS, desert steppe. Full-size DOI: 10.7717/peerj.8513/fig-6 Table 2 Summary of the effects of grazing intensity (GI), grassland types (TYPE), and the interactions between them in general linear models (GLMs) on the sensitivity, resilience, exposure, and vulnerabil- ity indices on the Qinghai-Tibet Plateau. Table 2 Summary of the effects of grazing intensity (GI), grassland types (TYPE), and the interactions between them in general linear models (GLMs) on the sensitivity, resilience, exposure, and vulnerabil- ity indices on the Qinghai-Tibet Plateau. Index Explicators d.f. SS MS F P %SS GI 1 7.80 7.84 0.13 0.72 0.10 Type 2 77.70 38.87 0.62 0.54 0.98 SI GI:Type 2 30.60 15.32 0.25 0.78 0.38 GI 1 55.34 55.34 3.77 0.05 1.07 Type 2 2993.90 1496.95 102.08 0.00 58.01 RI GI:Type 2 264.04 132.02 9.00 0.00 5.12 GI 1 2211.40 2211.44 23.42 0.00 14.98 Type 2 590.20 295.12 3.13 0.05 4.00 EI GI:Type 2 65.00 32.51 0.34 0.71 0.44 GI 1 2165.50 2165.45 26.97 0.00 15.36 Type 2 1457.50 728.75 9.08 0.00 10.34 VI GI:Type 2 357.70 178.86 2.23 0.11 2.54 Notes. d.f., degree of freedom; SS, sum squares; MS, mean squares; F, variance ratio; P, significance; %SS, percentage of the total sum of squares explained. 11/20 Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 The spatial pattern of grassland vulnerability We found that grasslands with a high vulnerability were mainly located on the edges of the plateau, while grasslands with a low vulnerability were distributed in the hinterlands of the plateau. The spatial pattern of vulnerability resembles the exposure pattern. The results indicate that the exposure index was the crucial component in determining the vulnerability of the grassland ecosystems on the Qinghai-Tibetan Plateau, which is partially consistent with previous studies. For example, Li et al. (2018) found that the vulnerability of ecosystems is mainly determined by the degree of exposure. Beaugrand et al. (2015) also suggested that the influences of exposure to biodiversity were stronger than those of sensitivity. Exposure is an indicator reflecting the ability of species to keep pace with the changing climates (Loarie et al., 2009), which highly depends on the degree of regional climate change that involves the range of species and habitats. Species need a survival environment to be within a particular range for vital growth and reproduction processes (Burrows et al., 2011). Once environmental variability exceeds the tipping point that species unable to cope with, the shift of biogeographic ranges may occur and the ecosystem may become increasingly vulnerable (Beisner, Haydon & Cuddington, 2003). Thus, environmental variability is particularly important for species richness, community structure, and biodiversity (Chen et al., 2011; Lindner et al., 2010), and thereby can determine the degree of ecosystem vulnerability. However, some differences between the spatial pattern of exposure and the spatial pattern of vulnerability also occurred in some sub-regions. For example, alpine grasslands in the southern Qinghai-Tibetan Plateau were exposed to low climate variation but with high vulnerability. This may be because the differences in resilience and sensitivity might result in various responses of an ecosystem to the same climate change. This result provides evidence to support the framework that ecosystem vulnerability assessment should integrate sensitivity and resilience simultaneously. As illustrated above, the effects of climate change on ecosystems were not only determined by the magnitude and distribution of perturbations but also influenced by the ability of the target ecosystem to resist perturbations. High ecological sensitivity is likely to exacerbate ecosystem vulnerability, whereas high ecological resilience is likely to alleviate ecosystem vulnerability. In this study, we found that vegetation dynamics in response to climate change varied among grassland types. Earlier studies also have revealed Li et al. The spatial pattern of grassland vulnerability (2020), PeerJ, DOI 10.7717/peerj.8513 that the sensitivity of grasslands on the Qinghai-Tibetan Plateau to the changing climate change is complex and often varies dramatically among regions and grassland types (Li et al., 2019a). For example, grasslands on the southeastern Qinghai-Tibetan Plateau are generally sensitive to temperature variations while northeastern grasslands show strong responses to precipitation variations (Huang et al., 2016; Sun, Qin & Yang, 2016). With regard to the differences in ecological sensitivity among three grassland types, we found that desert steppes were more sensitive to climate change than alpine meadows and alpine steppes. Our result is in line with previous studies showing that grasslands in drier and warmer regions are highly sensitive to precipitation (Li et al., 2019a). Li et al. (2019b) also suggested that desert steppes are likely to have high sensitivity to the timing variability of precipitation on the Northern Tibetan Plateau. One potential explanation is that desert steppes are always characterized by poor species and low vegetation productivity due to low precipitation (Wu, Shen & Zhang, 2014), but with high precipitation variation which can result in rapid changes in the key carbon cycle process (Knapp et al., 2002). The resilience metric in our study is a kind of engineering resilience, emphasizing the maintenance of ecosystem function effectiveness rather than the probability of an ecosystem switching to another state (De Keersmaecker et al., 2015). This kind of resilience is closely linked with ecosystem function and plant community composition (Hoover, Knapp & Smith, 2014). Xu et al. (2016) indicated that high productivity and adaptive dominant species might result in high resilience. Therefore, alpine meadows with higher productivity and larger species richness pool might have a higher ability to recover from perturbations than alpine steppes and desert steppes (Wu, Shen & Zhang, 2014; Zhu, Jiang & Zhang, 2016). Besides, we also found that desert steppes exhibited high sensitivity and low resilience, indicating that the variations in grassland biomass for desert steppes are higher than alpine meadows and alpine steppes if the ecosystems experiencing same extrinsic perturbations. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 Grassland type and grazing intensity aﬀect grassland vulnerability yp g g y g y In the GLM, we found that both grassland types and grazing intensity were significant in driving the spatial variation of grassland vulnerability. First, the vulnerability differed dramatically among grassland types. Desert steppes were significantly more vulnerable than alpine meadows and alpine steppes. The high vulnerability of desert steppes might result from high sensitivity, high exposure, and low resilience. For the three grassland types on the Qinghai-Tibetan Plateau, we found that high exposure was always connected with high sensitivity. This finding is consistent with the previous study, which suggests that ecosystems with high perturbations are more sensitive to climate change than ecosystems with low perturbations over short timescales (Kröel-Dulay et al., 2015). Moreover, in this study, human activities were not considered when calculating the vulnerability index; however, the alpine grassland’s vulnerability was significantly correlated with grazing intensity which is considered as the primary human activity. This finding indicates that anthropogenic disturbances to ecosystems should not be ignored. Grazing is the main form of land use for most grassland ecosystems, which also affects the vulnerability of ecosystems to climate change (Christensen et al., 2004; Izaurralde et al., 2011). Multiple experiments have demonstrated that grazing affects the stability and self-regulating ability Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 13/20 of the ecosystem by changing grasslands’ structure and function, such as grassland productivity, biodiversity, plant community composition and water-use efficiency (Wang et al., 2012; Zhang et al., 2015a). These ecosystem changes could result in changes in the sensitivity (Christensen et al., 2004) and the resilience (Li, 1997) of vegetation to climate change or even the degree of exposure. In our study, a significant U-shaped relation was found between grazing intensity and vulnerability index, indicating that moderate grazing intensity might play a crucial role in mitigating the grassland vulnerability. This result is in line with the intermediate disturbance hypothesis, which suggests that human influences in the moderate or middle range can promote community succession and maintain community structure and species diversity (Grime, 1973), and thereby prevent ecosystems becoming vulnerable. The interactions of climate change and livestock grazing play a significant role in shaping grassland functions (Watson et al., 2000). In the future, more frequent and intense climate changes and more pressures on grasslands for their resources are expected (Kharin et al., 2013; Venter et al., 2016). Grassland type and grazing intensity aﬀect grassland vulnerability Thus, it is necessary to investigate interactions between climate change and grazing, to prevent grasslands from being vulnerable and to maintain sustainable grazing systems. Implications for managers and policymakers More information about climate and human activities changes should be supplied to policymakers and herdsmen, to develop region-specific policies and sustainable management strategies. ForthevastgrasslandontheQinghai-Tibetan Plateau,itisnecessary to use simple ecological indicators that reflect the information about the status and health of ecosystems to identify the priority areas. We first recommend that policymakers should pay more attention to western desert steppes because of the high vulnerability resulting from high exposure, high sensitivity, and low resilience. Once experience strong climatic perturbations, grasslands in these regions are easy to collapse, and do not easily recover. Optimizing allocation of management, such as monitoring ecological processes and functioning of grassland ecosystems and curbing overgrazing, is needed to maximize the grassland resistance and resilience to perturbations. Second, we highlight the necessity to conduct studies to predict future vulnerability. Predicting how ecosystems respond to climate change is useful to make adaptation and mitigation strategies to alleviate the effects of climate change on ecosystems. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 Competing Interests Competing Interests The authors declare there are no competing interests. Author Contributions • Meng Li conceived and designed the experiments, performed the experiments, analyzed the data, prepared figures and/or tables, authored or reviewed drafts of the paper, and approved the final draft. • Xianzhou Zhang conceived and designed the experiments, prepared figures and/or tables, authored or reviewed drafts of the paper, and approved the final draft. • Yongtao He and Ben Niu analyzed the data, authored or reviewed drafts of the paper, and approved the final draft. • Jianshuang Wu analyzed the data, authored or reviewed drafts of the paper, and approved the final draft. Funding This research received joint support from the Ministry of Science and Technology of China (2016YFC0502001), Chinese Academy of Sciences (XDB03030401), and the Strategic Priority Research Program of the Chinese Academy of Sciences (XDA19050502). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. CONCLUSIONS Climate change is affecting the vulnerability of alpine grasslands on the Qinghai-Tibetan Plateau. In this study, we quantified the vulnerability of grasslands with the combination of the sensitivity, resilience, and exposure indices according to the definition by the IPCC. The vulnerable grasslands were mainly distributed on the sides of the Qinghai-Tibetan Plateau. Exposure was the dominant index for this vulnerability pattern; however, the sensitivity and resilience of alpine grasslands could also exacerbate or alleviate the degree of vulnerability. Although human activities were not considered in calculating vulnerability, the grazing intensity had a significant impact on the spatial pattern of grassland vulnerability. We Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 14/20 therefore suggest that anthropogenic factors should be considered in the assessment of ecosystem vulnerability in the future. Data Availability The following information was supplied regarding data availability: The raw measurements are available in a Supplemental File. The following information was supplied regarding data availability: The raw measurements are available in a Supplemental File. Supplemental Information Supplemental information for this article can be found online at http://dx.doi.org/10.7717/ peerj.8513#supplemental-information. Supplemental information for this article can be found online at http://dx.doi.org/10.7717/ peerj.8513#supplemental-information. Grant Disclosures The following grant information was disclosed by the authors: The following grant information was disclosed by the authors: Ministry of Science and Technology of China: 2016YFC0502001. Ministry of Science and Technology of China: 2016YFC0502001. Chinese Academy of Sciences: XDB03030401. Strategic Priority Research Program of the Chinese Academy of Sciences: XDA19050502. Strategic Priority Research Program of the Chinese Academy of Sciences: XDA19050502. REFERENCES Beaugrand G, Edwards M, Raybaud V, Goberville E, Kirby RR. 2015. Future vulnera- bility of marine biodiversity compared with contemporary and past changes. Nature Climate Change 5:695–701 DOI 10.1038/nclimate2650. Beaugrand G, Edwards M, Raybaud V, Goberville E, Kirby RR. 2015. Future vulnera- bility of marine biodiversity compared with contemporary and past changes. Nature Climate Change 5:695–701 DOI 10.1038/nclimate2650. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 Beisner BE, Haydon DT, Cuddington K. 2003. Alternative stable states in ecology. Frontiers in Ecology and the Environment 1:376–382 DOI 10.1890/1540-9295(2003)001[0376:assie]2.0.co;2. Burrows MT, Schoeman DS, Buckley LB, Moore P, Poloczanska ES, Brander KM, Brown C, Bruno JF, Duarte CM, Halpern BS, Holding J, Kappel CV, Kiessling W, O’Connor MI, Pandolfi JM, Parmesan C, Schwing FB, Sydeman WJ, Richardson AJ. 2011. The pace of shifting climate in marine and terrestrial ecosystems. Science 334:652–655 DOI 10.1126/science.1210288. Chen BX, Zhang XZ, Tao J, Wu JS, Wang JS, Shi PL, Zhang YJ, Yu CQ. 2014. The impact of climate change and anthropogenic activities on alpine grassland over the Qinghai-Tibet Plateau. Agricultural and Forest Meteorology 189:11–18 DOI 10.1016/j.agrformet.2014.01.002. Chen H, Zhu Q, Peng C, Wu N, Wang Y, Fang X, Gao Y, Zhu D, Yang G, Tian J, Kang X, Piao S, Ouyang H, Xiang W, Luo Z, Jiang H, Song X, Zhang Y, Yu G, Zhao X, Gong P, Yao T, Wu J. 2013. The impacts of climate change and human activities on biogeochemical cycles on the Qinghai-Tibetan Plateau. Global Change Biology 19:2940–2955 DOI 10.1111/gcb.12277. Chen IC, Hill JK, Ohlemuller R, Roy DB, Thomas CD. 2011. Rapid range shifts of species associated with high levels of climate warming. Science 333:1024–1026 DOI 10.1126/science.1206432. Chen J, Jönsson P, Tamura M, Gu Z, Matsushita B, Eklundh L. 2004. A simple method for reconstructing a high-quality NDVI time-series data set based on the Savitzky–Golay filter. Remote Sensing of Environment 91:332–344 DOI 10.1016/j.rse.2004.03.014. Christensen L, Coughenour MB, Ellis JE, Chen ZZ. 2004. Vulnerability of the Asian typical steppe to grazing and climate change. Climatic Change 63:351–368 DOI 10.1023/B:CLIM.0000018513.60904.fe. Crossman ND, Bryan BA, Summers DM. 2012. Identifying priority areas for reducing species vulnerability to climate change. Diversity and Distributions 18:60–72 DOI 10.1111/j.1472-4642.2011.00851.x. De Keersmaecker W, Lhermitte S, Tits L, Honnay O, Somers B, Coppin P. 2015. A model quantifying global vegetation resistance and resilience to short-term climate anomalies and their relationship with vegetation cover. Global Ecology and Biogeography 24:539–548 DOI 10.1111/geb.12279. REFERENCES Fan JW, Shao QQ, Liu JY, Wang JB, Harris W, Chen ZQ, Zhong HP, Xu XL, Liu RG. 2010. Assessment of effects of climate change and grazing activity on grassland yield in the Three Rivers Headwaters Region of Qinghai-Tibet Plateau, China. Environ- mental Monitoring and Assessment 170:571–584 DOI 10.1007/s10661-009-1258-1. Fu G, Shen ZX, Zhang XZ. 2018. Increased precipitation has stronger effects on plant production of an alpine meadow than does experimental warming in the Northern Tibetan Plateau. Agricultural and Forest Meteorology 249:11–21 DOI 10.1016/j.agrformet.2017.11.017. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 16/20 Geng S, Shi P, Song M, Zong N, Zu J, Zhu W. 2019. Diversity of vegetation composition enhances ecosystem stability along elevational gradients in the Taihang Mountains, China. Ecological Indicators 104:594–603 DOI 10.1016/j.ecolind.2019.05.038. Grime JP. 1973. Competitive exclusion in herbaceous vegetation. Nature 242:344–347 DOI 10.1038/242344a0. Harris RB. 2010. Rangeland degradation on the Qinghai-Tibetan plateau: a review of the evidence of its magnitude and causes. Journal of Arid Environments 74:1–12 DOI 10.1016/j.jaridenv.2009.06.014. Holling CS. 1996. Engineering resilience versus ecological resilience. Washington, D.C.: Island Press. Hoover DL, Knapp AK, Smith MD. 2014. Resistance and resilience of a grassland ecosystem to climate extremes. Ecology 95:2646–2656 DOI 10.1890/13-2186.1. Huang K, Zhang Y, Zhu J, Liu Y, Zu J, Zhang J. 2016. The influences of climate change and human activities on vegetation dynamics in the Qinghai-Tibet Plateau. Remote Sensing 8:Article 876 DOI 10.3390/rs8100876. Hutchinson M. 2004. Anusplin Version 4.3. Centre for Resource and Environmental Studies. Canberra: The Australian National University. IPCC. 2001. Climate Change 2001: impacts, adaptation, and vulnerability. Cambridge: Cambridge University Press. Izaurralde R, Thomson A, Morga JA, Fay PA, Polley H, Hatfield J. 2011. Climate impacts on agriculture: implications for forage and rangeland production. Agronomy Journal 103:371–381 DOI 10.2134/agronj2010.0304. Jiang L, Huang X, Wang F, Liu Y, An P. 2018. Method for evaluating ecological vul- nerability under climate change based on remote sensing: a case study. Ecological Indicators 85:479–486 DOI 10.1016/j.ecolind.2017.10.044. Kharin VV, Zwiers FW, Zhang X, Wehner M. 2013. Changes in temperature and precipitation extremes in the CMIP5 ensemble. Climatic Change 119:345–357 DOI 10.1007/s10584-013-0705-8. Knapp AK, Fay PA, Blair JM, Collins SL, Smith MD, Carlisle JD, Harper CW, Danner BT, Lett MS, McCarron JK. 2002. Rainfall variability, carbon cy- cling, and plant species diversity in a mesic grassland. Science 298:2202–2205 DOI 10.1126/science.1076347. Kozlov MV. 2008. REFERENCES Losses of birch foliage due to insect herbivory along geographical gradients in Europe: a climate-driven pattern? Climatic Change 87:107–117 DOI 10.1007/s10584-007-9348-y. Kröel-Dulay G, Ransijn J, Schmidt IK, Beier C, De Angelis P, De Dato G, Dukes JS, Emmett B, Estiarte M, Garadnai J, Kongstad J, Kovács-Láng E, Larsen KS, Liberati D, Ogaya R, Riis-Nielsen T, Smith AR, Sowerby A, Tietema A, Penuelas J. 2015. Increased sensitivity to climate change in disturbed ecosystems. Nature Communications 6:6682 DOI 10.1038/ncomms7682. Li B. 1997. The rangeland degradation in north china and its preventive strategy. Scientia Agricutura Sinica 30:1–10 (in Chinese). Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 Li D, Wu S, Liu L, Zhang Y, Li S. 2018. Vulnerability of the global terrestrial ecosystems to climate change. Global Change Biology 24:4095–4106 DOI 10.1111/gcb.14327. Li L, Yang S, Wang ZY, Zhu XD, Tang HY. 2010. Evidence of warming and wetting climate over the Qinghai-Tibet Plateau. Arctic Antarctic and Alpine Research 42:449–457 DOI 10.1657/1938-4246-42.4.449. Li L, Zhang Y, Wu J, Li S, Zhang B, Zu J, Zhang H, Ding M, Paudel B. 2019a. Increasing sensitivity of alpine grasslands to climate variability along an elevational gradi- ent on the Qinghai-Tibet Plateau. Science of The Total Environment 678:21–29 DOI 10.1016/j.scitotenv.2019.04.399. Li M, Wu JS, Song CQ, He YT, Niu B, Fu G, Tarolli P, Tietjen B, Zhang XZ. 2019b. Temporal variability of precipitation and biomass of Alpine Grasslands on the Northern Tibetan Plateau. Remote Sensing 11:Article 360 DOI 10.3390/Rs11030360. Lindner M, Maroschek M, Netherer S, Kremer A, Barbati A, Garcia-Gonzalo J, Seidl R, Delzon S, Corona P, Kolström M, Lexer MJ, Marchetti M. 2010. Climate change impacts, adaptive capacity, and vulnerability of European forest ecosystems. Forest Ecology and Management 259:698–709 DOI 10.1016/j.foreco.2009.09.023. Loarie SR, Duffy PB, Hamilton H, Asner GP, Field CB, Ackerly DD. 2009. The velocity of climate change. Nature 462:1052–1055 DOI 10.1038/nature08649. Long RJ, Apori SO, Castro FB, Ørskov ER. 1999. Feed value of native forages of the Tibetan Plateau of China. Animal Feed Science and Technology 80:101–113 DOI 10.1016/S0377-8401(99)00057-7. Neilson RP, Pitelka LF, Solomon AM, Nathan R, Midgley GF, Fragoso JMV, Lischke H, Thompson K. 2005. Forecasting regional to global plant migration in response to climate change. Bioscience 55:749–759 DOI 10.1641/0006-3568(2005)055[0749:frtgpm]2.0.co;2. Nguyen AK, Liou Y-A, Li M-H, Tran TA. 2016. Zoning eco-environmental vulnerability for environmental management and protection. Ecological Indicators 69:100–117 DOI 10.1016/j.ecolind.2016.03.026. Scheffer M, Carpenter S, Foley JA, Folke C, Walker B. 2001. REFERENCES Catastrophic shifts in ecosystems. Nature 413:591–596 DOI 10.1038/35098000. Seddon AW, Macias-Fauria M, Long PR, Benz D, Willis KJ. 2016. Sensitivity of global terrestrial ecosystems to climate variability. Nature 531:229–232 DOI 10.1038/nature16986. Shang ZH, Gibb MJ, Leiber F, Ismail M, Ding LM, Guo XS, Long RJ. 2014. The sustain- able development of grassland-livestock systems on the Tibetan plateau: problems, strategies and prospects. Rangeland Journal 36:267–296 DOI 10.1071/RJ14008. Shao Q, Cao W, Fan J, Huang L, Xu X. 2017. Effects of an ecological conservation and restoration project in the Three-River Source Region, China. Journal of Geographical Sciences 27:183–204 DOI 10.1007/s11442-017-1371-y. Shen MG, Piao SL, Cong N, Zhang GX, Janssens IA. 2015. Precipitation impacts on vegetation spring phenology on the Tibetan Plateau. Global Change Biology 21:3647–3656 DOI 10.1111/gcb.12961. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 18/20 18/20 Sun J, Qin X, Yang J. 2016. The response of vegetation dynamics of the different alpine grassland types to temperature and precipitation on the Tibetan Plateau. Environ- mental Monitoring and Assessment 188:1–11 DOI 10.1007/s10661-015-5014-4. Tilman D, Lehman C. 2001. Human-caused environmental change: impacts on plant diversity and evolution. Proceedings of the National Academy of Sciences of the United States of America 98:5433–5440 DOI 10.1073/pnas.091093198. Urruty N, Tailliez-Lefebvre D, Huyghe C. 2016. Stability, robustness, vulnerability and resilience of agricultural systems. A review. Agronomy for Sustainable Development 36:Article 15 DOI 10.1007/s13593-015-0347-5. Venter O, Sanderson EW, Magrach A, Allan JR, Beher J, Jones KR, Possingham HP, Laurance WF, Wood P, Fekete BM, Levy MA, Watson JE. 2016. Sixteen years of change in the global terrestrial human footprint and implications for biodiversity conservation. Nature Communications 7:Article 12558 DOI 10.1038/ncomms12558. Wang S, Duan J, Xu G, Wang Y, Zhang Z, Rui Y, Luo C, Xu B, Zhu X, Chang X, Cui X, Niu H, Zhao X, Wang W. 2012. Effects of warming and grazing on soil N availability, species composition, and ANPP in an alpine meadow. Ecology 93:2365–2376 DOI 10.1890/11-1408.1. Watson RT, Noble IR, Bolin B, Ravindranath NH, Verardo DJ, Dokken DJ, Estes JE, Ehlers M, Malingreau JP, Noble IR. 2000. IPCC special report on land use, land-use change, and forestry. Cambridge: Cambride University Press. Williams SE, Shoo LP, Isaac JL, Hoffmann AA, Langham G. 2008. Towards an inte- grated framework for assessing the vulnerability of species to climate change. PLOS Biology 6:2621–2626 DOI 10.1371/journal.pbio.0060325. Woznicki SA, Nejadhashemi AP, Tang Y, Wang L. 2016. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 REFERENCES Large-scale climate change vulnerability assessment of stream health. Ecological Indicators 69:578–594 DOI 10.1016/j.ecolind.2016.04.002. Wu J, Shen Z, Zhang X. 2014. Precipitation and species composition primarily determine the diversity–productivity relationship of alpine grasslands on the Northern Tibetan Plateau. Alpine Botany 124:13–25 DOI 10.1007/s00035-014-0125-z. Xu Y, Shen Z-H, Ying L-X, Ciais P, Liu H-Y, Piao S-L, Wen C, Jiang Y-X. 2016. The exposure, sensitivity and vulnerability of natural vegetation in China to climate thermal variability (1901-2013): an indicator-based approach. Ecological Indicators 63:258–272 DOI 10.1016/j.ecolind.2015.12.023. Yang G, Peng C, Chen H, Dong F, Wu N, Yang Y, Zhang Y, Zhu D, He Y, Shi S, Zeng X, Xi T, Meng Q, Zhu Q. 2017. Qinghai–Tibetan Plateau peatland sustainable utilization under anthropogenic disturbances and climate change. Ecosystem Health and Sustainability 3:e01263 DOI 10.1002/ehs2.1263. Yao T, Thompson LG, Mosbrugger V, Zhang F, Ma Y, Luo T, Xu B, Yang X, Joswiak DR, Wang W, Joswiak ME, Devkota LP, Tayal S, Jilani R, Fayziev R. 2012. Third Pole Environment (TPE). Environmental Development 3:52–64 DOI 10.1016/j.envdev.2012.04.002. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 19/20 You Q, Fraedrich K, Ren G, Pepin N, Kang S. 2013. Variability of temperature in the Tibetan Plateau based on homogenized surface stations and reanalysis data. International Journal of Climatology 33:1337–1347 DOI 10.1002/joc.3512. Zhang T, Zhang Y, Xu M, Zhu J, Wimberly MC, Yu G, Niu S, Xi Y, Zhang X, Wang J. 2015a. Light-intensity grazing improves alpine meadow productivity and adaption to climate change on the Tibetan Plateau. Scientific Reports 5:15949 DOI 10.1038/srep15949. Zhang Y, Gao QZ, Dong SK, Liu SL, Wang XX, Su XK, Li YY, Tang L, Wu XY, Zhao HD. 2015b. Effects of grazing and climate warming on plant diversity, productivity and living state in the alpine rangelands and cultivated grasslands of the Qinghai- Tibetan Plateau. Rangeland Journal 37:57–65 DOI 10.1071/RJ14080. Zhao DS, Wu SH. 2014. Vulnerability of natural ecosystem in China under regional climate scenarios: an analysis based on eco-geographical regions. Journal of Geo- graphical Sciences 24:237–248 DOI 10.1007/s11442-014-1085-3. Zheng D. 1996. The system of physico-geographical regions of the Qinghai-Xizang (Tibet) plateau. Science in China Series D-Earth Sciences 39:410–417 DOI 10.1360/yd1996-39-4-410. Zhu J, Jiang L, Zhang Y. 2016. Relationships between functional diversity and above- ground biomass production in the Northern Tibetan alpine grasslands. Scientific Reports 6:34105 DOI 10.1038/srep34105. Li et al. (2020), PeerJ, DOI 10.7717/peerj.8513 20/20
https://openalex.org/W3186303165	https://portal.findresearcher.sdu.dk/files/185029518/nutrients_13_02551_v2.pdf	English	null	Potential Benefits of Bovine Colostrum in Pediatric Nutrition and Health	Nutrients	2,021	cc-by	34,169	Download date: 24. Oct. 2024 Potential benefits of bovine colostrum in pediatric nutrition and health Sangild, Per Torp; Vonderohe, Caitlin; Melendez Hebib, Valeria; Burrin, Douglas G. Published in: Nutrients DOI: 10.3390/nu13082551 Publication date: 2021 Citation for pulished version (APA): Sangild, P. T., Vonderohe, C., Melendez Hebib, V., & Burrin, D. G. (2021). Potential benefits of bovine colostrum in pediatric nutrition and health. Nutrients, 13(8), Article 2551. https://doi.org/10.3390/nu13082551 Go to publication entry in University of Southern Denmark's Research Portal University of Southern Denmark Potential benefits of bovine colostrum in pediatric nutrition and health Sangild, Per Torp; Vonderohe, Caitlin; Melendez Hebib, Valeria; Burrin, Douglas G. Published in: Nutrients DOI: 10.3390/nu13082551 Publication date: 2021 Document version: Final published version Document license: University of Southern Denmark Citation for pulished version (APA): Sangild, P. T., Vonderohe, C., Melendez Hebib, V., & Burrin, D. G. (2021). Potential benefits of bovine colostrum in pediatric nutrition and health. Nutrients, 13(8), Article 2551. https://doi.org/10.3390/nu13082551 Review Potential Beneﬁts of Bovine Colostrum in Pediatric Nutrition and Health 1 Comparative Pediatrics & Nutrition, University of Copenhagen, DK-1870 Copenhagen, Denmark; pts@sund.ku.dk p 2 Department of Neonatology, Rigshospitalet, DK-1870 Copenhagen, Denmark 3 Department of Pediatrics, Odense University Hospital, DK-5000 Odense, Denmark 4 USDA-ARS Children’s Nutrition Research Center, Pediatrics, Gastroenterology & Nutrition, Baylor College of Medicine, Houston, TX 77030, USA; Caitlin.Vonderohe@bcm.edu (C.V.); Valeria.Melendez@bcm.edu (V.M.H.) * Correspondence: doug.burrin@usda.gov; Tel.: +1-713-798-7049 2 Department of Neonatology, Rigshospitalet, DK-1870 Copenhagen, Denmark 3 Department of Pediatrics, Odense University Hospital, DK-5000 Odense, Denmark 4 USDA-ARS Children’s Nutrition Research Center, Pediatrics, Gastroenterology & Nutrition, Baylor College of Medicine, Houston, TX 77030, USA; Caitlin.Vonderohe@bcm.edu (C.V.); Valeria.Melendez@bcm.edu (V.M.H.) * Correspondence: doug.burrin@usda.gov; Tel.: +1-713-798-7049 p gy, g p , p g , 3 Department of Pediatrics, Odense University Hospital, DK-5000 Odense, Denmark Abstract: Bovine colostrum (BC), the ﬁrst milk produced from cows after parturition, is increasingly used as a nutritional supplement to promote gut function and health in other species, including humans. The high levels of whey and casein proteins, immunoglobulins (Igs), and other milk bioactives in BC are adapted to meet the needs of newborn calves. However, BC supplementation may improve health outcomes across other species, especially when immune and gut functions are immature in early life. We provide a review of BC composition and its effects in infants and children in health and selected diseases (diarrhea, infection, growth-failure, preterm birth, necrotizing enterocolitis (NEC), short-bowel syndrome, and mucositis). Human trials and animal studies (mainly in piglets) are reviewed to assess the scientiﬁc evidence of whether BC is a safe and effective antimicrobial and immunomodulatory nutritional supplement that reduces clinical complications related to preterm birth, infections, and gut disorders. Studies in infants and animals suggest that BC should be supplemented at an optimal age, time, and level to be both safe and effective. Exclusive BC feeding is not recommended for infants because of nutritional imbalances relative to human milk. On the other hand, adverse effects, including allergies and intolerance, appear unlikely when BC is provided as a supplement within normal nutrition guidelines for infants and children. Larger clinical trials in infant populations are needed to provide more evidence of health beneﬁts when patients are supplemented with BC in addition to human milk or formula. Igs and other bioactive factors in BC may work in synergy, making it critical to preserve bioactivity with gentle processing and pasteurization methods. Citation: Sangild, P.T.; Vonderohe, C.; Melendez Hebib, V.; Burrin, D.G. Potential Beneﬁts of Bovine Colostrum in Pediatric Nutrition and Health. Nutrients 2021, 13, 2551. https://doi.org/10.3390/nu13082551 Academic Editors: Raymond Playford and Isabelle Luron Keywords: preterm infants; human milk; immunoglobulins; necrotizing enterocolitis; diarrhea Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. Terms of use Terms of use This work is brought to you by the University of Southern Denmark. Unless otherwise specified it has been shared according to the terms for self-archiving. If no other license is stated, these terms apply: This work is brought to you by the University of Southern Denmark. Unless otherwise specified it has been shared according to the terms for self-archiving. If no other license is stated these terms apply: • You may download this work for personal use only. y p y • You may not further distribute the material or use it for any profit-making activity or commercial Y f l di t ib t th URL id tif i thi i y p y You may not further distribute the material or use it fo y p y • You may not further distribute the material or use it for any profit-maki at this document breaches copyright please contact us providing details and we will investigate your claim. enquiries to puresupport@bib.sdu.dk If you believe that this document breaches copyright please contact us providing details and we will investigate your claim. Please direct all enquiries to puresupport@bib.sdu.dk If you believe that this document breaches copyright please contact us providing details and we will investigate your claim. Please direct all enquiries to puresupport@bib.sdu.dk Download date: 24. Oct. 2024 nutrients nutrients nutrients nutrients Review Potential Beneﬁts of Bovine Colostrum in Pediatric Nutrition and Health BC has the potential to become a safe and effective nutritional supplement for several pediatric subpopulations. Citation: Sangild, P.T.; Vonderohe, C.; Melendez Hebib, V.; Burrin, D.G. Potential Beneﬁts of Bovine Colostrum in Pediatric Nutrition and Health. Nutrients 2021, 13, 2551. https://doi.org/10.3390/nu13082551 Academic Editors: Raymond Playford and Isabelle Luron Received: 20 May 2021 Accepted: 13 July 2021 Published: 26 July 2021 1. Introduction Colostrum functions as a nutritional, immunological, and antimicrobial ‘bridge’ between the mostly sterile fetal life, sustained by the maternal umbilical nutrient supply, and the microbe-rich postnatal envi- ronment and enteral breast milk intake. Humans and other mammals often share the same external environment and are exposed to similar microbes, viruses, and fungi along the outer surfaces of the body (skin, lungs, and gut). Consequently, the antimicrobial activity of colostrum from one species should also be (at least partly) effective in another species. The ‘perinatal colostrum bridge’ is critical for survival and health for normal newborns in some mammals (e.g., large farm animals with lacking passive immunity transfer before birth), while, in others, the colostrum is important, although not essential for survival (e.g., term human infants). ) There are similarities in the composition of colostrum and milk among human and bovine species, but there are also some important differences in both nutrients (carbohy- drate, protein, lipid, minerals, and vitamins) and immunological factors, including Igs. Cow’s milk has been used as an important source of supplementary nutrition for infants for centuries in most countries around the world [6]. Beyond the ﬁrst year of life, other cow’s milk-based foods, such as milk, yogurt, and cheese, are also considered important sources of nutrition in children [7,8]. The potential functional and immunological beneﬁts of dairy products in infants and children, besides nutrition alone, are less clear. In the past decades, the accumulated evidence of clinical beneﬁts of human milk, either mother’s own breast milk or banked donor milk, have made human milk the recommended choice for infant nutrition, especially for preterm infants [9–11]. Evidence from preterm infants has raised serious concerns that processed formula products based on cow’s milk lead to more NEC, sepsis (late-onset sepsis, LOS), food intolerance (FI), allergies, and food-protein-induced enterocolitis syndrome (FPIES) in infants fed formula alone, or in combination with human milk [11–16]. This has led some clinicians to warn against use of bovine-milk products for all sensitive hospitalized infants [17]. It is unknown if the apparent adverse effects of formula products, relative to mother’s own milk, relate to their bovine origin, the indus- trial processing steps (i.e., serial heat treatment and ﬁltration steps) and/or addition of vegetable products (i.e., corn-based maltodextrin and vegetable oils), as part of commercial formula production. 1. Introduction Bovine colostrum (BC) is produced by cows in the ﬁrst days after parturition and provides nutrition and immunological protection of highly sensitive newborn calves [1]. The difference between colostrum and milk results from a partially open blood–milk barrier in the mammary gland around birth [2]. Besides the close link to the mother in which colostrum is produced, many elements of colostrum may have cross-species effects and could be used to support and protect newborns and growing offspring of other species when they are lacking their own mother’s colostrum and/or milk. Here we review the scientiﬁc literature related to composition and biological function of BC in pediatric nutrition and gastroenterology, and how this is supported by biomedical animal models of infants and children, most notably the piglet. The topic is timely because there is currently public and scientiﬁc debate about the possible risks and beneﬁts of bovine-based milk Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). https://www.mdpi.com/journal/nutrients Nutrients 2021, 13, 2551. https://doi.org/10.3390/nu13082551 Nutrients 2021, 13, 2551 2 of 41 products for infants and children when human milk is absent or inadequate, particularly for preterm infants and other highly sensitive hospitalized pediatric patients. products for infants and children when human milk is absent or inadequate, particularly for preterm infants and other highly sensitive hospitalized pediatric patients. p g y p p p As highlighted in the introductory article of this review series [3], the provision of mammary secretions in the form of colostrum and mature milk for newborns is an es- sential survival function that has evolved over millennia in mammal species. In humans, following the production of colostrum immediately after parturition, the subsequent two weeks postpartum is a time of secretory activation and increased volume secretion of what we eventually consider mature milk [4]. Colostrum serves a vital function for neonates by providing nutrients necessary for energy, growth, and development as they transition from placental supply of elemental nutrients, such as glucose and amino acids, before birth, to oral ingestion of more complex macromolecules, including lactose, proteins, and triglycerides. A second critical function of colostrum and milk after birth is to provide immune protection that supports elements of neonatal innate immunity until maturation of adaptive immunity related to the speciﬁc environment [5]. 2. Composition and Function of Colostrum in Humans and Animals 2.1. Macronutrients Nutritionally, bovine and human colostrum are similar, but they differ mainly in the relative concentration of speciﬁc macronutrients. A key feature of BC is its high protein content (~15%) relative to fat (4–6%) and lactose (3 to 4%) [22,23]. The protein and fat contents are higher than lactose in colostrum, but with advancing lactation, the relative content of protein and fat declines and lactose increases. Human and bovine early milk or colostrum contain relatively low amounts of lactose (1.2% and 2–2.9%, respectively), which increase with lactation to 7.0% and 4.8, respectively [24]. The increasing lactose content as milk and the offspring mature suggests an immunological and trophic primary role for colostrum rather than nutritional [4]. The production of lactose increases the movement of water into the secretory vesicles of the mammary epithelium; therefore, low levels of lactose in milk result in increased viscosity. The carbohydrate content of colostrum and milk also comprises oligosaccharides, and these are discussed in more detail in a later section. The protein content of BC and milk is higher than human colostrum and milk (Figure 1), and there are select differences in amino acid content [4,25–28]. Proteins in human and bovine milk are divided into whey and casein fractions, which comprise differing percentages of the total protein content in mammalian milk across species. In bovine milk, the casein and whey fraction comprise 80% and 20% of the total protein, respectively [29], whereas, in human milk, the ratio is whey predominant (40:60) [30]. However, the proteins within these fractions possess signiﬁcant homology among species. Among the most widely studied bioactive proteins in colostrum are Igs, lactoferrin (LF), lysozyme, α-lactalbumin, and growth factors. Several casein proteins have also been extensively studied mostly for their role in transporting calcium phosphate and promoting the bioavailability of other milk proteins [31]. p The nutritional value of milk proteins in colostrum is a function of the amino acid composition and digestion kinetics through the stomach and upper intestine [32]. Casein proteins form a clot in the stomach that slows release of amino acids for digestion and absorption into the circulation. In contrast, whey proteins are more soluble, in the stomach and empty rapidly into the small intestine for digestion and absorption. These differences between casein and whey protein digestion kinetics have been characterized as the slow vs. 1. Introduction Whether risk factors for infants differ between term and preterm infants, between bovine products and products from other mammals (donkeys, camels, and goats), and between milk and colostrum is unknown. The possible risks of feeding infants BC are discussed later in this review. Powdered products containing intact BC, or fractions thereof, have become increas- ingly available as health foods and their use as nutritional supplements to support gut health for children and adolescents is increasing. Additionally, their utility for newborn infants with limited access to mother’s own milk is currently being investigated. There is considerable speculation about beneﬁts in the marketing of BC products and our goal is to provide a review of the scientiﬁc evidence that is used to support these claims. This review highlights the use BC in pediatrics and neonatology, including its use for both as a Nutrients 2021, 13, 2551 3 of 41 3 of 41 nutritional and as a preventive or therapeutic supplement for severe pediatric diseases, such as diarrhea, NEC, inﬂammatory bowel disease (IBD), short-bowel syndrome, and chemotherapy-induced mucositis. Both clinical studies and animal model studies are reviewed and discussed. For detailed insights into BC constituents and applications in adults, readers are referred to other reviews [18–21] and the companion articles in the present review series in Nutrients [3], covering a wide range of possible applications of BC for humans, including gastrointestinal (GI) diseases, immune dysfunctions, and sports medicine. 2. Composition and Function of Colostrum in Humans and Animals 2.1. Macronutrients Human milk and colostrum is rich in α-lactalbumin, LF, and Figure 1. Composition of protein and oligosaccharides in human and bovine colostrum and milk. The ﬁgure illustrates the marked differences in total protein content and speciﬁc proteins, such as α-lactalbumin, β-lactoglobulin, and casein, between human and bovine colostrum and milk. gure 1 Composition of protein and oligosaccharides in human and bovine colostrum and milk The figure illustrates Figure 1. Composition of protein and oligosaccharides in human and bovine colostrum and milk. gure 1. Composition of protein and oligosaccharides in human and bovine colostrum and milk. The figure illustrates k d diff i l i d ifi i h l lb i β l l b li d i Figure 1. Composition of protein and oligosaccharides in human and bovine colostrum and milk. The ﬁgure illustrates the marked differences in total protein content and speciﬁc proteins such ure 1. Composition of protein and oligosaccharides in human and bovine colostrum and milk. The figure illustrates marked differences in total protein content and specific proteins, such as α-lactalbumin, β-lactoglobulin, and casein, ween human and bovine colostrum and milk. Human milk and colostrum is rich in α-lactalbumin, LF, and gosaccharides, relative to bovine milk and colostrum. Conversely, bovine colostrum and milk is rich in casein, β- toglobulin, and immunoglobulin G, relative to the human counterparts. All components are expressed as g/L. The nutritional value of milk proteins in colostrum is a function of the amino The ﬁgure illustrates the marked differences in total protein content and speciﬁc proteins, such as α-lactalbumin, β-lactoglobulin, and casein, between human and bovine colostrum and milk. Human milk and colostrum is rich in α-lactalbumin, LF, and oligosaccharides, relative to bovine milk and colostrum. Conversely, bovine colostrum and milk is rich in casein, β-lactoglobulin, and immunoglobulin G, relative to the human counterparts. All components are expressed as g/L. composition and digestion kinetics through the stomach and upper intestine [32]. Ca proteins form a clot in the stomach that slows release of amino acids for digestion absorption into the circulation. In contrast, whey proteins are more soluble, in the stom and empty rapidly into the small intestine for digestion and absorption. These differe between casein and whey protein digestion kinetics have been characterized as the s vs. 2. Composition and Function of Colostrum in Humans and Animals 2.1. Macronutrients fast protein metabolism concept [33] and has relevance for gut motility metabolism when bovine milk or colostrum products are used for pediatric pati Susceptibility to gut proteolysis also varies among these proteins, where whey prot have globular structures that are resistant to proteases, increasing their functi bioactivity throughout the GI tract. The susceptibility of these proteins to GI proteas also influenced by age; preterm infants have immature gastric acid production proteolytic digestive capacity. In preterm infants fed breast milk, both of these factors to increased survival of intact immunologically important proteins in the intestinal lum This latter point suggests that, in term, and especially preterm infants, colostral prot function not only as a source of amino acids for growth, but also as a source of immune protection in the gut. Differences in amino acid availability between human Few studies have tested the possible nutritional beneﬁts of intact BC or fractions thereof in sensitive newborn infants or as prevention or therapy of severe pediatric diseases (see later sections). To understand the potential and limitations of using BC for such condi- tions beyond the nutritional value of its constituents, it is relevant to brieﬂy review some examples of the numerous components in BC that could play a speciﬁc role in pediatric patients. These components can broadly be categorized as having nutritional, growth- stimulating, antimicrobial, and/or antimicrobial effects (Figure 2). A comprehensive review of all the possible bioactive components of BC is presented in the introductory article of this series [3], supplemented with many previous reviews on bioactive constituents in bovine and humane milk with potential effects in infants [36–38]. Here, we review ﬁve classes of bioactive components relevant to BC in the context of pediatric use: (1) Igs, (2) LF, lysozyme, and α-lactalbumin, (3) milk fat globule membrane proteins, (4) oligosaccharides, and (5) microRNA and stem cells (Figure 2). 2. Composition and Function of Colostrum in Humans and Animals 2.1. Macronutrients fast protein metabolism concept [33] and has relevance for gut motility and metabolism when bovine milk or colostrum products are used for pediatric patients. Susceptibility to gut proteolysis also varies among these proteins, where whey proteins have globular structures that are resistant to proteases, increasing their functional bioactivity throughout the GI tract. The susceptibility of these proteins to GI proteases is also inﬂuenced by age; preterm infants have immature gastric acid production and proteolytic digestive capacity. In preterm infants fed breast milk, both of these factors lead to increased survival of intact immunologically important proteins in the intestinal lumen. This latter point suggests that, in term, and especially preterm infants, colostral proteins function not only as a source of amino acids for growth, but also as a source of vital immune protection in the gut. Differences in amino acid availability between human milk and bovine-milk-based infant formulas may induce different plasma amino acid proﬁles, but also processing effects may affect digestion kinetics, as shown in vitro or in vivo piglet studies [34,35]. 4 of 41 4 Nutrients 2021, 13, 2551 Figure 1. Composition of protein and oligosaccharides in human and bovine colostrum and milk. The figure illustrate the marked differences in total protein content and specific proteins, such as α-lactalbumin, β-lactoglobulin, and casein between human and bovine colostrum and milk. Human milk and colostrum is rich in α-lactalbumin, LF, and oligosaccharides, relative to bovine milk and colostrum. Conversely, bovine colostrum and milk is rich in casein, β actoglobulin, and immunoglobulin G, relative to the human counterparts. All components are expressed as g/L. The nutritional value of milk proteins in colostrum is a function of the amino Figure 1. Composition of protein and oligosaccharides in human and bovine colostrum and milk. The ﬁgure illustrates the marked differences in total protein content and speciﬁc proteins, such as α-lactalbumin, β-lactoglobulin, and casein, between human and bovine colostrum and milk. Human milk and colostrum is rich in α-lactalbumin, LF, and oligosaccharides, relative to bovine milk and colostrum. Conversely, bovine colostrum and milk is rich in casein, β-lactoglobulin, and immunoglobulin G, relative to the human counterparts. All components are expressed as g/L. gure 1. Composition of protein and oligosaccharides in human and bovine colostrum and milk. The figure illustrates e marked differences in total protein content and specific proteins, such as α-lactalbumin, β-lactoglobulin, and casein, tween human and bovine colostrum and milk. and but 2.2. Immunoglobulins Here, we review five classes of bioactive components relevant to BC in the context of ediatric use: (1) Igs, (2) LF, lysozyme, and α-lactalbumin, (3) milk fat globule membrane roteins, (4) oligosaccharides, and (5) microRNA and stem cells (Figure 2). Figure 2. Nutritional and bioactive components present in bovine colostrum. The figure shows some of the key biological functions of bovine colostrum components related to their partially overlapping nutritional, immunomodulatory, antimicrobial, and cell-growth functions. (Created with BioRender.com; accessed on 7 June 2021). Figure 2. Nutritional and bioactive components present in bovine colostrum. The ﬁgure shows some of the key biolog- ical functions of bovine colostrum components related to their partially overlapping nutritional, immunomodulatory, antimicrobial, and cell-growth functions. (Created with BioRender.com; accessed on 7 June 2021). gure 2. Nutritional and bioactive components present in bovine colostrum. The figure shows some of the key biological unctions of bovine colostrum components related to their partially overlapping nutritional, immunomodulatory, ntimicrobial and cell-growth functions (Created with BioRender com; accessed on 7 June 2021) Figure 2. Nutritional and bioactive components present in bovine colostrum. The ﬁgure shows some of the key biolog- ical functions of bovine colostrum components related to their partially overlapping nutritional, immunomodulatory, antimicrobial, and cell-growth functions. (Created with BioRender.com; accessed on 7 June 2021). igure 2. Nutritional and bioactive components present in bovine colostrum. The figure shows some of the key biological unctions of bovine colostrum components related to their partially overlapping nutritional, immunomodulatory, ntimicrobial and cell-growth functions (Created with BioRender com; accessed on 7 June 2021) Figure 2. Nutritional and bioactive components present in bovine colostrum. The ﬁgure shows some of the key biolog- ical functions of bovine colostrum components related to their partially overlapping nutritional, immunomodulatory, antimicrobial, and cell-growth functions. (Created with BioRender.com; accessed on 7 June 2021). .2. Immunoglobulins The composition of immune components, such as Igs, LF, and growth factors, is ignificantly higher in colostrum of most species than in mature milk [4,19,22,39]. Most of hese components are well conserved in bovine, porcine, and human milk. Colostral Igs i id i i i hi h i l i i i b d The species-speciﬁc necessity for survival and immune protection of newborn mam- mals depends on whether there is prenatal transfer of maternal Igs (mainly immunoglobu- lin G, IgG) across the placenta, and/or postnatal transfer across the mucosa of the small intestine (Figure 3). and but 2.2. Immunoglobulins p g y g p studies [34,35]. Few studies have tested the possible nutritional benefits of intact BC or fract thereof in sensitive newborn infants or as prevention or therapy of severe pedi The composition of immune components, such as Igs, LF, and growth factors, is sig- niﬁcantly higher in colostrum of most species than in mature milk [4,19,22,39]. Most of these components are well conserved in bovine, porcine, and human milk. Colostral Igs function to provide immune protection within the gut against colonizing microbes and environmental toxins. In many species, they are also the primary transfer of passive immu- nity. Igs represent a major fraction of colostral protein and the main immune components, including the isotypes IgG, IgA, IgM, and IgD and their subclasses (IgG1–4 and IgA1–2). There are important differences in the relative composition of these isotypes in human, bovine, and porcine colostrum. Human colostrum is IgA dominant, whereas bovine and porcine colostrum are IgG dominant. IgA occurs as a monomer or dimer, with the latter comprising two IgA molecules joined together by a J-chain and a secretory component. 5 of 41 for 5 of 41 for Nutrients 2021, 13, 2551 This complex is called secretory IgA (sIgA). In humans, sIgA represents 90% of total im- munoglobulin compared with BC, where IgA is only 10% of total Ig. The glycosylation of both the Fc and Fab regions of sIgA play a key role in its protein structure, stability in the gut lumen, and especially its capacity to bind bacterial and host mucosal epitopes [40]. BC and milk contain IgG1 and IgG2, where IgG1 is the main isotype, comprising 70–80% of the immunoglobulin fraction in colostrum. IgG is secreted as a monomer composed of four peptide chains. The concentration of IgG in BC typically ranges from 50 to 100 mg/mL, whereas, in bovine serum, the proportions of IgG1 and IgG2 are relatively equal, and at lower total concentration (~20 mg/mL). n pediatric patients. These components can broadly be categorized as having nutritional, rowth-stimulating, antimicrobial, and/or antimicrobial effects (Figure 2). A omprehensive review of all the possible bioactive components of BC is presented in the ntroductory article of this series [3], supplemented with many previous reviews on ioactive constituents in bovine and humane milk with potential effects in infants [36–38]. and but 2.2. Immunoglobulins Moreover studies in preterm infants also show increased serum and urinar This may explain why sIgA is the dominant immunoglobulin in colostrum in humans, compared with IgG in domestic animal species. However, in infants born premature, there is incomplete placental IgG transfer and passive immunity that are proportional to the degree of prematurity [45]. Moreover, studies in preterm infants also show increased serum and urinar concentrations of milk proteins, such as sIgA, LF, and α-lactalbumin, compared with term and formula-fed infants [46–48]. This suggests that the immature intestine in preterm infants may be more permeable to absorption of macromolecules, such as Igs, in colostrum and mature milk. On the other hand, this increase in intestinal permeability due t immature structural integrity is biologically very different from the highly specialize Moreover, studies in preterm infants also show increased serum and urinary concen- trations of milk proteins, such as sIgA, LF, and α-lactalbumin, compared with term and formula-fed infants [46–48]. This suggests that the immature intestine in preterm infants may be more permeable to absorption of macromolecules, such as Igs, in colostrum and mature milk. On the other hand, this increase in intestinal permeability due to immature structural integrity is biologically very different from the highly specialized function of enterocytes in some newborn animals to absorb large molecules by endocytosis with or without involvement of speciﬁc (Fc) receptors. p p In contrast to infants, large domestic animals (e.g., cattle, pigs, horses, and sheep) lack maternal–placental transfer of Igs and depend almost exclusively on postnatal transfer of maternal IgG via colostrum intake and macromolecule uptake in the newborn gut [49–52]. Once absorbed into the blood, Igs have relatively slow turnover, and they have a half-life of about 2 weeks [42]. Regardless of species, IgG and sIgA in colostrum serve as a vital ﬁrst line of defense to neutralize and kill pathogenic microbes in the gut lumen and modulate mucosal immune function to limit inﬂammation in newborns [19,51]. The capacity for intestinal absorption of colostral immunoglobulin to provide systemic immunity is well established in pigs and calves (see Figure 3). This process occurs via a non-speciﬁc endocytotic process, together with active transport via Fc receptors in the mucosal epithelium [53,54]. Importantly, these studies demonstrate the capacity of porcine Fc receptors to transport bovine IgG in cell culture and in vivo. and but 2.2. Immunoglobulins In humans, passive transfer of immunoglobulin occurs largely through prenatal IgG1 transfer, mainly in the last trimester of pregnancy 6 of 41 ulin occur y Nutrients 2021, 13, 2551 Figure 3. Comparison of maternal immunoglobulin transfer to newborn infants and calves/piglets. The figure illustrates the species differences in the time and mode of maternal immunoglobulin isotype (IgG and IgA) transfer via fetal placental transfer in utero and/or postnatal consumption of colostrum and milk after birth [28,41–44]. The shaded box shows how premature birth may lead to incomplete maternal fetal–placental transfer of IgG in human infants (created with BioRender.com; accessed on 19 May 2021) [43]. Figure 3. Comparison of maternal immunoglobulin transfer to newborn infants and calves/piglets. The ﬁgure illustrates the species differences in the time and mode of maternal immunoglobulin isotype (IgG and IgA) transfer via fetal placental transfer in utero and/or postnatal consumption of colostrum and milk after birth [28,41–44]. The shaded box shows how premature birth may lead to incomplete maternal fetal–placental transfer of IgG in human infants (created with BioRender.com; accessed on 19 May 2021) [43]. Figure 3. Comparison of maternal immunoglobulin transfer to newborn infants and calves/piglets. The figure illustrates the species differences in the time and mode of maternal immunoglobulin isotype (IgG and IgA) transfer via fetal placental transfer in utero and/or postnatal consumption of colostrum and milk after birth [28,41–44]. The shaded box shows how premature birth may lead to incomplete maternal fetal–placental transfer of IgG in human infants (created with BioRender.com; accessed on 19 May 2021) [43]. Figure 3. Comparison of maternal immunoglobulin transfer to newborn infants and calves/piglets. The ﬁgure illustrates the species differences in the time and mode of maternal immunoglobulin isotype (IgG and IgA) transfer via fetal placental transfer in utero and/or postnatal consumption of colostrum and milk after birth [28,41–44]. The shaded box shows how premature birth may lead to incomplete maternal fetal–placental transfer of IgG in human infants (created with BioRender.com; accessed on 19 May 2021) [43]. This may explain why sIgA is the dominant immunoglobulin in colostrum i humans, compared with IgG in domestic animal species. However, in infants bor premature, there is incomplete placental IgG transfer and passive immunity that ar proportional to the degree of prematurity [45]. and but 2.2. Immunoglobulins This conﬁrms previous studies showing absorption of bovine IgG in neonatal piglets, albeit at lower rates than porcine IgG [55]. The Fc receptor is expressed in human fetal and neonatal intestine and has been shown to be important for bidirectional transport of IgG across the intestinal epithelium to neighboring dendritic cells for coordination of immune responses to luminal bacteria [56–58]. An important consideration regarding the use of BC in humans is whether the IgG present in BC is speciﬁc for microbial pathogens present in humans. Considerable evi- dence shows that bovine IgG can bind to a wide range of pathogenic bacteria, viruses 7 of 41 Nutrients 2021, 13, 2551 7 of 41 and allergens found in humans [19]. Furthermore, the speciﬁcity of bovine and human sIgA isolated from milk is shown to be similar for various pathogenic and commensal bacteria [59]. Thus, the antimicrobial actions of BC products are not restricted to pathogens only present in calves but would likely have effects across a range of microbes in different mammalian species. The situation may be different when it comes to BC interactions with host cells, including the various immune cells and the cells transporting IgG. The fact that BC is able to provide immune protection in piglets, both locally in the gut and systemically via absorbed IgG, suggests that basic functions and transport processes are not species-speciﬁc [51,60]. Bovine IgG can also bind to the human Fc receptor with higher afﬁnity than human IgG [61]. The speciﬁcity of BC IgG for pathogens mainly present in humans can be optimized by pathogen-speciﬁc vaccination of cows prior to collection of milk during lactation to produce ‘hyper-immune colostrum’ [20,62], as discussed further below. The applicability of bovine IgG as a medical supplement has also been developed by using serum-derived bovine immunoglobulin (SBI) that is enriched with >50% IgG. These SBI preparations have been shown to have anti-inﬂammatory actions on human intestinal epithelial and monocyte cell lines [63]. A few randomized controlled trials in patients with human-immunodeﬁciency-virus enteropathy and irritable-bowel-syndrome diarrhea, including children, given oral SBI supplements, reported no serious adverse events and modest improvements in GI symptoms [64–68]. In conclusion, the antimicro- bial and immunological effects of bovine colostral IgG and IgA are likely to act in both species-speciﬁc and species-unspeciﬁc ways, depending on the microbe in question and speciﬁc host-cell function. 2.3. Lactoferrin, Lysozyme, and α-Lactalbumin Additionally, LF’s ability to interact with lipopolysaccharides (LPS) on Gram-negative bacterial cell membrane and compete with LPS for the binding of CD14, a co-receptor for toll-like Nutrients 2021, 13, 2551 8 of 41 8 of 41 receptor 4 signaling [78], suggest that it can potentially regulate the host immune response to colonizing microbes. This is particularly interesting given that the pathogenesis of NEC is associated with the premature infant’s response to Gram-negative bacteria through TLR-4 [79]. Studies attempting to conﬁrm the effect of LF on host immune responses to commensal bacteria revealed that splenic and mesenteric lymph-node-derived cells from piglets fed bLF presented an anti-inﬂammatory cytokine proﬁle before and after ex vivo stimulation with LPS [80]. However, studies in preterm pigs also suggested that the NEC-preventive effects of bLF might be dose-dependent and that high doses of bLF could negatively affect immature epithelial cells via metabolic, apoptotic and inﬂammatory pathways [81,82]. Collectively, the ﬁndings suggest that bLF may function to prevent aberrant inﬂammatory responses in the intestinal epithelium of newborns. y p p Lysozyme, although known to be widely distributed in bodily ﬂuids and present in high concentrations in human breast milk (200–400 µg/mL), is present at a signiﬁ- cantly lower concentration in bovine milk (0.05–0.22 µg/mL). Alpha-lactalbumin exists in concentrations ranging from 1.2 to 1.5 mg/mL in bovine milk, has a primary role in the mammary lactose synthesis, and is a source of bioactive peptides and amino acids that support infant growth [83]. Alpha-lactalbumin also is thought to play a role in the development of the infant intestine and brain because of its unique amino acid composition, including tryptophan, lysine, branched-chain amino acids, and sulfur-containing amino acids [84]. Partly replacing the high amounts of β-lactoglobulin in bovine milk with more α-lactalbumin makes the amino acid composition of bovine milk products more similar to that of human milk. However, a recent study in preterm pigs failed to show effects of α-lactalbumin enrichment to a milk diet on growth, gut, immunity, and brain devel- opment [85]. Thus, the speciﬁc role of BC-derived α-lactalbumin in infant development remains to be elucidated, highlighting the need for further research. Among the rest of the whey fraction components in BC and milk, growth factors are important constituents as possible bioactive components (Figure 2). 2.3. Lactoferrin, Lysozyme, and α-Lactalbumin LF is an iron-binding glycoprotein found in many biological secretions but reaches particularly high concentrations in milk. In human colostrum, LF concentration ranges from 5 to 6 mg/mL and decreases to 1 mg/mL in mature milk [69]. In BC, the concentration of LF can range from 1.5 to 5 mg/mL, decreasing to 0.02–0.35 mg/mL in mature cow’s milk [70]. Milk LF has numerous biological functions that range from antioxidant to antitumor and antimicrobial properties. The cellular actions exerted by LF are mediated by the LF receptor (LFR) found in the brush border of the intestinal cell membrane with a sustained abundance in the jejunum throughout the ﬁrst months of life, as seen in the piglet intestine [71]. After binding to the LFR, LF can be translocated into the cell nucleus and regulate gene transcription, resulting in increased cellular proliferation in the intestine [72]. Beyond its implications in intestinal development, LF’s ability to sequester iron, a necessary source of nutrition for commensal and pathogenic bacteria, contributes to its antimicrobial activity. The homology of the human and bovine LF (bLF) amino acid sequences (69% shared amino sequence identity) [73] and the affordability of bLF have made it the most studied LF in both human and animal trials. Several human studies have tested its ability to prevent inﬂammatory diseases in premature infants, such as NEC and LOS [74]. Randomized controlled trials in preterm infants demonstrated that supplementing infant diet daily with 100 mg/kg/body weight (bw) of bLF (n = 472 infants) or 150 mg/kg/bw of recombinant human LF (n = 120 infants) protects against NEC and LOS [75]. In contrast, the largest trial (n = 2203 infants) of LF in preterm infants (150 mg/kg/day) failed to show any protection against either NEC or LOS [74,76]. These studies formed the basis for recommendations from a Cochrane meta-analysis that included 12 RCTs and 5425 participants given LF supplementation added to enteral feeds in preterm infants. This report found low-certainty evidence that LF supplementation of enteral feeds decreases LOS, but not NEC [77]. pp The mechanisms through which bLF may protect against LOS and NEC have been investigated in premature piglets. Using a porcine intestinal epithelial cell line, bLF exerted dose-dependent of anti-inﬂammatory effects on the cells in culture [75]. 2.3. Lactoferrin, Lysozyme, and α-Lactalbumin Many of these growth factors, such as EGF and IGF-1, have been investigated for their isolated roles on gut growth or to prevent disease in infants, children and adults [21,36]. However, these results have been mixed and not been uniformly positive, and only a few of the known bovine milk or colostrum-derived bioactive proteins have been developed for therapeutic use in infants and children. 2.4. Milk Fat Globule Membranes Milk fat globule (MFG) is a lipid droplet containing triacylglycerols (TAG) that buds from the endoplasmic reticulum into the cytoplasm of mammary gland alveolar epithelial cells [86]. These cytoplasmic lipid droplets are secreted from alveolar epithelial cells by fusing with the plasma membrane, acquiring a peripheral bilayer made of lipids and proteins referred to as the milk fat globule membrane (MFGM). This membrane contains a wide variety of bioactive molecules, such as amphipathic lipids, cerebrosides, gangliosides, mucins, lactadherin, butyrophilin, and glycosylated proteins [87], many of which have been found to have antimicrobial, anti-inﬂammatory, and anticarcinogenic activities [88]. The contents of MFGM, are highly inﬂuenced by environmental factors, as well as lactation and gestation period, maternal genetics, body composition, and diet [86]. Importantly, even though 98% of the milk fat is contained within the MFG, the MFGM constitutes only a minor fraction (1–4%) of the protein content [89] and of total fat (0.2–1%) [88] in human milk (HM) and bovine milk (BM). Bovine and human MFGM have signiﬁcant structural and functional homology; both are mainly composed of polar lipids that are present in very similar amounts in both species [86,88]. The major groups of lipids building up the MFGM are phospholipids, mainly glycerophospholipids that include phosphatidylcholine (25.2% in HM and 35–36% in BM of total phospholipids), phosphatidylethanolamine (28% in HM and 27–30% in BM), phosphatidylinositol (4.6% in HM and 5–11% in BM), phosphatidylserine (5.9% in HM and 3% in BM), and sphingolipids, which are mainly represented by sphingomyelin (35.7% in HM and 25% in BM). In the same way, the MFGM proteome is highly homologous, and it is composed mainly of adipophilin, butyrophilin, Nutrients 2021, 13, 2551 9 of 41 9 of 41 mucin 1, xanthine dehydrogenase/oxidase, mucins, lactadherin, and fatty acid-binding protein, some of which are present in higher amounts in bovine MFGM. mucin 1, xanthine dehydrogenase/oxidase, mucins, lactadherin, and fatty acid-binding protein, some of which are present in higher amounts in bovine MFGM. MFGM has been isolated successfully from dairy products, such as BC and cow’s milk, for supplementation in infant formula. The functional role of MFGM has been studied by supplementing milk and formula with MFGM-enriched protein fraction and it is suggested that MFGM protects against infections and modestly impacts the fecal microbiota [90]. Studies performed with premature infants have shown that consumption of sphingomyelin-fortiﬁed milk is beneﬁcial for neurobehavioral development [91]. 2.4. Milk Fat Globule Membranes An MFGM-enriched formula diet may improve lipid absorption, availability of essential fatty acids, and therefore neurodevelopment. However, a recent study in preterm pigs failed to show immediate effects of MFGM on the developing brain [92]. Other sphingolipids, such as gangliosides, have also been reported to play critical roles in neurodevelopment and the implications of their dietary consumption have been thoroughly discussed in other reviews [93,94]. Aside from the neurological implications, the digestion of sphingomyelin to sphingosine-1-phosphate (S1P) has been associated with the improvement of intestinal barrier function in vitro by increasing the production and localization of E-cadherin to cell–cell borders [95]. The bioactivity of the MFGM proteome ranges from the antimicrobial effects of mucin and xanthine oxidase to the anticancer effects of proteins such as fatty acid binding protein (FABP) [96,97]. Mucins and xanthine oxidases can act as decoy receptors for pathogens in the GI tract, therefore preventing direct bacterial interaction with the epithelium. Ad- ditionally, xanthine oxidase’s antimicrobial effects also stem from its ability to generate reactive oxygen and nitrogen species which have bactericidal effects. Other studies have highlighted the beneﬁts of bovine MFGM or MFGM-derived proteins and lipids by demon- strating that MFGM-supplemented formula fed to neonatal rodents improves intestinal growth, Paneth and goblet cell numbers, and tight junction protein patterns to be compara- ble to those seen in rat pups fed with mother’s milk [98]. In formula-fed infants, bovine MFGM supplementation reduced the levels of Moraxella catarrhalis, the leading cause of otitis media in children [99]. Although further studies on the effects of MFGM bioactive components on the health of the human infants are needed, the available literature suggest that BC-derived MFGM may indeed play an important role in infant development. 2.5. Oligosaccharides Beyond their capacity to drive the colonization of ‘beneﬁcial’ bacterial species in the colon, human and bovine colostrum and milk oligosaccharides may act as competitive inhibitors for pathogenic bacteria binding on the mucosal surface of the small intestine and colon, protecting the neonate from infection [102]. Human milk oligosaccharides (HMOs) have demonstrated anti-infective capabilities against a wide variety of pathogens such as H. pylori, M. meningitides, and inﬂuenza virus in various models [113]. Lane et al. showed that incubation of human epithelial cells with oligosaccharides found in BC reduced the internalization and growth of the devastating diarrheal agent C. jejuni [113]. Incubating human colonocytes with the dominant oligosaccharide in BC reduced adhesion of enteropathogenic E. coli by 50% [114]. Bovine and human milk oligosaccharides (HMO and BMO) have additional im- munomodulatory effects on the GI tract itself, beyond their efﬁcacy as prebiotics and pathogen binding agents. When HMO and BMOs are incubated with human epithelial cells, these cells show increased expression of cell surface receptors, chemokines and cytokines, indicating the potential for HMO and BMOs to improve the maturation of the cytokine response in neonates [115]. Ex vivo treatment of cultured peripheral blood mononuclear cells isolated from neonatal piglets with HMOs increased proliferative and anti-inﬂammatory activity, indicating potential direct immunomodulatory effects of HMO and BMO supplementation on the enteric immune system [116]. Studies in neonatal pigs challenged with rotavirus showed that feeding formula supplemented with HMOs or mix- tures of prebiotic oligosaccharides reduced the duration of diarrhea and enhanced T-helper type 1 interferon-gamma and IL-10 cytokines in the ileum [117]. HMO-fed pigs have twice as many natural killer (NK) cells, 36% more mesenteric lymph node effector memory T cells suggesting improved mucosal immune function [118]. Additional work supplementing puriﬁed oligosaccharides to neonatal calves receiving heat-treated colostrum demonstrated that oligosaccharides may also increase the absorption of the Igs, potentially improving health and growth outcomes [119]. g Recent research has investigated whether HMOs contribute to the protective effect of human milk feeding on the incidence of NEC in preterm infants. Rodent models of NEC demonstrated that the protective effects of HMO supplementation are highly HMO structure-speciﬁc [120]. These studies identiﬁed that disialyllacto-n-tetraose (DSLNT) was the most effective HMO in reducing NEC in rats, whereas a structurally similar and highly abundant HMO, α-1,2-fucosyllactose (2′-FL), has limited effects in both rodents and preterm pigs. 2.5. Oligosaccharides The major carbohydrates in bovine and human colostrum are lactose and oligosaccha- rides. Oligosaccharides are composed of more than three monosaccharide [100] units, with a core of lactose or N-acetyl lactosamine, and are classiﬁed as neutral or acidic depending on the presence of a sialic acid residue in the molecule [101,102]. Mammalian milks contain a variety of oligosaccharides that consist of different monosaccharides linked to the core lactose of N-acetyl lactosamine [103–106]. Oligosaccharides are largely undigested in the upper GI tract of the infant and are instead fermented by gut microbes in the distal small intestine and colon [102,107]. Oligosaccharide levels across species are highest in colostrum and decline postpartum [102]. Oligosaccharides are most abundant in humans, ranging from ~7 to 10 g/L, and represent 10% of the total caloric content of mature milk. Human colostrum contains 22–24 g/L oligosaccharides, while BC contains 1 g/L oligosaccharides, and this decreases substantially in the 48 h postpartum [100] (Figure 1). Despite the large differences in oligosaccharide concentrations, bovine and human milk and colostrum oligosaccharides contain similar oligosaccharide structures [108]. It remains unknown if the marked differences in oligosaccharide composition between species (bovine and human) have a speciﬁc role or if different species rely on different milk constituents for similar antimicrobial and/or immunomodulatory effects. Regardless, the much higher content of oligosaccharides in human milk versus bovine have led the formula industry to add select oligosaccharides into formula products in attempts to ‘humanize’ infant formula [27]. 10 of 41 10 of 41 Nutrients 2021, 13, 2551 Milk oligosaccharides have prebiotic activity, anti-adhesion effects, anti-inﬂammatory properties, and glycome-modifying activity, as well as a role in development of brain and intestinal cells [101]. The supplementation of sialic acid, an oligosaccharide found in both human and bovine colostrum, can improve neural development and memory in piglets [109]. Intact oligosaccharides serve as a prebiotic for bacteria, particularly in the colon [110]. Incubation of HT29 cells with an oligosaccharide found in both human and bovine colostrum resulted in increased adhesion of Biﬁdobacterium infantis [111]. B. infantis and other bacteria, growing readily in the presence of human and bovine oligosaccharides, reduce gut pH by producing volatile fatty acids. When epithelial cells were incubated with conditioned media from Infantis fermentation, there was a reduction in the release of pro-inﬂammatory cytokines [112]. 2.6. Insight into Novel Milk Components In a continuous effort to explain the beneﬁts of human milk, recent studies have focused on exosomes, their related extracellular vesicles (EV), micro-RNA (miRNA), and stem cells (Figure 2). MicroRNAs are small (18–25 nucleotides) non-coding strands of RNA that exert post-transcriptional regulation on a variety of tissues [127]. They are largely delivered to their target cell or tissue in extracellular vesicles and are found in most secretions from the body, including amniotic ﬂuid, tears, blood, saliva, and milk [128,129]. Moreover, miRNAs in milk and colostrum have two main purposes: ﬁrst, to maintain the functionality of the mammary gland, and second, to serve as a means of communication between mother and offspring [130]. A majority of miRNAs found in human, bovine, and porcine colostrum are associated with immunological pathways [128,131,132], but other work [130] has demonstrated that miRNAs play an important role in inﬂuencing the growth and development of the neonatal GI tract, particularly compared with mature milk. Extracellular vesicles in human milk and colostrum contain miRNAs associated with immune regulation and metabolism, proteins involved in the signal transduction, and inﬂammatory response. Milk and colostral extracellular vesicles may represent an anti- inﬂammatory mechanism underlying the prevention of NEC in preterm infants fed breast milk [133]. Proteomic work has demonstrated that exosomes and extracellular vesicles from BC are enriched with proteins associated with the immune response and growth, indicating their potential role in the regulation of these processes [134]. Co-incubation of bovine colostral vesicles with human macrophage cultures has shown that miRNAs and other proteins packaged in these vesicles have a profound effect on the metabolism, cell migration, and cellular response to LPS challenge [132]. Additionally, Baier et al. [135] showed that in vitro treatment of human cell culture from a variety of tissue origins with miRNAs from BC changes gene expression in human cell culture. Other work has shown that exosomes from bovine milk and BC can both be taken up by human enterocyte culture and maintain the cell cycle in these cells [136]. Although in vitro work has painted a compelling picture of the potential impact of bovine miRNA and exosomes on human tissue, in vivo experiments have been less consistent. The miRNAs packaged in exosomes are largely resistant to degradation in acidic environments, RNAse treatment, desiccation, and freezing [127,132]. Gu et al. 2.5. Oligosaccharides The latter studies in preterm pigs also tested whether adding the HMO, 2′-FL, was protective against an enterotoxigenic Escherichia coli F18 challenge, because 2′-FL has structural homology to bacterial adhesion sites in the intestine [121,122]. Both of these studies showed that the addition of 2′-FL to formula-fed preterm pigs did not reduce or prevent diarrhea or NEC incidence in the ﬁrst week of life. A study comparing a mixture of 4 vs. 25 different HMO blends found that these HMOs suppressed intestinal epithelial cell proliferation and had a modest immunomodulatory effect in the intestine in vivo, but did not prevent NEC or diarrhea when added to formula and fed to preterm pigs for 5–11 days after birth. Addition of mixtures of bovine milk oligosaccharides was well tolerated but did not improve any clinical outcomes in preterm pigs [121–125]. In 11 of 41 Nutrients 2021, 13, 2551 11 of 41 order to move beyond ﬁndings in preclinical NEC models, a recent study in matched human mother–infant cohorts correlated HMO composition in infants with healthy and NEC outcomes health [126]. This study showed that that DSLNT concentrations were signiﬁcantly lower in milk samples from NEC cases compared to controls. Moreover, the lower level of DSLNT was associated with lower relative abundance of Biﬁdobacterium longum and higher relative abundance of Enterobacter cloacae in infants with NEC. These studies highlight the speciﬁcity of HMOs to modulate NEC and the possible interaction between the immature intestine and developing microbiome community in preterm infants. These ﬁndings coupled with the structure and conﬁgurational speciﬁcity of the beneﬁcial oligosaccharides, indicate that additional work is warranted to explore the implications of milk oligosaccharide supplementation in diets for infants and children, particularly the highly sensitive preterm population. 3. Bovine Colostrum for Growth, Development and Immunity 3.1. Healthy Term Infants and Animal Models 3. Bovine Colostrum for Growth, Development and Immunity 3.1. Healthy Term Infants and Animal Models Limited information is available on the safety and efﬁcacy of providing BC for healthy term infants who have limited or no access to their own mother’s milk immediately after birth. In situations when this happens (e.g., maternal disease, agalactia, or no option to breastfeed), such infants would normally be fed infant formula, or maybe donor human milk if available. For infants beyond 6 months and toddlers beyond the ﬁrst year of life, commercially available follow-on or growing up formulas have been developed. These formulas were developed to meet the nutritional needs of infants and young toddlers whose complimentary food intake does meet speciﬁc nutritional requirements. The Ameri- can Academy of Pediatrics guidance indicates that follow-up formulas are nutritionally adequate but offer no nutritional advantage over infant formulas [142]. Others have raised questions about the nutritional rationale for follow-on formulas in healthy infants and toddlers [143]. Therefore, in healthy term infants, there would not seem to be a compelling rationale for supplementing BC, especially not in developed countries. pp g p y p Over the last decades, many adjustments have been made for conventional formulas to adapt their nutrient contents to age-related needs of infants and children (e.g., less overall protein and casein, more alpha-lactalbumin, and essential oils instead of bovine milk lipids). New ingredients continue to be added to infant formulas in order to make it more similar to that of human milk (‘humanized’ infant formula). More recently, speciﬁc components such as LF, osteopontin, lutein, oligosaccharides, MFGMs, and essential fatty acids have been added to specialized formulas [27]. The efforts are based on the assumption that human milk composition is closely adapted to the needs of infants, potentially even matched between individual infants and their own mother. Nevertheless, the biological effects of milk and colostrum may only partly be species-speciﬁc and closely matched between mother and offspring in the different species. Thus far, the ability to ‘humanize’ infant formula to accurately reﬂect nutritive and bioactive constituents of fresh human milk has failed, particularly regarding the highly unstable, heat-sensitive milk bioactive with immunological functions. This lack of progress is especially evident from numerous clinical studies showing the increased risk of diseases (e.g., NEC) in premature infants fed formula compared to human milk. 2.6. Insight into Novel Milk Components [128] demonstrated that miRNAs in extracellular porcine milk and colostrum are generally resistant to harsh conditions and may survive the acidic conditions of the stomach and pass intact into the small intestine. Baier et al. [135] demonstrated that healthy adults absorb miRNAs from mature cow’s milk, and other work has shown that humans may absorb animal- and plant-speciﬁc miRNAs. However, it remains unclear if the level of BC-derived absorbed miRNA and the systemic distribution thereof is sufﬁcient to have a meaningful biologic effect, as demonstrated in calves where expression of blood miRNAs correlated poorly with BC consumption [137]. Thus, it is possible that BC-derived miRNAs primarily exert a local effect on the GI tract and not transported to blood in large amounts. 12 of 41 12 of 41 Nutrients 2021, 13, 2551 Stem cells are present in human milk, and they are more abundant in human colostrum than in mature milk [138]. Interestingly, these cells are pluripotent, and work in rodent models has shown that human milk stem cells can be distributed to tissues as varied as the brain, thymus, pancreas, liver, spleen, and kidney. This wide distribution and additional characterization have indicated that there may be a role for these stem cells for regeneration of cells in innate immune system [139,140]. Similar cells have been found in bovine milk and colostrum, but it is unclear if they survive pasteurization and digestion in the human neonate, and how they may be bioactive in the small intestine. Additional research is needed to explore the implications of a pluripotent stem-cell population in BC, particularly if these cells are absorbed and remain pluripotent after absorption. p p p Extracellular vesicles, miRNA, and stem cells are all present in measurable amounts in human and bovine colostrum and mature milk. They are potential mechanisms to inﬂuence the growth and immune competence of the neonate [137,139–141]. However, it remains largely unclear if these effects are species-speciﬁc, or if neonatal humans are affected by these components in bovine milk and colostrum. Thus, additional work is required to explore the bioavailability and functions of colostral stem cells, particularly for sensitive neonates. 3. Bovine Colostrum for Growth, Development and Immunity 3.1. Healthy Term Infants and Animal Models On the other hand, it is possible that healthy term infants are less dependent on optimized levels of nutrients, growth factors, immunomodulatory components, and antimicrobials in human milk and colostrum than preterm, growth- restricted, or diseased infants. The short- and longer-term health beneﬁts of the numerous bioactive factors in fresh human milk or colostrum, and their varying concentrations among individuals remain poorly understood. Regardless, complete substitution of human milk Nutrients 2021, 13, 2551 13 of 41 13 of 41 by feeding BC to infants, even for shorter periods, is not recommended, due to the markedly different composition of nutrients and bioactives in BC versus both colostrum and human milk, but more research on dose-response relationships is needed (see later section). In human children and adults, some of the effects of bovine IgG are similar to those of IgA in human milk, e.g., binding to human-relevant pathogens, improved phagocytosis mediated through Fc receptors and prevention of infections [19]. In infants, bovine colostral IgG may pass through the gut undigested and have local antimicrobial or immune effects in the gut. Nevertheless, the speciﬁcity and effector functions of bovine IgG are not identical to human milk IgA and IgG, and it remains unclear if intact bovine IgG, provided in pure form (isolated from plasma or milk) or as part of BC, can fully or partly restore the lack of mother’s secretory IgA in formula-fed infants. Formula-fed infants receive minimal amounts of immunomodulatory proteins and Igs because processing technologies and heat treatment for microbiological safety normally denature these proteins. The development of more gentle milk-processing technologies in the future may enable inclusion of colostral Igs, or other fractions of BC, into standard infant formula. Longer-term trials are needed to conﬁrm if this reduces GI and respiratory infections, with or without changes to the incidence of allergies and asthma [19]. Bovine colostrum and its components have been extensively investigated for their potential use as a nutritional, growth-stimulating, immunological, and/or antimicrobial supplement for newborns of several animal species when their own maternal milk is deﬁcient or absent. Such animals may even beneﬁt from a short period of exclusive BC feeding, in contrast to infants. Beneﬁcial effects on growth and development have been reported for domestic animals (piglets, foals, and lambs) [42,51,144,145] and some pet- animal species (dogs, cats, and hamsters) [146–149], either as exclusive diet or supplement. 3. Bovine Colostrum for Growth, Development and Immunity 3.1. Healthy Term Infants and Animal Models Results from such studies suggest that BC may also be a useful nutritional and bioactive supplement for newborn infants, especially for gut functions, as its biological effects are only partly species-speciﬁc. The mechanisms whereby supplementary BC may induce effects in the gut of another species via nutritional, growth-stimulating, immunological, and/or antimicrobial factors (Figure 2) may be similar to possible mechanisms for effects of colostrum/milk versus formula to infants, as illustrated in Figure 4. g The most widespread use of BC in non-bovine animal species is for piglets. Modern pig breeds often produce more offspring than the number of functional teats on sows, and thus there is a need to provide nutrition and passive immunization of newborn piglets that do not get adequate sow’s colostrum [150]. Supplementing such piglets with intact BC increases their survival, but it remains lower than for species- and herd-speciﬁc sow’s colostrum provided via their own mother or foster mothers [144]. As in the human infant, supplementation of BC in piglets has the potential to affect host immunity initially via its interaction with gut pathogens and mucosal epithelial cells. Similarly, supplementation of BC with porcine plasma may further improve GI health and development in newborn pigs, demonstrating that this as a partly species-speciﬁc substitution for porcine colostrum, relative to formula [60]. In these studies, the gut trophic and enzyme maturation effects of exclusive BC feeding were similar to, or even exceeded, the effects of porcine colostrum. 3.2. Preterm Infants and Animal Models In humans, preterm birth (<90% gestation or <37 weeks in humans) occurs in 5–15% of all pregnancies [151]. The most immature, very preterm infants (<32 weeks gestation) suffer most from complications in pregnancy (e.g., inﬂammation, infection, and placen- tal dysfunction) or immediate postnatal period, resulting in growth failure and various longer-term maladaptation syndromes [152,153]. In the weeks after birth, very preterm infants have high susceptibility to systemic infections (bacteremia and LOS), gut disor- ders (FI and NEC), lung complications (bronchopulmonary dysplasia, BPD), and brain damage (intraventricular hemorrhage, IVH; cerebral palsy). Not surprisingly, these infants need specialized clinical care, but despite the challenges, survival, health, and growth of preterm infants, even when birth occurs as early as 60–70% gestation (24–28 weeks), have Nutrients 2021, 13, 2551 14 of 41 14 of 41 increased dramatically in the past decades. Improved care for these infants has come partly via advances in nutritional care (parenteral nutrition, donor milk, specialized formulas, and better feeding routines) and immunological protection (pro- and antibiotics, and im- munomodulatory drugs). Especially in preterm infants, intact BC or fractions thereof may have direct and indirect health effects by modulating the gut immune system, reducing gut inﬂammation and enhancing mucosal integrity and tissue repair (see Figure 4 for overview of possible mechanisms). VIEW 15 of 42 Figure 4. Possible differences between effects of maternal colostrum/milk and formula on the neonatal gut. The figure illustrates some of the proposed functional effects of maternal colostrum/milk bioactive factors on the dominant gut microbiota communities, mucosal epithelium, and immune cells, relative to formula diets. Enterobacteriaceae, Bifidobacteraceae, and Clostridiaceae families represent common bacteria groups found in the developing neonatal gut. The bacterial community composition can be affected by maternal milk factors, such as Igs (IgA and IgG), LF, bovine and human oligosaccharides (BMO/HMO), and a number of growth factors (e.g., EGF). Igs and LF have antimicrobial properties that function to limit epithelial inflammation and apoptosis resulting from activation of Toll-like receptor 4 (TLR4) by Enterobacteriaceae. BMO/HMO may serve as a substrate for growth and colonization of Bifidobacteraceae, a family of bacteria associated with gut health. The figure also illustrates that key peripheral organs, including the brain and lungs, directly or indirectly may be impacted by colostrum/milk-induced improved gut microbial activity and gut mucosal immune defense. (Created with BioRender.com; accessed on 7 June 2021.) The studies indicated no adverse clinical effects of BC supplementation, increas Figure 4. 3.2. Preterm Infants and Animal Models Enterobacteriaceae, Biﬁdobacteraceae, and Clostridiaceae families represent common bacteria groups found in the developing neonatal gut. The bacterial community composition can be affected by maternal milk factors, such as Igs (IgA and IgG), LF, bovine and human oligosaccharides (BMO/HMO), and a number of growth factors (e.g., EGF). Igs and LF have antimicrobial properties that function to limit epithelial inﬂammation and apoptosis resulting from activation of Toll-like receptor 4 (TLR4) by Enterobacteriaceae. BMO/HMO may serve as a substrate for growth and colonization of Biﬁdobacteraceae, a family of bacteria associated with gut health. The ﬁgure also illustrates that key peripheral organs, including the brain and lungs, directly or indirectly may be impacted by colostrum/milk-induced improved gut microbial activity and gut mucosal immune defense. (Created with BioRender.com; accessed on 7 June 2021.) enteral protein intake (when feeding with human milk), and/or a shortened time to reach full enteral feeding (when feeding formula). However, elevated plasma tyrosine suggested that excessive protein intake from BC may be a concern, especially in the first week of life (see also later section). These pilot studies suggest that a growth-stimulating effect of adding BC to human milk or formula may relate to indirect effects via the developing gut microbiota, potentially via Igs, which, in turn, affect plasma amino acid levels to a composition more favorable for growth [156]. Studies indicate that intestinal The potential for using BC speciﬁcally to support nutrition and immunological pro- tection of this highly sensitive infant population has remained largely unexplored until recently. It is possible that preterm infants, because they have an immature gut, immune system, and metabolism, have some physiological and immunological similarities with newborn farm animals, for whom colostrum intake just after birth is absolutely essential for survival. Recently, stepwise pilot-phase safety trials that involved using a powdered intact BC product during the ﬁrst 2 weeks of life were completed in Denmark and China [154,155]. p g [ ] absorption of human colostral proteins can occur [46–48,157], yet the capacity of preterm infants to absorb Igs from BC is very limited or absent [154,155]. A larger study is ongoing to confirm these initial observations, feeding a maximum of 50 mL/kg/d BC (ColoDan, Biofiber, Denmark) as a supplement to formula for very preterm infants in the first 2 weeks of life, with time to full enteral feeding as the primary outcome (n = 350, ClinicalTrials.gov: NCT03085277). 3.2. Preterm Infants and Animal Models Possible differences between effects of maternal colostrum/milk and formula on the neonatal gut. The ﬁgure illustrates some of the proposed functional effects of maternal colostrum/milk bioactive factors on the dominant gut microbiota communities, mucosal epithelium, and immune cells, relative to formula diets. Enterobacteriaceae, Biﬁdobacteraceae, and Clostridiaceae families represent common bacteria groups found in the developing neonatal gut. The bacterial community composition can be affected by maternal milk factors, such as Igs (IgA and IgG), LF, bovine and human oligosaccharides (BMO/HMO), and a number of growth factors (e.g., EGF). Igs and LF have antimicrobial properties that function to limit epithelial inﬂammation and apoptosis resulting from activation of Toll-like receptor 4 (TLR4) by Enterobacteriaceae. BMO/HMO may serve as a substrate for growth and colonization of Biﬁdobacteraceae, a family of bacteria associated with gut health. The ﬁgure also illustrates that key peripheral organs, including the brain and lungs, directly or indirectly may be impacted by colostrum/milk-induced improved gut microbial activity and gut mucosal immune defense. (Created with BioRender.com; accessed on 7 June 2021.) Figure 4. Possible differences between effects of maternal colostrum/milk and formula on the neonatal gut. The figure illustrates some of the proposed functional effects of maternal colostrum/milk bioactive factors on the dominant gut microbiota communities, mucosal epithelium, and immune cells, relative to formula diets. Enterobacteriaceae, Bifidobacteraceae, and Clostridiaceae families represent common bacteria groups found in the developing neonatal gut. The bacterial community composition can be affected by maternal milk factors, such as Igs (IgA and IgG), LF, bovine and human oligosaccharides (BMO/HMO), and a number of growth factors (e.g., EGF). Igs and LF have antimicrobial properties that function to limit epithelial inflammation and apoptosis resulting from activation of Toll-like receptor 4 (TLR4) by Enterobacteriaceae. BMO/HMO may serve as a substrate for growth and colonization of Bifidobacteraceae, a family of bacteria associated with gut health. The figure also illustrates that key peripheral organs, including the brain and lungs, directly or indirectly may be impacted by colostrum/milk-induced improved gut microbial activity and gut mucosal immune defense. (Created with BioRender.com; accessed on 7 June 2021.) The studies indicated no adverse clinical effects of BC supplementation increas Figure 4. Possible differences between effects of maternal colostrum/milk and formula on the neonatal gut. The ﬁgure illustrates some of the proposed functional effects of maternal colostrum/milk bioactive factors on the dominant gut microbiota communities, mucosal epithelium, and immune cells, relative to formula diets. 3.2. Preterm Infants and Animal Models Further studies demonstrated that bovine and porcine colostrum were equally effective in inducing body and gut growth and in protecting against inﬂammatory conditions in preterm pigs (NEC and LOS; see later disease section [163–166]). We also showed that BC-fed preterm piglets had a remarkable capacity to rapidly adapt their gut and immune development to that in term pigs (within 1 to 2 weeks) [167,168], while brain and neurodevelopment were slower in reaching normal levels (within 3 to 4 weeks) [169–172] when fed BC during the ﬁrst week. These studies clearly indicated the potential to use BC in states of immaturity with a high sensitivity to gut and immune disorders. Importantly, we subsequently showed that human donor milk was also relatively effective in protecting preterm pig against NEC [173]. However, when fed in a minimal enteral feeding protocol, BC was more effective than donor human milk to reduce the density of mucosa-associated bacteria and putative pathogens [174]. These studies in preterm pigs suggest that common component(s) in porcine and bovine colostrum and in human donor milk provide protection against the development of NEC and importantly that these factors are not species-speciﬁc. p p y p p The above studies highlight the value of the preterm pig as a preclinical model to evaluate the function and nutritional availability of milk diets, supplements, and novel ingredients in the clinical support of preterm human infants. It is also notable that, be- sides gut complications (feeding intolerance and NEC), this model incorporates a range of the complications that are commonly known from moderately and very preterm infants, such as respiratory insufﬁciency; impaired growth; dysmotility; high sepsis sensitivity; metabolic derangements; and kidney, liver, and brain dysfunctions [158,159,175,176]. The integration of all of these complications into the same clinical model is a great advantage over other (rodent) models in neonatology, because any intervention, including the ﬁrst milk diet, is likely to have multi-organ effects. Due to the similar size (0.6–1 kg) and physiol- ogy of 90% gestation preterm pigs and 70% preterm infants, clinical tools and interventions can be made similar (e.g., respiratory care, parenteral/enteral nutrition, use of diagnostic imaging techniques, repeated blood sampling from indwelling catheters, and surgical interventions) [158]. 3.2. Preterm Infants and Animal Models The interest to use BC for nutritional and immunological support of preterm infants was sparked by a large series of studies in preterm pigs over 15 years. The neonatal pig has been used as a model for human infant nutrition and gastroenterology for decades p g p [ ] The studies indicated no adverse clinical effects of BC supplementation, increased enteral protein intake (when feeding with human milk), and/or a shortened time to reach full enteral feeding (when feeding formula). However, elevated plasma tyrosine suggested that excessive protein intake from BC may be a concern, especially in the ﬁrst week of life (see also later section). These pilot studies suggest that a growth-stimulating effect of adding BC to human milk or formula may relate to indirect effects via the developing gut microbiota, potentially via Igs, which, in turn, affect plasma amino acid levels to a composition more favorable for growth [156]. Studies indicate that intestinal absorption of human colostral proteins can occur [46–48,157], yet the capacity of preterm infants to Nutrients 2021, 13, 2551 15 of 41 15 of 41 absorb Igs from BC is very limited or absent [154,155]. A larger study is ongoing to conﬁrm these initial observations, feeding a maximum of 50 mL/kg/d BC (ColoDan, Bioﬁber, Denmark) as a supplement to formula for very preterm infants in the ﬁrst 2 weeks of life, with time to full enteral feeding as the primary outcome (n = 350, ClinicalTrials.gov: NCT03085277). ) The interest to use BC for nutritional and immunological support of preterm infants was sparked by a large series of studies in preterm pigs over 15 years. The neonatal pig has been used as a model for human infant nutrition and gastroenterology for decades based on homologies with regards to physiology, anatomy, and metabolism [158–160]. We ﬁrst began by studying pigs delivered by caesarean section at 90% gestation and observed that feeding porcine colostrum to newborn preterm versus term pigs induced a marked trophic and functional gut response, for some parameters even more than for term animals [161]. Subsequent studies showed that preterm pig intestines were highly sensitive to formula feeding, even more than preterm infants, and spontaneously developed diet- and microbiota-dependent NEC in the ﬁrst 1 to 2 weeks of life [162,163]. 3.2. Preterm Infants and Animal Models Potentially, preterm pigs may also be used to test the interacting effects of BC for gut and skin healing in preterm newborns and the interacting effects with non-medical clinical routines, such as maternal singing, skin-to-skin contact, and kangaroo care [177]. A recent study demonstrated reduced gut complications in preterm pigs being co-bedded, facilitating sibling skin-to-skin contact [178]. The possibility that the immuno- logical properties of BC may beneﬁt the immature skin of preterm newborns, having altered cell differentiation and perturbed barrier functions [179], remains unexplored. 3.3. Growth-Restricted Infants, Children, or Animal Models There is a potential to use BC as a supplement in conditions of growth-restriction in developing countries, mainly because the non-nutritive immunoprotective components may alleviate clinical risks of infection and inﬂammation and thus limit negative impacts Nutrients 2021, 13, 2551 16 of 41 16 of 41 on growth at birth or later in infancy and childhood. However, it is not recommended to feed BC as the sole diet to promote growth (see later section). In clinical practice, the speciﬁc effects of nutrition-induced growth restriction (e.g., poor placental function before birth and deﬁcient nutrient intake after birth) can be difﬁcult to separate from associated increases in inﬂammation and infections. A number of studies performed mainly in developed countries indicate that providing hyperimmune supplemental BC reduced severity of diarrhea in children that have evidence of GI infection (see more in later section) [180,181]. Whether colostrum supplementation increases growth and development in growth-restricted infants born at full term, without associated diseases (infectious, inﬂammatory, or other), is unknown. For infants, it is well-known that growth restriction at birth is a risk factor for many later diseases, especially if birth also occurs prematurely, making the infant both under- weight and immature. A large proportion of preterm infants are born intra-uterine growth restricted (IUGR) and after birth they may continue to experience slow growth related to their increased postnatal complications, resulting in extra-uterine growth restriction (EUGR). Human milk is relatively deﬁcient in some nutrients, especially protein and select minerals (e.g., Ca, P, Fe, and Zn), to support optimal growth of such infants from 1 to 2 weeks of age. Thus, nutrient fortiﬁers to human milk are needed and a number of prod- ucts based on typical formula products are available on the market [182]. Due to its high content of both nutrients (protein) and immunomodulatory factors, BC has been speculated to have a potential as a nutrient fortiﬁer to human milk. As a basis for the human trials, we have tested the efﬁcacy of BC as a fortiﬁer to donor human milk in preterm pigs in the ﬁrst weeks of life. In pigs fed donor human milk, fortiﬁcation with BC was superior to formula-based fortiﬁers to support growth, gut function, nutrient absorption, and mucosal defense [183,184]. 3.3. Growth-Restricted Infants, Children, or Animal Models The studies are signiﬁcant because they suggest that, even though BC and donor human milk do not contain pig- and species-speciﬁc immune components (e.g., Igs and LF), their local gut effects and non-speciﬁc systemic immune effects appear sufﬁcient to support health and development in preterm pigs. Reduced growth and health challenges of newly weaned piglets can be viewed as an animal model for the nutrition and immune challenges often observed in infants weaned early from their mother and fed alternative milk or vegetable diets, particularly when reared in low-sanitary environments in developing countries. When pigs were weaned later (3 to 4 weeks) on to vegetable-based diets, growth, intestinal function, and immunity were improved (local and systemic IgA, Th1 and Th2 cytokines, and nutrient digestion and absorption) by supplementing small amounts of intact BC or colostrum whey pow- der (1–10 g/kg body weight per day). Effects occurred mainly during the immediate post-weaning period when digestive complications, microbial perturbations, and adverse immune responses were most pronounced [185–190]. A few days of exclusive BC feeding post-weaning (40–45 g/kg/d) reduced diarrhea, E. coli (Enterobacteriaceae) density in intesti- nal contents and tissue, and Gram-negative mucosal immune responses (TLR4 and IL-2), but it also increased short chain fatty acid production, in part due to excessive protein supply and fermentation [187,188]. Collectively, the studies in young piglets support that supplemental BC indirectly improves body growth via improved gut functions and immu- nity, especially in conditions of stress and inﬂammation, such as the weaning transition, probably reﬂecting the mechanisms illustrated in Figure 4. As for infants and children, the optimal BC intake to stimulate body growth in various clinical conditions is unclear. 4. Bovine Colostrum to Prevent or Treat Speciﬁc Pediatric Diseases 4.1. Gut and Lung Infections in Children and Animal Models One of the most extensively studied applications of BC is treatment of gut infections in children, including rotavirus, enteropathogenic and enterotoxigenic E. coli, Shigella, and Helicobacter pylori infections (for detailed reviews, see References [19,23,62]). Most of these studies have been conducted by using hyperimmune BC, which is produced by immuniz- ing cows with select pathogens or toxins prior to lactation, with the intention to obtain 17 of 41 Nutrients 2021, 13, 2551 17 of 41 BC with enriched titers of pathogen- or antigen-speciﬁc IgG antibodies. Studies also have used different forms of immunoglobulin-enriched products, including IgG-rich colostrum, IgG-isolates from colostrum or milk, and serum-derived IgG. The products used in the clinical studies were tested for prophylactic or therapeutic effects in ﬁeld settings, as well as in controlled pathogen challenge models. In two double-blind placebo-controlled studies and two controlled studies designed to treat rotavirus diarrhea, children aged 4–30 months given hyperimmune BC or concentrated antibodies showed signiﬁcant clinical reductions in duration of diarrhea, stool frequency and duration of virus shedding [191–193]. How- ever, a similar study showed only modest beneﬁts [194]. In controlled studies to treat enteropathogenic E. coli (EPEC), children with diarrhea were given hyperimmune bovine milk immunoglobulin concentrate for 10 days. Negative EPEC cultures were found in 84% of treated cases, but in only 11% of control children [195]. In another double-blind placebo-controlled study, children with E. coli–induced diarrhea were treated with milk concentrate from cows hyperimmunized with enterotoxigenic E. coli (ETEC), but no ben- eﬁcial effects on duration of diarrhea or stool frequency were observed [196]. In other placebo-controlled studies, children infected with Shighella were treated with hyperimmune BC with one study showing a reduction in stool frequency [197] and another study no effect [198]. Studies in adults and pathogen challenge studies with Clostridium difﬁcile, ETEC, and Cryptosporidium parvum showed that hyperimmune BC products can indeed reduce diarrhea and the presence of pathogens in stool [19,62]. p p g Studies designed to prevent gut infections with rotavirus in children have shown a protective effect of feeding hyperimmune BC or immune concentrates [199–201]. However, in another large controlled ﬁeld study, where immune fractions from cows immunized with rotavirus and E. coli were added to infant formula, there was no protection against diarrhea [202]. Thus, a summary of the studies in children infected with either rotavirus or E. 4.2. Necrotizing Enterocolitis in Preterm Infants and Animal Models 4.2. Necrotizing Enterocolitis in Preterm Infants and Animal Models Necrotizing enterocolitis is the leading cause of infant death from GI disease in infants, affecting 3–10% of the hospitalized preterm infants around the world [158,207]. NEC has a mortality as high as 50%, and surgical intervention is necessary in 20–40% of cases, leading to increased morbidity. Three key features necessary for NEC pathogenesis are infant prematurity, the presence of gut microbiota, and enteral feeding, especially infant formula. Importantly, feeding mother’s own breast milk has been shown to effectively reduce the incidence of NEC in preterm infants [9]. The fact that feeding breast milk prevents NEC, especially in preterm infants, suggests that two key features of NEC, namely the microbiota and enteral feeding, may be neutralized by the important immune factors present in human colostrum and milk, but their absence in infant formula. A recent meta-analysis summarized results from eight RCT studies (n = 394 infants) including mostly preterm infants given human or BC via oropharyngeal route during the ﬁrst 48 h of life demonstrated no effect on NEC incidence or all-cause mortality, but a trend to reduce culture proven sepsis and reduced time to full feeds [208]. The ongoing large clinical trials on very preterm infants in Denmark and China (<1500 g, total n = 700) will soon provide more evidence for safety and efﬁcacy of providing BC to preterm infants to protect against NEC, but also against milder gut immaturity complications, such as constipation, feeding intolerance, and diarrhea [154,155,209]. However, in none of these trials was NEC the primary outcome, because of its low prevalence in the above countries (<5% [207]), requiring a very large sample size to verify NEC effects. Indirect evidence to support BC for protection against NEC comes from studies showing that BC contains IgG and IgA antibodies directed against pathogens that have been associated with NEC, such as Klebsiella, Citrobacter, Enterobacter, and Serratia [210]. A pilot clinical study in India (total n = 86) showed no beneﬁts of providing a processed BC product to very preterm infants, and there were even indications (although not signiﬁcant) of increased gut inﬂammation, indicated by elevated IL-6 in stool samples and radiological features of NEC [211]. 4. Bovine Colostrum to Prevent or Treat Speciﬁc Pediatric Diseases 4.1. Gut and Lung Infections in Children and Animal Models coli suggest that they are not uniformly positive, and are heterogeneous with regard to dose, duration of treatment and form of colostrum product. However, a majority of the studies found a clinical beneﬁt on diarrhea outcomes when treated with hyperimmune BC. Likewise, the studies targeting prevention of rotavirus by feeding hyperimmune BC showed positive effects of reduced diarrhea outcomes. Importantly, hyperimmune BC directed against speciﬁc pathogens, in this case rotavirus, has clear therapeutic potentials in children and better effects than against pathogenic bacterial species. A recent meta- analysis including 213 children mostly supplemented with hyperimmune BC indicates reduced severity of diarrhea in children that have evidence of GI infection with E. coli and rotavirus [180]. The clinical evidence may suggest that non-immune BC is less effective in controlling gut diseases than hyperimmune colostrum derived from pathogen-speciﬁc immunized cows. However, a recent double-blind RCT with 160 children with evidence of GI infection performed in Egypt showed that providing supplemental non-immune BC reduced severity of diarrhea [181]. An important observation from these studies, in children as young as 3 months of age, is the absence of any adverse effects of BC. This point raises the question of whether BC can be safely fed to even younger term infants and those born preterm. Possibly, BC can be particularly effective in preventing bacterial translocation and additional immune protection in immature states of gut microbial colonization (e.g., low species abundance and diversity) and mucosal immunity (e.g., mucous production and immune-cell responses). In addition to binding GI tract associated pathogens, IgG from BC can bind to res- piratory pathogens, such as human Respiratory Syncytial Virus (RSV), inﬂuenza virus and Streptococcus pneumonia [203]. This immunomodulation explains why raw milk consumption (provision of more intact IgG) is associated with fewer upper respiratory tract infections and otitis media [204]. While the studies suggest that BC can prevent upper respiratory tract infections, the many open, non-controlled prospective studies should be interpreted with caution. A role for IgG from BC in preventing or ameliorating viral respiratory tract infections is possible, but whether this protective effect will also affect allergy prevalence remains to be established [19]. 4. Bovine Colostrum to Prevent or Treat Speciﬁc Pediatric Diseases 4.1. Gut and Lung Infections in Children and Animal Models The reduced prevalence of allergy in Nutrients 2021, 13, 2551 18 of 41 18 of 41 farming families in many countries [205] may be related to intake of unpasteurized milk (containing more intact IgG and immunological factors) by both infants, children as well as their pregnant and lactating mothers. Again, important lessons are available from studying the effects of intact or frac- tionated BC in developing pigs with infections. Effects on lung infections are poorly documented, but a number of reports show speciﬁc abilities of BC to protect gut epithelial cells against infections, even without prior immunization of pregnant cow’s against speciﬁc pig pathogens. Thus, the increased in vitro membrane permeability caused by piglet ETEC bacteria was decreased by three different BC fractions [206]. The fractions contained widely different amounts of Igs and growth factors, suggesting that individual BC factors do not alone explain the protective effects, but antimicrobial compounds such as LF, lysozyme, and lactoperoxidase may all be involved in synergy [206]. In vivo, exclusive feeding with intact BC prevented diarrhea, relative to formula-fed piglets, probably via inducing a higher ratio of lactic acid bacteria to hemolytic E. Coli, and lower expression of intestinal Toll-like receptor-4 and IL-2 [187,188]. Similar antibacterial and immune modulating effects in newly weaned piglets were found by other investigators after supplementation with intact BC [185,186,189], but whether such effects extend to organs distant to the gut (e.g., lungs) remains unknown It is noteworthy that small BC supplements to weanling 3-to-4-week-old pigs (0.5–1 g/kg/day) slightly increased systemic IgA levels (likely gut-derived) [185,186], potentially supporting epithelial protection throughout the body, while an exclusive BC diet from day 3 of life (40–45 g/kg/day) had no effects on systemic IgG, IgA, or IgM [188]. 4.2. Necrotizing Enterocolitis in Preterm Infants and Animal Models Relatively large amounts (5–8 g/kg/d) of product (Pedimmune, Merck, India) were fed for up to 3 weeks with mixed feedings, and its safety related to other constituents than BC alone was questioned (e.g., excessive osmolality). In another recent study on preterm infants (total n = 80, <34 weeks gestation), an intact BC product (Immuguard, Dulex-Lab Pharmaceutical, Egypt) increased systemic T-regulatory (Treg) cell number Nutrients 2021, 13, 2551 19 of 41 19 of 41 and showed a clear tendency to improved feeding tolerance, growth and resistance to NEC [212]. In this study, maximum 20 mL/kg BC ﬂuid (e.g., <1 g/kg/d) was supplemented to daily formula meals and was compared to formula alone for a maximum of 2 weeks, while gradually transitioning infants to full formula feeding when mother’s own milk was unavailable. Beyond the few published clinical studies described above, no studies have examined the effects of BC supplementation on NEC. However, there are a number of clinical studies that have tested whether feeding human immunoglobulin preparations can prevent the incidence of NEC [213–217]. Three of the ﬁve clinical studies, based on a total of 2095 infants, were reviewed, and it was concluded the current evidence does not support admin- istration of oral immunoglobulin for NEC prevention [218]. The earliest reported studies, showing a protective effect of human Igs against NEC, were small, the outcome assess- ments not blinded, and infants that received human breast milk were excluded [213,215]. This contrasts with the more recent large randomized placebo-controlled double-blind study where 90% of the infants received breast milk, and with no effect of immunoglobulin supplementation [214]. Here, the immune protection provided by breast milk itself may explain that additional immunoglobulin had no effect. Additionally, Eibl et al. (1988) [213] and Rubaltelli et al. (1991) [215] administered the oral immunoglobulin within the ﬁrst 24 h following birth, whereas Lawrence et al. (2001) [214] only began oral supplemen- tation after initiation of enteral feeding. The Lawrence et al. (2001) and Rubaltelli et al. (1991) studies used preparations containing predominately IgG, whereas the study by Eibl et al. (1988) used an immunoglobulin mixture containing 73% IgA and 26% IgG. The Lawrence et al. (2001) study also fed a higher dose of immunoglobulin (1200 mg/kg/bw) than the Eibl et al. (1988) and Rubaltelli et al. (1991) studies (600 mg/kg/bw). 4.2. Necrotizing Enterocolitis in Preterm Infants and Animal Models Thus, from the available studies in preterm infants, is would appear that supplementing BC may be safe, but the beneﬁts on endpoints of growth, morbidity and disease outcomes are lacking given the relatively small number of studies reported. g y p The available evidence from pigs suggest that Igs, and a wide range of antibacterial and immunomodulatory factors in BC [3], may work across species, and therefore have protective effects against pathogens and inﬂammatory reactions in the infant gut. Using preterm pigs as a model of infant NEC [52,158,163,219], BC has been shown to effectively reduce NEC incidence, compared to different feeding regimens of infant formulas, human milk or even amniotic ﬂuid (the natural fetal ‘enteral diet’) [163,164,174,183,184,219–238]. The effects were most pronounced within the ﬁrst week after preterm birth [174,229,239], with similar efﬁcacy as porcine colostrum [164] and may ameliorate damage induced by ear- lier or later formula feeding [223,233,239,240], at least when representing a major part of the daily diet [240]. Importantly, BC maintained its in vivo efﬁcacy to protect against NEC in preterm pigs following the spray-drying and heat-pasteurization required for long shelf life, easy handling, and product sterility [232]. Preterm pigs have provided detailed insight into the possible modes of action of BC to prevent NEC development. The effects of BC include stimulation of physical activity [171], feeding-induced mucosal growth [164,223,229], diges- tive enzyme support [174,184,229,233,240], better nutrient absorption [173,174,224,229,241], improved enteric nervous system development [227,234,242], dampened bacterial over- growth [163,164,173,174,223,233], reduced intestinal cytokine responses [219,229,230,233], less intestinal permeability [173,174,224,229,233], prevention of pathogen adherence to the gut epithelium [173,184,236], greater production of intestinal mucus [163,243], and reduced bacterial fermentation of nutrients to lactate and short chain fatty acids. It is not clear if BC mediates these effects on NEC by altering gut bacterial colonization. We found no consistent changes in microbiota community proﬁles associated with BC-induced NEC protection, particularly not during the ﬁrst week after birth when colonization is chaotic and highly variable [164,173,174,219,223,229,244]. In week two after birth, BC feeding appeared to produce a more robust impact on the gut microbiota, preventing overall bacterial adherence to the mucosa, and proliferation of potentially pathogenic strains of Campylobacter, Helicobacter, Enterococcus, and some Clostridia [174]. 20 of 41 20 of 41 Nutrients 2021, 13, 2551 4.3. 4.2. Necrotizing Enterocolitis in Preterm Infants and Animal Models Fetal Infection and Neonatal Sepsis in Infants and Animal Models A large fraction of the infants born prematurely are born in response to maternal infection (chorioamnionitis) with exposure of the developing fetus to gut, lung and systemic inﬂammation in utero. Such prenatal insults are associated with infections just after birth (early onset sepsis, EOS) and altered immune system development with a range of adverse short- and longer-term outcomes across many organs (e.g., NEC, BPD, and IVH) [245]. The health consequences of fetal inﬂammation for preterm infants are highly dependent on the type, length and level of fetal infection-induced inﬂammation, from inducing a precocious (potentially beneﬁcial) immune maturation, to adverse immune defects with subsequent widespread inﬂammatory insults [246,247]. No studies are available in infants but in theory, BC supplementation may ameliorate the postnatal consequences of fetal inﬂammation by better control of bacterial colonization after birth and modulation of local and systemic immune responses. In preterm pigs, postnatal BC feeding was clearly more effective than formula feeding to dampen the gut inﬂammatory effects of experimental intra-uterine infection of the fetus [238,248]. Even without fetal inﬂammatory insults, newborn preterm infants are highly sus- ceptible to LOS) following systemic infections caused by entry of bacteria across their permeable barriers (skin, lung, and gut) or via indwelling catheters for parenteral nutrition and intravenous medication. Between 10 and 40% of hospitalized preterm infants expe- rience one or more periods of systemic infection [183,249,250]. Preterm infants are more predisposed to LOS because of immature function of circulating innate and adaptive im- mune cells, and an inability to mount an effective immune response to eliminate invading microorganisms. They rely on a disease-tolerant rather than a disease-resistant strategy to combat infections [251]. There is no doubt that mother’s own milk, via IgA and other immunomodulatory components, decreases sepsis sensitivity [252]. y p p y Whether BC supplementation decreases the risk of sepsis, especially in formula-fed preterm infants, is unclear. A colostrum-like product showed no effect on sepsis in the pre- viously mentioned Indian study [211], while the Egyptian study reported a clear tendency of fewer LOS cases. The latter study showed more systemic Treg cells, inducing a more disease-tolerant state, when formula-fed preterm infants were supplemented with BC from birth [212]. 4.2. Necrotizing Enterocolitis in Preterm Infants and Animal Models Blood T-regulatory cells (Tregs) are critical factors in regulating the newborn innate and adaptive host response to infection and prevent excessive tissue inﬂammatory damage whilst allowing development of the immune system components, necessary to combat infections (Th1). The results are promising and may suggest a preventive effect of BC against LOS, but results need to be conﬁrmed in more immature infants receiving variable amounts of human milk and formula. In preterm pigs, heat treatment (pasteurization) reduced the capacity of human milk to reduce systemic infection [253], yet the addition of BC to pasteurized donor milk improved its capacity to prevent infection [184]. Direct antimicrobial effects of BC are supported by studies in vitro, both in combination with formula [254] and human milk [250]. Conversely, potential indirect effects, such as improved gut integrity, reduced bacterial transloca- tion, and systemic immunity, in immature newborns are supported by studies in both piglets [229] and rodents [255]. Such luminal effects of BC, facilitating systemic immune protection, may explain why BC feeding markedly reduced adverse clinical symptoms and brain defects in a preterm pig sepsis model [166]. However, reduced systemic neutropenia and bacterial clearance, relative to formula feeding, were observed mainly when BC was fed in the ﬁrst days after preterm birth, not later [165,239]. Because preterm pigs are able to absorb intact Igs during the ﬁrst day after birth, some of these effects may be mediated via circulating bovine Igs, and these effects could be pig-speciﬁc, and not relevant for human infants. Regardless, enteral feeding with BC may reduce sepsis sensitivity in both infants and pigs, especially just after preterm birth. 21 of 41 21 of 41 Nutrients 2021, 13, 2551 4.4. Inﬂammatory Bowel Disease in Adolescents or Animal Models Inﬂammatory bowel disease (IBD) comprises multiple GI disorders, including Crohn’s disease (CD) and ulcerative colitis (UC), and has multifactorial etiology that involves an exaggerated immune response in genetically select groups in response to an environmental factor [256]. The incidence of pediatric IBD is increasing and it develops mainly in adoles- cence (ages 10–17), but can also occur in younger children less than 5 years of age. Common treatments of pediatric IBD are corticosteroids, amino salicylate, and immunomodulatory agents, such as thiopurines, methotrexate, and anti-TNF antibody drugs. Many of these treatments have important adverse side effects that can reduce body growth and skeletal development during this critical stage of life. 4.2. Necrotizing Enterocolitis in Preterm Infants and Animal Models Key factors that are commonly altered in cases of IBD and its treatment are the gut microbiota composition and nutrition, since both elements can interact to affect the mucosal immune response [257]. Since nutrition plays a key role in pediatric growth and can also modulate IBD morbidity, dietary approaches have been used in treatment, including use of exclusive and partial enteral nutrition as well as dietary exclusion. The use of exclusive enteral nutrition involves nutrition inter- vention using a complete liquid formula and is recommended as a ﬁrst-line therapy for mild-to-moderate IBD by the European Society of Pediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN). There are no clinical studies in pediatric patients and only one study that has examined the effect of BC in adult IBD patients with distal colitis [258]. This study involved a small randomized double-blind controlled protocol in middle-aged (45 years) patients with mild to moderately severe distal colitis, given twice daily rectal enemas of either a puriﬁed BC product (100 mL of 10% solution) or a control bovine serum albumin solution for four weeks. Several of the patients in both groups were concurrently taking amino salicylate. The BC enemas were well tolerated, and the results showed a signiﬁcant reduction in clinical symptom score after two weeks in BC-treated patients with no improvement in controls. Further, BC-treated patients’ biopsies showed improvement in histological score. , p p p g The above human study is supported by studies in a mouse DSS colitis model show- ing that relatively low daily oral gavage of BC (20 mg/kg/bw) improved occult blood, stool consistency, and clinical recovery from colitis, but without preventing the initial weight loss [259]. Conversely, human colostral sIgA (1 to 2 mg/kg/bw) improved re- covery of weight loss induced by colitis [259]. Oral gavage with higher daily intakes of BC (100 mg/kg/bw but not 500 mg/kg/bw) in DSS-treated mice also showed protec- tion against tissue injury, and the effects were partially mediated by the IgG present in BC. In a mouse model of TNBS-induced colitis, a relatively high daily oral BC intake (100 mg/mouse) for 21 days reduced colonic tissue injury, proinﬂammatory cytokine ex- pression, and abundance E. coli and Enterococci [260]. Taken together, despite only a single human adult IBD study, the results in mouse colitis models suggest that oral BC offers some protection against mucosal inﬂammation and tissue injury. 4.2. Necrotizing Enterocolitis in Preterm Infants and Animal Models This experimental evidence encourages further well-designed, clinical trials in pediatric IBD patients to explore the beneﬁts of BC to reduce disease morbidity. 4.5. Short-Bowel Syndrome in Infants or Animal Models 4.5. Short-Bowel Syndrome in Infants or Animal Models Short bowel syndrome (SBS) is the clinical condition resulting from surgical resection of large portion of intestine due to congenital defects (intestinal atresia and gastroschisis) or disease-associated loss of absorption, leading to an inability to maintain nutrient balance when fed a normal diet. In pediatrics, disease-related SBS is prevalent, resulting from severe NEC lesions, requiring removal the affected (necrotic) parts of the intestine and/or colon. The clinical condition is particularly severe after removal of the distal part of the small intestine, as demonstrated in both patients and animal models, possibly due to the dense localization of immune and enteroendocrine cells in the ileum [261]. Preterm infants with SBS after NEC and intestinal resection depend on small volumes of enteral food, in addition to parenteral nutrition, to stimulate gut growth and adaptation, without overloading the immature, remaining parts of the resected gut. Human milk is the Nutrients 2021, 13, 2551 22 of 41 22 of 41 recommended diet for all preterm infants, including SBS infants, but when the mother’s own milk or donor human milk is not available, formulas are used. Again, a protein-rich and growth-factor-rich non-human milk diet, such as BC, could be a relevant supplement to this diet. In a pilot study, children (>1 year) with SBS due to a previous episode of NEC and surgery did not improve their intestinal functions (nutrient balance) after inclusion of BC in the enteral diet [262]. Enteral BC supplementation was well tolerated by preterm infants shortly after surgery, with no signs of cow’s milk allergy [263], but more studies are required to substantiate if and how BC should be used for this highly sensitive patient population. A number of both infant- and feeding-related factors (e.g., gestational age, postnatal, resection type, feeding volume, and other diets) may affect responses, making it very difﬁcult to fully document beneﬁts or possible harm in infant studies. y y p The clinical complications observed after 50% distal intestine resection in preterm pigs are similar to those in infants subjected to intestinal resection for reasons other than NEC [261]. However, surgical resection in preterm pigs following extensive NEC lesions is difﬁcult due to both practical and ethical limitations in the care of such individuals. 4.5. Short-Bowel Syndrome in Infants or Animal Models Importantly, the postsurgical adaptation responses to enteral formula feeding are reduced in preterm versus term SBS pigs, and the clinical complications, such as hemodynamic instability, hypothermia, intestinal dysmotility, dehydration, respiratory distress, and peritonitis, are more severe [264]. Preterm SBS pigs have a blunted postsurgical increase in intestinal protein synthesis, villus height, crypt depth, and digestive enzyme activities when compared with the corresponding intestinal segment in unresected control pigs. In such piglets, there is limited effect of providing enteral diets, either as infant formula or BC [263], yet the enteroendocrine growth factor glucagon-like peptide 2 (GLP-2) had marked effects immediately after resection [265]. The diminished adaptation in preterm and term newborn pigs contrasts with the structural and functional adaptation of the remnant distal part of the intestine in slightly older (4–8-week-old) suckling pigs that were subjected to resection. In these older pigs, a BC concentrate signiﬁcantly increased adaptation and circulating levels of insulin-like growth factor 1 (IGF-1) [266,267] and GLP-2 [268]. These ﬁndings suggest that the effect of BC in developing SBS individuals is highly dependent on the stage of maturation, as well as many other confounding variables. 5. Processing of Bovine Colostrum for Use in Pediatrics Infants, especially those born preterm or otherwise immunocompromised, may beneﬁt from the antibacterial, immunomodulatory, and growth-stimulating effects of BC (Figure 2). Therefore, the adverse effects of processing colostrum to its constituents (e.g., bioactives, macronutrients, and micronutrients) are more likely to impact the response of these infants to BC supplementation, relative to more robust term infants and children. Processing pro- cedures differentially reduce the content of many bioactive factors in BC [225,232,254,276], and while this may decrease in vivo bioactivity, it does not eliminate it, partly due to the high amounts present in native BC before processing. Igs and LF are relatively sen- sitive to heat damage, while other bioactive peptides, such as IGF-1 and TGF-β, are less sensitive [276–278]. Consequently, IgG and LF are often used as markers of BC bioactivity. q y g y There are similar challenges in the preservation of the natural bioactivity of human or bovine milk concentrates. In NEC-sensitive preterm pigs, untreated or gently sterilized (ultraviolet-light lower-temperature sterilization) human donor milk or bovine whey pro- tein concentrates have superior effects to protect against NEC, intestinal inﬂammation and systemic infections, relative to products subjected to more intense heat [225,253,279,280]. The same may be true for BC. Thus, care should be taken not to eliminate colostral bioac- tivity during industrial processing, despite the need to provide a microbiologically safe product, for infection-sensitive newborn infants. The product- and host-related effects of different processing steps have been exten- sively studied for dairy products (i.e., freezing, thawing, ultra-ﬁltration, heat- or freeze- drying, food matrix effects, heat pasteurization, ultrahigh heat treatment, irradiation, and storage conditions with regards to temperature, oxygen pressure, and moisture). Much less is known about colostrum, but such processing steps are critical to ensure microbi- ological safety. The processing effects may differ between mature milk and colostrum due to widely differing composition of both nutrients and heat-sensitive bioactive factors (Figure 3) [3]. Concern for damage to BC by standard heat treatment is not primarily related to macronutrient levels (e.g., protein, lipid, carbohydrate, minerals, and vitamins), but to effects on heat-sensitive proteins that may lose their biological function. Thus, denaturation of colostral proteins by standard heat treatment will not likely reduce amino acid absorption kinetics, which is determined more by protein composition (casein and whey), as shown in piglets [35]. 4.6. Chemotherapy-Induced Mucositis Childhood leukemia is the most common cancer found in children. Gastrointestinal mucositis is a common adverse advent of cytotoxic anticancer treatment that increase mor- bidity and mortality. Chemotherapy-induced mucositis (CIM) is an inﬂammatory process that affects mucosal surfaces and submucosal layers, and is often considered the most serious side effect of cancer treatment. Beyond the problem of mucosal inﬂammation itself, CIM may adversely affect nutrient and ﬂuid absorption, inﬂuence endocrine functions of the intestine, and increase intestinal permeability. CIM also affects innate and adaptive immune functions of the developing intestine, leading to malnutrition, microbial translo- cation, and systemic inﬂammatory complications [269]. Considering these complications, coupled with the immune-depressing effects of the cancer itself (especially in leukemia), it is not surprising that BC has been speculated to alleviate the complications of CIM in pediatric patients. In a recent study on childhood acute lymphoblastic leukemia (ALL), children with newly diagnosed ALL were randomized to receive daily BC or placebo supplements during 4 weeks of chemotherapy induction treatment (total n = 62). No differences were found for days with fever (primary outcome), bacteremia, or inﬂammation, but peak severity of oral mucositis was reduced by BC supplementation, despite that only small amounts of BC (0.5–1 g/kg/d) were ingested by these severely ill pediatric patients [270]. g g g y y p p In piglet models of pediatric CIM (without the associated cancer), whole BC had beneﬁcial effects on gut responses after chemotherapy compared with infant formula, but less when compared with intact bovine milk. BC ameliorated the short-term effects of doxorubicin [271–273] and also of the more intense myeloablative chemotherapy treatment Nutrients 2021, 13, 2551 23 of 41 23 of 41 using busulfan and cyclophosphamide [274,275]. Thus, BC may support the general nutritional status, but it may also promote speciﬁc gut mucosal integrity via antimicrobial and endotoxin-neutralizing effects, suppression of gut inﬂammation, and promotion of mucosal tissue repair. 5. Processing of Bovine Colostrum for Use in Pediatrics In individuals with immature digestive function (preterm infants), protein digestibility and amino acid absorption may be improved by heat-induced protein denaturation, as shown in infants after ultrahigh temperature (UHT) treatment of formula [281]. Differential clinical responses to BC supplementation of infants and children in the previously reported clinical trials (e.g., gut, immunity, infections, and growth) may arise from processing-related differences in the contents of BC bioactives. Bioactivity of dif- ferent commercially available BC products varies markedly, as measured in vitro (up to 6-fold) [282]. This inconsistency in bioactivity is a great concern for use of BC as therapy for hospitalized infants and children. The factors in BC that support gut development, immunity and survival in other species are heat-sensitive. When whey protein concen- trates are severely heated (80–85 ◦C, 20–30 min), a reduction in gut-protective effects is observed [279,280]. In contrast, raw, frozen, and mildly heat-treated BC may retain its bioactivity to protect against gut disorders and inﬂammation, as demonstrated in newborn term pigs [144,224,283,284]. In addition to heat pasteurization, milk and colostrum components are exposed to signiﬁcant heat during the spray-drying process to obtain a powdered dry product for easy handling and long shelf life. This process, with or without vacuum-concentration, Nutrients 2021, 13, 2551 24 of 41 24 of 41 induces relatively minor damage to milk proteins [285], probably due to the large surface area of milk particles during the evaporation and drying process. Both in vitro cell stud- ies and in vivo studies in immature pigs, analyzing fresh and spray-dried BC, support this [224,232,254,276]. Conversely, one or several pasteurization steps, especially using the fast high-temperature method (15 s, 72 ◦C), relative to the more gentle, slow low- temperature Holder pasteurization method (30 min, 63 ◦C), induces much greater damage to many colostral proteins [254,276]. In our studies, attempts to secure a near-sterile in- tact BC product for preterm infants by exposing the product to additional irradiation, induced further protein and amino acid modiﬁcations, potentially affecting both nutrient availability and product bioactivity. On the other hand, it can eliminate common low- density pathogens in processed milk powders, such as Bacillus cereus, that often escape heat pasteurization steps [254]. While it remains unclear if different processing steps reduce BC bioactivity for infants and children in vivo, it remains critical to balance the need for microbiological safety against processing-related damage to milk proteins. 5. Processing of Bovine Colostrum for Use in Pediatrics In the ongoing preterm infant studies [154,155,209], we used a low-temperature pasteur- ized, gamma-irradiated, spray-dried BC product, carefully characterized for its bioactivity in vitro [250,254,276] and in vivo in preterm pigs [232]. The obligation to secure microbiological safety of human donor milk (typically by heat pasteurization) for sensitive preterm newborn infants may decrease its natural protective bioactivity. This may present an option to use BC as a supplement or a fortiﬁer to human donor milk. Pasteurization-induced loss of milk bioactivity will occur for both ﬂuids, but because BC is much richer in many bioactive components, its protective effects may be retained after processing, as shown in preterm pigs studies [232]. Among the available dietary choices for preterm infants, it is also clear that infant formula is inferior in its gut-protective effects to both BC and human donor milk [173]. Alternative methods to secure microbiological safety, such as exposure to ultraviolet light, have been tested, and provide superior quality of human milk for preterm pigs [253], but the general use of this method is still limited. Adverse effects of pasteurization techniques, interacting with long and inappropriate storage conditions (warm, moist, and light), may further accelerate protein modiﬁcations that inevitably decrease nutritional value and bioactivity of BC. In recent decades, ultrahigh temperature (UHT)-treated ready-to-feed liquid milk formulas have been widely used for hospitalized infants. Studies indicate that the extra heat-treatment steps for such products, combined with long storage, induce signiﬁcant damage to milk bioactivity by modifying amino acids and high levels of Maillard reaction to components [286,287]. Ultrahigh-temperature treatment of liquid BC is therefore not a preferred method as part of BC processing for sensitive infants, despite the beneﬁts for easy handling, feeding, and long-term storage. 6. Product Consistency and Quality Any medical food or nutritional product used in patients, within or outside hospitals, must have a consistent product composition and quality, especially with respect to nutrient and microbiological proﬁles, but also environmental contaminants. This is a challenge for BC because its composition changes markedly over the ﬁrst two days of lactation after calving, with rapidly decreasing amounts of many bioactive components [3,277]. Unless the collection time from cows is tightly controlled and restricted, with ﬁnal products pooled from many farms and cows, this may result in varying BC product compositions, as indi- cated in a recent survey [282]. In addition, BC composition will vary among different cattle breeds, seasons of the year, cow nutrition, cow parity, and herd health conditions, together affecting product consistency for both nutrients and bioactive constituents. Importantly, it cannot be excluded that cows reared in an environment with pollutants entering the food chain (e.g., heavy metals or persistent organic pollutants, POPs [288]), would also excrete these pollutants in their colostrum. These would likely be excreted in higher amounts in colostrum than in the later mature milk due to the leaky mammary epithelium at parturi- tion [2]. In this perspective, it is critical that colostrum collected from cows treated with Nutrients 2021, 13, 2551 25 of 41 25 of 41 antibiotics is tested for antibiotic residues, because any antibiotic residue may be excreted particularly in colostrum. This seems unlikely since most dairies and companies collecting BC for human use need to adhere to strict guidelines for screening antibiotic residues [289]. Thus, bovine colostrum products need to be carefully analyzed to prevent infants and children from exposure to antibiotic residues [290]. Further studies are needed to establish when unwanted components in colostrum, human, or bovine induce health or growth complications short- and longer-term, particularly in sensitive preterm infants [288]. antibiotics is tested for antibiotic residues, because any antibiotic residue may be excreted particularly in colostrum. This seems unlikely since most dairies and companies collecting BC for human use need to adhere to strict guidelines for screening antibiotic residues [289]. Thus, bovine colostrum products need to be carefully analyzed to prevent infants and children from exposure to antibiotic residues [290]. Further studies are needed to establish when unwanted components in colostrum, human, or bovine induce health or growth complications short- and longer-term, particularly in sensitive preterm infants [288]. 7. Possible Health Risks of Using Bovine Colostrum in Pediatrics Any novel dietary supplement fed to infants and children must be carefully checked for its short- and longer-term safety. Safety concerns regarding BC are particularly rele- vant for the most vulnerable newborns and children with serious diseases. Some com- mon reservations against use of BC in pediatrics include widely varying product supply, quality, purity and microbiological safety (see previous section), increased child allergy risk, excessive supply of nutrients (protein), nutrient deﬁciencies (certain amino acids, oligosaccharides, and minerals), excessive supply of growth factors, and inhibition of drug absorption. Most of these concerns are poorly documented in the literature, but it remains important to consider how BC can be used safely, in the optimal formulation and intake levels, at the right time and age, and only in pediatric patients that will in fact beneﬁt from BC supplementation. Several commercial colostrum products are available and were cited in this review and previous reviews [282]. Some products available online are marketed for use in infants and toddlers, but the evidence to support the safety of these products is usually lacking. A commercial bovine colostrum product produced by a US based biotechnology company recently self-afﬁrmed GRAS (Generally Recognized as Safe) after being reviewed by an independent expert panel of scientists. For a substance to be given GRAS determination, scientiﬁc data and information about its use and its chemical, toxicological, and microbiological properties must be widely known, and the safety under the conditions of use are approved by consensus among qualiﬁed experts. 6. Product Consistency and Quality The highly variable production conditions add to the concern for adequate micro- biological safety, with possible contamination of the product at farms and/or during processing and packaging (see previous section). Use of BC in pediatrics should normally meet the microbiological safety standards for milk powders. In some countries, microbio- logical safety standards may be particularly strict for certain microbes when powders are used for sensitive newborn infants (e.g., E. sakazakii, E. coli, and Bacillus cereus). Variable composition, risk of microbiological contaminants, and possible pollutants are factors that require strict quality control with BC products for use in human patients, especially infants and children. In this regard, quality control of BC should be similar to that of infant formula and products derived from human donor milk [291–293]. 7.1. Cow’s Milk Allergy Cow’s milk allergy (CMA) is a very well-studied pediatric condition and affects 2 to 3% of infants in developed countries [14,294] and decreases rapidly after the ﬁrst years of life. Clinical CMA symptoms are typically seen within few hours of ingesting formula based on cow’s milk and can be IgE or non-IgE mediated. FPIES [16] is part of a larger cluster of non-IgE mediated allergic disease conditions with a prevalence <1%, often related to cow’s milk protein, and affecting the GI tract (e.g., rectal bleeding) [295]. While these allergic conditions in term children may show similarity with NEC in preterm infants, they are likely distinct diseases with limited overlap in predisposing factors and disease mechanisms [15,296]. There is currently no reason to suspect that the risk of developing CMA is greater for BC than for formula or intact cow’s milk. Many bovine milk proteins (>25) can cause allergy and casein subtypes; alpha-lactalbumin and beta-lactoglobulin are potent allergens [297]. These groups are present in similar amounts in BC and milk, while the large amount of non-allergenic immunoglobulin G [298] is present only in BC (Figure 1). Nutrients 2021, 13, 2551 26 of 41 26 of 41 In our pilot preterm infant studies, bovine immunoglobulin G (bIgG) was not detected in the plasma of preterm infants supplemented with BC [155], and no IgE-mediated allergy was detected for BC-supplemented infants with short-bowel syndrome [263]. Even if bIgG is detected systemically in infants, this would not likely cause allergic reactions. Many human vaccines are developed from Igs raised in other species [42]. In weanling piglets fed BC-rich diets, no anti-bovine immunoglobulin could be detected in plasma although BC modulated the lymphocyte populations locally in the gut [185]. It remains unknown why a commercial IgG-containing BC-like product (6% IgG) supplemented to the diet of preterm infants in India (400 mg IgG/kg/day) was associated with a trend to more gut inﬂammation [211], but this is not likely to result from bIgG-induced NEC or FPIES symp- toms. Preterm infants have an immature immune system that is skewed towards tolerance to foreign proteins, potentially making them more tolerant to introduction of bovine milk, and thereby also colostral proteins from cows, than normal infants. 7.2. Recommended Safe and Effective Intake Levels 7.2. Recommended Safe and Effective Intake Levels Widely differing intakes of BC have been applied when powdered, whole, or frac- tionated BC products were supplemented to formula or mother’s own milk for infants and children (0.1–3 g/kg/d, 5–75% of recommended daily protein intake). Typically, the highest intake levels were used in the ﬁrst week of life for preterm or underweight in- fants [155,209,211,212], having a deﬁcient supply of own mother’s milk. These infants also have the highest daily protein requirements for metabolism and growth (~4 g/kg/d). Sup- plementary BC (containing 50–70% protein mainly as IgG) is an attractive way to provide more enteral protein for gut and body growth and metabolism in preterm infants reared on human milk, which is normally considered deﬁcient in protein. However, because the highly enriched IgG may be partly undigested to maintain bioactivity [19], the bioavailabil- ity of amino acids in colostrum protein may be less than in highly digestible hydrolyzed protein in infant formulas. This potential issue of amino acid bioavailability of BC proteins highlight a gap in our current knowledge and warrants further investigation, especially in preterm infants with a high protein requirement and an immature digestive capacity. While a low level of supplementary BC may be nutritionally and immunologically in- effective, excessive BC intake could raise concerns about possible protein toxicity, especially for preterm infants [300]. In such infants, excessive protein may pose a risk for metabolic disturbance, azotemia, and acidosis, in part related to immature metabolic, hepatic, and renal functions. This dysfunction may, in turn, induce brain defects via toxicity of certain amino acids (phenylalanine, glycine, and methionine [301]). In preterm infants, excessive protein supply may ﬁrst raise plasma levels of an aromatic amino acid, tyrosine. While such feeding-induced ‘hypertyrosinemia’ is not generally considered toxic [302], tyrosine levels may be used as a marker of excessive BC supply to preterm infants during the ﬁrst weeks after birth [154,155]. Plasma tyrosine levels increase in preterm pigs following BC supplementation during the ﬁrst week, consistent with the relatively high tyrosine levels in BC [229]. Despite these concerns, amino acid toxicity is a rare condition, even in preterm infants, and normally excess amino acid supply would be effectively excreted as urea [300]. 7.1. Cow’s Milk Allergy Consequently, the prevalence of CMA in preterm infants tends to be lower, not higher, than for normal infants, indicating that preterm infants are not at increased risk of developing CMA after exposure to BC or milk [299]. Advice against using BC for infants and children, based on CMA concerns, is therefore relevant mainly for individuals already diagnosed with CMA. 7.2. Recommended Safe and Effective Intake Levels In preterm infants, moderately high blood urea levels (e.g., 5–10 mM) may indicate immature liver and kidney functions and/or excess protein supply, but such urea levels are not toxic, and blood urea level alone is a poor marker of adequate protein intake for growth and protein intake in preterm infants [301]. Collectively, protein toxicity is unlikely to be a problem following BC supplementation to human milk or formula, at least not within the normal upper limits of daily protein intake for infants. Normal gut motility is diet-dependent and critical to avoid maldigestion, constipation, and diarrhea and GI inﬂammatory reactions. For sensitive pediatric patients, too much Nutrients 2021, 13, 2551 27 of 41 27 of 41 BC-derived casein may adversely affect GI motility and delay gastric emptying due to casein clotting in the stomach. Thus, casein-based pediatric formulas often consist of partly hydrolyzed proteins [303,304] to avoid feeding intolerance, gastric residuals, and constipa- tion, as often observed in preterm infants [305–307]. Low-dose BC supplementation within normal protein limits of preterm infants is unlikely to induce casein-related dysmotility and ongoing clinical trials will test this [154,209,224]. In preterm pigs, exclusive BC feeding was associated with accumulation of casein clots, and more gastric mucosal lesions in the ﬁrst week of life when the gut was still very immature [164,173,174,233]. However, neither casein supplementation to whey-based formula nor rapid advancement of a diet with casein (intact or hydrolyzed) increased the NEC sensitivity in preterm pigs [304,308]. In fact, a moderate BC-induced retention of digesta in the stomach may promote digestion and gut maturation, relative to whey-based formulas, as indicated from food transit studies in preterm pigs [309]. The available literature on from infants and children indicate that the optimal time, duration, and intake level of BC supplementation may differ among different clinical conditions, and much research is lacking to deﬁne optimal treatments. Pediatric animal models may help to evaluate if excessive or prolonged BC supplementation induces nutritional deﬁciencies and/or health risks for infants and children. Consistent with its biological role to secure nutrition and immunological protection of newborn calves, there may be limited nutritional beneﬁt of BC when provided in high amounts beyond the ﬁrst weeks of life although in preterm pigs exclusive BC feeding is effective to promote growth, gut, immunity, and microbiota development beyond 10 days, at least relative to formula [174,239]. 7.2. Recommended Safe and Effective Intake Levels In the ﬁrst week of life, BC supplementation to formula-fed preterm pigs is most NEC-protective when fed in amounts of >50% of the total diet [240], but optimal doses later are less clear, as also shown in chemotherapy-treated term piglets [273]. In a single study, preterm pigs fed exclusive BC in high amounts for seven days showed reduced NEC incidence but abnormal body growth and blood chemistry values (e.g., high blood Na, urea, and lipids) [310]. In all pediatric studies, much less BC was provided than in these animal studies. Relatively high intake levels (8 g/kg/day) of a BC-like highly processed product (6% IgG) for 3 weeks to very preterm infants in India may have led to high osmolality and the reported tendency to adverse NEC effects, but limited product information was provided in the published report [211]. Clearly, more research is required to deﬁne when and how much BC, provided in intact or fractionated form and after different processing, should be given to speciﬁc pediatric patients to optimize their nutrition, growth, and health outcomes. 8. Conclusions This review provides an extensive overview of the human and animal studies that have examined the potential beneﬁts of feeding BC to infants and children, mainly with regards to its effect in the GI tract. A number of studies in term infants and children have fed BC as a supplement to infant formula or donor human milk with limited adverse effects. More recent studies in preterm infants fed a low supplemental BC intake showed no adverse effects, whereas one study with high intake of a processed BC product, with relatively low IgG levels, produced a tendency to adverse intestinal inﬂammatory effects. The limited number of studies in infants, and several studies in term and preterm piglets, suggest that feeding exclusive pasteurized BC, or as a supplement to infant formula, protects against several GI diseases, such as rotavirus diarrhea, NEC, sepsis, and CIM mucositis. The evidence suggests beneﬁts and safety especially from hyperimmune BC products directed towards speciﬁc pathogens in infants and children. The mechanisms explored to explain the beneﬁts of BC on GI diseases frequently suggest antimicrobial and immune-stimulating functions of Igs, potentially in synergy with other bioactives, such as LF. Nevertheless, it remains unclear how all of these BC bioactive factors act independently or in synergy to promote growth and gut mucosal protection. While safe and effective use of single BC constituents, especially Igs, is reported, the complexity of BC and the variety Nutrients 2021, 13, 2551 28 of 41 28 of 41 of clinical conditions provide an argument for using intact BC, rather than fractions or isolated components from BC for infants and children. of clinical conditions provide an argument for using intact BC, rather than fractions or isolated components from BC for infants and children. The few existing infant studies, together with numerous preclinical animal and cell culture studies, provide a framework to stimulate further investigation of feeding BC in human infants and children. This option has become increasingly possible with the commercial availability of BC preparations to be used as medical foods. This question should also be addressed considering the fact that processed bovine milk products (infant formula) have been fed to infants for >100 years and to very preterm infants for >50 years. Supplementation with BC should only be considered when mother’s own milk or donor human milk are not available or clearly insufﬁcient in nutrients or protective factors. 8. Conclusions We need safe, effective, and economically viable alternatives to processed infant formulas, which, for preterm infants, is well documented to increase the risk of many diseases, including NEC and sepsis. It is exceedingly difﬁcult to prove both safety and clear health-promoting effects of new clinical procedures, drugs, or nutritional therapies for hospitalized infants and children. Generally, patient populations are small and the associated complications to test safety and efﬁcacy of new interventions develop with low frequency. Another therapy that has recently been considered for infants and children is probiotics, which only a few decades ago was considered unsafe for immunocompromised preterm infants. Based on the evidence from dozens of randomized controlled trials, many neonatal units are feeding probiotics to hospitalized preterm infants, even though they are not ofﬁcially approved for such use, in either Europe or United States. The same applies for numerous drugs given to fragile and critically ill infants and children that have not been tested in randomized controlled trials for safety of the intended indication. Given that bovine-milk- based formulas have been given exclusively, or as supplements, to term infants and children for centuries, we suggest that further studies are warranted to test whether pasteurized BC can be a safe and effective nutritional supplement for a number of pediatric patient groups, providing improved immunity and gut health. At the same time, research is needed to better identify the most responsive patient groups and target diseases (e.g., preterm infants with high NEC/sepsis sensitivity, and children with infections and gut complications); the optimal timing, age, and dose of BC; and the optimal BC product for each condition (e.g., optimally processed intact or fractionated BC). Author Contributions: Conceptualization, writing—original draft preparation and writing—review and editing, P.T.S., C.V., V.M.H. and D.G.B. All authors have read and agreed to the published version of the manuscript. Funding: We acknowledge the ﬁnancial support of Innovation Fund Denmark, Bioﬁber Damino, and University of Copenhagen, together sponsoring the majority of all studies related to bovine colostrum (P.T.S.). Finalizing this manuscript was supported by funds from PanTheryx to University of Copen- hagen (P.T.S.) and by federal funds from the USDA, Agricultural Research Service under Cooperative Agreement Number 58-6250-6-001 and the National Institutes of Health Grant DK-094616 (D.G.B.). 8. Conclusions Acknowledgments: We acknowledge the generous support and critical comments of all the persons involved in the many studies on BC products and use of BC in piglet models and human infants, including (in alphabetical order): Lise Aunsholt, Stine Bering, Randal Buddington, Ole Bæk, Dereck Chatterton, Malene Cilieborg, Fei Gao, Gorm Greisen, Michael Ladegaard, Duc Ninh Nguyen, Xiaoyu Pan, Rene Shen, Susanne Søndergaard, Thomas Thymann, Xudong Yan, Peng Yuemeng, Gitte Zacchariasen, and Mette Østergaard. We thank Gunner Jakobsen, Arne Høst, and Murali Premkumar for critical review of this manuscript. Conﬂicts of Interest: Caitlin Vonderohe has received an honorarium for preparation of this manuscript. All other authors declare that they have not received any personal remuneration for scientiﬁc or pri- vate work related to bovine colostrum products and their biological effects. University of Copenhagen has received research grants related to use of bovine colostrum for pediatric patients (see above). The University of Copenhagen holds a patent on the use of bovine colostrum for pediatric patients. Per T. Sangild is listed as a sole inventor but has declined any share of potential revenue arising from commercial exploitation of such a patent. Other authors declare no other conﬂict of interest. 29 of 41 29 of 41 Nutrients 2021, 13, 2551 References 1. Hammon, H.M.; Liermann, W.; Frieten, D.; Koch, C. Review: Importance of colostrum supply and milk feeding intensity on gastrointestinal and systemic development in calves. Animal 2020, 14, s133–s143. [CrossRef] g y p 2. Wall, S.K.; Gross, J.J.; Kessler, E.C.; Villez, K.; Bruckmaier, R.M. Blood-derived proteins in milk at start of lactation: Indicators of active or passive transfer. J. Dairy Sci. 2015, 98, 7748–7756. [CrossRef] 3. Playford, R.J.; Weiser, M.J. Bovine Colostrum: Its Constituents and Uses. Nutrients 2021, 13, 265. [CrossRef] [PubMed] 4. Ballard, O.; Morrow, A.L. Human milk composition: Nutrients and bioactive factors. Pediatr. Clin. North Am. 2013, 60, 49–74. [CrossRef] [PubMed] [ ] [ ] 5. Goldman, A.S. Evolution of the mammary gland defense system and the ontogeny of the immune system. J. Mammary Gland. Biol. Neoplasia 2002, 7, 277–289. [CrossRef] 6. Stevens, E.E.; Patrick, T.E.; Pickler, R. A history of infant feeding. J. Perinat. Educ. 2009, 18, 32–39. [Cro 7. Riley, L.K.; Rupert, J.; Boucher, O. Nutrition in Toddlers. Am. Fam. Physician 2018, 98, 227–233. [PubMe y p y 8. Clark, D.C.; Cifelli, C.J.; Pikosky, M.A. Growth and Development of Preschool Children (12–60 Months): A Review of the Effect of Dairy Intake. Nutrients 2020, 12, 3556. [CrossRef] y 9. Lucas, A. Scientiﬁc Evidence for Breastfeeding. Nestle Nutr. Inst. Workshop Ser. 2019, 90, 1–12. [CrossRef] 10 S ti B tf di B tf di d th f h ilk P di t i 2012 129 827 841 [C R f] [P bM d] 9. Lucas, A. Scientiﬁc Evidence for Breastfeeding. Nestle Nutr. Inst. Workshop Ser. 2019, 90, 1–12. [CrossRef] 10. Section on Breastfeeding. Breastfeeding and the use of human milk. Pediatrics 2012, 129, e827–e841. [CrossRef] [PubMed] Lucas, A. Scientiﬁc Evidence for Breastfeeding. N 10. Section on Breastfeeding. Breastfeeding and the use of human milk. Pediatrics 2012, 129, e827–e841. [CrossRef] [PubMed] 11. Maffei, D.; Schanler, R.J. Human milk is the feeding strategy to prevent necrotizing enterocolitis! Semin. Perinatol. 2017, 41, 36–40. [CrossRef] g g 11. Maffei, D.; Schanler, R.J. Human milk is the feeding strategy to prevent necrotizing enterocolitis! Semin. Perinatol. 2017, 41, 36–40. [CrossRef] 12. Quigley, M.; Embleton, N.D.; McGuire, W. Formula versus donor breast milk for feeding preterm or low birth weight infants. Cochrane Database Syst. Rev. 2018, 6, Cd002971. [CrossRef] [PubMed] 13. Ananthan, A.; Balasubramanian, H.; Rao, S.; Patole, S. Human Milk-Derived Fortiﬁers Compared with Bovine Milk-Derived Fortiﬁers in Preterm Infants: A Systematic Review and Meta-Analysis. References Adv. Nutr. 2020, 11, 1325–1333. [CrossRef] Cow’s milk allergy: Where have we come from and where are we going? Endocr. Metab. Immune Disord. Drug [CrossRef] 14. Høst, A.; Halken, S. Cow’s milk allergy: Where have we come from and where are we going? Endocr Targets 2014, 14, 2–8. [CrossRef] g 15. Burris, A.D.; Burris, J.; Järvinen, K.M. Cow’s Milk Protein Allergy in Term and Preterm Infants: Clinical Manifestations, Immunologic Pathophysiology, and Management Strategies. NeoReviews 2020, 21, e795–e808. [CrossRef] g p y gy g g 16. Michelet, M.; Schluckebier, D.; Petit, L.M.; Caubet, J.C. Food protein-induced enterocolitis syndrome—A review of the literature with focus on clinical management. J. Asthma Allergy 2017, 10, 197–207. [CrossRef] [PubMed] 17. Canvasser, J.; Hair, A.B.; Kim, J.H.; Taylor, S.N. Parent and Provider Perspectives on the Imprecise Label of “Human Milk Fortiﬁer” in the NICU. Nutrients 2020, 12, 720. [CrossRef] [PubMed] 18. Korhonen, H.; Marnila, P.; Gill, H.S. Bovine milk antibodies for health. Br. J. Nutr. 2000, 84 (Suppl. S1), S135–S146. [CrossRef] 19. Ulfman, L.H.; Leusen, J.H.W.; Savelkoul, H.F.J.; Warner, J.O.; van Neerven, R.J.J. Effects of Bovine Immunoglobulins on Immune Function, Allergy, and Infection. Front. Nutr. 2018, 5, 52. [CrossRef] [PubMed] , ; , ; , J , ( pp ), [ ] 19. Ulfman, L.H.; Leusen, J.H.W.; Savelkoul, H.F.J.; Warner, J.O.; van Neerven, R.J.J. Effects of Bovine Immunoglobulins on Immune Function, Allergy, and Infection. Front. Nutr. 2018, 5, 52. [CrossRef] [PubMed] 19. Ulfman, L.H.; Leusen, J.H.W.; Savelkoul, H.F.J.; Warner, J.O.; van Neerven, R.J.J. Effects of Bov Function, Allergy, and Infection. Front. Nutr. 2018, 5, 52. [CrossRef] [PubMed] gy 20. Rathe, M.; Müller, K.; Sangild, P.T.; Husby, S. Clinical applications of bovine colostrum therapy: A systematic review. Nutr. Rev. 2014, 72, 237–254. [CrossRef] 21. Playford, R.J.; Macdonald, C.E.; Johnson, W.S. Colostrum and milk-derived peptide growth factors for the treatment of gastroin- testinal disorders. Am. J. Clin. Nutr. 2000, 72, 5–14. [CrossRef] 22. Stelwagen, K.; Carpenter, E.; Haigh, B.; Hodgkinson, A.; Wheeler, T.T. Immune components of bovine colostrum and milk. J. Anim. Sci. 2009, 87, 3–9. [CrossRef] ; Theil, P.K. Perspectives on immunoglobulins in colostrum and milk. Nutrients 2011, 3, 442–474. [CrossRef] 23. Hurley, W.L.; Theil, P.K. Perspectives on immunoglobulins in colostrum and milk. Nutrients 2011, 3, 44 23. Hurley, W.L.; Theil, P.K. Perspective 24. Puppel, K.; Golebiewski, M.; Grodkowski, G.; Slosarz, J.; Kunowska-Slosarz, M.; Solarczyk, P.; Lukasiewicz, M.; Balcerak, M.; Przysucha, T. p p j 32. Dupont, D.; Tomé, D. Milk proteins: Digestion and absorption in the gastrointestinal tract. In Milk Proteins; Academic Press: Cambridge, MA, USA, 2020; pp. 701–714. [CrossRef] 31. Bhat, M.Y.; Dar, T.A.; Singh, L.R. Casein Proteins: Structural and Functional Aspects. In Milk Proteins—From Structure to Biological Properties and Health Aspects; InTech: Rijeka, Croatia, 2016. [CrossRef] é lk d b h l ilk i A d References [CrossRef] [PubMed] g 48. Axelsson, I.; Jakobsson, I.; Lindberg, T.; Polberger, S.; Benediktsson, B.; Raiha, N. Macromolecular absorption in preterm and term infants. Acta Paediatr. 1989, 78, 532–537. [CrossRef] [PubMed] g 48. Axelsson, I.; Jakobsson, I.; Lindberg, T.; Polberger, S.; Benediktsson, B.; Raiha, N. Macromolecular a infants. Acta Paediatr. 1989, 78, 532–537. [CrossRef] [PubMed] 48. Axelsson, I.; Jakobsson, I.; Lindberg, T.; Polberger, S.; Benediktsson infants. Acta Paediatr. 1989, 78, 532–537. [CrossRef] [PubMed] 49. Simister, N.E. Placental transport of immunoglobulin G. Vaccine 2003, 21, 3365–3369. [CrossRef] 50. Van de Perre, P. Transfer of antibody via mother’s milk. Vaccine 2003, 21, 3374–3376. [CrossRef] 50. Van de Perre, P. Transfer of antibody via mother’s milk. Vaccine 2003, 21, 3374–3376. [CrossRef] y 51. Sangild, P.T. Uptake of colostral immunoglobulins by the compromised newborn farm animal. Acta Vet. Scandinavica. Suppl. 2003, 98, 105–122. [CrossRef] Sangild, P.T. Gut responses to enteral nutrition in preterm infants and animals. Exp. Biol. Med. 2006, 231, 1 52. Sangild, P.T. Gut responses to enteral nutrition in preterm infants and animals. Exp. Biol. Med. 2006, 231, 1695–1711. [CrossRef] 53. Westrom, B.; Arevalo Sureda, E.; Pierzynowska, K.; Pierzynowski, S.G.; Perez-Cano, F.J. The Immature Gut Barrier and Its Importance in Establishing Immunity in Newborn Mammals. Front. Immunol. 2020, 11, 1153. [CrossRef] g p p p 53. Westrom, B.; Arevalo Sureda, E.; Pierzynowska, K.; Pierzynowski, S.G.; Perez-Cano, F.J. The Immature Gut Barrier and Its Importance in Establishing Immunity in Newborn Mammals. Front. Immunol. 2020, 11, 1153. [CrossRef] p g y [ ] 54. Stirling, C.M.; Charleston, B.; Takamatsu, H.; Claypool, S.; Lencer, W.; Blumberg, R.S.; Wileman, T.E. Characterization of the porcine neonatal Fc receptor–potential use for trans-epithelial protein delivery. Immunology 2005, 114, 542–553. [CrossRef] 55. Drew, M.D.; Owen, B.D. The provision of passive immunity to colostrum-deprived piglets by b munoglobulins. Can. J. Anim Sci. 1988, 68, 1277–1284. [CrossRef] g 56. Pyzik, M.; Sand, K.M.K.; Hubbard, J.J.; Andersen, J.T.; Sandlie, I.; Blumberg, R.S. The Neonatal Fc Receptor (FcRn): A Misnomer? Front. Immunol. 2019, 10, 1540. [CrossRef] 57. Shah, U.; Dickinson, B.L.; Blumberg, R.S.; Simister, N.E.; Lencer, W.I.; Walker, W.A. Distribution of the IgG Fc receptor, FcRn, in the human fetal intestine. Pediatr. Res. 2003, 53, 295–301. [CrossRef] 58. Yoshida, M.; Kobayashi, K.; Kuo, T.T.; Bry, L.; Glickman, J.N.; Claypool, S.M.; Kaser, A.; Nagaishi, T.; Higgins, D.E.; Mizoguchi, E.; et al. Neonatal Fc receptor for IgG regulates mucosal immune responses to luminal bacteria. J. Clin. Investig. References Composition and Factors Affecting Quality of Bovine Colostrum: A Review. Animals 2019, 9, 1070. [CrossRef] 25. Davis, T.A.; Nguyen, H.V.; Garcia-Bravo, R.; Fiorotto, M.L.; Jackson, E.M.; Reeds, P.J. Amino acid composition of the milk of some mammalian species changes with stage of lactation. Br. J. Nutr. 1994, 72, 845–853. [CrossRef] 26. Lonnerdal, B.; Erdmann, P.; Thakkar, S.K.; Sauser, J.; Destaillats, F. Longitudinal evolution of true p bioactive proteins in breast milk: A developmental perspective. J. Nutr. Biochem. 2017, 41, 1–11. [Cross 27. Ahern, G.J.; Hennessy, A.A.; Ryan, C.A.; Ross, R.P.; Stanton, C. Advances in Infant Formula Science. Annu. Rev. Food Sci. Technol. 2019, 10, 75–102. [CrossRef] [PubMed] 28. Godden, S.M.; Lombard, J.E.; Woolums, A.R. Colostrum Management for Dairy Calves. Vet. Clin. N. Am. Food Anim. Pract. 2019, 35, 535–556. [CrossRef] 29. West, D.W.D.; Mitchell, C.J. Tracking the Fate of Milk Proteins: Better in Whole or in Part? J. Nutr. 2020, 150, 2001–2002. [CrossRef] [PubMed] 30. Martin, C.R.; Ling, P.R.; Blackburn, G.L. Review of Infant Feeding: Key Features of Breast Milk and Infant Formula. Nutrients 2016, 8, 279. [CrossRef] 31. Bhat, M.Y.; Dar, T.A.; Singh, L.R. Casein Proteins: Structural and Functional Aspects. In Milk Proteins—From Structure to Biological Properties and Health Aspects; InTech: Rijeka, Croatia, 2016. [CrossRef] 32 Dupont D ; Tomé D Milk proteins: Digestion and absorption in the gastrointestinal tract In Milk Proteins; Academic Press: 31. Bhat, M.Y.; Dar, T.A.; Singh, L.R. Casein Proteins: Structural and Functional Aspects. In Milk Proteins—F Properties and Health Aspects; InTech: Rijeka, Croatia, 2016. [CrossRef] 30 of 41 Nutrients 2021, 13, 2551 33. Boirie, Y.; Dangin, M.; Gachon, P.; Vasson, M.P.; Maubois, J.L.; Beaufrere, B. Slow and fast dietary proteins differently modulate postprandial protein accretion. Proc. Natl. Acad. Sci. USA 1997, 94, 14930–14935. [CrossRef] p p p 34. Bourlieu, C.; Ménard, O.; Bouzerzour, K.; Mandalari, G.; Macierzanka, A.; Mackie, A.R.; Dupont, D. Speciﬁcity of Infant Digestive Conditions: Some Clues for Developing Relevant In Vitro Models. Crit. Rev. Food Sci. Nutr. 2014, 54, 1427–1457. [CrossRef] [PubMed] 35. Welch-Jernigan, R.J.; Abrahamse, E.; Stoll, B.; Smith, O.; Wierenga, P.A.; van de Heijning, B.J.M.; Renes, I.B.; Burrin, D.G. Postprandial Amino Acid Kinetics of Milk Protein Mixtures are Affected by Composition, But Not Denaturation, in Neonatal Piglets. Curr. Dev. Nutr. 2019, 3, nzy102. [CrossRef] [PubMed] 36. Chatterton, D.E.; Nguyen, D.N.; Bering, S.B.; Sangild, P.T. Anti-inﬂammatory mechanisms of bioactive milk proteins in the intestine of newborns. Int. J. Biochem. References Cell Biol. 2013, 45, 1730–1747. [CrossRef] [PubMed] gical effects of novel bovine milk fractions. Nestle Nutr. Workshop Ser. Paediatr. Program. 2011, 67, 41–54 37. Lönnerdal, B. Biological effects of novel bovine milk fractions. Nestle Nutr. Workshop Ser. Paedia [CrossRef] 38. Wada, Y.; Lönnerdal, B. Bioactive peptides derived from human milk proteins—Mechanisms of action. J. Nutr. Biochem. 2014, 25, 503–514. [CrossRef] [ ] 39. Korhonen, H.; Marnila, P.; Gill, H.S. Milk immunoglobulins and complement factors. Br. J. Nutr. 2000, 84 (Suppl. S1), S75–S80. [CrossRef] 40. Cakebread, J.A.; Humphrey, R.; Hodgkinson, A.J. Immunoglobulin A in Bovine Milk: A Potential Functional Food? J. Agric. Food Chem. 2015, 63, 7311–7316. [CrossRef] 41. Klobasa, F.; Werhahn, E.; Butler, J.E. Composition of sow milk during lactation. J. Anim. Sci. 1987, 64, 1458–1466. [CrossRef] 42. Hedegaard, C.J.; Heegaard, P.M. Passive immunisation, an old idea revisited: Basic principles and application to modern animal production systems. Vet. Immunol. Immunopathol. 2016, 174, 50–63. [CrossRef] J p g J [ ] 42. Hedegaard, C.J.; Heegaard, P.M. Passive immunisation, an old idea revisited: Basic principles and application to modern animal production systems. Vet. Immunol. Immunopathol. 2016, 174, 50–63. [CrossRef] p y p [ ] 43. Malek, A.; Sager, R.; Kuhn, P.; Nicolaides, K.H.; Schneider, H. Evolution of maternofetal transport of immunoglobulins during human pregnancy. Am. J. Reprod. Immunol. 1996, 36, 248–255. [CrossRef] [PubMed] 43. Malek, A.; Sager, R.; Kuhn, P.; Nicolaides, K.H.; Schneider, H. Evolution of maternofetal transpor human pregnancy. Am. J. Reprod. Immunol. 1996, 36, 248–255. [CrossRef] [PubMed] 44. Palmeira, P.; Quinello, C.; Silveira-Lessa, A.L.; Zago, C.A.; Carneiro-Sampaio, M. IgG placental transfer in healthy and pathological pregnancies. Clin. Dev. Immunol. 2012, 2012, 985646. [CrossRef] p g 45. Van den Berg, J.P.; Westerbeek, E.A.; van der Klis, F.R.; Berbers, G.A.; van Elburg, R.M. Transplacental transport of IgG antibodies to preterm infants: A review of the literature. Early Hum. Dev. 2011, 87, 67–72. [CrossRef] Lee, J.; Kim, H.S.; Jung, Y.H.; Choi, K.Y.; Shin, S.H.; Kim, E.K.; Choi, J.H. Oropharyngeal colostrum admin t i f t A RCT P di t i 2015 135 357 366 [C R f] 46. Lee, J.; Kim, H.S.; Jung, Y.H.; Choi, K.Y.; Shin, S.H.; Kim, E.K.; Choi, J.H. Orop premature infants: An RCT. Pediatrics 2015, 135, e357–e366. [CrossRef] p 47. Goldblum, R.M.; Schanler, R.J.; Garza, C.; Goldman, A.S. Human milk feeding enhances the urinary excretion of immunologic factors in low birth weight infants. Pediatr. Res. 1989, 25, 184–188. References 2006, 116, 2142–2151. [CrossRef] 59. Hodgkinson, A.J.; Cakebread, J.; Callaghan, M.; Harris, P.; Brunt, R.; Anderson, R.C.; Armstrong, K.M.; Haigh, B. Comparative innate immune interactions of human and bovine secretory IgA with pathogenic and non-pathogenic bacteria. Dev. Comp. Immunol. 2017, 68, 21–25. [CrossRef] [PubMed] nsen, A.R.; Elnif, J.; Burrin, D.G.; Sangild, P.T. Development of intestinal immunoglobulin absorption and en onatal pigs is diet dependent. J. Nutr. 2001, 131, 3259–3265. [CrossRef] [PubMed] ; Burrin, D.G.; Sangild, P.T. Development of intestinal immunoglobulin absorption and enzyme activities in dependent. J. Nutr. 2001, 131, 3259–3265. [CrossRef] [PubMed] p g p 61. Nash, G.S.; MacDermott, R.P.; Schloemann, S.; Bertovich, M.J.; O’Neal, J.; Porter, L.; Kulczycki, A. Bovine IgG1, but not IgG2, binds to human B cells and inhibits antibody secretion. Immunology 1990, 69, 361–366. [PubMed] 31 of 41 31 of 41 Nutrients 2021, 13, 2551 62. Bagwe-Parab, S.; Yadav, P.; Kaur, G.; Tuli, H.S.; Buttar, H.S. Therapeutic Applications of Human and Bovine Colostrum in the Treatment of Gastrointestinal Diseases and Distinctive Cancer Types: The Current Evidence. Front. Pharmacol. 2020, 11, 01100. [CrossRef] [PubMed] 63. Detzel, C.J.; Horgan, A.; Henderson, A.L.; Petschow, B.W.; Warner, C.D.; Maas, K.J.; Weaver, E.M. Bovine immunoglobulin/protein isolate binds pro-inﬂammatory bacterial compounds and prevents immune activation in an intestinal co-culture model. PLoS ONE 2015, 10, e0120278. [CrossRef] 64. Arrouk, R.; Herdes, R.E.; Karpinski, A.C.; Hyman, P.E. Serum-derived bovine immunoglobulin for children with diarrhea- predominant irritable bowel syndrome. Pediatr. Health Med. Ther. 2018, 9, 129–133. [CrossRef] [PubMed] 65. Asmuth, D.M.; Hinkle, J.E.; LaMarca, A.; Fichtenbaum, C.J.; Somsouk, M.; Utay, N.S.; Shaw, A.L.; Petschow, B.W.; Detzel, C.J.; Weaver, E.M. Evaluation of oral serum-derived bovine immunoglobulins in HIV-infected patients with chronic idiopathic diarrhea. HIV Clin. Trials 2017, 18, 205–213. [CrossRef] [PubMed] 66. Petschow, B.W.; Blikslager, A.T.; Weaver, E.M.; Campbell, J.M.; Polo, J.; Shaw, A.L.; Burnett, B.P.; Klein, G.L.; Rhoads, J.M. Bovine immunoglobulin protein isolates for the nutritional management of enteropathy. World J. Gastroenterol. 2014, 20, 11713–11726. [CrossRef] [PubMed] 67. Petschow, B.W.; Burnett, B.; Shaw, A.L.; Weaver, E.M.; Klein, G.L. Serum-derived bovine immunoglobulin/protein isolate: Postulated mechanism of action for management of enteropathy. Clin. Exp. Gastroenterol. 2014, 7, 181–190. [CrossRef] 68. Van Arsdall, M.; Haque, I.; Liu, Y.; Rhoads, J.M. Is There a Role for the Enteral Administration of Serum-Derived Immunoglobulins in Human Gastrointestinal Disease and Pediatric Critical Care Nutrition? Adv. Nutr. 2016, 7, 535–543. [CrossRef] 69. Montagne, P.; Culliere, M.L.; Bene, M.C.; Faure, G. References Lactoferrin as a natural immune modulator. Curr. Pharm. Des. 2009, 15, 1956–1973. [CrossRef] 78. Actor, J.K.; Hwang, S.; Kruzel, M.L. Lactoferrin as a natural immune modulator. Curr. Pharm. Des. 2009, 15, 1956–1973. [CrossRef] 79. Afrazi, A.; Sodhi, C.P.; Richardson, W.; Neal, M.; Misty, G.; Siggers, R.; Hackman, D.J. New insights into pathogenesis and 78. Actor, J.K.; Hwang, S.; Kruzel, M.L. Lactoferrin as a natural immune modulator. Curr. Pharm. Des. 2009, 15, 1956–1973. [CrossRef] 79. Afrazi, A.; Sodhi, C.P.; Richardson, W.; Neal, M.; Misty, G.; Siggers, R.; Hackman, D.J. New insights into pathogenesis and treatment of necrotizing enterocolitis: Toll-like receptors and beyond. Pediatr. Res. 2011, 69, 183–188. [CrossRef] 80. Comstock, S.S.; Reznikov, E.A.; Contractor, N.; Donovan, S.M. Dietary bovine lactoferrin alters mucosal and systemic immune cell responses in neonatal piglets. J. Nutr. 2014, 144, 525–532. [CrossRef] p p g 81. Nguyen, D.N.; Jiang, P.; Stensballe, A.; Bendixen, E.; Sangild, P.T.; Chatterton, D.E. Bovine lactoferrin regulates cell survival, apoptosis and inﬂammation in intestinal epithelial cells and preterm pig intestine. J. Proteom. 2016, 139, 95–102. [CrossRef] 82 N D N Li Y S ild PT B i S B Ch tt t D E Eff t f b i l t f i th i t i i t ti p p g 81. Nguyen, D.N.; Jiang, P.; Stensballe, A.; Bendixen, E.; Sangild, P.T.; Chatterton, D.E. Bovine lactoferrin regulates cell survival, apoptosis and inﬂammation in intestinal epithelial cells and preterm pig intestine. J. Proteom. 2016, 139, 95–102. [CrossRef] 82. Nguyen, D.N.; Li, Y.; Sangild, P.T.; Bering, S.B.; Chatterton, D.E. Effects of bovine lactoferrin on the immature porcine intestine. Br. J. Nutr. 2014, 111, 321–331. [CrossRef] apoptosis and inﬂammation in intestinal epithelial cells and preterm pig intestine. J. Proteom. 2016, 139, 95 102. [CrossRef] 82. Nguyen, D.N.; Li, Y.; Sangild, P.T.; Bering, S.B.; Chatterton, D.E. Effects of bovine lactoferrin on the immature porcine intestine. Br. J. Nutr. 2014, 111, 321–331. [CrossRef] 83. Layman, D.K.; Lonnerdal, B.; Fernstrom, J.D. Applications for alpha-lactalbumin in human nutrition. Nutr. Rev. 2018, 76, 444–460. [CrossRef] 84. Krissansen, G.W. Emerging health properties of whey proteins and their clinical implications. J. Am. Coll. Nutr. 2007, 26, 713s–723s. [CrossRef] [PubMed] 85. Nielsen, C.H.; Hui, Y.; Nguyen, D.N.; Ahnfeldt, A.M.; Burrin, D.G.; Hartmann, B.; Heckmann, A.B.; Sangild, P.T.; Thymann, T.; Bering, S.B. Alpha-Lactalbumin Enriched Whey Protein Concentrate to Improve Gut, Immunity and Brain Development in Preterm Pigs. Nutrients 2020, 12, 245. References Changes in lactoferrin and lysozyme levels in human milk during the ﬁrst twelve weeks of lactation. In Bioactive Components of Human Milk; Springer: Boston, MA, USA, 2001. 70. Donovan, S.M.; Odle, J. Growth factors in milk as mediators of infant development. Annu. Rev. Nutr. 1994, 14, 147–167. [CrossRef] 71. Gislason, J.; Iyer, S.; Hutchens, T.W.; Lonnerdal, B. Lactoferrin receptors in piglet small intestine: Lactoferrin binding properties, ontogeny, and regional distribution in the gastrointestinal tract. J. Nutr. Biochem. 1993, 4, 528–533. [CrossRef] d regional distribution in the gastrointestinal tract. J. Nutr. Biochem. 1993, 4, 528–533. [CrossRef] 72. Donovan, S.M. The Role of Lactoferrin in Gastrointestinal and Immune Development and Function: A Preclinical Perspective. J. Pediatr. 2016, 173, S16–S28. [CrossRef] 73. Zhang, J.L.; Han, X.; Shan, Y.J.; Zhang, L.W.; Du, M.; Liu, M.; Yi, H.X.; Ma, Y. Effect of bovine lactofer the proliferative activity of the osteoblast cell line MC3T3-E1 in vitro. J. Dairy Sci. 2018, 101, 1827–1 73. Zhang, J.L.; Han, X.; Shan, Y.J.; Zhang, L.W.; Du, M.; Liu, M.; Yi, H.X.; Ma, Y. Effect of bovine lactoferrin and human lactoferrin on the proliferative activity of the osteoblast cell line MC3T3-E1 in vitro. J. Dairy Sci. 2018, 101, 1827–1833. [CrossRef] p y y 74. Embleton, N.D.; Berrington, J.E. Clinical Trials of Lactoferrin in the Newborn: Effects on Infection an Nutr. Inst. Workshop Ser. 2020, 94, 141–151. [CrossRef] rington, J.E. Clinical Trials of Lactoferrin in the Newborn: Effects on Infection and the Gut Microbiome. Nestle Ser. 2020, 94, 141–151. [CrossRef] 75. Manzoni, P.; Meyer, M.; Stolﬁ, I.; Rinaldi, M.; Cattani, S.; Pugni, L.; Romeo, M.G.; Messner, H.; Decembrino, L.; Laforgia, N.; et al. Bovine lactoferrin supplementation for prevention of necrotizing enterocolitis in very-low-birth-weight neonates: A randomized clinical trial. Early Hum. Dev. 2014, 90, S60–S65. [CrossRef] y 76. Manzoni, P.; Dall’Agnola, A.; Tomé, D.; Kaufman, D.A.; Tavella, E.; Pieretto, M.; Messina, A.; De Luca, D. et al. Role of Lactoferrin in Neonates and Infants: An Update. Am. J. Perinatol. 2018, 35, 561–565. [Cros nola, A.; Tomé, D.; Kaufman, D.A.; Tavella, E.; Pieretto, M.; Messina, A.; De Luca, D.; Bellaiche, M.; Mosca, A. rrin in Neonates and Infants: An Update. Am. J. Perinatol. 2018, 35, 561–565. [CrossRef] 77. Pammi, M.; Suresh, G. Enteral lactoferrin supplementation for prevention of sepsis and necrotizing enterocolitis in preterm infants. Cochrane Database Syst. Rev. 2020, 3, CD007137. [CrossRef] y 78. Actor, J.K.; Hwang, S.; Kruzel, M.L. References [CrossRef] g 86. Lee, H.; Padhi, E.; Hasegawa, Y.; Larke, J.; Parenti, M.; Wang, A.; Hernell, O.; Lonnerdal, B.; Slupsky, C. Compositional Dynamics of the Milk Fat Globule and Its Role in Infant Development. Front. Pediatr. 2018, 6, 313. [CrossRef] p 87. Koletzko, B. Human Milk Lipids. Ann. Nutr. Metab. 2016, 69 (Suppl. S2), 27–40. [CrossRef] p pp 88. Manoni, M.; Di Lorenzo, C.; Ottoboni, M.; Tretola, M.; Pinotti, L. Comparative Proteomics of Milk Fat Globule Membrane (MFGM) Proteome across Species and Lactation Stages and the Potentials of MFGM Fractions in Infant Formula Preparation. Foods 2020, 9, 1251. [CrossRef] [PubMed] 32 of 41 32 of 41 Nutrients 2021, 13, 2551 89. Brink, L.R.; Lonnerdal, B. Milk fat globule membrane: The role of its various components in infant health and development. J. Nutr. Biochem. 2020, 85, 108465. [CrossRef] [ ] 90. Hernell, O.; Lonnerdal, B.; Timby, N. Milk Fat Globule Membranes: Effects on Microbiome, Metabolome, and Infections in Infants and Children. Nestle Nutr. Inst. Workshop Ser. 2020, 94, 133–140. [CrossRef] [PubMed] 91. Tanaka, K.; Hosozawa, M.; Kudo, N.; Yoshikawa, N.; Hisata, K.; Shoji, H.; Shinohara, K.; Shimizu, T. The pilot study: Sphingomyelin-fortiﬁed milk has a positive association with the neurobehavioural development of very low birth weight infants during infancy, randomized control trial. Brain Dev. 2013, 35, 45–52. [CrossRef] g y 92. Henriksen, N.L.; Aasmul-Olsen, K.; Venkatasubramanian, R.; Nygaard, M.K.E.; Sprenger, R.R.; Heckmann, A.B.; Ostenfeld, M.S.; Ejsing, C.S.; Eskildsen, S.F.; Müllertz, A.; et al. Dairy-Derived Emulsiﬁers in Infant Formula Show Marginal Effects on the Plasma Lipid Proﬁle and Brain Structure in Preterm Piglets Relative to Soy Lecithin. Nutrients 2021, 13, 718. [CrossRef] p g y 93. Palmano, K.; Rowan, A.; Guillermo, R.; Guan, J.; McJarrow, P. The role of gangliosides in neurodevelo 3891–3913. [CrossRef] [PubMed] 94. Rueda, R.; Maldonado, J.; Narbona, E.; Gil, A. Neonatal dietary gangliosides. Early Hum. Dev. 1998, 53, S135–S147. [CrossRef] 94. Rueda, R.; Maldonado, J.; Narbona, E.; Gil, A. Neonatal dietary gangliosides. Early Hum. Dev. 1998, 53, S135–S147. [CrossRef] 95. Greenspon, J.; Li, R.; Xiao, L.; Rao, J.N.; Sun, R.; Strauch, E.D.; Shea-Donohue, T.; Wang, J.Y.; Turner, D.J. Sphingosine-1-phosphate regulates the expression of adherens junction protein E-cadherin and enhances intestinal epithelial cell barrier function Dig Dis 94. Rueda, R.; Maldonado, J.; Narbona, E.; Gil, A. Neonatal dietary gangliosides. Early Hum. Dev. 1998, 53, S135–S147. [CrossRef] 95. References Annotation and structu milk oligosaccharides and determination of novel fucosylated structures. Glycobiology 2013, 23, 664–67 g milk oligosaccharides and determination of novel fucosylated structures. Glycobiology 2013, 23, 664–676. [CrossRef] 109. Wang, B.; Yu, B.; Karim, M.; Hu, H.; Sun, Y.; McGreevy, P.; Petocz, P.; Held, S.; Brand-Miller, J. Dietary sialic acid supplementation improves learning and memory in piglets Am J Clin Nutr 2007 85 561–569 [CrossRef] milk oligosaccharides and determination of novel fucosylated structures. Glycobiology 2013, 23, 664–676. [CrossRef] 109. Wang, B.; Yu, B.; Karim, M.; Hu, H.; Sun, Y.; McGreevy, P.; Petocz, P.; Held, S.; Brand-Miller, J. Dietary sialic acid supplementation improves learning and memory in piglets. Am. J. Clin. Nutr. 2007, 85, 561–569. [CrossRef] g y y gy 109. Wang, B.; Yu, B.; Karim, M.; Hu, H.; Sun, Y.; McGreevy, P.; Petocz, P.; Held, S.; Brand-Miller, J. Dietary sia improves learning and memory in piglets. Am. J. Clin. Nutr. 2007, 85, 561–569. [CrossRef] 110. Harmsen, H.J.; Wildeboer-Veloo, A.C.; Raangs, G.C.; Wagendorp, A.A.; Klijn, N.; Bindels, J.G.; Welling, G.W. Analysis of intestinal ﬂora development in breast-fed and formula-fed infants by using molecular identiﬁcation and detection methods. J. Pediatr. Gastroenterol. Nutr. 2000, 30, 61–67. [CrossRef] [PubMed] 111. Morrin, S.T.; Lane, J.A.; Marotta, M.; Bode, L.; Carrington, S.D.; Irwin, J.A.; Hickey, R.M. Bovine colostru intestinal epithelial cells for increased commensal colonisation. Appl. Microbiol. Biotechnol. 2019, 103, 2 111. Morrin, S.T.; Lane, J.A.; Marotta, M.; Bode, L.; Carrington, S.D.; Irwin, J.A.; Hickey, R.M. Bovine colostrum-driven modulation of intestinal epithelial cells for increased commensal colonisation. Appl. Microbiol. Biotechnol. 2019, 103, 2745–2758. [CrossRef] 112. Davis, E.C.; Wang, M.; Donovan, S.M. The role of early life nutrition in the establishment of gastrointestinal microbial composition and function. Gut Microbes 2017, 8, 143–171. [CrossRef] [PubMed] p pp [ ] 112. Davis, E.C.; Wang, M.; Donovan, S.M. The role of early life nutrition in the establishment of gastrointestinal microbial composition and function. Gut Microbes 2017, 8, 143–171. [CrossRef] [PubMed] 112. Davis, E.C.; Wang, M.; Donovan, S.M. The role of early life nutrition in and function. Gut Microbes 2017, 8, 143–171. [CrossRef] [PubMed] 113. Lane, J.A.; Mariño, K.; Naughton, J.; Kavanaugh, D.; Clyne, M.; Carrington, S.D.; Hickey, R.M. Anti-infective bovine colostrum oligosaccharides: Campylobacter jejuni as a case study. Int. J. Food Microbiol. 2012, 157, 182–188. [CrossRef] 114. Angeloni, S.; Ridet, J.L.; Kusy, N.; Gao, H.; Crevoisier, F.; Guinchard, S.; Kochhar, S.; Sigrist, H.; Sprenger, N. References Greenspon, J.; Li, R.; Xiao, L.; Rao, J.N.; Sun, R.; Strauch, E.D.; Shea-Donohue, T.; Wang, J.Y.; Turner, D.J. Sphingosine-1-phosphate regulates the expression of adherens junction protein E-cadherin and enhances intestinal epithelial cell barrier function. Dig. Dis. Sci. 2011, 56, 1342–1353. [CrossRef] 96. Hernell, O.; Timby, N.; Domellof, M.; Lonnerdal, B. Clinical Beneﬁts of Milk Fat Globule Membranes for Infants and Children. J. Pediatr. 2016, 173, S60–S65. [CrossRef] 97. Spitsberg, V.L. Invited Review: Bovine Milk Fat Globule Membrane as a Potential Nutraceutical. J. Dairy Sci. 2005, 88, 2289–2294. [CrossRef] 98. Bhinder, G.; Allaire, J.M.; Garcia, C.; Lau, J.T.; Chan, J.M.; Ryz, N.R.; Bosman, E.S.; Graef, F.A.; Crowley, S.M.; Celiberto, L.S.; et al. Milk Fat Globule Membrane Supplementation in Formula Modulates the Neonatal Gut Microbiome and Normalizes Intestinal Development. Sci. Rep. 2017, 7, 45274. [CrossRef] p p [ ] 99. Timby, N.; Domellof, M.; Holgerson, P.L.; West, C.E.; Lonnerdal, B.; Hernell, O.; Johansson, I. Oral Microbiota in Infants Fed a Formula Supplemented with Bovine Milk Fat Globule Membranes—A Randomized Controlled Trial. PLoS ONE 2017, 12, e0169831. [CrossRef] [ ] 100. Urashima, T.; Tauﬁk, E.; Fukuda, K.; Asakuma, S. Recent advances in studies on milk oligosaccharides of cows and other domestic farm animals. Biosci. Biotechnol. Biochem. 2013, 77, 455–466. [CrossRef] 101. Albrecht, S.; Lane, J.A.; Mariño, K.; Al Busadah, K.A.; Carrington, S.D.; Hickey, R.M.; Rudd, P.M. A comparative study of free oligosaccharides in the milk of domestic animals. Br. J. Nutr. 2014, 111, 1313–1328. [CrossRef] 102. Gopal, P.K.; Gill, H.S. Oligosaccharides and glycoconjugates in bovine milk and colostrum. Br. J. Nutr. 2000, 84 (Suppl. S1), S69–S74. [CrossRef] 103. Donovan, S.M.; Comstock, S.S. Human Milk Oligosaccharides Inﬂuence Neonatal Mucosal and System Metab. 2016, 69 (Suppl. S2), 42–51. [CrossRef] 104. Pacheco, A.R.; Barile, D.; Underwood, M.A.; Mills, D.A. The impact of the milk glycobiome on the neonate gut microbiota. Annu. Rev. Anim. Biosci. 2015, 3, 419–445. [CrossRef] [PubMed] 105. Smilowitz, J.T.; Lebrilla, C.B.; Mills, D.A.; German, J.B.; Freeman, S.L. Breast milk oligosaccharides: Structure-function relation- ships in the neonate. Annu. Rev. Nutr. 2014, 34, 143–169. [CrossRef] 106. Bode, L.; Jantscher-Krenn, E. Structure-function relationships of human milk oligosaccharides. Adv. Nutr. (Bethesda Md.) 2012, 3, 383S–391S. [CrossRef] 107. Newburg, D.S. Neonatal protection by an innate immune system of human milk consisting of oligosaccharides and glycans. J. Anim. Sci. 2009, 87, 26–34. [CrossRef] 108. Aldredge, D.L.; Geronimo, M.R.; Hua, S.; Nwosu, C.C.; Lebrilla, C.B.; Barile, D. 89. Brink, L.R.; Lonnerdal, B. Milk fat globule membrane: The role of its various components in infant health and development. J. Nutr. Biochem. 2020, 85, 108465. [CrossRef] References Minimal short-term effect of dietary 2′-fucosyllactose on bacterial colonisation, intestinal function and necrotising enterocolitis in preterm pigs. Br. J. Nutr. 2016, 116, 834–841. [CrossRef] 123. Obelitz-Ryom, K.; Rendboe, A.K.; Nguyen, D.N.; Rudloff, S.; Brandt, A.B.; Nielsen, D.S.; Heckmann, A.B.; Chichlowski, M.; Sangild, P.T.; Thymann, T.; et al. Bovine Milk Oligosaccharides with Sialyllactose for Preterm Piglets. Nutrients 2018, 10, 1489. [CrossRef] 124. Rasmussen, S.O.; Martin, L.; Østergaard, M.V.; Rudloff, S.; Roggenbuck, M.; Nguyen, D.N.; Sangild, P.T.; Bering, S.B. Human milk oligosaccharide effects on intestinal function and inﬂammation after preterm birth in pigs. J. Nutr. Biochem. 2017, 40, 141–154. [CrossRef] 125. Bering, S.B. Human Milk Oligosaccharides to Prevent Gut Dysfunction and Necrotizing Enterocolitis in Preterm Neonates. Nutrients 2018, 10, 1461. [CrossRef] [PubMed] 126. Masi, A.C.; Embleton, N.D.; Lamb, C.A.; Young, G.; Granger, C.L.; Najera, J.; Smith, D.P.; Hoffman, K.L.; Petrosino, J.F.; Bode, L.; et al. Human milk oligosaccharide DSLNT and gut microbiome in preterm infants predicts necrotising enterocolitis. Gut 2020. [CrossRef] 127. Kim, S.Y.; Yi, D.Y. Components of human breast milk: From macronutrient to microbiome and microRNA. Clin. Exp. Pediatr. 2020, 63, 301–309. [CrossRef] [PubMed] , M.; Wang, T.; Liang, Y.; Zhong, Z.; Wang, X.; Zhou, Q.; Chen, L.; Lang, Q.; He, Z.; et al. Lactation-related n proﬁles of porcine breast milk exosomes. PLoS ONE 2012, 7, e43691. [CrossRef] 128. Gu, Y.; Li, M.; Wang, T.; Liang, Y.; Zhong, Z.; Wang, X.; Zhou, Q.; Chen, L.; Lang, Q.; He, Z.; et al. expression proﬁles of porcine breast milk exosomes. PLoS ONE 2012, 7, e43691. [CrossRef] 129. Zempleni, J.; Sukreet, S.; Zhou, F.; Wu, D.; Mutai, E. Milk-Derived Exosomes and Metabolic Regulation. Annu. Rev. Anim. Biosci. 2019, 7, 245–262. [CrossRef] 130. Van Hese, I.; Goossens, K.; Vandaele, L.; Opsomer, G. Invited review: MicroRNAs in bovine colostrum—Focus on their origin and potential health beneﬁts for the calf. J. Dairy Sci. 2020, 103, 1–15. [CrossRef] 131. Özdemir, S. Identiﬁcation and comparison of exosomal microRNAs in the milk and colostrum of two different cow breeds. Gene 2020, 743, 144609. [CrossRef] [PubMed] 132. Sun, Q.; Chen, X.; Yu, J.; Zen, K.; Zhang, C.Y.; Li, L. Immune modulatory function of abundant immune-related microRNAs in microvesicles from bovine colostrum. Protein Cell 2013, 4, 197–210. [CrossRef] 133. O’Reilly, D.; Dorodnykh, D.; Avdeenko, N.V.; Nekliudov, N.A.; Garssen, J.; Elolimy, A.A.; Petrou, L.; Simpson, M.R.; Yeruva, L.; Munblit, D. Perspective: The Role of Human Breast-Milk Extracellular Vesicles in Child Health and Disease. References Glycoproﬁling with micro-arrays of glycoconjugates and lectins. Glycobiology 2005, 15, 31–41. [CrossRef] y g y j g 115. Lane, J.A.; O’Callaghan, J.; Carrington, S.D.; Hickey, R.M. Transcriptional response of HT-29 intestinal epithelial cells to human and bovine milk oligosaccharides. Br. J. Nutr. 2013, 110, 2127–2137. [CrossRef] [PubMed] 33 of 41 33 of 41 Nutrients 2021, 13, 2551 116. Comstock, S.S.; Wang, M.; Hester, S.N.; Li, M.; Donovan, S.M. Select human milk oligosaccharides directly modulate peripheral blood mononuclear cells isolated from 10-d-old pigs. Br. J. Nutr. 2014, 111, 819–828. [CrossRef] [PubMed] p g 117. Li, M.; Monaco, M.H.; Wang, M.; Comstock, S.S.; Kuhlenschmidt, T.B.; Fahey, G.C., Jr.; Miller, M.J.; Kuhlenschmidt, M.S.; Donovan, S.M. Human milk oligosaccharides shorten rotavirus-induced diarrhea and modulate piglet mucosal immunity and colonic microbiota. ISME J. 2014, 8, 1609–1620. [CrossRef] 118. Comstock, S.S.; Li, M.; Wang, M.; Monaco, M.H.; Kuhlenschmidt, T.B.; Kuhlenschmidt, M.S.; Donovan, S.M. Dietary Human Milk Oligosaccharides but Not Prebiotic Oligosaccharides Increase Circulating Natural Killer Cell and Mesenteric Lymph Node Memory T Cell Populations in Noninfected and Rotavirus-Infected Neonatal Piglets. J. Nutr. 2017, 147, 1041–1047. [CrossRef] [PubMed] [ ] 119. Short, D.M.; Moore, D.A.; Sischo, W.M. A Randomized Clinical Trial Evaluating the Effects of Oligosaccharides on Transfer of Passive Immunity in Neonatal Dairy Calves. J. Vet. Intern. Med. 2016, 30, 1381–1389. [CrossRef] Short, D.M.; Moore, D.A.; Sischo, W.M. A Randomized Clinical Trial Evaluating the Effects of Oligosacch P i I it i N t l D i C l J V t I t M d 2016 30 1381 1389 [C R f] 119. Short, D.M.; Moore, D.A.; Sischo, W.M. A Randomized Clinical Trial Evaluating the Effects of Passive Immunity in Neonatal Dairy Calves. J. Vet. Intern. Med. 2016, 30, 1381–1389. [CrossRef] , ; , ; , g g Passive Immunity in Neonatal Dairy Calves. J. Vet. Intern. Med. 2016, 30, 1381–1389. [CrossRef] 120. Bode, L. Human Milk Oligosaccharides in the Prevention of Necrotizing Enterocolitis: A Journey from in vitro and in vivo Models to Mother-Infant Cohort Studies. Front. Pediatr. 2018, 6, 385. [CrossRef] 121. Cilieborg, M.S.; Sangild, P.T.; Jensen, M.L.; Østergaard, M.V.; Christensen, L.; Rasmussen, S.O.; Mørbak, A.L.; Jørgensen, C.B.; Bering, S.B. α1,2-Fucosyllactose Does Not Improve Intestinal Function or Prevent Escherichia coli F18 Diarrhea in Newborn Pigs. J. Pediatr. Gastroenterol. Nutr. 2017, 64, 310–318. [CrossRef] 122. Cilieborg, M.S.; Bering, S.B.; Østergaard, M.V.; Jensen, M.L.; Krych, Ł.; Newburg, D.S.; Sangild, P.T. References Adv. Nutr. 2021, 12, 59–70. [CrossRef] 134. Samuel, M.; Chisanga, D.; Liem, M.; Keerthikumar, S.; Anand, S.; Ang, C.S.; Adda, C.G.; Versteegen, E.; Jois, M.; Mathivanan, S. Bovine milk-derived exosomes from colostrum are enriched with proteins implicated in immune response and growth. Sci. Rep. 2017, 7, 5933. [CrossRef] 135. Baier, S.R.; Nguyen, C.; Xie, F.; Wood, J.R.; Zempleni, J. MicroRNAs are absorbed in biologically meaningful amounts from nutritionally relevant doses of cow milk and affect gene expression in peripheral blood mononuclear cells, HEK-293 kidney cell cultures, and mouse livers. J. Nutr. 2014, 144, 1495–1500. [CrossRef] J 136. Ross, M.; Atalla, H.; Karrow, N.; Mallard, B.A. The bioactivity of colostrum and milk exosomes of high, average, and low immune responder cows on human intestinal epithelial cells. J. Dairy Sci. 2020, 104, 2499–2510. [CrossRef] p p y 137. Kirchner, B.; Buschmann, D.; Paul, V.; Pfafﬂ, M.W. Postprandial transfer of colostral extracellular vesicles and their protein and miRNA cargo in neonatal calves. PLoS ONE 2020, 15, e0229606. [CrossRef] 137. Kirchner, B.; Buschmann, D.; Paul, V.; Pfafﬂ, M.W. Postprandial transfer of co miRNA cargo in neonatal calves. PLoS ONE 2020, 15, e0229606. [CrossRef] 138. Goudarzi, N.; Shabani, R.; Ebrahimi, M.; Baghestani, A.; Dehdashtian, E.; Vahabzadeh, G.; Soleimani, M.; Moradi, F.; Katebi, M. Comparative phenotypic characterization of human colostrum and breast milk-derived stem cells. Hum. Cell 2020, 33, 308–317. [CrossRef] 34 of 41 Nutrients 2021, 13, 2551 139. Ghosh, A. Breast Milk Stem Cell Survival in Neonate’s Gut, Entery into Neonate Circulation and Adaption by the Body. Curr. Stem Cell Res. Ther. 2020, 15, 98–101. [CrossRef] 140. Cacho, N.T.; Lawrence, R.M. Innate Immunity and Breast Milk. Front. Immunol. 2017, 8, 584. [CrossRef] 141. Admyre, C.; Johansson, S.M.; Qazi, K.R.; Filén, J.J.; Lahesmaa, R.; Norman, M.; Neve, E.P.; Scheynius, A.; with immune modulatory features are present in human breast milk. J. Immunol. 2007, 179, 1969–1978 y p J , , [ ] 142. Martinez, J.A.; Ballew, M.P. Infant formulas. Pediatr. Rev. 2011, 32, 179–189. [CrossRef] y p 142. Martinez, J.A.; Ballew, M.P. Infant formulas. Pediatr. Rev. 2011, 32, 179–189. [CrossRef] 143. Przyrembel, H.; Agostoni, C. Growing-up milk: A necessity or marketing? World Rev. Nutr. Diet. 2 y , ; g , g p y g , , [ ] 144. Gomez, G.G.; Phillips, O.; Goforth, R.A. Effect of immunoglobulin source on survival, growth, and hematological and immuno- logical variables in pigs. J. Anim. Sci. References 1998, 76, 1–7. [CrossRef] g p g 145. Staak, C. Bovine colostrum and protection of young animals. Berl. Munch. Tierarztl. Wochenschr. 1992, 105, 219–224. 145. Staak, C. Bovine colostrum and protection of young animals. Berl. 146. Heidebrecht, H.J.; Weiss, W.J.; Pulse, M.; Lange, A.; Gisch, K.; Kliem, H.; Mann, S.; Pfafﬂ, M.W.; Kulozik, U.; von Eichel-Streiber, C. Treatment and Prevention of Recurrent Clostridium difﬁcile Infection with Functionalized Bovine Antibody-Enriched Whey in a Hamster Primary Infection Model. Toxins 2019, 11, 98. [CrossRef] Treatment and Prevention of Recurrent Clostridium difﬁcile Infection with Functionalized Bovine Antibody-Enriched Whey in a Hamster Primary Infection Model. Toxins 2019, 11, 98. [CrossRef] y 147. Lyerly, D.M.; Bostwick, E.F.; Binion, S.B.; Wilkins, T.D. Passive immunization of hamsters against disease caused by Clostridium difﬁcile by use of bovine immunoglobulin G concentrate. Infect. Immun. 1991, 59, 2215–2218. [CrossRef] 148. Satyaraj, E.; Reynolds, A.; Pelker, R.; Labuda, J.; Zhang, P.; Sun, P. Supplementation of diets with bov immune function in dogs. Br. J. Nutr. 2013, 110, 2216–2221. [CrossRef] 149. Casal, M.L.; Jezyk, P.F.; Giger, U. Transfer of colostral antibodies from queens to their kittens. Am. J. Vet. Res. 1996, 57, 1653–1658. [PubMed] [ ] 150. Kirkden, R.D.; Broom, D.M.; Andersen, I.L. Invited review: Piglet mortality: Management solutions. J. Anim. Sci. 2013, 91, 3361–3389. [CrossRef] 151. Blencowe, H.; Cousens, S.; Chou, D.; Oestergaard, M.; Say, L.; Moller, A.B.; Kinney, M.; Lawn, J. Born too soon: The global epidemiology of 15 million preterm births. Reprod. Health 2013, 10 (Suppl. S1), S2. [CrossRef] [PubMed] owe, H.; Cousens, S.; Chou, D.; Oestergaard, M.; Say, L.; Moller, A.B.; Kinney, M.; Lawn, J. Born too soo encowe, H.; Cousens, S.; Chou, D.; Oestergaard, M.; Say, L.; Moller, A.B.; Kinney, M.; Lawn, J. Born too idemiology of 15 million preterm births Reprod Health 2013 10 (Suppl S1) S2 [CrossRef] [PubMed] Blencowe, H.; Cousens, S.; Chou, D.; Oestergaard, M.; Say, L.; Moller, A.B.; Kinney, M.; Lawn, J. Born epidemiology of 15 million preterm births. Reprod. Health 2013, 10 (Suppl. S1), S2. [CrossRef] [PubMed] g y y J idemiology of 15 million preterm births. Reprod. Health 2013, 10 (Suppl. S1), S2. [CrossRef] [PubMed] 152. Lawn, J.E.; Blencowe, H.; Waiswa, P.; Amouzou, A.; Mathers, C.; Hogan, D.; Flenady, V.; Frøen, J.F.; Q et al. Stillbirths: Rates, risk factors, and acceleration towards 2030. Lancet 2016, 387, 587–603. [Cros 152. References Lawn, J.E.; Blencowe, H.; Waiswa, P.; Amouzou, A.; Mathers, C.; Hogan, D.; Flenady, V.; Frøen, J.F.; Qureshi, Z.U.; Calderwood, C.; et al. Stillbirths: Rates, risk factors, and acceleration towards 2030. Lancet 2016, 387, 587–603. [CrossRef] 153. Lee, A.C.; Katz, J.; Blencowe, H.; Cousens, S.; Kozuki, N.; Vogel, J.P.; Adair, L.; Baqui, A.H.; Bhutta, Z.A.; Caulﬁeld, L.E.; et al. National and regional estimates of term and preterm babies born small for gestational age in 138 low-income and middle-income countries in 2010. Lancet Glob. Health 2013, 1, e26–e36. [CrossRef] 154. Li, Y.; Juhl, S.M.; Ye, X.; Shen, R.L.; Iyore, E.O.; Dai, Y.; Sangild, P.T.; Greisen, G.O. A Stepwise, Pilot Study of Bovine Colostrum to Supplement the First Enteral Feeding in Preterm Infants (Precolos): Study Protocol and Initial Results. Front. Pediatr. 2017, 5, 42. [CrossRef] [PubMed] 155. Juhl, S.M.; Ye, X.; Zhou, P.; Li, Y.; Iyore, E.O.; Zhang, L.; Jiang, P.; van Goudoever, J.B.; Greisen, G.; Sangild, P.T. Bovine Colostrum for Preterm Infants in the First Days of Life: A Randomized Controlled Pilot Trial. J. Pediatr. Gastroenterol. Nutr. 2018, 66, 471–478. [CrossRef] 156. Jiang, P.P.; Muk, T.; Krych, L.; Nielsen, D.S.; Khakimov, B.; Li, Y.; Juhl, S.M.; Greisen, G.; Sangild, P.T. Gut colonization in preterm infants supplemented with bovine colostrum in the ﬁrst week of life: An explorative pilot study. J. Parenter. Enter. Nutr. 2021. [CrossRef] [PubMed] 157. Hutchens, T.W.; Henry, J.F.; Yip, T.T. Structurally intact (78-kDa) forms of maternal lactoferrin puriﬁed from urine of preterm infants fed human milk: Identiﬁcation of a trypsin-like proteolytic cleavage event in vivo that does not result in fragment dissociation. Proc. Natl. Acad. Sci. USA 1991, 88, 2994–2998. [CrossRef] [ ] 158. Burrin, D.; Sangild, P.T.; Stoll, B.; Thymann, T.; Buddington, R.; Marini, J.; Olutoye, O.; Shulman, R.J. Translational Advances in Pediatric Nutrition and Gastroenterology: New Insights from Pig Models. Annu. Rev. Anim. Biosci. 2020, 8, 321–354. [CrossRef] 159. Sangild, P.T.; Thymann, T.; Schmidt, M.; Stoll, B.; Burrin, D.G.; Buddington, R.K. Invited review: The preterm pig as a model in 159. Sangild, P.T.; Thymann, T.; Schmidt, M.; Stoll, B.; Burrin, D.G.; Buddington, R.K. Invited review: The p pediatric gastroenterology. J. Anim. Sci. 2013, 91, 4713–4729. [CrossRef] [PubMed] pediatric gastroenterology. J. Anim. Sci. 2013, 91, 4713–4729. [CrossRef] [PubMed] 160. Odle, J.; Lin, X.; Jacobi, S.K.; Kim, S.W.; Stahl, C.H. The suckling piglet as an agrimedical model for the study of pediatric nutrition and metabolism. Annu. Rev. Anim. Biosci. References 2014, 2, 419–444. [CrossRef] p g gy 160. Odle, J.; Lin, X.; Jacobi, S.K.; Kim, S.W.; Stahl, C.H. The suckling piglet as an agrimedical model for the st and metabolism. Annu. Rev. Anim. Biosci. 2014, 2, 419–444. [CrossRef] 161. Sangild, P.T.; Petersen, Y.M.; Schmidt, M.; Elnif, J.; Petersen, T.K.; Buddington, R.K.; Greisen, G.; Michaelsen, K.F.; Burrin, D.G. Preterm birth affects the intestinal response to parenteral and enteral nutrition in newborn pigs. J. Nutr. 2002, 132, 3786–3794. [CrossRef] [PubMed] 162. Bjornvad, C.R.; Schmidt, M.; Petersen, Y.M.; Jensen, S.K.; Offenberg, H.; Elnif, J.; Sangild, P.T. Preterm birth makes the immature intestine sensitive to feeding-induced intestinal atrophy. Am. J. Physiol. Regul. Integr. Comp. Physiol. 2005, 289, R1212–R1222. [CrossRef] [PubMed] 163. Sangild, P.T.; Siggers, R.H.; Schmidt, M.; Elnif, J.; Bjornvad, C.R.; Thymann, T.; Grondahl, M.L.; Hansen, A.K.; Jensen, S.K.; Boye, M.; et al. Diet- and colonization-dependent intestinal dysfunction predisposes to necrotizing enterocolitis in preterm pigs. Gastroenterology 2006, 130, 1776–1792. [CrossRef] gy 164. Bjornvad, C.R.; Thymann, T.; Deutz, N.E.; Burrin, D.G.; Jensen, S.K.; Jensen, B.B.; Mølbak, L.; Boye, M.; Larsson, L.I.; Schmidt, M.; et al. Enteral feeding induces diet-dependent mucosal dysfunction, bacterial proliferation, and necrotizing enterocolitis in preterm pigs on parenteral nutrition. Am. J. Physiol. Gastrointest. Liver Physiol. 2008, 295, G1092–G1103. [CrossRef] 35 of 41 35 of 41 Nutrients 2021, 13, 2551 165. Bæk, O.; Brunse, A.; Nguyen, D.N.; Moodley, A.; Thymann, T.; Sangild, P.T. Diet Modulates the High Sensitivity to Systemic Infection in Newborn Preterm Pigs. Front. Immunol. 2020, 11, 1019. [CrossRef] [PubMed] g 166. Brunse, A.; Worsøe, P.; Pors, S.E.; Skovgaard, K.; Sangild, P.T. Oral Supplementation With Bovine Colostrum Prevents Septic Shock and Brain Barrier Disruption During Bloodstream Infection in Preterm Newborn Pigs. Shock 2019, 51, 337–347. [CrossRef] [PubMed] [ ] 167. Ren, S.; Hui, Y.; Obelitz-Ryom, K.; Brandt, A.B.; Kot, W.; Nielsen, D.S.; Thymann, T.; Sangild, P.T.; Nguyen, D.N. Neonatal gut and immune maturation is determined more by postnatal age than by postconceptional age in moderately preterm pigs. Am. J. Physiol. Gastrointest. Liver Physiol. 2018, 315, G855–G867. [CrossRef] [PubMed] 168. Hansen, C.F.; Thymann, T.; Andersen, A.D.; Holst, J.J.; Hartmann, B.; Hilsted, L.; Langhorn, L.; Jelsing, J.; Sangild, P.T. Rapid gut growth but persistent delay in digestive function in the postnatal period of preterm pigs. Am. J. Physiol. Gastrointest. Liver Physiol. 2016, 310, G550–G560. [CrossRef] [PubMed] 169. References Andersen, A.D.; Sangild, P.T.; Munch, S.L.; van der Beek, E.M.; Renes, I.B.; Ginneken, C.; Greisen, G.O.; Thymann, T. Delayed growth, motor function and learning in preterm pigs during early postnatal life. Am. J. Physiol. Regul. Integr. Comp. Physiol. 2016, 310, R481–R492. [CrossRef] 170. Bergström, A.; Kaalund, S.S.; Skovgaard, K.; Andersen, A.D.; Pakkenberg, B.; Rosenørn, A.; van Elburg, R.M.; Thymann, T.; Greisen, G.O.; Sangild, P.T. Limited effects of preterm birth and the ﬁrst enteral nutrition on cerebellum morphology and gene expression in piglets. Physiol. Rep. 2016, 4, e12871. [CrossRef] [PubMed] 171. Cao, M.; Andersen, A.D.; Van Ginneken, C.; Shen, R.L.; Petersen, S.O.; Thymann, T.; Jing, J.; Sangild, P.T. Physical activity level is impaired and diet dependent in preterm newborn pigs. Pediatr. Res. 2015, 78, 137–144. [CrossRef] [PubMed] 172. Holme Nielsen, C.; Bladt Brandt, A.; Thymann, T.; Obelitz-Ryom, K.; Jiang, P.; Vanden Hole, C.; van Ginneken, C.; Pankratova, S.; Sangild, P.T. Rapid Postnatal Adaptation of Neurodevelopment in Pigs Born Late Preterm. Dev. Neurosci. 2018, 40, 586–600. [CrossRef] 173. Jensen, M.L.; Sangild, P.T.; Lykke, M.; Schmidt, M.; Boye, M.; Jensen, B.B.; Thymann, T. Similar efﬁcacy of human banked milk and bovine colostrum to decrease incidence of necrotizing enterocolitis in preterm piglets. Am. J. Physiol. Regul. Integr. Comp. Physiol. 2013, 305, R4–R12. [CrossRef] 174. Rasmussen, S.O.; Martin, L.; Østergaard, M.V.; Rudloff, S.; Li, Y.; Roggenbuck, M.; Bering, S.B.; Sangild, P.T. Bovine colostrum improves neonatal growth, digestive function, and gut immunity relative to donor human milk and infant formula in preterm pigs. Am. J. Physiol. Gastrointest. Liver Physiol. 2016, 311, G480–G491. [CrossRef] p g y y 175. Oosterloo, B.C.; Premkumar, M.; Stoll, B.; Olutoye, O.; Thymann, T.; Sangild, P.T.; Burrin, D.G. Dual purpose use of preterm piglets as a model of pediatric GI disease. Vet. Immunol. Immunopathol. 2014, 159, 156–165. [CrossRef] [PubMed] 176. Bæk, O.; Cilieborg, M.S.; Bering, S.B.; Nguyen, D.N.; Thymann, T.; Sangild, P.T. Sex-speciﬁc survival, growth, immunity and organ development in preterm pigs as models for immature newborns. Front. Pediatr. 2021, 9, 63. [CrossRef] [PubMed] 177. Meder, U.; Tarjanyi, E.; Kovacs, K.; Szakmar, E.; Cseko, A.J.; Hazay, T.; Belteki, G.; Szabo, M.; Jermendy, in preterm infants during maternal singing combined with skin-to-skin care. Pediatr. Res. 2020. [CrossR 178. Brunse, A.; Peng, Y.; Li, Y.; Lykkesfeldt, J.; Sangild, P.T. Co-bedding of preterm newborn pigs reduces necrotizing enterocolitis incidence independent of vital functions and cortisol levels. Front. Pediatr. 2021, 9, 216. [CrossRef] [PubMed] p 179. 187. Poulsen, A.R.; de Jonge, N.; Sugiharto, S.; Nielsen, J.L.; Lauridsen, C.; Canibe, N. The microbial community of the gut differs between piglets fed sow milk, milk replacer or bovine colostrum. Br. J. Nutr. 2017, 117, 964–978. [CrossRef] [PubMed] References Hyperimmune bovine colostrum in the treatment of shigellosis in children: A double-blind, randomized, controlled trial. Acta Paediatr. 2001, 90, 1373–1378. [CrossRef] 199 Ebi T S A U K I hid N Oh S Ohi i A Aik K K i i S K hi N I i A l Ebina, T.; Sato, A.; Umezu, K.; Ishida, N.; Ohyama, S.; Ohizumi, A.; Aikawa, K.; Katagiri, S.; Katsushim 199. Ebina, T.; Sato, A.; Umezu, K.; Ishida, N.; Ohyama, S.; Ohizumi, A.; Aikawa, K.; Katagiri, S.; Katsu Prevention of rotavirus infection by cow colostrum antibody against human rotaviruses. Lancet 1983, 2 Prevention of rotavirus infection by cow colostrum antibody against human rotaviruses. Lancet 1983, 2, 1029–1030. [CrossRef] 200. Davidson, G.P.; Whyte, P.B.; Daniels, E.; Franklin, K.; Nunan, H.; McCloud, P.I.; Moore, A.G.; Moore, D.J. Passive immunisation of children with bovine colostrum containing antibodies to human rotavirus. Lancet 1989, 2, 709–712. [CrossRef] y y g 200. Davidson, G.P.; Whyte, P.B.; Daniels, E.; Franklin, K.; Nunan, H.; McCloud, P.I.; Moore, A.G.; Moore, D.J. Passive immunisation of children with bovine colostrum containing antibodies to human rotavirus. Lancet 1989, 2, 709–712. [CrossRef] 201. Turner, R.B.; Kelsey, D.K. Passive immunization for prevention of rotavirus illness in healthy infants. Pediatr. Infect. Dis. J. 1993, 12, 718–722. [CrossRef] 202. Brunser, O.; Espinoza, J.; Figueroa, G.; Araya, M.; Spencer, E.; Hilpert, H.; Link-Amster, H.; Brussow, H. Field trial of an infant formula containing anti-rotavirus and anti-Escherichia coli milk antibodies from hyperimmunized cows. J. Pediatr. Gastroenterol. Nutr. 1992, 15, 63–72. [CrossRef] [PubMed] 203. Den Hartog, G.; Jacobino, S.; Bont, L.; Cox, L.; Ulfman, L.H.; Leusen, J.H.; van Neerven, R.J. Speciﬁcity and Effector Functions of Human RSV-Speciﬁc IgG from Bovine Milk. PLoS ONE 2014, 9, e112047. [CrossRef] 204. Loss, G.; Depner, M.; Ulfman, L.H.; van Neerven, R.J.; Hose, A.J.; Genuneit, J.; Karvonen, A.M.; Hyvärinen, A.; Kaulek, V.; Roduit, C.; et al. Consumption of unprocessed cow’s milk protects infants from common respiratory infections. J. Allergy Clin. Immunol. 2015, 135, 56–62. [CrossRef] [PubMed] 205. Von Mutius, E.; Vercelli, D. Farm living: Effects on childhood asthma and allergy. Nat. 205. Von Mutius, E.; Vercelli, D. Farm living: Effects on childhood asthma and allergy. Nat. Rev. Immunol. 2010, 10, 861–868. [CrossRef] 205. Von Mutius, E.; Vercelli, D. Farm living: Effects on childhood asthma and allergy. Nat. Rev. Immunol. 2010, 10, 861–868. [CrossRef] 206. Roselli, M.; Britti, M.S.; Le Huërou-Luron, I.; Marfaing, H.; Zhu, W.Y.; Mengheri, E. References 2012, 90, 1513–1520. [CrossRef] pp 190. Lallès, J.P.; Bosi, P.; Janczyk, P.; Koopmans, S.J.; Torrallardona, D. Impact of bioactive substances on the performance of weaned piglets: A review. Animal 2009, 3, 1625–1643. [CrossRef] p p g 191. Hilpert, H.; Brussow, H.; Mietens, C.; Sidoti, J.; Lerner, L.; Werchau, H. Use of bovine milk concentrate containing antibody to rotavirus to treat rotavirus gastroenteritis in infants. J. Infect. Dis. 1987, 156, 158–166. [CrossRef] [PubMed] 192. Mitra, A.K.; Mahalanabis, D.; Ashraf, H.; Unicomb, L.; Eeckels, R.; Tzipori, S. Hyperimmune cow colostrum reduces diarrhoea due to rotavirus: A double-blind, controlled clinical trial. Acta Paediatr. 1995, 84, 996–1001. [CrossRef] [PubMed] 192. Mitra, A.K.; Mahalanabis, D.; Ashraf, H.; Unicomb, L.; Eeckels, R.; Tzipori, S. Hyperimmune cow colostrum reduces diarrhoea d i A d bl bli d ll d li i l i l A t P di t 1995 84 996 1001 [C R f] [P bM d] o rotavirus: A double-blind, controlled clinical trial. Acta Paediatr. 1995, 84, 996–1001. [CrossRef] [PubMed] 193. Sarker, S.A.; Casswall, T.H.; Mahalanabis, D.; Alam, N.H.; Albert, M.J.; Brussow, H.; Fuchs, G.J.; Hammerstrom, L. Successful treatment of rotavirus diarrhea in children with immunoglobulin from immunized bovine colostrum. Pediatr. Infect. Dis. J. 1998, 17, 1149–1154. [CrossRef] 194. Ylitalo, S.; Uhari, M.; Rasi, S.; Pudas, J.; Leppaluoto, J. Rotaviral antibodies in the treatment of acute rotaviral gastroenteritis. Acta Paediatr. 1998, 87, 264–267. [CrossRef] 195. Mietens, C.; Keinhorst, H.; Hilpert, H.; Gerber, H.; Amster, H.; Pahud, J.J. Treatment of infantile E. coli gastroenteritis with speciﬁc bovine anti-E. coli milk immunoglobulins. Eur. J. Pediatr. 1979, 132, 239–252. [CrossRef] [PubMed] 196. Casswall, T.H.; Sarker, S.A.; Faruque, S.M.; Weintraub, A.; Albert, M.J.; Fuchs, G.J.; Alam, N.H.; Dahlstrom, A.K.; Link, H.; Brussow, H.; et al. Treatment of enterotoxigenic and enteropathogenic Escherichia coli-induced diarrhoea in children with bovine immunoglobulin milk concentrate from hyperimmunized cows: A double-blind, placebo-controlled, clinical trial. Scand. J. Gastroenterol. 2000, 35, 711–718. [CrossRef] 197. Huppertz, H.I.; Rutkowski, S.; Busch, D.H.; Eisebit, R.; Lissner, R.; Karch, H. Bovine colostrum ameliorates diarrhea in infection with diarrheagenic Escherichia coli, shiga toxin-producing E. Coli, and E. coli expressing intimin and hemolysin. J. Pediatr. Gastroenterol. Nutr. 1999, 29, 452–456. [CrossRef] [PubMed] 198. Ashraf, H.; Mahalanabis, D.; Mitra, A.K.; Tzipori, S.; Fuchs, G.J. Hyperimmune bovine colostrum in t children: A double-blind, randomized, controlled trial. Acta Paediatr. 2001, 90, 1373–1378. [CrossRe 198. Ashraf, H.; Mahalanabis, D.; Mitra, A.K.; Tzipori, S.; Fuchs, G.J. References Reed, R.C.; Johnson, D.E.; Nie, A.M. Preterm Infant Skin Structure Is Qualitatively and Quantitatively Different from That of Term Newborns. Pediatr. Dev. Pathol. 2021, 24, 96–102. [CrossRef] 180. Li, J.; Xu, Y.W.; Jiang, J.J.; Song, Q.K. Bovine colostrum and product intervention associated with relief of childhood infectious diarrhea. Sci. Rep. 2019, 9, 3093. [CrossRef] 181. Barakat, S.H.; Meheissen, M.A.; Omar, O.M.; Elbana, D.A. Bovine Colostrum in the Treatment of Acute Diarrhea in Children: A Double-Blinded Randomized Controlled Trial. J. Trop. Pediatr. 2020, 66, 46–55. [CrossRef] [PubMed] 182. Premkumar, M.H.; Massieu, L.A.; Anderson, D.M.; Gokulakrishnan, G. Human Milk Supplemen Current Controversies. Clin. Perinatol. 2020, 47, 355–368. [CrossRef] [PubMed] 183. Sun, J.; Li, Y.; Nguyen, D.N.; Mortensen, M.S.; van den Akker, C.H.P.; Skeath, T.; Pors, S.E.; Pankratova, S.; Rudloff, S.; Sørensen, S.J.; et al. Nutrient Fortiﬁcation of Human Donor Milk Affects Intestinal Function and Protein Metabolism in Preterm Pigs. J. Nutr. 2018, 148, 336–347. [CrossRef] g 184. Sun, J.; Li, Y.; Pan, X.; Nguyen, D.N.; Brunse, A.; Bojesen, A.M.; Rudloff, S.; Mortensen, M.S.; Burrin, D.G.; Sangild, P.T. Human Milk Fortiﬁcation with Bovine Colostrum Is Superior to Formula-Based Fortiﬁers to Prevent Gut Dysfunction, Necrotizing Enterocolitis, and Systemic Infection in Preterm Pigs. J. Parenter. Enter. Nutr. 2019, 43, 252–262. [CrossRef] [PubMed] y g 185. Boudry, C.; Buldgen, A.; Portetelle, D.; Collard, A.; Théwis, A.; Dehoux, J.P. Effects of oral supplementation with bovine colostrum on the immune system of weaned piglets. Res. Vet. Sci. 2007, 83, 91–101. [CrossRef] [PubMed] y p g 186. Boudry, C.; Dehoux, J.P.; Wavreille, J.; Portetelle, D.; Théwis, A.; Buldgen, A. Effect of a bovine colostrum whey supplementation on growth performance, faecal Escherichia coli population and systemic immune response of piglets at weaning. Animal 2008, 2, 730–737. [CrossRef] 187. Poulsen, A.R.; de Jonge, N.; Sugiharto, S.; Nielsen, J.L.; Lauridsen, C.; Canibe, N. The microbial community of the gut differs between piglets fed sow milk, milk replacer or bovine colostrum. Br. J. Nutr. 2017, 117, 964–978. [CrossRef] [PubMed] Nutrients 2021, 13, 2551 36 of 41 188. Sugiharto, S.; Poulsen, A.S.; Canibe, N.; Lauridsen, C. Effect of bovine colostrum feeding in comparison with milk replacer and natural feeding on the immune responses and colonisation of enterotoxigenic Escherichia coli in the intestinal tissue of piglets. Br. J. Nutr. 2015, 113, 923–934. [CrossRef] 189. Huguet, A.; Le Dividich, J.; Le Huërou-Luron, I. Improvement of growth performance and sanitary statu bovine colostrum-supplemented diet. J. Anim. Sci. References Effect of different plant extracts and natural substances (PENS) against membrane damage induced by enterotoxigenic Escherichia coli K88 in pig intestinal cells. Toxicol. In Vitro 2007, 21, 224–229. [CrossRef] 207. De Waard, M.; Li, Y.; Zhu, Y.; Ayede, A.I.; Berrington, J.; Bloomﬁeld, F.H.; Busari, O.O.; Cormack, B.E.; Embleton, N.D.; van Goudoever, J.B.; et al. Time to Full Enteral Feeding for Very Low-Birth-Weight Infants Varies Markedly Among Hospitals Worldwide But May Not Be Associated With Incidence of Necrotizing Enterocolitis: The NEOMUNE-NeoNutriNet Cohort Study. J. Parenter. Enter. Nutr. 2019, 43, 658–667. [CrossRef] [PubMed] ghirad, B.; Morgan, R.L.; Zeraatkar, D.; Zea, A.M.; Couban, R.; Johnston, B.C.; Florez, I.D. Human and Bovine 208. Sadeghirad, B.; Morgan, R.L.; Zeraatkar, D.; Zea, A.M.; Couban, R.; Johnston, B.C.; Florez, I.D. Hum Prevention of Necrotizing Enterocolitis: A Meta-analysis. Pediatrics 2018, 142. [CrossRef] ad, B.; Morgan, R.L.; Zeraatkar, D.; Zea, A.M.; Couban, R.; Johnston, B.C.; Florez, I.D. Human and Bovine C n of Necrotizing Enterocolitis: A Meta-analysis. Pediatrics 2018, 142. [CrossRef] g g Prevention of Necrotizing Enterocolitis: A Meta-analysis. Pediatrics 2018, 142. [CrossRef] 209. Ahnfeldt, A.M.; Hyldig, N.; Li, Y.; Kappel, S.S.; Aunsholdt, L.; Sangild, P.T.; Zachariassen, G. FortiColos—A multicentre study using bovine colostrum as a fortiﬁer to human milk in very preterm infants: Study protocol for a randomised controlled pilot trial. Trials 2019, 20, 279. [CrossRef] 210. Brooks, H.J.; McConnell, M.A.; Corbett, J.; Buchan, G.S.; Fitzpatrick, C.E.; Broadbent, R.S. Potential prophylactic value of bovine colostrum in necrotizing enterocolitis in neonates: An in vitro study on bacterial attachment, antibody levels and cytokine production. FEMS Immunol. Med. Microbiol. 2006, 48, 347–354. [CrossRef] 37 of 41 37 of 41 Nutrients 2021, 13, 2551 211. Balachandran, B.; Dutta, S.; Singh, R.; Prasad, R.; Kumar, P. Bovine Colostrum in Prevention of Necrotizing Enterocolitis and Sepsis in Very Low Birth Weight Neonates: A Randomized, Double-blind, Placebo-controlled Pilot Trial. J. Trop. Pediatr. 2017, 63, 10–17. [CrossRef] [PubMed] [ ] [ ] 212. Ismail, R.I.H.; Awad, H.A.; Imam, S.S.; Gad, G.I.; Aboushady, N.M.; Abdou, R.M.; Eissa, D.S.; Azzam, N.T.; Barakat, M.M.; Yassin, M.M.; et al. Gut priming with bovine colostrum and T regulatory cells in preterm neonates: A randomized controlled trial. Pediatr. Res. 2021. [CrossRef] 213. Eibl, M.M.; Wolf, H.M.; Furnkranz, H.; Rosenkranz, A. Prevention of necrotizing enterocolitis in low-birth-weight infants by IgA-IgG feeding. N. Engl. J. Med. 1988, 319, 1–7. [CrossRef] [PubMed] 214. References Lawrence, G.; Tudehope, D.; Baumann, K.; Jeffery, H.; Gill, A.; Cole, M.; Drew, J.; McPhee, A.; Ratcliffe, J.; Reynolds, G.; et al. Enteral human IgG for prevention of necrotising enterocolitis: A placebo-controlled, randomised trial. Lancet 2001, 357, 2090–2094. [CrossRef] [ ] 215. Rubaltelli, F.F.; Benini, F.; Sala, M. Prevention of necrotizing enterocolitis in neonates at risk by oral administration of monomeric IgG. Dev. Pharmacol. Ther. 1991, 17, 138–143. [CrossRef] g 216. Fast, C.; Rosegger, H. Necrotizing enterocolitis prophylaxis: Oral antibiotics and lyophilized enterobacteria vs. oral immunoglob- ulins. Acta Paediatr. 1994, 83, 86–90. [CrossRef] [ ] n, P.; Pohlandt, F. Prevention of necrotizing enterocolitis in extremely low birth weight infants by IgG feeding? 157, 924–925. [CrossRef] [PubMed] 217. Richter, D.; Bartmann, P.; Pohlandt, F. Prevention of necrotizing enterocolitis in extremely low birth w Eur. J. Pediatr. 1998, 157, 924–925. [CrossRef] [PubMed] 218. Foster, J.P.; Seth, R.; Cole, M.J. Oral immunoglobulin for preventing necrotizing enterocolitis in preterm and low birth weight neonates. Cochrane Database Syst. Rev. 2016. [CrossRef] y 219. Siggers, J.; Sangild, P.T.; Jensen, T.K.; Siggers, R.H.; Skovgaard, K.; Støy, A.C.; Jensen, B.B.; Thymann, T.; Bering, S.B.; Boye, M. Transition from parenteral to enteral nutrition induces immediate diet-dependent gut histological and immunological responses in preterm neonates. Am. J. Physiol. Gastrointest. Liver Physiol. 2011, 301, G435–G445. [CrossRef] p y y 220. Danielsen, M.; Thymann, T.; Jensen, B.B.; Jensen, O.N.; Sangild, P.T.; Bendixen, E. Proteome proﬁles of mucosal immunoglobulin uptake in inﬂamed porcine gut. Proteomics 2006, 6, 6588–6596. [CrossRef] [PubMed] h h h ld d d 221. Che, L.; Thymann, T.; Bering, S.B.; Le Huerou-Luron, I.; D’inca, R.; Zhang, K.; Sangild, P.T. IUGR does not predispose to necrotizing enterocolitis or compromise postnatal intestinal adaptation in preterm pigs. Pediatr. Res. 2010, 67, 54–59. [CrossRef] [PubMed] [ ] 222. Cilieborg, M.S.; Boye, M.; Mølbak, L.; Thymann, T.; Sangild, P.T. Preterm birth and necrotizing enterocolitis alter gut colonization in pigs. Pediatr. Res. 2011, 69, 10–16. [CrossRef] [PubMed] 223. Cilieborg, M.S.; Boye, M.; Thymann, T.; Jensen, B.B.; Sangild, P.T. Diet-dependent effects of minimal enteral nutrition on intestinal function and necrotizing enterocolitis in preterm pigs. J. Parenter. Enter. Nutr. 2011, 35, 32–42. [CrossRef] 224. Li, Y.; Jensen, M.L.; Chatterton, D.E.; Jensen, B.B.; Thymann, T.; Kvistgaard, A.S.; Sangild, P.T. Raw bovine milk improves gut responses to feeding relative to infant formula in preterm piglets. Am. J. Physiol. Gastrointest. Liver Physiol. 2014, 306, G81–G90. [CrossRef] [ ] 225. References Li, Y.; Østergaard, M.V.; Jiang, P.; Chatterton, D.E.; Thymann, T.; Kvistgaard, A.S.; Sangild, P.T. Whey protein processing inﬂuences formula-induced gut maturation in preterm pigs. J. Nutr. 2013, 143, 1934–1942. [CrossRef] 226. Møller, H.K.; Thymann, T.; Fink, L.N.; Frokiaer, H.; Kvistgaard, A.S.; Sangild, P.T. Bovine colostrum is superior to enriched formulas in stimulating intestinal function and necrotising enterocolitis resistance in preterm pigs. Br. J. Nutr. 2011, 105, 44–53. [CrossRef] [ ] 227. Oste, M.; Van Ginneken, C.J.; Van Haver, E.R.; Bjornvad, C.R.; Thymann, T.; Sangild, P.T. The intestinal trophic response to enteral food is reduced in parenterally fed preterm pigs and is associated with more nitrergic neurons. J. Nutr. 2005, 135, 2657–2663. [CrossRef] 228. Oste, M.; Van Haver, E.; Thymann, T.; Sangild, P.; Weyns, A.; Van Ginneken, C.J. Formula induces in pigs within a few hours of feeding. J. Parenter. Enter. Nutr. 2010, 34, 271–279. [CrossRef] [PubMed] 229. Shen, R.L.; Thymann, T.; Østergaard, M.V.; Støy, A.C.; Krych, Ł.; Nielsen, D.S.; Lauridsen, C.; Hartmann, B.; Holst, J.J.; Burrin, D.G.; et al. Early gradual feeding with bovine colostrum improves gut function and NEC resistance relative to infant formula in preterm pigs. Am. J. Physiol. Gastrointest. Liver Physiol. 2015, 309, G310–G323. [CrossRef] p p g y y 230. Siggers, J.; Ostergaard, M.V.; Siggers, R.H.; Skovgaard, K.; Mølbak, L.; Thymann, T.; Schmidt, M.; Møller, H.K.; Purup, S.; Fink, L.N.; et al. Postnatal amniotic ﬂuid intake reduces gut inﬂammatory responses and necrotizing enterocolitis in preterm neonates. Am. J. Physiol. Gastrointest. Liver Physiol. 2013, 304, G864–G875. [CrossRef] [PubMed] 231. Siggers, R.H.; Thymann, T.; Jensen, B.B.; Mølbak, L.; Heegaard, P.M.; Schmidt, M.; Buddington, R.K.; Sangild, P.T. Elective cesarean delivery affects gut maturation and delays microbial colonization but does not increase necrotizing enterocolitis in preterm pigs. Am. J. Physiol. Regul. Integr. Comp. Physiol. 2008, 294, R929–R938. [CrossRef] [PubMed] 232. Sty, A.C.; Sangild, P.T.; Skovgaard, K.; Thymann, T.; Bjerre, M.; Chatterton, D.E.; Purup, S.; Boye, M.; Heegaard, P.M. Spray Dried, Pasteurised Bovine Colostrum Protects Against Gut Dysfunction and Inﬂammation in Preterm Pigs. J. Pediatr. Gastroenterol. Nutr. 2016, 63, 280–287. [CrossRef] 233. Støy, A.C.; Heegaard, P.M.; Thymann, T.; Bjerre, M.; Skovgaard, K.; Boye, M.; Stoll, B.; Schmidt, M.; Jensen, B.B.; Sangild, P.T. Bovine colostrum improves intestinal function following formula-induced gut inﬂammation in preterm pigs. Clin. Nutr. 2014, 33, 322–329. [CrossRef] 38 of 41 38 of 41 Nutrients 2021, 13, 2551 234. References [CrossRef] p g g J [ ] 241. Buddington, R.K.; Bering, S.B.; Thymann, T.; Sangild, P.T. Aldohexose malabsorption in preterm pigs is directly related to the it f ti i t liti P di t R 2008 63 382 387 [C R f] [P bM d] p g g 241. Buddington, R.K.; Bering, S.B.; Thymann, T.; Sangild, P.T. Aldohexose malabsorption in preterm pigs is directly related to the severity of necrotizing enterocolitis. Pediatr. Res. 2008, 63, 382–387. [CrossRef] [PubMed] 242. Oste, M.; De Vos, M.; Van Haver, E.; Van Brantegem, L.; Thymann, T.; Sangild, P.; Weyns, A.; Van Ginneken, C. Parenteral and enteral feeding in preterm piglets differently affects extracellular matrix proteins, enterocyte proliferation and apoptosis in the small intestine. Br. J. Nutr. 2010, 104, 989–997. [CrossRef] 243. Puiman, P.J.; Jensen, M.; Stoll, B.; Renes, I.B.; de Bruijn, A.C.; Dorst, K.; Schierbeek, H.; Schmidt, M.; Boehm, G.; Burrin, D.G.; et al. Intestinal threonine utilization for protein and mucin synthesis is decreased in formula-fed preterm pigs. J. Nutr. 2011, 141, 1306–1311. [CrossRef] [PubMed] e, M.; Sangild, P.T. Bacterial colonization and gut development in preterm neonates. Early Hum. Dev. 2012, 88 . [CrossRef] [PubMed] 244. Cilieborg, M.S.; Boye, M.; Sangild, P.T. Bacterial colonization and gut development in preterm neona (Suppl. S1), S41–S49. [CrossRef] [PubMed] 245. Wolfs, T.G.; Jellema, R.K.; Turrisi, G.; Becucci, E.; Buonocore, G.; Kramer, B.W. Inﬂammation-induced immune suppression of the fetus: A potential link between chorioamnionitis and postnatal early onset sepsis. J. Matern. Fetal Neonatal Med. 2012, 25 (Suppl. S1), 8–11. [CrossRef] [PubMed] enhoven, R.; Klein, L.; Jellema, R.K.; Westerlaken, R.J.J.; Hütten, M.C.; Vermeulen, J.; Wassink, G.; Gunn, A.J. ain Injury, Antenatal Triggers, and Therapeutics: Timing Is Key. Cells 2020, 9, 1871. [CrossRef] 246. Ophelders, D.; Gussenhoven, R.; Klein, L.; Jellema, R.K.; Westerlaken, R.J.J.; Hütten, M.C.; Vermeule Wolfs, T. Preterm Brain Injury, Antenatal Triggers, and Therapeutics: Timing Is Key. Cells 2020, 9, 1 247. Gussenhoven, R.; Westerlaken, R.J.J.; Ophelders, D.; Jobe, A.H.; Kemp, M.W.; Kallapur, S.G.; Zimmermann, L.J.; Sangild, P.T.; Pankratova, S.; Gressens, P.; et al. Chorioamnionitis, neuroinﬂammation, and injury: Timing is key in the preterm ovine fetus. J. Neuroinﬂamm. 2018, 15, 113. [CrossRef] [PubMed] ﬂ 248. Cilieborg, M.S.; Schmidt, M.; Skovgaard, K.; Boye, M.; Weber, N.R.; Heegaard, P.M.; Burrin, D.G.; Sangild, P.T. Fetal lipopolysac- charide exposure modulates diet-dependent gut maturation and sensitivity to necrotising enterocolitis in pre-term pigs. Br. J. Nutr. 2011, 106, 852–861. [CrossRef] [PubMed] 249. References Van Haver, E.R.; de Vooght, L.; Oste, M.; Sangild, P.T.; Thymann, T.; Weyns, A.L.; van Ginneken, C.J. Postnatal and diet-dependent increases in enteric glial cells and VIP-containing neurones in preterm pigs. Neurogastroenterol. Motil. 2008, 20, 1070–1079. [CrossRef] 235. Van Haver, E.R.; Oste, M.; Thymann, T.; Sys, S.U.; Lamers, W.H.; Weyns, A.L.; Sangild, P.T.; van Ginneken, C.J. Enteral feeding reduces endothelial nitric oxide synthase in the caudal intestinal microvasculature of preterm piglets. Pediatr. Res. 2008, 63, 137–142. [CrossRef] [ ] 236. Van Haver, E.R.; Sangild, P.T.; Oste, M.; Siggers, J.L.; Weyns, A.L.; Van Ginneken, C.J. Diet-dependent mucosal colonization and interleukin-1beta responses in preterm pigs susceptible to necrotizing enterocolitis. J. Pediatr. Gastroenterol. Nutr. 2009, 49, 90–98. [CrossRef] [PubMed] 237. Willems, R.; Krych, L.; Rybicki, V.; Jiang, P.; Sangild, P.T.; Shen, R.L.; Hensel, K.O.; Wirth, S.; Postberg, J.; Jenke, A.C. Introducing enteral feeding induces intestinal subclinical inﬂammation and respective chromatin changes in preterm pigs. Epigenomics 2015, 7, 553–565. [CrossRef] 238. Ren, S.; Hui, Y.; Goericke-Pesch, S.; Pankratova, S.; Kot, W.; Pan, X.; Thymann, T.; Sangild, P.T.; Nguyen, D.N. Gut and immune effects of bioactive milk factors in preterm pigs exposed to prenatal inﬂammation. Am. J. Physiol. Gastrointest. Liver Physiol. 2019, 317, G67–G77. [CrossRef] 239. Li, Y.; Pan, X.; Nguyen, D.N.; Ren, S.; Moodley, A.; Sangild, P.T. Bovine Colostrum Before or After Fo Systemic Immune Protection and Gut Function in Newborn Preterm Pigs. Front. Immunol. 2019, 10, 30 239. Li, Y.; Pan, X.; Nguyen, D.N.; Ren, S.; Moodley, A.; Sangild, P.T. Bovine Colostrum Before or After Formula Feeding Improves Systemic Immune Protection and Gut Function in Newborn Preterm Pigs. Front. Immunol. 2019, 10, 3062. [CrossRef] 240. Yan, X.; Sangild, P.T.; Peng, Y.; Li, Y.; Bering, S.B.; Pan, X. Supplementary bovine colostrum feeding to formula-fed preterm pigs improves gut function and reduces necrotizing enterocolitis J Pediatr Gastroenterol Nutr 2021 [CrossRef] g y y g g p Systemic Immune Protection and Gut Function in Newborn Preterm Pigs. Front. Immunol. 2019, 10, 3062. [CrossRef] 240. Yan, X.; Sangild, P.T.; Peng, Y.; Li, Y.; Bering, S.B.; Pan, X. Supplementary bovine colostrum feeding to formula-fed preterm pigs improves gut function and reduces necrotizing enterocolitis. J. Pediatr. Gastroenterol. Nutr. 2021. [CrossRef] y g 240. Yan, X.; Sangild, P.T.; Peng, Y.; Li, Y.; Bering, S.B.; Pan, X. Supplementary bovine colostrum feeding to formula-fed preterm pigs improves gut function and reduces necrotizing enterocolitis. J. Pediatr. Gastroenterol. Nutr. 2021. References Coggins, S.A.; Laskin, B.; Harris, M.C.; Grundmeier, R.W.; Passarella, M.; McKenna, K.J.; Srinivasan, L. Acute Kidney Injury Associated with Late-Onset Neonatal Sepsis: A Matched Cohort Study. J. Pediatr. 2021, 231, 185–192. [CrossRef] A.B.; Nguyen, D.N. Fortiﬁcation with bovine colostrum enhances antibacterial activity of human milk. J 2020. [CrossRef] 250. Gao, X.; Li, Y.; Olin, A.B.; Nguyen, D.N. Fortiﬁcation with bovine colostrum enhances antibacteri Parenter. Enter. Nutr. 2020. [CrossRef] [ ] 251. Harbeson, D.; Francis, F.; Bao, W.; Amenyogbe, N.A.; Kollmann, T.R. Energy Demands of Early Life Drive a Disease Tolerant Phenotype and Dictate Outcome in Neonatal Bacterial Sepsis. Front. Immunol. 2018, 9, 1918. [CrossRef] yp p 252. Trend, S.; Strunk, T.; Hibbert, J.; Kok, C.H.; Zhang, G.; Doherty, D.A.; Richmond, P.; Burgner, D.; Simmer, K.; Davidson, D.J.; et al. Antimicrobial protein and Peptide concentrations and activity in human breast milk consumed by preterm infants at risk of late-onset neonatal sepsis. PLoS ONE 2015, 10, e0117038. [CrossRef] p 253. Li, Y.; Nguyen, D.N.; de Waard, M.; Christensen, L.; Zhou, P.; Jiang, P.; Sun, J.; Bojesen, A.M.; Lauridsen, C.; Lykkesfeldt, J.; et al. Pasteurization Procedures for Donor Human Milk Affect Body Growth, Intestinal Structure, and Resistance against Bacterial Infections in Preterm Pigs. J. Nutr. 2017, 147, 1121–1130. [CrossRef] 254. Nguyen, D.N.; Currie, A.; Ren, S.; Bering, S.; Sangild, P. Heat treatment and irradiation reduce anti-bacterial and immune- modulatory properties of bovine colostrum. J. Funct. Foods 2019, 57, 182–189. [CrossRef] y p p 255. Playford, R.J.; Garbowsky, M.; Marchbank, T. Pasteurized Chicken Egg Powder Stimulates Proliferation and Migration of AGS, RIE1, and Caco-2 Cells and Reduces NSAID-Induced Injury in Mice and Colitis in Rats. J. Nutr. 2020, 150, 1434–1442. [CrossRef] [PubMed] 39 of 41 39 of 41 Nutrients 2021, 13, 2551 256. Ye, Y.; Manne, S.; Treem, W.R.; Bennett, D. Prevalence of Inﬂammatory Bowel Disease in Pediatric and Adult Populations: Recent Estimates from Large National Databases in the United States, 2007–2016. Inﬂamm. Bowel Dis. 2020, 26, 619–625. [CrossRef] [PubMed] no, C.; Dipasquale, V. Diet and Nutrition in Pediatric Inﬂammatory Bowel Diseases. Nutrients 2021, 13, 655 d] [ ] 257. Cucinotta, U.; Romano, C.; Dipasquale, V. Diet and Nutrition in Pediatric Inﬂammatory Bowel Disease [CrossRef] [PubMed] [ ] [ ] 258. Khan, Z.; Macdonald, C.; Wicks, A.C.; Holt, M.P.; Floyd, D.; Ghosh, S.; Wright, N.A.; Playford, R.J. Use of the ‘nutriceutical’, bovine colostrum, for the treatment of distal colitis: Results from an initial study. References Aliment. Pharmacol. Ther. 2002, 16, 1917–1922. [CrossRef] 259. Bodammer, P.; Kerkhoff, C.; Maletzki, C.; Lamprecht, G. Bovine colostrum increases pore-forming claudin-2 protein expression but paradoxically not ion permeability possibly by a change of the intestinal cytokine milieu. PLoS ONE 2013, 8, e64210. [CrossRef] [PubMed] 260. Filipescu, I.E.; Leonardi, L.; Menchetti, L.; Guelﬁ, G.; Traina, G.; Casagrande-Proietti, P.; Piro, F.; Quattrone, A.; Barbato, O.; Brecchia, G. Preventive effects of bovine colostrum supplementation in TNBS-induced colitis in mice. PLoS ONE 2018, 13, e0202929. [CrossRef] 261. Sangild, P.T.; Ney, D.M.; Sigalet, D.L.; Vegge, A.; Burrin, D. Animal models of gastrointestinal and liver diseases. Animal models of infant short bowel syndrome: Translational relevance and challenges. Am. J. Physiol. Gastrointest. Liver Physiol. 2014, 307, G1147–G1168. [CrossRef] [ ] 262. Aunsholt, L.; Jeppesen, P.B.; Lund, P.; Sangild, P.T.; Ifaoui, I.B.; Qvist, N.; Husby, S. Bovine colostrum to children with short bowel syndrome: A randomized, double-blind, crossover pilot study. J. Parenter. Enter. Nutr. 2014, 38, 99–106. [CrossRef] 263. Aunsholt, L.; Qvist, N.; Sangild, P.T.; Vegge, A.; Stoll, B.; Burrin, D.G.; Jeppesen, P.B.; Eriksen, T.; Husby, S.; Thymann, T. Minimal Enteral Nutrition to Improve Adaptation After Intestinal Resection in Piglets and Infants. J. Parenter. Enter. Nutr. 2018, 42, 446–454. [CrossRef] 264. Aunsholt, L.; Thymann, T.; Qvist, N.; Sigalet, D.; Husby, S.; Sangild, P.T. Prematurity Reduces Functional Adaptation to Intestinal Resection in Piglets. J. Parenter. Enter. Nutr. 2015, 39, 668–676. [CrossRef] 265. Vegge, A.; Thymann, T.; Lund, P.; Stoll, B.; Bering, S.B.; Hartmann, B.; Jelsing, J.; Qvist, N.; Burrin, D.G.; Jeppesen, P.B.; et al. Glucagon-like peptide-2 induces rapid digestive adaptation following intestinal resection in preterm neonates. Am. J. Physiol. Gastrointest. Liver Physiol. 2013, 305, G277–G285. [CrossRef] 266. Nagy, E.S.; Paris, M.C.; Taylor, R.G.; Fuller, P.J.; Sourial, M.; Justice, F.; Bines, J.E. Colostrum protein conce adaptation after massive small bowel resection in juvenile pigs. J. Pediatr. Gastroenterol. Nutr. 2004, 39, 266. Nagy, E.S.; Paris, M.C.; Taylor, R.G.; Fuller, P.J.; Sourial, M.; Justice, F.; Bines, J.E. Colostrum protein concentrate enhances intestinal adaptation after massive small bowel resection in juvenile pigs. J. Pediatr. Gastroenterol. Nutr. 2004, 39, 487–492. [CrossRef] 267. Pereira-Fantini, P.M.; Thomas, S.L.; Taylor, R.G.; Nagy, E.; Sourial, M.; Fuller, P.J.; Bines, J.E. Colostrum supplementation restores insulin-like growth factor -1 levels and alters muscle morphology following massive small bowel resection. J. Parenter. Enter. Nutr. 2008, 32, 266–275. [CrossRef] p j p g J , , [ ] 267. References Pereira-Fantini, P.M.; Thomas, S.L.; Taylor, R.G.; Nagy, E.; Sourial, M.; Fuller, P.J.; Bines, J.E. Colostrum supplementation restores insulin-like growth factor -1 levels and alters muscle morphology following massive small bowel resection. J. Parenter. Enter. Nutr. 2008, 32, 266–275. [CrossRef] 268. Paris, M.C.; Fuller, P.J.; Carstensen, B.; Nagy, E.; Taylor, R.G.; Sourial, M.; Holst, J.J.; Hartmann, B.; Binesm, J.E. Plasma GLP-2 levels and intestinal markers in the juvenile pig during intestinal adaptation: Effects of different diet regimens. Dig. Dis. Sci. 2004, 49, 1688–1695. [CrossRef] [PubMed] 269. Sangild, P.T.; Shen, R.L.; Pontoppidan, P.; Rathe, M. Animal models of chemotherapy-induced mucositis: Translational relevance and challenges. Am. J. Physiol. Gastrointest. Liver Physiol. 2018, 314, G231–G246. [CrossRef] [PubMed] 270. Rathe, M.; De Pietri, S.; Wehner, P.S.; Frandsen, T.L.; Grell, K.; Schmiegelow, K.; Sangild, P.T.; Husby, S.; Muller, K. Bovine Colostrum Against Chemotherapy-Induced Gastrointestinal Toxicity in Children With Acute Lymphoblastic Leukemia: A Randomized, Double-Blind, Placebo-Controlled Trial. J. Parenter. Enter. Nutr. 2020, 44, 337–347. [CrossRef] 271. Martin, J.; Howard, S.C.; Pillai, A.; Vogel, P.; Naren, A.P.; Davis, S.; Ringwald-Smith, K.; Buddington, K.; Buddington, R.K. The weaned pig as a model for Doxorubicin-induced mucositis. Chemotherapy 2014, 60, 24–36. [CrossRef] [PubMed] 272. Shen, R.L.; Pontoppidan, P.E.; Rathe, M.; Jiang, P.; Hansen, C.F.; Buddington, R.K.; Heegaard, P.M.; Müller, K.; Sangild, P.T. Milk diets inﬂuence doxorubicin-induced intestinal toxicity in piglets. Am. J. Physiol. Gastrointest. Liver Physiol. 2016, 311, G324–G333. [CrossRef] [PubMed] 273. Shen, R.L.; Rathe, M.; Jiang, P.; Pontoppidan, P.E.; Heegaard, P.M.; Müller, K.; Sangild, P.T. Doxorubicin-Induced Gut Toxicity in Piglets Fed Bovine Milk and Colostrum. J. Pediatr. Gastroenterol. Nutr. 2016, 63, 698–707. [CrossRef] [PubMed] 274. Pontoppidan, P.E.; Shen, R.L.; Cilieborg, M.S.; Jiang, P.; Kissow, H.; Petersen, B.L.; Thymann, T.; Heilmann, C.; Muller, K.; Sangild, P.T. Bovine Colostrum Modulates Myeloablative Chemotherapy-Induced Gut Toxicity in Piglets. J. Nutr. 2015, 145, 1472–1480. [CrossRef] [PubMed] 275. Pontoppidan, P.L.; Shen, R.L.; Petersen, B.L.; Thymann, T.; Heilmann, C.; Müller, K.; Sangild, P.T. Intestinal response to myeloablative chemotherapy in piglets. Exp. Biol. Med. 2014, 239, 94–104. [CrossRef] 276. Chatterton, D.E.W.; Aagaard, S.; Hesselballe Hansen, T.; Nguyen, D.N.; De Gobba, C.; Lametsch, R.; Sangild, P.T. Bioactive proteins in bovine colostrum and effects of heating, drying and irradiation. Food Funct. 2020, 11, 2309–2327. [CrossRef] p g y g 277. Elfstrand, L.; Lindmark-Månsson, H.; Paulsson, M.; Nyberg, L.; Åkesson, B. Immunoglobulins, gro hormone in bovine colostrum and the effects of processing. Int. Dairy J. 2002, 12, 879–887. [CrossRef] 278. References Foster, D.M.; Poulsen, K.P.; Sylvester, H.J.; Jacob, M.E.; Casulli, K.E.; Farkas, B.E. Effect of high-pressure processing of bovine colostrum on immunoglobulin G concentration, pathogens, viscosity, and transfer of passive immunity to calves. J. Dairy Sci. 2016, 99, 8575–8588. [CrossRef] 40 of 41 40 of 41 Nutrients 2021, 13, 2551 279. Navis, M.; Muncan, V.; Sangild, P.T.; Møller Willumsen, L.; Koelink, P.J.; Wildenberg, M.E.; Abrahamse, E.; Thymann, T.; van Elburg, R.M.; Renes, I.B. Beneﬁcial Effect of Mildly Pasteurized Whey Protein on Intestinal Integrity and Innate Defense in Preterm and Near-Term Piglets. Nutrients 2020, 12, 1125. [CrossRef] [PubMed] g 280. Navis, M.; Schwebel, L.; Soendergaard Kappel, S.; Muncan, V.; Sangild, P.T.; Abrahamse, E.; Aunsholt, L.; Thymann, T.; van Elburg, R.M.; Renes, I.B. Mildly Pasteurized Whey Protein Promotes Gut Tolerance in Immature Piglets Compared with Extensively Heated Whey Protein. Nutrients 2020, 12, 3391. [CrossRef] y y 281. Lönnerdal, B.; Hernell, O. Effects of feeding ultrahigh-temperature (UHT)-treated infant formula with different protein concentra- tions or powdered formula, as compared with breast-feeding, on plasma amino acids, hematology, and trace element status. Am. J. Clin. Nutr. 1998, 68, 350–356. [CrossRef] J [ ] 282. Playford, R.J.; Cattell, M.; Marchbank, T. Marked variability in bioactivity between commercially available bovine colostrum for human use; implications for clinical trials. PLoS ONE 2020, 15, e0234719. [CrossRef] 282. Playford, R.J.; Cattell, M.; Marchbank, T. Marked variability in bioactivity between com human use; implications for clinical trials. PLoS ONE 2020, 15, e0234719. [CrossRef] 283. Klobasa, F.; Butler, J.E.; Habe, F. Maternal-neonatal immunoregulation: Suppression of de novo synthesis of IgG and IgA, but not IgM, in neonatal pigs by bovine colostrum, is lost upon storage. Am. J. Vet. Res. 1990, 51, 1407–1412. 284. Agbokounou, A.M.; Ahounou, G.S.; Youssao Abdou Karim, I.; Mensah, G.A.; Koutinhouin, B.; Hornick, J.L. Effect of cow colostrum on the performance and survival rate of local newborn piglets in Benin Republic. Trop. Anim. Health Prod. 2017, 49, 287–294. [CrossRef] [PubMed] 285. Liu, Y.; Zhang, W.; Han, B.; Zhang, L.; Zhou, P. Changes in bioactive milk serum proteins during milk powder processing. Food Chem. 2020, 314, 126177. [CrossRef] [PubMed] 286. Langhendries, J.P.; Hurrell, R.F.; Furniss, D.E.; Hischenhuber, C.; Finot, P.A.; Bernard, A.; Battisti, O.; Bertrand, J.M.; Senterre, J. Maillard reaction products and lysinoalanine: Urinary excretion and the effects on kidney function of preterm infants fed heat-processed milk formula. J. Pediatr. Gastroenterol. Nutr. 1992, 14, 62–70. [CrossRef] p 287. References Cattaneo, S.; Masotti, F.; Pellegrino, L. Liquid infant formulas: Technological tools for limiting heat damage. J. Agric. Food Chem. 2009, 57, 10689–10694. [CrossRef] [PubMed] 288. Cano-Sancho, G.; Alexandre-Gouabau, M.C.; Moyon, T.; Royer, A.L.; Guitton, Y.; Billard, H.; Darmaun, D.; Rozé, J.C.; Boquien, C.Y.; Le Bizec, B.; et al. Simultaneous exploration of nutrients and pollutants in human milk and their impact on preterm infant growth: An integrative cross-platform approach. Environ. Res. 2020, 182, 109018. [CrossRef] g g p pp 289. Andrew, S.M. Effect of composition of colostrum and transition milk from Holstein heifers on speciﬁcity rates of antibiotic residue tests. J. Dairy Sci. 2001, 84, 100–106. [CrossRef] 290. an Wattum, J.J.; Leferink, T.M.; Wilffert, B.; Ter Horst, P.G.J. Antibiotics and lactation: An overview of relative infant doses and a systematic assessment of clinical studies. Basic Clin. Pharmacol. Toxicol. 2019, 124, 5–17. [CrossRef] 291. Hair, A.B.; Blanco, C.L.; Moreira, A.G.; Hawthorne, K.M.; Lee, M.L.; Rechtman, D.J.; Abrams, S.A. Randomized trial of human milk cream as a supplement to standard fortiﬁcation of an exclusive human milk-based diet in infants 750-1250 g birth weight. J. Pediatr. 2014, 165, 915–920. [CrossRef] 292. Rechtman, D.J.; Lee, M.L.; Berg, H. Effect of environmental conditions on unpasteurized donor human milk. Breastfeed. Med. Off. J. Acad. Breastfeed. Med. 2006, 1, 24–26. [CrossRef] f 293. Salcedo, J.; Karav, S.; Le Parc, A.; Cohen, J.L.; de Moura Bell, J.; Sun, A.; Lange, M.C.; Barile, D. Applihairations of industrial treatments to donor human milk: Inﬂuence of pasteurization treatments, storage temperature, and time on human milk gangliosides. NPJ Sci. Food 2018, 2, 5. [CrossRef] g g 294. Høst, A. Cow’s milk protein allergy and intolerance in infancy. Some clinical, epidemiological and immunological aspects. Pediatr. Allergy Immunol. 1994, 5, 5–36. [CrossRef] 295. Barni, S.; Giovannini, M.; Mori, F. Epidemiology of non-IgE-mediated food allergies: What can we learn from that? Curr. Opin. Allergy Clin. Immunol. 2021, 21, 188–194. [CrossRef] d protein-induced enterocolitis syndrome epidemiology. Ann. Allergy Asthma Immunol. 2021, 126, 469–477 d] 296. Cianferoni, A. Food protein-induced enterocolitis syndrome epidemiology. Ann. Allergy Asthma [CrossRef] [PubMed] 297. Matsumoto, N.; Okochi, M.; Matsushima, M.; Kato, R.; Takase, T.; Yoshida, Y.; Kawase, M.; Isobe, K.; Kawabe, T.; Honda, H. Peptide array-based analysis of the speciﬁc IgE and IgG4 in cow’s milk allergens and its use in allergy evaluation. Peptides 2009, 30, 1840–1847. [CrossRef] [PubMed] liveira, M.; Mafra, I. Bovine Milk Allergens: A Comprehensive Review. Compr. Rev. Food Sci. References Food Saf. 2018 Ref] 298. Villa, C.; Costa, J.; Oliveira, M.; Mafra, I. Bovine Milk Allergens: A Comprehensive Review. Compr. Rev. Food Sci. Food Saf. 2018, 17, 137–164. [CrossRef] 299. Zachariassen, G.; Faerk, J.; Esberg, B.H.; Fenger-Gron, J.; Mortensen, S.; Christesen, H.T.; Halken, S. Alle preterm infants according to nutrition after hospital discharge. Pediatr. Allergy Immunol. 2011, 22, 515– 299. Zachariassen, G.; Faerk, J.; Esberg, B.H.; Fenger-Gron, J.; Mortensen, S.; Christesen, H.T.; Halken, S. Allergic diseases among very preterm infants according to nutrition after hospital discharge. Pediatr. Allergy Immunol. 2011, 22, 515–520. [CrossRef] 300. Thureen, P.J.; Melara, D.; Fennessey, P.V.; Hay, W.W., Jr. Effect of low versus high intravenous amino acid intake on very low birth weight infants in the early neonatal period Pediatr Res 2003 53 24 32 [CrossRef] p g p g gy 300. Thureen, P.J.; Melara, D.; Fennessey, P.V.; Hay, W.W., Jr. Effect of low versus high intravenous amino acid intake on very low birth weight infants in the early neonatal period. Pediatr. Res. 2003, 53, 24–32. [CrossRef] g y p 301. Hay, W.W., Jr. Strategies for feeding the preterm infant. Neonatology 2008, 94, 245–254. [CrossRef] [PubMed] 302. Clark, D.; Henderson, M.; Smith, M.; Dear, P.R. Plasma amino acid concentrations in parenterally fed preterm infants. Arch. Dis. g y p 301. Hay, W.W., Jr. Strategies for feeding the preterm infant. Neonatology 2008, 94, 245–254. [CrossRef] [Pub g y p 301. Hay, W.W., Jr. Strategies for feeding the preterm infant. Neonatology 2008, 94, 245–254. [CrossRef] [PubMed] 302 Cl k D H d M S ith M D PR Pl i id t ti i t ll f d t i f t A h Di y g g p gy 302. Clark, D.; Henderson, M.; Smith, M.; Dear, P.R. Plasma amino acid concentrations in parenterally fe Child. 1989, 64, 939–942. [CrossRef] 303. Meyer, R.; Foong, R.X.; Thapar, N.; Kritas, S.; Shah, N. Systematic review of the impact of feed protein type and degree of hydrolysis on gastric emptying in children. BMC Gastroenterol. 2015, 15, 137. [CrossRef] 41 of 41 Nutrients 2021, 13, 2551 304. Ghoneim, N.; Bauchart-Thevret, C.; Oosterloo, B.; Stoll, B.; Kulkarni, M.; de Pipaon, M.S.; Zamora, I.J.; Olutoye, O.O.; Berg, B.; Wittke, A.; et al. Delayed initiation but not gradual advancement of enteral formula feeding reduces the incidence of necrotizing enterocolitis (NEC) in preterm pigs. PLoS ONE 2014, 9, e106888. [CrossRef] [PubMed] p p g 305. 304. Ghoneim, N.; Bauchart-Thevret, C.; Oosterloo, B.; Stoll, B.; Kulkarni, M.; de Pipaon, M.S.; Zamora, I.J.; Olutoye, O.O.; Berg, B.; Wittke, A.; et al. Delayed initiation but not gradual advancement of enteral formula feeding reduces the incidence of necrotizing enterocolitis (NEC) in preterm pigs. PLoS ONE 2014, 9, e106888. [CrossRef] [PubMed] References Corvaglia, L.; Mariani, E.; Aceti, A.; Galletti, S.; Faldella, G. Extensively hydrolyzed protein form esophageal reﬂux in symptomatic preterm infants. Early Hum. Dev. 2013, 89, 453–455. [CrossRef] p g y p p y 306. Ferreira, C.H.F.; Martinez, F.E.; Crott, G.C.; Belik, J. Gavage Feed Volume Determines the Gastric Em Infants. J. Pediatr. Gastroenterol. Nutr. 2018, 67, e43–e46. [CrossRef] [PubMed] 307. Perrella, S.L.; Hepworth, A.R.; Gridneva, Z.; Simmer, K.N.; Hartmann, P.E.; Geddes, D.T. Gastric Emptying and Curding of Pasteurized Donor Human Milk and Mother’s Own Milk in Preterm Infants. J. Pediatr. Gastroenterol. Nutr. 2015, 61, 125–129. [CrossRef] 308. Thymann, T.; Støy, C.A.; Bering, S.B.; Mølbak, L.; Sangild, P.T. Casein addition to a whey-based formula function in preterm pigs. J. Anim. Sci. 2012, 90 (Suppl. S4), 378–380. [CrossRef] 309. Kappel, S.S.; Sangild, P.T.; Scheike, T.; Friborg, C.R.; Gormsen, M.; Aunsholt, L. Radiographic Imaging to Evaluate Food Passage Rate in Preterm Piglets as a Model for Preterm Infants. Front. Pediatr. 2020, 8, 624915. [CrossRef] [PubMed] 310. Buddington, R.K.; Davis, S.L.; Buddington, K.K. The Risk of Necrotizing Enterocolitis Differs Among Preterm Pigs Fed Formulas With Either Lactose or Maltodextrin. J. Pediatr. Gastroenterol. Nutr. 2018, 66, e61–e66. [CrossRef]
https://openalex.org/W2521519883	https://revistas.javeriana.edu.co/index.php/vnimedica/article/download/17038/13577, https://www.redalyc.org/pdf/2310/231046892007.pdf	es		Infección por chikunguña materno-fetal asociada con miocarditis neonatal	Universitas Medica	2,016	cc-by	4,301	236 REPORTES DE CASO doi: 10.11144/Javeriana.umed57-2.icmf Infección por chikunguña materno-fetal asociada con miocarditis neonatal Silvia Catalina Corrales Martínez1, Carlos Fonseca2, Doris Salgado3 Cómo citar: Corrales Martínez SC, Fonseca C, Salgado D. Infección por chikunguña materno-fetal asociada con miocarditis neonatal. Univ Med. 2016;57(2):236-48. doi: http://dx.doi.org/10.11144/Javeriana.umed57-2.icmf Resumen El virus de chikunguña es un arbovirus de la familia Togaviridae y fue aislado por primera vez en Tanzania y Uganda, en 1953. Esta infección se ha desarrollado de manera endémica y es transmitida por el mosquito del género Aedes. El mecanismo exacto de la transmisión de madre a hijo aún no se tiene claro, pero se han identificado potenciales complicaciones de la transmisión vertical, como malformaciones congénitas, mortinatos, restricción del crecimiento intrauterino y partos pretérmino. En este artículo se presentan dos casos clínicos de neonatos que presentaron arritmias cardiacas en sus primeros días de vida, hijos de madres con síndrome febril en curso por chikunguña. Se diagnosticó miocarditis por chikunguña y transmisión vertical, confirmado por reacción de cadena de polimerasa. Ambos pacientes recibieron manejo médico, incluidas inmunoglobulinas, con mejoría del cuadro clínico, sin desenlaces fatales. Palabras clave: chikunguña, arbovirus, recién nacido, embarazo, exposición transplacentaria. 1 Médica residente de primer año de Pediatría, Facultad de Salud, Universidad Surcolombiana-Hospital Universitario Hernando Moncaliano, Neiva, Colombia. 2 Médico pediatra, Unidad de Cuidado Intensivo Neonatal-Hospital Universitario Hernando Moncaliano, Neiva, Colombia. 3 Médica pediatra, Departamento Infectología Pediátrica, Facultad de Salud, Universidad Surcolombiana-Hospital Universitario Hernando Moncaliano, Neiva, Colombia. Recibido: 13/08/2015 Revisado: 13/10/2015 Aceptado: 09/11/2015 Silvia Catalina Corrales Martínez et al. Infección por chikunguña materno-fetal asociada con miocarditis neonatal 237 Title: Maternal-fetal Chikungunya Infection Associated with Neonatal Myocarditis Abstract Chikungunya is an arbovirus of the Togaviridae family first isolated in Tanzania and Uganda, in 1953. This infection has been developed with endemic and is transmitted by the Aedes mosquito. The mechanism of transmission from mother to child is not yet clear, but it has identified potential complications from vertical transmission, such as birth defects, stillbirths, intrauterine growth retardation and preterm deliveries. Then two cases of infants who developed cardiac arrhythmias in their first days of life, children of mothers with febrile syndrome in progress by Chikungunya are presented. Myocarditis vertical transmission Chikungunya and confirmed by polymerase chain reaction was diagnosed. Both patients received medical management including immunoglobulins, with clinical improvement without complications or dead. Key words: Chikungunya arbovirus, newborn, pregnancy, maternal-fetal exchange. Introducción El virus del chikunguña es transmitido por artrópodos que causan una enfermedad caracterizada por fiebre, erupciones cutáneas y artralgias. Ha vuelto a surgir en la última década y ha causado numerosos brotes en todo el mundo [1]. El chikunguña se infecta a través de su inoculación en la piel por picaduras de mosquitos infectados. Afecta células residentes de la piel, lo que desemboca en un fase de infección cutánea; posteriormente, en estas células se lleva a cabo la replicación inicial. La inoculación directa del virus en la circu- lación sanguínea también puede ocurrir a través de las picaduras de mosquitos. El virus logra difundirse a los ganglios linfáticos, donde continua su ciclo de replicación para expandirse a los órganos periféricos, como el bazo y los tejidos muscular, hepático, neurológico y articulares [1]. El chikunguña ha llegado a las costas americanas y ha causado más de un millón de casos [1]. Desde finales de 2004, el chikunguña ha surgido en las islas del sudoeste del océano Índico. Entre enero y marzo de 2005 se registraron más de 5000 casos en las Comoras. En la isla Reunión, un departamento francés de ultramar, con un total de población de 787.836 habitantes, la transmisión fue moderada hasta la temporada de lluvias, cuando se calculó una epidemia de magnitud sin precedentes, al reportarse 300.000 casos, con una incidencia máxima de más de 45.000 casos [2]. Durante este brote, estuvo afectada más de un tercio de la población [2]. En junio de 2005, se identificó el primer caso de transmisión del chikunguña de madre a hijo [2]. Durante este brote, el 70 %-80 % de las madres infectadas eran sintomáticas. Estudios posteriores informaron que, durante el brote, la enfermedad por chikunguña cursó sin efectos observables en los resultados del embarazo y no afectó a los recién nacidos si la madre estaba infectada antes del periodo cercano al parto [3]. Sin embargo, otro estudio encontró que si la madre era sintomática entre dos días antes del parto o después de este, Univ. Méd. ISSN 0041-9095. Bogotá (Colombia), 57 (2): 236-248, abril-junio, 2016 238 los recién nacidos de estas madres también desarrollaban síntomas pocos días después del nacimiento [3]. Este evento se conoce como transmisión vertical del chikunguña, y ocurre durante brotes de la enfermedad, asociado a viremia significativa en la madre en el último trimestre. Durante este periodo, cerca del 50 % de los neonatos puede presentar infección por chikunguña con cuadros severos y afectación en diferentes órganos vitales [4]. Con un amplio espectro de manifestaciones, la infección neonatal por chikunguña incluye anomalías del sistema nervioso, alteraciones hemorrágicas y manifestaciones cardiacas [5], con una alta tasa de morbilidad neonatal [6]. Cuando existe afectación del sistema nervioso central, puede causar enfermedad grave con discapacidad a lo largo de toda la vida del recién nacido [4]. El conocimiento sobre la evolución clínica de la enfermedad y la confirmación rápida con pruebas virológicas mejoran la gestión de los entes de salud, lo que puede contribuir a disminuir las complicaciones [6]. La ausencia de un tratamiento específico justifica las medidas actuales de protección en mujeres embarazadas, sobre todo con gestaciones cercanas al término [6]. A pesar de todos los esfuerzos en investigación sobre el tema, hay poca información sobre la seroprevalencia de infección por chikunguña y la inci- dencia de transmisión vertical [3]. La exposición prenatal a los anticuerpos maternos determina el paso trasplacentario y el tiempo de latencia de la inmunoglobulina G (IgG) específica, durante la primera infancia. Este paso de anticuerpos predomina en el final del embarazo, en un embarazo a término [6]. Un estudio de seroprevalencia después de una epidemia, basada en el distrito Thepa en la provincia de Songkhla, Tailandia, utilizando pruebas para IgM e IgG en sangre materna, en las que dieron positivo para IgM, también se estudiaron muestras del cordón umbilical [3]. La seroprevalencia de IgM positiva fue de 227 con IgG positiva en 319 casos, interpretado como un 71 % de seroconversión. Con una seroprevalencia previa al brote (IgM e IgG positiva) de 43, para un total del 6 % de seroconversión. Ninguno de los recién nacidos cuyas madres eran IgM tenían la sangre del cordón positiva [3]. La permeabilidad de la barrera placentaria para chikunguña se demostró in vitro [4] y la hipótesis de la transmisión vertical se centra en la baja penetrancia de los anticuerpos protectores inmaduros de origen materno, que no logran el paso al recién nacido. Ello permite la transmisión a través de la placenta durante el parto, principalmente en madres con alta viremia [6]. Por esta razón, no se recomienda la cesárea como intervención obstétrica de rutina. Existen varias propuestas para explicar los mecanismos de transferen- Silvia Catalina Corrales Martínez et al. Infección por chikunguña materno-fetal asociada con miocarditis neonatal 239 cia placentaria [6]; sin embargo, no es claro por qué puede existir infección tardía, ya que la mayor transferencia de anticuerpos maternos ocurre cerca del final del tercer trimestre, lo que debería suministrarle protección al recién nacido [6]. En diferentes investigaciones se ha relacionado el tiempo de seroconversión de anticuerpos IgG con el momento de la exposición durante el embarazo [7]. Más del 75 % de los niños tienen anticuerpos IgG persistentes más allá de un año después de la infección materna durante el primer trimestre, más del 30 % en el segundo trimestre, y menos del 1 % si la exposición ocurre en el tercer trimestre [7]. La transmisión vertical se ha definido como la aparición de signos clínicos de afectación neonatal. En los casos que se describen en este artículo se diagnosticó miocarditis por chikunguña, la cual resulta de la infección por un virus cardiotrópico seguido por la destrucción inflamatoria del miocardio [8]. La miocarditis también puede extenderse al pericardio y endocardio. Puede ser causado por una variedad de mecanismos como la infección viral o medicamentos cardiotóxicos [9]. Como resultado final de esta patología se encuentra cierto grado de disfunción cardiaca, desde leves efectos subclínicos, como arritmias, insuficiencia cardiaca, choque cardiogénico, hasta muerte súbita [9]. Con base en la gravedad de la enfermedad, la miocarditis se clasifica en tres ca- tegorías clínicas: aguda, fulminante y crónica. Los casos subclínicos más leves pueden tener resolución espontánea y pasar inadvertidos. Caso contrario en la enfermedad fulminante, que produce en los pacientes choque cardiogénico. En los casos crónicos puede progresar a miocardiopatía dilatada [9]. Como entidad, la caracterización de la miocarditis tiene presentaciones clínicas y etiologías diversas [8]. La miocarditis y la consecuente miocardiopatía que pudiera ocurrir es causada por agentes infecciosos y no infecciosos [8]. En las últimas cuatro décadas, la infección viral se ha reconocido como la etiología más común. Utilizando la reacción en cadena de la polimerasa y las técnicas de hibridación in situ, ARN y ADN viral, se han identificado pacientes con afección del miocardio. El tratamiento de la miocarditis sigue siendo en gran medida de apoyo sintomático. El tratamiento inmunosupresor puede ser favorable en pacientes con miocarditis de células gigantes y sarcoidosis. El uso de la terapia inmunomoduladora y antiviral permanece, en gran parte, en investigación en este momento. Se necesitan ensayos con tamaños de muestra mayores y aleatorizados controlados para determinar su papel en el tratamiento de la cardiopatía inflamatoria [8]. El presente reporte de caso demuestra la transmisión vertical del chikunguña. Por lo tanto, se sugiere que en Univ. Méd. ISSN 0041-9095. Bogotá (Colombia), 57 (2): 236-248, abril-junio, 2016 240 los casos en que las maternas presenten fiebre en trabajo de parto, en zonas endémicas, debe considerarse una emergencia médica, con indicación de hospitalización y, de tener la opción, incluir toma de muestras para estudios virológicos, clasificándolos como casos sospechosos de infección por arbovirus [5]. El bienestar fetal debe ser monitorizado regularmente y, si es necesario, tratar a la paciente como un parto de emergencia. Esto porque la transmisión vertical se observa casi exclusivamente en el contexto de la viremia materna intraparto, que conduce a infección neonatal grave, y estos recién nacidos prematuros o de término deben mantenerse en la unidad de cuidado intensivo neonatal durante al menos una semana, pues esta entidad representa un riesgo significativo para la sobrevida y salud de los neonatos, e implicaciones en la salud pública [5]. Informe del caso Se presentan los casos de dos pacientes neonatales en quienes se evidenciaron manifestaciones cardiacas dadas por bradiarritmias. Uno pretérmino y otro a término, ambos hijos de madres que cursaron con síndrome febril y sintomatología probablemente secundarios a chikunguña. En un monitoreo materno fetal del tercer trimestre se demostró bienestar cardiaco fetal. Caso 1 El caso corresponde a un recién nacido de sexo masculino, de 36 semanas de gestación al nacer. Tuvo un peso adecuado para la edad gestacional de 2890 gramos y talla de 51 centímetros. La madre, de 33 años de edad, era multípara, con pobres controles prenatales, quien no tuvo ruptura prematura de membranas. Los resultados de sus exámenes paraclínicos fueron negativos para enfermedades infecciosas: sífilis, toxoplasma, rubéola, citomegalovirus y herpes (STORCH), hepatitis B y VIH, durante la gestación. Durante la anamnesis, la madre refirió que nueve días previos al parto presentó un episodio de fiebre no cuantificada asociado a náuseas, cefalea y mialgias. Por este motivo consultó el servicio de urgencias, donde no identificaron signos de alarma; sin embargo, se le realizó un monitoreo fetal reactivo, cuyo resultado no evidenció alteraciones en la frecuencia cardiaca del bebé que sugiriera sufrimiento o patología fetal (figura 1). Se interpretó como síndrome febril por chikunguña. El día del parto se obtuvo un recién nacido por cesárea de urgencia, debido a presencia de taquicardia fetal mayor de 200 latidos por minuto (lpm) y perfil biofísico que evidenciaba arritmias y pausas cardiacas que alternaban con taquicardia. Silvia Catalina Corrales Martínez et al. Infección por chikunguña materno-fetal asociada con miocarditis neonatal 241 a) b) Figura 1. a) Monitoreo fetal. b) Perfil biofísico Al nacer, el paciente presentó bradicardia con paro cardiorrespiratorio, que requirió una adaptación inducida. Luego de la intubación orotraqueal se obtuvo una frecuencia cardiaca (FC) de 160 lpm y saturación de oxígeno del 96 % con fracción inspirada de oxígeno (FIO2) del 35 %. Durante su estancia en la unidad de cuidado intensivo neonatal (UCIN), se obtuvo en el electrocardiograma (EKG) taquicardia sinusal con FC 150 lpm, con cifras tensionales normales (figura 2a). Se realizó ecocardiograma que demostró disfunción biventricular severa de predominio derecho e hipertensión pulmonar con insuficiencia tricuspídea grado 3 (figura 2b). También se le realizó al neonato un EKGHolter, que evidenció variabilidad de la frecuencia cardiaca, afectada de forma severa y escasas extrasístoles supraventriculares (figura 2c). Por ello se inició manejo médico con digitálicos tipo be- tametildigoxina a dosis de 1 gota cada 12 horas (26 µg/kg) para mejorar la contractibilidad cardiaca; así como sildenafil a dosis de 5,80 mg cada 6 horas (2 mg/kg) para manejo de la hipertensión pulmonar. Por las características clínicas de la madre con enfermedad febril, con alta sospecha de chikunguña, se diagnosticó miocarditis de origen viral y se decidió realizar tratamiento con IgG (KIOVIG) a dosis de 3 g. Se ampliaron estudios paraclínicos con prueba para dengue negativa, y pruebas de función miocárdica con creatinina cinasa total, creatinina cinasa fracción MB y troponina I, las cuales fueron positivas. Se descartaron alteraciones metabólicas o electrolíticas (tabla 1). En un nuevo control ecocardiográfico, se evidenció mejoría clínica y recuperación de la función sistodiastólica ventricular izquierda, cierre del conducto arterioso y Univ. Méd. ISSN 0041-9095. Bogotá (Colombia), 57 (2): 236-248, abril-junio, 2016 242 disminución de la hipertensión pulmonar (figura 2d). El paciente evolucionó satisfactoriamente, en que un último reporte de exámenes paraclínicos estuvieron de parámetros normales. Se le dio egreso al neonato con manejo ambulatorio con digitálicos y control con cardiología e infectología pediátrica. Durante el cuadro clínico no se observó otra sintomatología. Figura 2b. Ecocardiograma inicial. Disfunción miocárdica biventricular severa, conducto arterioso permeable hipertensivo, hipertensión pulmonar severa, seno coronario dilatado. FE 49 %. Flujo del istmo de 0,70 m/s, gasto cardiaco de 107 ml/kg minuto Figura 2a. Electrocardiograma derivación DII largo, caso 1. Taquicardia sinusal FC 150 lpm Figura 2d. Ecocardiograma de control. Buena función sistodiastólica ventricular izquierda, discinesia del tabique anteromedial. Seno coronario dilatado. FE 58 %. Flujo del istmo de 1,08 m/s, gasto cardiaco de 302 ml/kg minuto Caso 2 Figura 2c. Holter de 24 horas. Ritmo sinusal con frecuencia ventricular máxima de 175 latidos por minuto. Arritmia sinusal con escasas extrasístoles supraventriculares. Variabilidad de la frecuencia cardiaca comprometida de forma severa Este caso corresponde a una recién nacida a término de 40 semanas y 3235 gramos de peso, de una madre multípara con controles prenatales adecuados, quien fue remitida de una institución de salud de primer nivel por un cuadro clínico que había iniciado cuatro días previos al parto, consistente en fiebre no Silvia Catalina Corrales Martínez et al. Infección por chikunguña materno-fetal asociada con miocarditis neonatal 243 cuantificada asociada a cefalea, mialgias y artralgias. En la mujer se había evidenciado edema, rubor, calor y dolor localizado en la región maleolar interna del miembro inferior derecho y eritema en el ab- domen. Con nexo epidemiológico para chikunguña, se le diagnosticó síndrome febril de posible etiología viral por chikunguña. Al tercer día de fiebre materna, presentó actividad uterina y parto vaginal sin complicaciones y sin nuevas distermias durante el puerperio. Tabla 1. Reporte de exámenes paraclínicos Caso 1 Caso 2 Unidades PCR 0,1 0,1 mg/dl Leucocitos 15,0 10,7 103 × mm3 Neutrófilos 60,7 54,5 % Monocitos 8,9 … % Linfocitos 28,1 33,8 % RBC 5,0 18,3 103 × mm3 Hemoglobina 16,7 16,0 g/dl Hematocrito 48,1 45,0 % Plaquetas 386,0 400,0 103 × mm3 BUN 12,5 6,9 mg/dl Creatinina 0,7 0,4 mg/dl CPK 507,6 191,1 U/L CK-MB 33,3 49,9 U/L Troponina I 0,06 0,1-0,9* ng/ml Dengue IgM Negativo Negativo Dengue IgG Positivo Positivo Negativo Negativo Dengue NS1Ag * Troponina inicial. Troponina de control 48 horas después. ** PCR: proteína C reactiva; RBC: Red Blood Cells; BUN: nitrógeno ureico; CPK: creatina fosfocinasa; CK-MB: creatina cinasa fracción MB; IgM: inmunoglobulina M; IgG: inmunoglobulina G; NS1 AG: antígeno contra proteína no estructural 1; TGO: transaminasa glutámico oxaloacética; TGP: transaminasa glutámico pirúvica; PT: tiempo de protrombina; PTT: tiempo parcial de tromboplastina; INR: International Normalized Ratio. Fuente: elaboración propia. Univ. Méd. ISSN 0041-9095. Bogotá (Colombia), 57 (2): 236-248, abril-junio, 2016 244 En el parto se obtuvo una recién nacida con adaptación inmediata espontánea. En el segundo día de vida presentó un episodio de somnolencia e hipotonía asociado a bradicardia afebril e hipoglucemia transitoria. Se le realizó un EKG donde se observó ritmo sinusal y ondas T aplanadas con presencia de ondas Q (figura 3a). Además de troponina I positiva (véase tabla 1). Un ecocardiograma fue compatible con hipertensión pulmonar leve, sin deficiencia miocárdica evidente, por lo que se indicó manejo expectante con oxígeno a bajo flujo y monitorización continua. Dados los hallazgos clínicos y paraclínicos, se sospechó miocarditis viral neonatal secundaria a probable transmisión vertical del chikunguña. A la recién nacida se le tomó un nuevo EKG que mostró presencia de ritmo sinusal y persistencia de ondas Q y T planas (figura 3a). En un nuevo ecocardiograma se registró frecuencia cardiaca limítrofe, disminución del gasto cardiaco, disfunción contráctil ventricular izquierda, diámetros diastólicos hacia la dilatación, hipertensión pulmonar leve e insuficiencia tricuspídea grado 2 (figura 3b). Figura 3a. Electrocardiograma derivación DII larga. Ritmo sinusal. FC 112 xm, presencia de ondas Q y T planas Figura 3b. Ecocardiograma inicial (izquierda). Buena función sistodiastólica ventricular izquierda. Gasto cardiaco de 209 ml/kg/min. Insuficiencia tricuspídea para una PSPE de 45-47 mmHg y mitral mínima. FE 57 %. FC 110 xm. Ecocardiograma control (derecha). Disfunción contráctil ventricular izquierda, con frecuencia cardiaca limítrofe. Diámetros diastólicos ventriculares hacia la dilatación, gasto cardiaco de 188 ml/kg/min FE 52 %. FC 99 xm Silvia Catalina Corrales Martínez et al. Infección por chikunguña materno-fetal asociada con miocarditis neonatal 245 Dada la afectación cardiaca se administró IgG (INTRATEC) a dosis de 1 g/kg. Posteriormente, en un control con exámenes paraclínicos se observó una ligera disminución de marcadores cardiacos (tabla 1). Al quinto día de edad, tercer día de enfermedad, presentó exantema difuso generalizado sin ningún otro síntoma (figura 4). La bebé continuó con una evolución clínica satisfactoria sin requerimientos de apoyo inotrópico y se le dio egreso con control en cardiología e infectología pediátrica. Durante el curso de la enfermedad se realizaron toma de muestras para serología por chikunguña y niveles de anticuerpos. Se obtuvo PCR positiva en ambos casos, con IgM negativas. Figura 4. Fotografía de la paciente con exantema difuso generalizado al tercer día del inicio de los síntomas Discusión La enfermedad viral por chikunguña es transmitida a los humanos por la picadura de mosquitos infectados y se ha vuelto emergente en las Américas [10]. Los vectores involucrados son Aedes aegypti y Aedes albopictus, distribuidos en países tropicales, subtropicales e, incluso, en climas templados [11]. En este trabajo se describió otra de las formas de transmisión que pueden suceder en el periodo cercano al parto o durante este. La transmisión vertical en el 48,7 % de las gestantes se asocia con enfermedad neonatal grave [6]. Durante el inicio del embarazo, la infección es inusual; pero si ocurre, puede ser causa de abortos [11]. Esta transmisión vertical se encuentra asociada a enfermedad neonatal grave y es relativamente poco frecuente; se ha descrito una prevalencia global de infecciones materno-fetales después de las 22 semanas de gestación del 0,25 %. En contraste, durante el parto, la tasa de transmisión en madres con viremia alta es cercana al 50 %, momento en que destaca el periodo intraparto como el momento crítico para transmisión al neonato [7]. En la literatura se encontró que la barrera placentaria es eficaz en la prevención de la transmisión de chikunguña materno-fetal durante el trabajo de parto [7]. Pero la transmisión maternofetal se observa con frecuencia en madres con viremia hasta el término del Univ. Méd. ISSN 0041-9095. Bogotá (Colombia), 57 (2): 236-248, abril-junio, 2016 246 embarazo [2]. Cuando la madre es altamente virémica, con una carga viral media de 1,5 millones copias por mililitro en plasma, se reporta que es posible el paso de la barrera placentaria. Con la aparición de contracciones uterinas durante el parto, la tasa de transmisión vertical puede llegar a ser del 48,7 %, en los recién nacidos expuestos al virus durante el parto. Este valor podría ser predictivo para transmisión confirmada. Se ha demostrado la presencia de IgG específica para chikunguña durante más de 6 meses en más del 50 % de los recién nacidos infectados. Estos anticuerpos podrían ser transferidos pasivamente y exhibir una actividad neutralizante y protectora [2]. Cuando existe transmisión vertical, se encuentra una mayor tasa de bradiarritmias fetales y asfixia neonatal [2]. Los dos casos reportados presentaron arritmias y disfunción miocárdica que se resolvieron con la administración de Ig. El diagnóstico se realiza con base en los signos y síntomas clínicos y el nexo epidemiológico, y se confirma por criterios de laboratorio. El diagnóstico de laboratorio se hace mediante la detección del virus, del ARN viral y de anticuerpos específicos [12]. El principal hallazgo de laboratorio es linfopenia, cuando el recuento de linfocitos es menor de 1000/mm3. Este hecho no se observó en los casos presentados. Otras anomalías de laboratorio incluyen trombocitopenia, alteración de la función hepática y alteraciones hidroelectrolíticas, como hipocalcemia [12]. El diagnóstico de laboratorio de la infección por chikunguña se basa en dos tipos de pruebas: la detección de ARN viral mediante la transcripción inversa de la cadena de polimerasa específica e IgM específica para el virus. En ambos casos se demostró la presencia de la infección aguda con reacción en cadena de la polimerasa para el virus. La transcripción inversa de la cadena de polimerasa es positiva durante la fase de viremia y la semana después del comienzo de los síntomas. Los anticuerpos específicos IgM son detectados al quinto día de la aparición de la enfermedad y pueden persistir hasta tres meses, mientras que la IgG específica aparece en el día 15 y permanece por años [12]. La IgG persiste de por vida, sin conferir una protección duradera con seguridad [6]. En uno de los dos casos reportados se observó que los recién nacidos expuestos no fueron sintomáticos al nacer, con aparición de síntomas antes del séptimo día, por lo que se recomienda mantenerlos en estancia hospitalaria al menos una semana, mientras se toman los exámenes paraclínicos [2]. Como consecuencia del reciente aumento y distribución del vector en todo el mundo [10], el chikunguña tiene el potencial de causar brotes masivos en el futuro y de transformarse en un problema de salud pública; además, en pediatría representa un papel fundamental en el Silvia Catalina Corrales Martínez et al. Infección por chikunguña materno-fetal asociada con miocarditis neonatal 247 riesgo de transmisión materno-fetal, donde se debe hacer diagnóstico y tratamiento temprano, dado la afectación multisistémica, que puede generar esta infección. En ambos casos reportados, se requirió manejo en UCIN, y por los hallazgos cardiacos, ambos pacientes recibieron manejos adecuados, incluyendo Ig, que permitió una evolución satisfactoria y una mejoría evidente de la inflamación del miocardio, de los parámetros hemodinámicos, así como la resistencia vascular sistémica y el gasto cardíaco con el uso de la inmunomodulación [8]. mother-to-child chikungunya virus infections on the island of La Réunion. PLoS Med. 2008;5(3):0413-23. 3. Laoprasopwattana K, Suntharasaj T, Petmanee P, Suddeaugrai O, Geater a. Chikungunya and dengue virus infections during pregnancy: seroprevalence, seroincidence and maternal-fetal transmission, southern Thailand, 2009-2010. Epidemiol Infect [Internet]. 2015;1-8. Disponible en: http:// www.journals.cambridge.org/abstract_ S0950268815001065 4. Gérardin P, Sampériz S, Ramful D, Boumahni B, Bintner M, Alessandri JL, et al. Neurocognitive outcome of children exposed to perinatal mother-to-child chikungunya virus infection: The CHIMERE Cohort Study on Reunion Island. PLoS Negl Trop Dis. 2014;8(7). 5. Shrivastava A, Waqar Beg M, Gujrati C, Gopalan N, Rao PVL. Management of a vertically transmitted neonatal Chikungunya thrombocytopenia. Indian J Pediatr. 2011;78(8):1008-9. 6. Ramful D, Carbonnier M, Pasquet M, Bouhmani B, Ghazouani J, Noormahomed T, et al. Mother-to-child transmission of Chikungunya virus infection. Pediatr Infect Dis J. 2007;26(9):811-5. 7. Ramful D, Samperiz S, Fritel X, Michault a., Jaffar-Bandjee M-C, Rollot O, et al. Antibody kinetics in infants exposed to chikungunya virus infection during pregnancy reveals absence of congenital infection. J Infect Dis [Internet]. 2014;209(11):172630. Disponible en: http://jid.oxfordjournals. org/lookup/doi/10.1093/infdis/jit814 8. Pollack A, Kontorovich AR, Fuster V, Dec GW. Viral myocarditis—diagnosis, treatment options, and current controversies. Nat Rev Cardiol [Internet]. 2015;1-12. Disponible en: http://www.nature.com/ doifinder/10.1038/nrcardio.2015.108 Agradecimientos Agradecemos al personal de la Unidad de Cuidados Intensivos Neonatales del Hospital Universitario Hernando Moncaliano Perdomo, de Neiva, Colombia, por su colaboración. Conflictos de interés Los autores no registran conflictos de interés. Referencias 1. 2. Lum F-M, Ng LFP. Cellular and molecular mechanisms of chikungunya pathogenesis. Antiviral Res [internet]. 2015;120:165-74. Disponibl en: http:// linkinghub.elsevier.com/retrieve/pii/ S0166354215001400 Gérardin P, Barau G, Michault A, Bintner M, Randrianaivo H, Choker G, et al. Multidisciplinary prospective study of Univ. Méd. ISSN 0041-9095. Bogotá (Colombia), 57 (2): 236-248, abril-junio, 2016 248 9. Durani Y, Giordano K, Goudie BW. Myocarditis and pericarditis in children. Pediatr Clin North Am [Internet]. 2010;57(6):1281-303. Disponible en: http://dx.doi.org/10.1016/j. pcl.2010.09.012 10. Weaver SC, Lecuit M. Chikungunya virus and the global spread of a mosquito-borne disease. N Engl J Med. 2015;372:12311239. doi: 10.1056/NEJMra1406035 11. Fritel X, Rollot O, Gérardin P, Gaüzère BA, Bideault J, Lagarde L, et al. Chikungunya virus infection during pregnancy, Réunion, France, 2006. Emerg Infect Dis. 2010;16(3):418-25. 12. Jaramillo Restrepo BN. Infección por el virus del chikungunya. Rev CES Med [internet]. 2014;28(2):313-23. Disponible en: http://www.scielo.org.co/pdf/cesm/ v28n2/v28n2a16.pdf Correspondencia Silvia Catalina Corrales Martínez catalinasc.corrales@gmail.com Silvia Catalina Corrales Martínez et al. Infección por chikunguña materno-fetal asociada con miocarditis neonatal
https://openalex.org/W2045398367	https://zenodo.org/records/1259889/files/article.pdf	English	null	GCF2: expression and molecular analysis of repression	Biochimica et biophysica acta, N. Gene structure and expression/Biochimica et biophysica acta. Gene structure and expression	2,003	public-domain	7,310	0167-4781/$ - see front matter. Published by Elsevier B.V. doi:10.1016/S0167-4781(03)00156-8 * Corresponding author. Tel.: +1-301-496-3224; fax: +1-301-480- 4667. E-mail address: acjohnson@nih.gov (A.C. Johnson). Abstract GC-binding factor 2 (GCF2) is a transcriptional repressor that decreases activity of the epidermal growth factor receptor (EGFR) and other genes. We have mapped the gene for GCF2 by fluorescence in situ hybridization (FISH) to chromosome 2q37. Sequence analysis of the GCF2 gene and cDNA showed that the gene consists of eight exons and introns and spans 73 kbp of DNA. Northern blot analysis showed that GCF2 mRNA was differentially expressed in many human tissues and cell lines. GCF2 mRNA was expressed as a 4.2 kb mRNA in most human tissues with the highest expression level in peripheral blood leukocytes and lowest expression in brain and testis. Additional transcripts of 6.6, 2.9 and 2.4 kb were found in some tissues but the only transcript detected in cancer cell lines was 4.2 kb with high levels found in seven Burkitts’ lymphoma cell lines. Western blot analysis showed that GCF2 protein is present at high levels in Burkitts’ lymphoma and several other cancer cell lines. GCF2 was found in both nuclear and cytoplasmic compartments in cells. Deletion mutants of GCF2 revealed that amino acids 429–528 are required for both DNA binding and repression of the EGFR promoter. Furthermore, GCF2 was able to substantially decrease activator protein 2 (AP2) enhancement of the EGFR promoter. Thus, GCF2 is a transcriptional repressor overexpressed in cancer cell lines with a role in regulating expression of the EGFR. Published by Elsevier B.V. Keywords: Epidermal growth factor receptor; Transcription; Repression; Promoter; Lymphoma GCF2: expression and molecular analysis of repression Toshiki Rikiyamaa, Joseph Curtisa, Masaya Oikawaa, Drazen B. Zimonjicb, Nicholas Popescub, Barbara A. Murphya, Melissa A. Wilsona, Alfred C. Johnsona,* aLaboratory of Molecular Biology, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Room 5002, Building 37, 37 Convent Drive MSC-4264, Bethesda, MD 20892-4264, USA bLaboratory of Experimental Carcinogenesis, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Room 5002, Building 37, 37 Convent Drive MSC-4264, Bethesda, MD 20892-4264, USA Received 9 May 2003; received in revised form 10 July 2003; accepted 17 July 2003 www.bba-direct.com 1. Introduction In addition to the EGFR promoter, GCF2 is also able to repress the activity of the Rous sarcoma virus (RSV) and simian virus 40 (SV40) promoters. Cells treated with phorbol esters demonstrate an inverse correlation of GCF2 mRNA and EGFR mRNA expression. In this study, we have initiated characterization of the GCF2 gene and its expression in human tissues and cancer cell lines. We further characterized GCF2 with respect to chromosomal and cellular localization and expression pat- tern in normal tissues and cancer cell lines. Using deletion mutants of GCF2, we have identified the region required for binding and repression. Furthermore, we determined that GCF2 is able to partially inhibit AP2 enhancement of EGFR promoter activity. identity (305/309) in this region. The deduced protein sequence has a calculated molecular weight of 83 kDa but migrates as a 160 kDa protein when subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS- PAGE). GCF2 binds to a specific site in the EGFR promoter and represses transcription of the gene. When the GCF2 binding site in the promoter is mutated, the EGFR promoter is no longer repressed by GCF2, correlating GCF2 binding and repression [26]. In addition to the EGFR promoter, GCF2 is also able to repress the activity of the Rous sarcoma virus (RSV) and simian virus 40 (SV40) promoters. Cells treated with phorbol esters demonstrate an inverse correlation of GCF2 mRNA and EGFR mRNA expression. or mutants/His were transformed to BL21 cells (Promega, Madison, WI) and correct clones were used to inoculate cultures of LB-ampicillin plus glucose and induced with IPTG (1 mM final concentration) at A600 = 0.7. The GST/ GCF2 and mutants/His fusion proteins were obtained by batch purification with GST-Sepharose according to the manufacturer’s instructions and were subjected to a second purification using Ni-NTA resin. Protein concentrations were determined using the Bio-Rad protein assay kit (Bio-Rad Laboratories, Hercules, CA). The purity and size of the mutant proteins were verified by SDS-PAGE and Coomassie staining. For antibody production, the GST–GCF2 fusion protein prepared from IPTG induced cells was dialyzed against saline and used to inoculate rabbits. Antiserum was raised in New Zealand rabbits and tested for its ability to immu- noprecipitate GCF2 made in vitro. The crude antiserum was subjected to protein A chromatography using ImmunoPure immobilized Recomb Protein A according to the manufac- turer’s protocol (Pierce, Rockford, IL). 2.1. Chromosomal localization A genomic clone in a BAC vector was obtained from CHORI (Oakland, CA). The presence of the GCF2 gene was confirmed by PCR analysis with primers specific for the GCF2 cDNA. A 75-kbp-long GCF2 probe was labeled with biotin or digoxigenin (Random Primed DNA Labeling Kit, Boehringer-Mannheim, Indianapolis, IN), and used for in situ hybridization of human chromosomes derived from methotrexate-synchronized normal peripheral lymphocyte cultures. The conditions of hybridization, the detection of hybridization signals, digital-image acquisition, processing and analysis as well as the procedure for direct visualization of fluorescent signals on banded chromosomes were carried out as previously described [27]. 2. Materials and methods From cultured cells, total RNA was isolated by guanidi- nium–thiocyanate–phenol–chloroform extraction [28]. Poly (A)+ RNA was selected from the total RNA population by oligo (dT)-cellulose chromatography [29]. RNA was fractionated on a 1% formaldehyde–agarose gel and trans- ferred to nitrocellulose [30]. RNA was UV crosslinked to the nitrocellulose filter, prehybridized, hybridized and washed as previously described [31]. Labeled cDNA probes were prepared by random primer extension of PCR gener- ated fragments [32]. Tissue and cancer cell line blots were purchased (Clontech, Palo Alto, CA) and probed according to manufacturer’s instructions. 1. Introduction The eluted IgG fraction was adjusted to 0.3 mg/ml with bovine serum albumin and dialyzed against phosphate buffered saline. In this study, we have initiated characterization of the GCF2 gene and its expression in human tissues and cancer cell lines. We further characterized GCF2 with respect to chromosomal and cellular localization and expression pat- tern in normal tissues and cancer cell lines. Using deletion mutants of GCF2, we have identified the region required for binding and repression. Furthermore, we determined that GCF2 is able to partially inhibit AP2 enhancement of EGFR promoter activity. 1. Introduction well-characterized transcription factors bind to the promoter region and influence transcription including specificity protein 1 (Sp1), AP2, p53, and Wilms’ Tumor (WT1) [13–18]. The p53 homologue p63 has also been shown to regulate EGFR expression [19]. Less characterized tran- scription factors are also implicated in EGFR gene regula- tion [11,20–22]. Also, two repressors that bind to GC-rich sequences bind to the EGFR promoter [23,24]. More recently, a role for the early growth response gene (Egr-1) in upregulation of EGFR expression during hypoxia has been determined [25]. The epidermal growth factor receptor (EGFR) must be correctly regulated to ensure proper cell growth. Overex- pression of the receptor is found in many tumors and cancer cell lines [1,2]. Overexpression occurs by gene amplifica- tion with or without gene rearrangement. Overexpression of EGFR in breast cancer correlates with failure on endocrine therapy [3,4]. The EGFR is considered a marker for cell transformation, and therefore it is an attractive target for clinical intervention [5,6]. High-level expression is also seen when there is no amplification in many cancer cell lines [7]. A 3.6 kbp cDNA that lacks a portion of the 3V untrans- lated sequences was cloned and corresponds to a 4.2 kb mRNA. This cDNA was termed GC-binding factor 2 (GCF2) based on its homology to GC-binding factor (GCF) [26]. Characterization of this cDNA revealed a region of homology with GCF that includes the amino terminal (DNA binding) region and the 5V untranslated region. The homology extends for 309 bp and has 98% The regulatory region of the EGFR gene has been cloned and characterized [8–11]. Mechanisms have been proposed for regulation of the EGFR gene in breast cancer [12]. Many 0167-4781/$ - see front matter. Published by Elsevier B.V. doi:10.1016/S0167-4781(03)00156-8 T. Rikiyama et al. / Biochimica et Biophysica Acta 1629 (2003) 15–25 16 identity (305/309) in this region. The deduced protein sequence has a calculated molecular weight of 83 kDa but migrates as a 160 kDa protein when subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS- PAGE). GCF2 binds to a specific site in the EGFR promoter and represses transcription of the gene. When the GCF2 binding site in the promoter is mutated, the EGFR promoter is no longer repressed by GCF2, correlating GCF2 binding and repression [26]. 2.4. In vitro transcription/translation and immunoprecipitation (B) Exon/intron structure of GCF2 with amino acids encoded by each exon. Fig. 1. Chromosomal localization and gene structure of GCF2. (A) Chromosome localization of the GCF2 gene by FISH. Digital image of a metaphase derived from methotrexate-synchronized normal human peripheral leukocytes after hybridization with GCF2 probe and DAPI counterstaining. Both chromosomes 2 have signals on of the long arm of chromosome 2 at band q36-37. (B) Exon/intron structure of GCF2 with amino acids encoded by each exon. for 1 h. The gel was fixed in 20% methanol/10% acetic acid, immersed in Enlightningk (Perkin-Elmer, Boston, MA) for 30 min, dried and exposed to film at 80 jC. for 1 h. The gel was fixed in 20% methanol/10% acetic acid, immersed in Enlightningk (Perkin-Elmer, Boston, MA) for 30 min, dried and exposed to film at 80 jC. at 37 jC for 10 min and samples centrifuged at 12,000 g for 5 min. The supernatant was collected and protein determination made using the Bio-Rad reagent. Cell fractionation was performed as described by Dyer and Herzog [34]. Briefly, cells were pelleted and resus- pended in sucrose buffer containing NP-40. The lysate was centrifuged at 500 g to pellet nuclei. The superna- tant was taken and processed to isolate cytoplasmic proteins. The pellet was resuspended in sucrose buffer, washed with a low salt buffer and lysed in high salt 2.4. In vitro transcription/translation and immunoprecipitation The open reading frame of GCF2 was amplified by the polymerase chain reaction (PCR) and subcloned into pCI- TE2A (Novagen, Madison, WI). GCF2 was synthesized in vitro into the presence of 35S-methionine with the coupled transcription translation system (TNT) from Promega. The in vitro prepared proteins were analyzed on SDS-polyacryl- amide gels [33]. To test the GCF2 antiserum, in vitro translated products (10 Al) were incubated with GCF2 antiserum (20 Al) on ice in RIPA buffer (100 Al) of for 3 h. Fifty microliters of a 10% suspension of Staphylococcus aureus was added and incubation continued for 30 min. The samples were pelleted and washed four times with RIPA buffer. The samples were resuspended in 20 Al SDS gel loading buffer, boiled for 3 min, loaded onto a 6% SDS polyacrylamide gel and subjected to electrophoresis at 200 V 2.2. Expression of GCF2 fusion proteins and antibody generation Plasmid pGST/His was prepared by ligation of the BamHI and EcoRI fragment from pQE60, which contains several restriction sites and 6xHis, into pGEX-1ET (Amer- sham, Piscataway, NJ). The full-length GCF2 cDNA except stop codon was subcloned into the BamHI and XbaI sites of pGST/His to generate pGST/GCF2/His, all other pGST/GCF2 mutants/His were amplified by PCR and subcloned into pGST/His in the same way. Plasmid DNAs were sequenced using the Applied Biosystem model 373A sequencer to confirm that the fusion was in frame and that no mutations were present. Plasmid pGST/GCF2 17 T. Rikiyama et al. / Biochimica et Biophysica Acta 1629 (2003) 15–25 for 1 h The gel was fixed in 20% methanol/10% acetic acid at 37 jC for 10 min and samples Fig. 1. Chromosomal localization and gene structure of GCF2. (A) Chromosome localization of the GCF2 gene by FISH. Digital image of from methotrexate-synchronized normal human peripheral leukocytes after hybridization with GCF2 probe and DAPI counterstaining. B have signals on of the long arm of chromosome 2 at band q36-37. (B) Exon/intron structure of GCF2 with amino acids encoded by ea y p y ( ) Fig. 1. Chromosomal localization and gene structure of GCF2. (A) Chromosome localization of the GCF2 gene by FISH. Digital image of a metaphase derived from methotrexate-synchronized normal human peripheral leukocytes after hybridization with GCF2 probe and DAPI counterstaining. Both chromosomes 2 have signals on of the long arm of chromosome 2 at band q36-37. 2.6. Western blotting Protein extracts were subjected to gel electrophoresis on a 6% SDS-PAGE. The proteins were transferred to nitro- cellulose and the nitrocellulose incubated in TPBS for 30 min. The GCF2 antiserum was added and incubation continued for 1 h at 37 jC. The nitrocellulose was washed three times with TPBS and the blot developed using the VectaStain ABC method according to the manufacturer’s protocol or the ECL plus detection kit (Amersham). 2.5. Preparation of cell lysates and cell fractionation Cells were harvested by scraping or centrifugation and resuspended in TD buffer at 106 cells per 100 Al. The cells were subjected to three cycles of freeze/thaw fol- lowed by sonication for 30 s. The samples were incubated 18 T. Rikiyama et al. / Biochimica et Biophysica Acta 1629 (2003) 15–25 Fig. 2. GCF2 mRNA expression in normal human tissues. Northern hybridization analysis was performed using tissue RNA blots purchased from Clontech. A radiolabeled 1.1 kbp GCF2 cDNA fragment was hybridized according to Clontech’s protocol. The tissues from which the RNAs were obtained are indicated above the figure. Sizes of hybridizing GCF2 RNAs are on the left of the figure. Results from reprobing the blot with actin are shown below the figure. Fig. 2. GCF2 mRNA expression in normal human tissues. Northern hybridization analysis was performed using tissue RNA blots purchased from Clontech. A radiolabeled 1.1 kbp GCF2 cDNA fragment was hybridized according to Clontech’s protocol. The tissues from which the RNAs were obtained are indicated above the figure. Sizes of hybridizing GCF2 RNAs are on the left of the figure. Results from reprobing the blot with actin are shown below the figure. buffer. A nuclear extract was produced and stored at 80 jC. For phorbol ester experiments, nuclear extracts were prepared from cells treated with DMSO (control) or PMA (40 ng/ml for 2 h) by the method of Dignam et al. [35]. otide was end-labeled with P-32 using T4 polynucleotide kinase and gamma ATP. For the gel shift analysis, the end- labeled double-stranded oligonucleotide was incubated with GCF2-His and/or AP2 at room temperature (23 jC) for 15 min in the presence of 10 mM Tris pH 7.5, 1 mM MgCl2, 0.5 mM EDTA, 0.5 mM DTT, 50 mM NaCl, 50 Ag/ml poly (dI–dC)(poly dI–dC) and 4% glycerol. Samples (10 Al) were loaded onto a 5% polyacrylamide gel and subjected to electrophoresis at 150 V for 2 h using 0.5 TBE (1 TBE = 89 mM Tris, 8 mM boric acid and 2 mM EDTA, pH 8.3) as running buffer. After electrophoresis, gels were transferred to Whatman 3 MM paper and exposed to Kodak XAR film with intensifying screens at 70 jC. 2.7. Electrophoretic mobility shift assay NIH3T3 (2 105) cells grown in triplicate in 35-mm- diameter wells were transfected with the expression vectors by lipofectamine (Invitrogen). The EGFR luciferase reporter construct (pER1-Luc) containing the EGFR promoter was prepared by ligation of the 1.1 kbp HindIII promoter fragment from pEGFR-CAT1 into pGL3-Basic (Promega). The reporter plasmid (0.1 Ag) was cotransfected with the indicated amount of expression vectors. DNA concentration was kept constant by addition of herring sperm DNA. Cells were harvested 24 h after transfection and cell extracts prepared according to the protocol from Analytical Lumi- nescence Laboratory (San Diego, CA). All luciferase activ- ities were normalized for protein concentration and Mobility shift assays were performed as described pre- viously [13]. A double-stranded oligonucleotide containing the GCF2 binding site was prepared by annealing two complementary oligonucleotides containing nucleotides 249 to 229 of the EGFR promoter in a buffer contain- ing 10 mM Tris, pH 8.0, 500 mM NaCl and 1 mM EDTA. Equimolar amounts of the complementary oligonucleotides were mixed in a 1.5 ml Eppendorf centrifuge tube and placed in a heat block at 95 jC. The heat block was allowed to cool to room temperature and the sample was desalted on a G-25 Sephadex column. The double-stranded oligonucle- 19 T. Rikiyama et al. / Biochimica et Biophysica Acta 1629 (2003) 15–25 NA expression in cancer cell lines. (A) Northern hybridization analysis was performed using the cancer cell line RNA blot from Clontech. A bp GCF2 cDNA fragment was hybridized according to Clontech’s protocol. The names of the cell lines and type of cancers from which they indicated. RNA marker sizes are indicated on the left. (B) GCF2 mRNA expression in Burkitts’ lymphoma cell lines. Northern blot analysis in A with RNA isolated from seven Burkitts’ lymphoma cell lines (lanes 2–7). RNA from D551, a normal human fibroblast cell line, was Actin levels did not significantly change for any of the samples (data not shown). (C) GCF2 mRNA expression in breast cancer cell lines. ysis was performed as in A with RNA isolated from eight breast cancer cell lines. Actin levels did not significantly change for any of the shown). Fig. 3. GCF2 mRNA expression in cancer cell lines. (A) Northern hybridization analysis was performed using the cancer cell line RNA blot from Clontech. A radiolabeled 1.1 kbp GCF2 cDNA fragment was hybridized according to Clontech’s protocol. 2.7. Electrophoretic mobility shift assay The names of the cell lines and type of cancers from which they were obtained are indicated. RNA marker sizes are indicated on the left. (B) GCF2 mRNA expression in Burkitts’ lymphoma cell lines. Northern blot analysis was performed as in A with RNA isolated from seven Burkitts’ lymphoma cell lines (lanes 2–7). RNA from D551, a normal human fibroblast cell line, was present in lane 1. Actin levels did not significantly change for any of the samples (data not shown). (C) GCF2 mRNA expression in breast cancer cell lines. Northern blot analysis was performed as in A with RNA isolated from eight breast cancer cell lines. Actin levels did not significantly change for any of the samples (data not shown). transfection efficiency using B-galactosidase. All experi- ments were performed in triplicate. somes 2. This signal was localized on DAPI counterstained metaphases at the region 2q36-37, where we assign the location of the GCF2 gene (Fig. 1A). The cDNA sequence for GCF2 was compared to the human genome sequence using Blast (National Center for Biotechnology Information, Bethesda, MD). The GCF2 cDNA matched sequences located at chromosome 2q37. The comparative analysis revealed the presence of eight exons and seven introns (Fig. 1B). Exon 8 is the largest and encodes for 511 of the 752 amino acids. 3. Results (A) In vitro translated GCF2, prepared by coupled transcription/translation, was immunoprecipitated with antiserum from rabbits immunized with a GST-GCF2 fusion protein or with preimmune serum from the same rabbit. The immunoprecipitated products were analyzed on a 6% SDS-polyacrylamide gel. Lanes: (1) Translation products without DNA; (2) total translation products using GCF2 DNA; (3) immunoprecipitation with preimmune serum; and (4) immunoprecipitation with GCF2 antiserum. Molecular weight markers are shown. (B) Total protein extracts were prepared from cell lines indicated above the blot and subjected to electrophoresis on a 6% SDS-polyacrylamide gel. Proteins were transferred to nitrocellulose and GCF2 detected using the IgG fraction from the antiserum. Fig. 4. Characterization of GCF2 antiserum and Western blot analysis. (A) In vitro translated GCF2, prepared by coupled transcription/translation, was immunoprecipitated with antiserum from rabbits immunized with a GST-GCF2 fusion protein or with preimmune serum from the same rabbit. The immunoprecipitated products were analyzed on a 6% SDS-polyacrylamide gel. Lanes: (1) Translation products without DNA; (2) total translation products using GCF2 DNA; (3) immunoprecipitation with preimmune serum; and (4) immunoprecipitation with GCF2 antiserum. Molecular weight markers are shown. (B) Total protein extracts were prepared from cell lines indicated above the blot and subjected to electrophoresis on a 6% SDS-polyacrylamide gel. Proteins were transferred to nitrocellulose and GCF2 detected using the IgG fraction from the antiserum. RNA from all seven lines we examined (Fig. 3B). Again, only the 4.2 kb mRNA was detected. addition to the 4.2 kb mRNA, some tissues expressed a 2.4 and/or a 2.9 kb mRNAs. Brain and testis showed very little expression of the 4.2 kb mRNA although other tissues except heart and skeletal muscle had an abundant amount. Heart and skeletal muscle both predominantly expressed a 2.9 kb mRNA. Peripheral blood leukocytes expressed very high levels of the 4.2 kb mRNA but also two larger RNAs of approximately 6.6 and 9 kb. The 2.4 kb mRNA was expressed highest in ovary, colon and small intestine. We next examined expression of GCF2 mRNA in various cancer cell lines (Fig. 3A). Unlike normal tissues, only the 4.2 kb mRNA was detected. The highest level of GCF2 mRNA was found in Raji cells derived from a Burkitts’ lymphoma. Leukemia and epidermoid cancer cell lines expressed relative lower levels of GCF 2 mRNA. We examined additional Burkitts’ lymphoma cell lines for GCF2 mRNA. 3. Results Chromosome preparations obtained from cultured pe- ripheral lymphocytes derived from two healthy donors were hybridized with a GCF2 probe. The majority, if not all of the metaphases with informative signal and a minimal nonspe- cific background, had symmetrical fluorescent signal on the terminal band of the long arm of chromosome 2, the largest submetacentric chromosome readily identifiable by size. Thirty labeled chromosome spreads recorded from both samples exhibited signal at the same site on both chromo- Since GCF2 acts to repress EGFR promoter activity, we were prompted to examine the expression of this gene in normal tissues and cancer cell lines. Northern blot hybrid- ization was used to detect GCF2 mRNA. As shown in Fig. 2, GCF2 mRNA expression varied in normal tissues. In erization of GCF2 antiserum and Western blot analysis. (A) In vitro translated GCF2, prepared by coupled transcription/translation, was tated with antiserum from rabbits immunized with a GST-GCF2 fusion protein or with preimmune serum from the same rabbit. The tated products were analyzed on a 6% SDS-polyacrylamide gel. Lanes: (1) Translation products without DNA; (2) total translation products NA; (3) immunoprecipitation with preimmune serum; and (4) immunoprecipitation with GCF2 antiserum. Molecular weight markers are shown. n extracts were prepared from cell lines indicated above the blot and subjected to electrophoresis on a 6% SDS-polyacrylamide gel. Proteins d to nitrocellulose and GCF2 detected using the IgG fraction from the antiserum. T. Rikiyama et al. / Biochimica et Biophysica Acta 1629 (2003) 15–25 20 T. Rikiyama et al. / Biochimica et Biophysica Acta 1629 (2003) 15–25 Fig. 4. Characterization of GCF2 antiserum and Western blot analysis. (A) In vitro translated GCF2, prepared by coupled transcription/translation, was immunoprecipitated with antiserum from rabbits immunized with a GST-GCF2 fusion protein or with preimmune serum from the same rabbit. The immunoprecipitated products were analyzed on a 6% SDS-polyacrylamide gel. Lanes: (1) Translation products without DNA; (2) total translation products using GCF2 DNA; (3) immunoprecipitation with preimmune serum; and (4) immunoprecipitation with GCF2 antiserum. Molecular weight markers are shown. (B) Total protein extracts were prepared from cell lines indicated above the blot and subjected to electrophoresis on a 6% SDS-polyacrylamide gel. Proteins were transferred to nitrocellulose and GCF2 detected using the IgG fraction from the antiserum. Fig. 4. Characterization of GCF2 antiserum and Western blot analysis. 3. Results High-level expression was detected in To examine protein expression of GCF2, an antiserum was raised against a GST–GCF2 fusion protein. The antiserum was tested by immunoprecipitation of in vitro translated GCF2 (Fig. 4A). Antiserum but not preimmune serum was able to specifically immunoprecipitate GCF2. The IgG fraction of the antiserum was purified by protein A Sepharose chromatography and used in Western blotting experiments to detect GCF2 expression in cells (Fig. 4B). GCF2 was detected as a 160 kDa protein and was found expressed at high levels in A498 cells (derived from a kidney carcinoma), A549 cells (derived from a lung carci- noma), AGS cells (derived from a gastric adenocarcinoma) and U118MG cells (derived from a glioblastoma). To determine the location of GCF2 within the cell, Western blotting was performed with nuclear and cytosolic 21 T. Rikiyama et al. / Biochimica et Biophysica Acta 1629 (2003) 15–25 tion of GCF2. (A) HUT 102 cells were fractionated into nuclear and cytosolic components as presented in Materials and methods. ubjected to Western blot analysis with GCF2 antiserum. Lanes: (1) GCF2 His fusion protein as a control; (2) total protein extract; (3) (4) nuclear fraction. (B) A431 cells were fractionated into nuclear and cytosolic components as presented in Materials and methods. bjected to Western blot analysis with GCF2 antiserum. Nuc = nuclear, Cyt = cytosolic. (C) Nuclear extract were prepared from the cell gure after treatment with or without PMA. Western blot analysis was performed using GCF2 antibody. Fig. 5. Cellular localization of GCF2. (A) HUT 102 cells were fractionated into nuclear and cytosolic components as presented in Materials and methods. Protein extracts were subjected to Western blot analysis with GCF2 antiserum. Lanes: (1) GCF2 His fusion protein as a control; (2) total protein extract; (3) cytosolic fraction; and (4) nuclear fraction. (B) A431 cells were fractionated into nuclear and cytosolic components as presented in Materials and methods. Protein extracts were subjected to Western blot analysis with GCF2 antiserum. Nuc = nuclear, Cyt = cytosolic. (C) Nuclear extract were prepared from the cell lines listed above the figure after treatment with or without PMA. Western blot analysis was performed using GCF2 antibody. fractions prepared from HUT-102 cells that express high levels of GCF2. The majority of GCF2 was found in the nuclear fraction but some GCF2 was detected in the cytosol (Fig. 5A). 3. Results In A431 cell, GCF2 is distributed almost equally in nuclear and cytosolic fractions (Fig. 5B). We had previ- ously shown that GCF2 levels decreased when EGFR expression increased due to phorbol ester treatment of cells. We examined the consequence of phorbol ester treatment on GCF2 cellular location. Nuclear extracts from Raji, Hela and K562 cells were examined for GCF2 nuclear expres- sion. All nuclear extracts showed decreased GCF2 nuclear expression after phorbol ester treatment (40 ng/ml for 2 h) (Fig. 5C). GCF2 represses EGFR transcriptional activity by binding to a specific site in the promoter. To determine the region of GCF2 required for binding and repression of promoter activity, we performed electrophoretic mobility shift assays and cotransfection experiments with deletion mutants of GCF2. GCF2 deletion mutants that contained amino acids 428–752 (lanes 2–5 and 12) were able to bind the GCF2 binding site located in the EGFR promoter whereas deletion mutants that contained amino acids 529– 752 (lane 6) could not (Fig. 6). A GCF2 deletion mutant expressing amino acids 429–528 was also able to bind the GCF2 binding site oligonucleotide (lane 18). Consistent with the hypothesis that GCF2 binding is required for repression, cotransfection assays showed that amino acids 429–528 were necessary for full repression of EGFR promoter activity (Fig. 7). Whereas full-length GCF2 re- duced EGFR promoter greater than 80%, GCF2 429–528 reduced activity by greater than 70%, indicating the require- ment of this region for DNA binding and repression. We observed that the GCF2 binding site from 249 to 233 in the EGFR promoter overlapped with an AP2 binding site located 253 to 232. We used cotransfec- tion experiments to determine if GCF2 could inhibit AP2 enhancement of the EGFR promoter. Transfections of NIH3T3 cells with increasing concentrations of an AP2 expression vector increased EGFR promoter activity in a dose-dependent manner. Conversely, transfection with a GCF2 expression vector decreased EGFR promoter activity in a dose-dependent manner. Cotransfection with AP2 and GCF-2 expression vectors resulted in decreased EGFR 22 T. Rikiyama et al. / Biochimica et Biophysica Acta 1629 (2003) 15–25 Fig. 6. Electrophoretic mobility shift with GCF2 deletion mutants and EGFR GCF2 binding site. The GCF2 binding site oligonucleotide was incubated with equivalent amounts of GCF2 His proteins (indicated above each lane) and assays performed as described in Materials and methods. The bound area indicates the region of the shifted bands. 4. Discussion We report here that GCF2 is differentially expressed in human cell lines and tissues. At least three RNA species from human tissues hybridized to GCF2 cDNA. The 4.2 kb mRNA reported previously was found in all tissues exam- ined with very low levels detected in brain and testis. Fetal liver, bone marrow, thymus, spleen, lymph node, pancreas, liver, placenta and kidney expressed the 4.2 kb mRNA as the predominant species of RNA. Heart and skeletal muscle had very low levels of the 4.2 kb mRNA but possessed an abundant level of a 2.9 kb mRNA. In addition, a high level of a 2.4 kb mRNA was found in RNA from ovary, small intestine and colon. Peripheral blood leukocytes expressed not only a high level of the 4.2 kb mRNA but also significant levels of a 6.6 and a 9 kb mRNA. The relation- ship between these species of RNA remains to be deter- mined. In cancer cell lines, the only mRNA that hybridized to GCF2 cDNA was 4.2 kb. In terms of regulation, GCF2 expression decreases upon phorbol ester treatment of cells [26]. It is increased, how- ever, by treatment of PC12 cells with nerve growth factor [39]. In both of these instances, EGFR expression is inversely correlated. In cells, the nuclear level of GCF2 is decreased by factors that activate EGFR expression. This allows for increased EGFR expression that is returned to steady state levels as GCF2 levels are increased. We speculate that posttranslational modifications such as phos- phorylation may be the trigger for translocation of GCF2 between nuclear and cytoplasmic compartments. GCF2 has been shown to repress transcription of the EGFR and PDGF-A chain promoters [26,40]. GCF2 binds to the PDGF-A chain promoter and competes with several zinc-finger transcription factors. We examined the effect of expressing both GCF2 and AP2, a known activator of EGFR expression, in cells. GCF2 partially inhibited AP2 enhancement of EGFR expression. However, the combi- nation of GCF2 and AP2 did not alter the DNase I footprint in size or intensity (data not shown). These results suggest that GCF2 represses EGFR promoter activ- ity by competition with positive activators such as AP2. In this report, we have shown that GCF2 requires amino acids 429–528, which contains a lysine-rich motif, for repression and DNA binding. The position of the GCF2 gene on chromosome 2 overlaps with an aphidicholine-inducible common fragile site (FS). Fig. 6. Electrophoretic mobility shift with GCF2 deletion mutants and EGFR GCF2 binding site. The GCF2 binding site oligonucleotide was incubated with equivalent amounts of GCF2 His proteins (indicated above each lane) and assays performed as described in Materials and methods. The bound area indicates th i f th hift d b d Fig. 7. Repression activity of GCF2 deletion mutants. Eukaryotic expression constructs (2 Ag) containing the open reading frame for GCF2 amino acids region i di t d t th l ft f th fi t f t d i t NIH3T3 ll ith ER1 L d ib d i M t i l d th d L if ti it d Fig. 7. Repression activity of GCF2 deletion mutants. Eukaryotic expression constructs (2 Ag) containing the open reading frame for GCF2 amino acids regions indicated to the left of the figure were cotransfected into NIH3T3 cells with pER1 Luc as described in Materials and methods. Luciferase activity was measured in lysates prepared 24 h after transfection and is expressed relative to the activity of pCDNA3. Activity is corrected for protein concentration and transfection efficiency. Fig. 7. Repression activity of GCF2 deletion mutants. Eukaryotic expression constructs (2 Ag) containing the open reading frame for GCF2 amino acids regions indicated to the left of the figure were cotransfected into NIH3T3 cells with pER1 Luc as described in Materials and methods. Luciferase activity was measured in lysates prepared 24 h after transfection and is expressed relative to the activity of pCDNA3. Activity is corrected for protein concentration and transfection efficiency. Fig. 7. Repression activity of GCF2 deletion mutants. Eukaryotic expression constructs (2 Ag) containing the open reading frame for GCF2 amino acids regions indicated to the left of the figure were cotransfected into NIH3T3 cells with pER1 Luc as described in Materials and methods. Luciferase activity was measured in lysates prepared 24 h after transfection and is expressed relative to the activity of pCDNA3. Activity is corrected for protein concentration and transfection efficiency. T. Rikiyama et al. / Biochimica et Biophysica Acta 1629 (2003) 15–25 23 Fig. 8. GCF2 inhibits AP2 enhancement of the EGFR promoter. NIH3T3 cells were transfected with pER1 Luc (0.1 Ag) and indicated concentrations of pRSV AP2 and pCMV GCF2. Activity is relative to the sample containing no GCF2 or AP2 plasmid. Varying amounts of control vector was used to keep the amount of DNA constant. drome (2q31-ter), cerebrotendinous xanthomatosis (CTX) (2q33-ter) and Klein–Waardenburg syndrome (2q35-37) as well as in alveolar rhabdomyosarcoma with a specific trans- location t(2;13) [37,38]. Molecular analyses might reveal a role of GCF2 gene in these diseases. Significantly, in B and T cell lymphoma cell lines, we detected higher levels of GCF2 expression than in other tumor cell lines. High-level expres- sion may be associated with chromosomal alterations and suggests that GCF2 gene is important in pathogenesis of malignant lymphomas. The level of GCF2 was also high in kidney, lung, gastric and glioblastomas cancer cell lines with variable in a panel of breast cancer cell lines. The expression level of GCF2 may reflect differences in the regulatory proteins that control GCF2 expression in these cells. GCF2 was found primarily in the nuclear compartment within the cell but a measurable amount was also cyto- plasmic. The percentage of GCF2 found in the nuclear or cytoplasmic fractions was variable within different cell lines. For example, GCF2 was predominantly in the nuclear fraction of extracts from prepared from HUT- 102 cells but was equally distributed in A431 cells (derived from an epidermoid carcinoma) (Fig. 5). Several protein bands of approximately 160 kDa were reactive with the GCF2 antibody. We suspect that posttranslational modifications may be involved in the determination of cellular location of GCF2. We have shown that phorbol esters reduce nuclear GCF2 levels in several cell lines. Apparently, GCF2 levels are reduced prior to PMA- enhanced EGFR expression. This is consistent with a previous kinetic analysis of GCF2 and EGFR expression after phorbol ester treatment of cells [26]. Fig. 8. GCF2 inhibits AP2 enhancement of the EGFR promoter. NIH3T3 cells were transfected with pER1 Luc (0.1 Ag) and indicated concentrations of pRSV AP2 and pCMV GCF2. Activity is relative to the sample containing no GCF2 or AP2 plasmid. Varying amounts of control vector was used to keep the amount of DNA constant. promoter activation, indicating that GCF2 repressed AP2- mediated promoter activation (Fig. 8). Therefore, these results support a competition for binding model for GCF2 repression. References [1] J. Jankowski, D. Hopwood, K.G. Wormsley, Expression of epidermal growth factor, transforming growth factor alpha and their receptor in gastro-oesophageal diseases, Dig. Dis. 11 (1993) 1–11. [21] A.C. Johnson, Y. Jinno, G.T. Merlino, Modulation of epidermal growth factor receptor proto-oncogene transcription by a promoter site sensitive to S1 nuclease, Mol. Cell. Biol. 8 (1988) 4174–4184. [2] L. Jardines, M. Weiss, B. Fowble, M. Greene, neu(c-erbB-2/HER2) and the epidermal growth factor receptor (EGFR) in breast cancer, Pathobiology 61 (1993) 268–282. [22] X. Hou, A.C. Johnson, M.R. Rosner, Induction of epidermal growth factor receptor gene transcription by transforming growth factor beta 1: association with loss of protein binding to a negative regulatory element, Cell Growth Differ. 5 (1994) 801–809. [3] C.F. LeMaistre, C. Meneghetti, L. Howes, C.K. Osborne, Targeting the EGF receptor in breast cancer treatment, Breast Cancer Res. Treat. 32 (1994) 97–103. [23] R. Kageyama, I. Pastan, Molecular cloning and characterization of a human DNA binding factor that represses transcription, Cell 59 (1989) 815–825. [4] R.I. Nicholson, R.A. McClelland, J.M. Gee, D.L. Manning, P. Can- non, J.F. Robertson, I.O. Ellis, R.W. Blamey, Epidermal growth factor receptor expression in breast cancer: association with response to endocrine therapy, Breast Cancer Res. Treat. 29 (1994) 117–125. [24] X. Hou, A.C. Johnson, M.R. Rosner, Identification of an epidermal growth factor receptor transcriptional repressor, J. Biol. Chem. 269 (1994) 4307–4312. [5] A. Budillon, Molecular genetics of cancer. Oncogenes and tumor suppressor genes, Cancer 76 (1995) 1869–1873. [25] H. Nishi, K.H. Nishi, A.C. Johnson, Early growth response-1 gene mediates up-regulation of epidermal growth factor receptor expres- sion during hypoxia, Cancer Res. 62 (2002) 827–834. [6] F. Iacopino, G. Ferrandina, G. Scambia, P. Benedetti-Panici, S. Man- cuso, G. Sica, Interferons inhibit EGF-stimulated cell growth and reduce EGF binding in human breast cancer cells, Anticancer Res. 16 (1996) 1919–1924. [26] A.L. Reed, H. Yamazaki, J.D. Kaufman, Y. Rubinstein, B. Murphy, A.C. Johnson, Molecular cloning and characterization of a transcrip- tion regulator with homology to GC-binding factor, J. Biol. Chem. 273 (1998) 21594–21602. [7] C.R. King, M.H. Kraus, L.T. Williams, G.T. Merlino, I.H. Pastan, S.A. Aaronson, Human tumor cell lines with EGF receptor gene amplification in the absence of aberrant sized mRNAs, Nucleic Acids Res. 13 (1985) 8477–8486. [27] D.B. Zimonjic, L. Rezanka, J.A. DiPaolo, N.C. 4. Discussion Over the last several years, compelling evidences linking FS to the process of tumor development have accumulated. Chromosome translocations, amplification of proto-onco- genes, deletion of tumor suppressor genes, and integration of oncogenic viruses all result from the specific breakage of genomic DNA at FSs, suggesting a causative role for FSs in cancer [36]. Chromosome abnormalities involving region 2q36-37 have been described in Buschke–Ollendorff syn- T. Rikiyama et al. / Biochimica et Biophysica Acta 1629 (2003) 15–25 24 transcription by phorbol 12-myristate 13-acetate is mediated by acti- vator protein 2, J. Biol. Chem. 271 (1996) 3033–3038. transcription by phorbol 12-myristate 13-acetate is mediated by acti- vator protein 2, J. Biol. Chem. 271 (1996) 3033–3038. GCF2 contains 752 amino acids of which 100 appear involved in DNA binding and repression. The amino terminal domain contain coil–coil motif with homology to those found in flightless [41–43]. These domains allow for protein–protein interaction and thus may provide an addi- tional activity for GCF2. Further analysis is necessary to more thoroughly define the mechanism(s) by which GCF2 is able to effect changes in gene expression. [15] R. Kageyama, G.T. Merlino, I. Pastan, Epidermal growth factor (EGF) receptor gene transcription. Requirement for Sp1 and an EGF receptor-specific factor, J. Biol. Chem. 263 (1988) 6329–6336. [16] S.P. Deb, R.M. Munoz, D.R. Brown, M.A. Subler, S. Deb, Wild-type human p53 activates the human epidermal growth factor receptor promoter, Oncogene 9 (1994) 1341–1349. [17] M.S. Sheikh, F. Carrier, A.C. Johnson, S.E. Ogdon , A.J. Fornace Jr., Identification of an additional p53-responsive site in the human epi- dermal growth factor receptor gene promotor, Oncogene 15 (1997) 1095–1101. [18] C. Englert, X. Hou, S. Maheswaran, P. Bennett, C. Ngwu, G.G. Re, A.J. Garvin, M.R. Rosner, D.A. Haber, WT1 suppresses synthesis of the epidermal growth factor receptor and induces apoptosis, EMBO J. 14 (1995) 4662–4675. Acknowledgements The authors would like to thank Dr. Trevor Williams for the AP2 expression construct. They also would like to thank Dr. Hirotaka Nishi for critical suggestions and comments. [19] H. Nishi, M. Senoo, K.H. Nishi, B. Murphy, T. Rikiyama, Y. Matsu- mura, S. Habu, A.C. Johnson, p53 Homologue p63 represses epider- mal growth factor receptor expression, J. Biol. Chem. 276 (2001) 41717–41724. [20] L.L. Chen, M.L. Clawson, S. Bilgrami, G. Carmichael, A sequence- specific single-stranded DNA-binding protein that is responsive to epidermal growth factor recognizes an S1 nuclease-sensitive region in the epidermal growth factor receptor promoter, Cell Growth Differ. 4 (1993) 975–983. References Popescu, Refined localization of the erbB-3 proto-oncogene by direct visualization of FISH signals on LUT-inverted and contrast-enhanced digital images of DAPI-banded chromosomes, Cancer Genet. Cytogenet. 80 (1995) 100–102. [8] S. Ishii, Y.H. Xu, R.H. Stratton, B.A. Roe, G.T. Merlino, I. Pastan, Characterization and sequence of the promoter region of the human epidermal growth factor receptor gene, Proc. Natl. Acad. Sci. U. S. A. 82 (1985) 4920–4924. [28] P. Chomczynski, N. Sacchi, Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol–chloroform extraction, Anal. Biochem. 162 (1987) 156–159. [9] J. Haley, N. Whittle, P. Bennet, D. Kinchington, A. Ullrich, M. Water- field, The human EGF receptor gene: structure of the 110 kb locus and identification of sequences regulating its transcription, Oncog. Res. 1 (1987) 375–396. [29] H. Aviv, P. Leder, Purification of biologically active globin messenger RNA by chromatography on oligothymidylic acid-cellulose, Proc. Natl. Acad. Sci. U. S. A. 69 (1972) 1408–1412. [10] L.G. Hudson, J.B. Santon, G.N. Gill, Regulation of epidermal growth factor receptor gene expression, Mol. Endocrinol. 3 (1989) 400–408. [30] P.S. Thomas, Hybridization of denatured RNA and small DNA frag- ments transferred to nitrocellulose, Proc. Natl. Acad. Sci. U. S. A. 77 (1980) 5201–5205. [11] L.G. Hudson, J.B. Santon, C.K. Glass, G.N. Gill, Ligand-activated thyroid hormone and retinoic acid receptors inhibit growth factor receptor promoter expression, Cell 62 (1990) 1165–1175. [31] A.C. Johnson, R. Kageyama, N.C. Popescu, I. Pastan, Expression and chromosomal localization of the gene for the human transcriptional repressor GCF, J. Biol. Chem. 267 (1992) 1689–1694. [12] S.A. Chrysogelos, R.I. Yarden, A.H. Lauber, J.M. Murphy, Mecha- nisms of EGF receptor regulation in breast cancer cells, Breast Cancer Res. Treat. 31 (1994) 227–236. [13] A.C. Johnson, S. Ishii, Y. Jinno, I. Pastan, G.T. Merlino, Epidermal growth factor receptor gene promoter. Deletion analysis and identifi- cation of nuclear protein binding sites, J. Biol. Chem. 263 (1988) 5693–5699. [32] A.P. Feinberg, B. Vogelstein, A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity, Adden- dum, Anal. Biochem. 137 (1984) 266–267. [33] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680–685. [14] A.C. Johnson, Activation of epidermal growth factor receptor gene T. Rikiyama et al. / Biochimica et Biophysica Acta 1629 (2003) 15–25 25 [34] R.B. Dyer, N.K. Herzog, Isolation of intact nuclei for nuclear extract preparation from a fragile B-lymphocyte cell line, BioTechniques 19 (1995) 192–195. [40] L.M. References Khachigian, F.S. Santiago, L.A. Rafty, O.L. Chan, G.J. Del- bridge, A. Bobik, T. Collins, A.C. Johnson, GC factor 2 represses platelet-derived growth factor A-chain gene transcription and is itself induced by arterial injury, Circ. Res. 84 (1999) 1258–1267. [35] J.D. Dignam, R.M. Lebovitz, R.G. Roeder, Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei, Nucleic Acids Res. 11 (1983) 1475–1489. [41] H.D. Campbell, S. Fountain, I.G. Young, C. Claudianos, J.D. Hohei- sel, K.S. Chen, J.R. Lupski, Genomic structure, evolution, and ex- pression of human FLII, a gelsolin and leucine-rich-repeat family member: overlap with LLGL, Genomics 42 (1997) 46–54. [36] N.C. Popescu, Genetic alterations in cancer cells as a result of break- age at fragile sites, Cancer Lett. 192 (2003) 1–17. [37] V.A. McKusick, J.S. Amberger, The morbid anatomy of the human genome: chromosomal location of mutations causing disease, J. Med. Genet. 30 (1993) 1–26. [42] C. Claudianos, H.D. Campbell, The novel flightless-I gene brings together two gene families, actin-binding proteins related to gelsolin and leucine-rich-repeat proteins involved in Ras signal transduction, Mol. Biol. Evol. 12 (1995) 405–414. [38] F. Mitelman, Y. Kakeno, R. Berger, in: A.J. Cuticchia (Ed.), Human Gene Mapping, Johns Hopkins University, Baltimore, 1995, pp. 1332–1351. [43] H.D. Campbell, T. Schimansky, C. Claudianos, N. Ozsarac, A.B. Kasprzak, J.N. Cotsell, I.G. Young, H.G. de Couet, G.L. Miklos, The Drosophila melanogaster flightless-I gene involved in gastrula- tion and muscle degeneration encodes gelsolin-like and leucine-rich repeat domains and is conserved in Caenorhabditis elegans and hu- mans, Proc. Natl. Acad. Sci. U. S. A. 90 (1993) 11386–11390. [39] M. Shibutani, P. Lazarovici, A.C. Johnson, Y. Katagiri, G. Guroff, Transcriptional down-regulation of epidermal growth factor receptors by nerve growth factor treatment of PC12 cells, J. Biol. Chem. 273 (1998) 6878–6884.
https://openalex.org/W3049456510	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0238847&type=printable	English	null	Leave No Child Behind: Using Data from 1.7 Million Children from 67 Developing Countries to Measure Inequality Within and Between Groups of Births and to Identify Left Behind Populations	medRxiv (Cold Spring Harbor Laboratory)	2,020	cc-by	13,145	PLOS ONE PLOS ONE RESEARCH ARTICLE Received: September 12, 2019 Accepted: August 25, 2020 Published: October 14, 2020 Peer Review History: PLOS recognizes the benefits of transparency in the peer review process; therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. The editorial history of this article is available here: https://doi.org/10.1371/journal.pone.0238847 Copyright: © 2020 Ramos et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Leave no child behind: Using data from 1.7 million children from 67 developing countries to measure inequality within and between groups of births and to identify left behind populations Antonio P. RamosID1,2, Martin J. Flores3, Robert E. WeissID1,2 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 1 Department of Biostatistics, Fielding School of Public Health, UCLA, Los Angeles, CA, United States of America, 2 California Center for Population Research, UCLA, Los Angeles, CA, United States of America, 3 Department of General Internal Medicine, UCLA David Geffen School of Medicine, Los Angeles, CA, United States of America TOMRAMOS@UCLA.EDU * TOMRAMOS@UCLA.EDU * TOMRAMOS@UCLA.EDU Background Goal 3.2 from the Sustainable Development Goals (SDG) calls for reductions in national averages of Under-5 Mortality. However, it is well known that within countries these reductions can coexist with left behind populations that have mortality rates higher than national averages. To measure inequality in under-5 mortality and to identify left behind populations, mortality rates are often disaggregated by socioeconomic status within coun- tries. While socioeconomic disparities are important, this approach does not quantify within group variability since births from the same socioeconomic group may have differ- ent mortality risks. This is the case because mortality risk depends on several risk factors and their interactions and births from the same socioeconomic group may have different risk factor combinations. Therefore mortality risk can be highly variable within socioeco- nomic groups. We develop a comprehensive approach using information from multiple risk factors simultaneously to measure inequality in mortality and to identify left behind populations. Editor: Bruno Masquelier, University of Louvain, BELGIUM OPEN ACCESS Citation: Ramos AP, Flores MJ, Weiss RE (2020) Leave no child behind: Using data from 1.7 million children from 67 developing countries to measure inequality within and between groups of births and to identify left behind populations. PLoS ONE 15(10): e0238847. https://doi.org/10.1371/journal. pone.0238847 Editor: Bruno Masquelier, University of Louvain, BELGIUM Received: September 12, 2019 Accepted: August 25, 2020 Published: October 14, 2020 Interpretation Competing interests: The authors have declared that no competing interests exist. While important, differences in under-5 mortality across socioeconomic groups do not explain most of overall inequality in mortality risk because births from the same socioeco- nomic groups have different mortality risks. Similarly, policy makers can reach the highest risk children by targeting births based on several risk factors (socioeconomic status, residing in rural areas, having a previous death of a child and more) instead of using a single risk fac- tor such as socioeconomic status. We suggest that researchers and policy makers monitor inequality in under-5 mortality using multiple risk factors simultaneously, quantifying inequal- ity as a function of several risk factors to identify left behind populations in need of policy interventions and to help monitor progress toward the SDG. Findings via the following URL: https://www.dhsprogram. com/data/new-user-registration.cfm. For all countries there is more variability in mortality within socioeconomic groups than between them. Within countries, socioeconomic membership usually explains less than 20% of the total variation in mortality risk. In contrast, country of birth explains 19% of the total variance in mortality risk. Targeting the 20% highest risk children based on our model better identifies under-5 deaths than targeting the 20% poorest. For all surveys, we report efficiency gains from 26% in Mali to 578% in Guyana. High risk births tend to be births from mothers who are in the lowest socioeconomic group, live in rural areas and/or have already experienced a prior death of a child. Funding: The corresponding author, Dr. Antonio Pedro Ramos, has his worked funded by by the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the National Institutes of Health under Award Number K99HD088727. The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication. Methods We use Demographic and Health Surveys (DHS) data on 1,691,039 births from 182 different surveys from 67 low and middle income countries, 51 of which had at least two surveys. We estimate mortality risk for each child in the data using a Bayesian hierarchical logistic regres- sion model. We include commonly used risk factors for monitoring inequality in early life mortality for the SDG as well as their interactions. We quantify variability in mortality risk within and between socioeconomic groups and describe the highest risk sub-populations. Data Availability Statement: All relevant data are available on the Demographic and Health Surveys website and publicly accessible upon registering 1 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 Measure inequality within and between groups of births and to identify left behind populations in LMIC PLOS ONE via the following URL: https://www.dhsprogram. com/data/new-user-registration.cfm. PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 1 Introduction [21, 22] In Burkina Faso, families enrolled in conditional cash transfer schemes were required to obtain quarterly child growth monitoring at local health clinics for all children under 60 months of age. [23] In India, the randomized controlled trial (RCT) Lentils for Vaccines targeted the poor, as do most RCTs that aim to increase vaccine uptake, good nutrition, or child health more generally [24]. same country [6–10]. Studies have also documented significant under-5 mortality inequities across other demographic categories such as race, ethnicity, and geographic location [11–13]. Public health policies seeking to reduce inequality in early life mortality often target births Public health policies seeking to reduce inequality in early-life mortality often target births from an easily defined group with a high average mortality rates, usually the poorest. [9, 14– 19] A recent meta-analysis shows that most targeted interventions aiming to improve maternal and child health often address economic disparities through various incentive schemes like conditional cash transfers and voucher schemes. [20] For example, Cash Transfer Programs (CTP), currently implemented in many low and middle income countries (LMIC), often improve infant and child health. [21, 22] In Burkina Faso, families enrolled in conditional cash transfer schemes were required to obtain quarterly child growth monitoring at local health clinics for all children under 60 months of age. [23] In India, the randomized controlled trial (RCT) Lentils for Vaccines targeted the poor, as do most RCTs that aim to increase vaccine uptake, good nutrition, or child health more generally [24]. One important assumption underlying these approaches to measure inequality and target populations is that most of the variability in mortality risk exists between groups of births, not within them. If that is the case, (a) comparing average mortality rates between groups provides us with a complete picture of the inequality in mortality risk faced by children in the popula- tion and (b) targeting the group with the highest average mortality risk will reach most high risk births in the population and reduce overall inequalities. However, if the grouping factors used to monitor inequality have high levels of within-group variation in mortality risk, then monitoring inequality based solely on between group comparisons will miss most of the vari- ability in mortality risk and monitors will not be able to identify important left behind popula- tions that require intervention. 2 Methods Births are the units of our analysis. We first estimate mortality risk for each child in our data and then we use these estimates as inputs in our subsequent equity analysis. 1 Introduction [7] Using data from India a recent study shows that most of the variation in mortality risk exists within groups, not between groups, and that program target- ing based on poverty alone can be inefficient. [25] This makes sense as it is well known that multiple risk factors are associated with under-5 mortality risk. In this paper we develop a novel framework to monitor disparities in mortality risk and to identify high risk subpopulations that cannot be identified otherwise. Our novel approach uses data from several demographic variables and a Bayesian hierarchical model to estimate mortal- ity risk for each birth in our data set. We use these estimates to investigate within and between group variability across several commonly used demographic stratifiers that are used to moni- tor progress toward the SDG’s and make international comparisons in inequality in under-5 mortality. We identify children with the highest mortality risk in the population and show how to construct a targetable group that contains more deaths than other targetable groups of the same size that are based on only one risk factor, such as poverty. We identify the groups at highest risk in each country to gain insight on their needs. Our methodology supports UN rec- ommendations to disaggregate health indicators by demographic stratifiers to guide inequality monitoring so that countries can meet SDG targets with equity. We offer a more comprehen- sive approach that considers the effects of multiple risk factors and their interactions on mor- tality risk. 1 Introduction Goal 3.2 from the Sustainable Development Goals (SDG) requires reductions in under-5 mor- tality (http://www.un.org/sustainabledevelopment/health/). However, these reductions can co- exist with socioeconomic inequalities within countries where some groups have much higher mortality risk than others. [1] Studies have suggested that some of the Millennium Develop- ment Goals, which preceded the SDG, have not been achieved within many countries because of high levels of inequality. [2] Monitoring and reducing inequities in under-5 mortality requires the identification of births that are at highest risk of death such that policy interven- tions can target them. [3] The United Nations (UN) General Assembly Resolution 68/261, which highlights the Sustainable Development Indicators as a central framework for making progress on reducing early-life mortality, recommends that health indicators should be disag- gregated, where relevant, by income, sex, age, and other characteristics. [4, 5] Disaggregation of inequality by several demographic groups has a clear policy implication: leave no one behind. The literature that monitors progress towards SDG often quantifies gaps in either key health outcomes, such as neonatal or under-5 mortality, or in the coverage of health services, such as prenatal care or sanitation. Researchers and policy makers monitor progress toward SDG by evaluating mortality rates broken down by stratifiers, including wealth quintiles, rural/urban residence, maternal education, maternal age, gender of the child and geographic location (see https://www.equidade.org/indicators). [5] Even outside SDG monitoring, equity based strategies to reduce under-5 mortality usually measure gaps in average mortality rates between large groups of births, such as births from different socioeconomic groups within the PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 2 / 23 PLOS ONE Measure inequality within and between groups of births and to identify left behind populations in LMIC same country [6–10]. Studies have also documented significant under-5 mortality inequities across other demographic categories such as race, ethnicity, and geographic location [11–13]. Public health policies seeking to reduce inequality in early-life mortality often target births from an easily defined group with a high average mortality rates, usually the poorest. [9, 14– 19] A recent meta-analysis shows that most targeted interventions aiming to improve maternal and child health often address economic disparities through various incentive schemes like conditional cash transfers and voucher schemes. [20] For example, Cash Transfer Programs (CTP), currently implemented in many low and middle income countries (LMIC), often improve infant and child health. 2.2 Estimating mortality risk Mortality risk is a latent variable that must be estimated from data. Given our goal to improve inequality monitoring of the SDG, we base our estimation on predictors that are commonly used in studies that quantify progress toward SDG (https://www.equidade.org/indicators): maternal age, wealth, gender, year of birth, place of residence (urban/rural), maternal educa- tion in years. The probability density functions (pdf) of the the original wealth index scores do not have a common range across countries. To make them more comparable across surveys we transform these pdf’s into cumulative distribution functions (cdf). This approach gives wealth scores from different countries and surveys a common range, the unit interval (0,1) and makes the results interpretable in terms of relative wealth, a proxy for socioeconomic status within the countries. Details of the transformation are given in the appendix. We also include three other variables that are available in DHS surveys and could aid inequality monitoring and targeting. Geographical locations are well known risk factors for mortality, as mortality risk tends to be geographically clustered. Using sampling clusters from DHS in our model allows us to capture unmeasured variables at the local level that were not otherwise recorded in the data. Further, geographic locations can potentially be targeted by policy makers. Similarly, we also construct a 0 −1 indicator variable for whether a child was born to a mother that had already experienced a death of a previous child. Prior death summa- rizes a number of risk factors at the maternal level that are not measured by existing variables. It is a forward looking variable because it only uses information on prior births to inform risk for the current birth. In particular, information on future siblings deaths are not used to pre- dict past deaths and it is coded zero for a mother’s first birth. It is also an actionable risk factor because policy makers can potentially target births from those mothers, as they are identifiable. Finally, we include birth order, coded as a continuous variable. We estimate child mortality for each birth in our data as a function of these predictors and their interactions in a Bayesian hierarchical logistic regression model. We fit one model to the data from each survey. To avoid model misspecification and allow for all important interac- tions among the risk factors, we include all two-way, three-way, and four-way interaction terms for all covariates in the model. 2.1 Data sources The data used in this study comes from multiple Demographic and Health Surveys (DHS) (https://dhsprogram.com/). These are nationally representative surveys that have been 3 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 PLOS ONE Measure inequality within and between groups of births and to identify left behind populations in LMIC conducted in more than 100 low and middle income countries since 1984 [26, 27]. We analyze under-5 mortality and we exclude births that did not occur at least five years prior to the sur- vey. We exclude all births that happen 10 years or more before the date of the survey to mini- mize measurement error and censoring issues. The final data set includes information on 1,691,039 births from a total of 182 different surveys from 67 countries, 51 of which had at least two surveys. 2.2 Estimating mortality risk We include piecewise linear splines to capture non-linear trends in mortality as a function of the continuous variables. To aid in the estimation and avoid overfitting, we place increasingly restrictive priors on the variance parameters of the ran- dom effects for the higher order interaction terms, which shrink effects toward zero. We incor- porate a location random effect to model differences in risk between births from different locations. 2.3 Equity analysis We use estimates of the posterior distribution of mortality risk for each child in our data to feed our equity analysis. We use 1000 Markov Chain Monte Carlo (MCMC) samples from our model todo so. For the boxplots we use these samples to calculate the expected mortality risk for each child and then we plot these quantities. 4 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 PLOS ONE Measure inequality within and between groups of births and to identify left behind populations in LMIC We use box plots to display the within and between group variability in fitted mortality risk stratified by the DHS-assigned wealth quintile. We formally quantify how much of the variabil- ity in mortality risk is explained by the wealth quintiles using a Bayesian ANOVA, which allows us to get point and interval estimates of the R2. Details of the ANOVA methods are given in the appendix. Finally, we investigate whether using multiple risk factors simultaneously can help to iden- tify high risk births that should be targeted by policy interventions. Using the last survey from each country, we compare how many actual deaths occur among the 20% highest risk births from our model versus the 20% poorest births based on the wealth CDF variable. Under the assumption that intervention has the same cost for each birth, we calculate the efficiency gain in targeting the highest risk births versus the poorest births by dividing the difference in mor- tality rates between highest risk births and poorest births by mortality rates among the poorest times 100. We thus define the efficiency gain as ðHRDeathsPoorDeathsÞ PoorDeaths 100, where “HRDeaths” is mortality among the 20% highest risk births and “PoorDeaths” is defined as mortality among the poorest 20% of births. For each survey, we compare births in the high risk group to births not in the high risk group based on the following covariates: wealth, maternal education, maternal age, place of residency (urban/rural), whether the birth was born to a mother who has experienced a prior death of another child. We compare lower and higher mortality risk groups by using either risk ratios for categorical risk factors or mean risk difference for contin- uous risk factors. 2.3.1 Incorporating uncertainty in the equity analysis. We use estimates of the posterior distribution of mortality risk for each child in our data to feed our equity analysis. 2.3 Equity analysis We use 1000 Markov Chain Monte Carlo (MCMC) samples from our model to do so. For the boxplots we use these samples to calculate the expected mortality risk for each child and then we plot these quantities. For ANOVA and other tabulations, we calculate a quantity for each MCMC sample so that we have a distribution of these quantities that can be used to calculate posterior means and intervals. These also allow us to implement significant tests. 2.4 Role of the funding source We acknowledge financial support from the Eunice Kennedy Shriver National Institute Of Child Health & Human Development of the National Institutes of Health under Award Num- ber K99HD088727 and CCPR’s Population Research Infrastructure Grant P2C from NICHD: P2C-HD041022. The sponsor of the study had no role in study design, data analysis, data col- lection, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study; all authors had final responsibility for the decision to submit for publication. PLOS ONE PLOS ONE Table 1. N is the survey sample size used in our analysis. U5MR is the under-5 mortality rates by age five for each survey. Non-poor deaths (NPD) is the fraction of deaths from the top 80% wealth quintile. DRC is Democratic Republic of Congo, DR is Dominican Republic, and CAR is Central African Republic. The first quintile is the poorest births and the fifth quantile is the richest births. Table 1. N is the survey sample size used in our analysis. U5MR is the under-5 mortality rates by age five for each survey. Non-poor deaths (NPD) is the fraction of deaths from the top 80% wealth quintile. DRC is Democratic Republic of Congo, DR is Dominican Republic, and CAR is Central African Republic. The first quintile is the poorest births and the fifth quantile is the richest births p q p g p p q p births and the fifth quantile is the richest births. 3.1 Mortality by wealth quintile in the raw data All results use individual births as the unit of the analysis. Summaries of the Demographic and Health Surveys (DHS) are presented in Table 1. Each row presents data for one survey. From left to right, the columns in Table 1 are the number of births in each survey (N); the under-5 mortality rate (U5MR), defined as the fraction of births who die before age five, both overall and for each wealth quintile; and the proportion of deaths that occurred to the top 80% in wealth, which we call the non-poor deaths (NPD) fraction. If there is perfect equity in mortal- ity across socioeconomic groups, then the NPD would be exactly 80%. If the poorest 20% con- tain more than their share of deaths, then the NPD would be lower than 80%. Under-5 PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 5 / 23 Measure inequality within and between groups of births and to identify left behind populations in LMIC PLOS ONE PLOS ONE Table 1. PLOS ONE Country (Survey Year) N U5MR U5MR by Wealth Quintile NPD First Second Third Fourth Fifth Albania (2009) 2,481 0.028 0.049 0.022 0.030 0.009 0.023 0.551 Angola (2011) 5,812 0.109 0.149 0.105 0.127 0.103 0.088 0.819 Armenia (2000) 2,602 0.057 0.058 0.059 0.056 0.066 0.043 0.736 Armenia (2010) 1,545 0.028 0.028 0.052 0.023 0.013 0.022 0.814 Azerbaijan (2006) 2,739 0.063 0.071 0.072 0.056 0.052 0.055 0.680 Bangladesh (2000) 9,061 0.127 0.164 0.158 0.110 0.109 0.082 0.717 Bangladesh (2004) 7,261 0.101 0.120 0.112 0.094 0.091 0.080 0.725 Bangladesh (2007) 6,929 0.083 0.100 0.105 0.091 0.073 0.046 0.733 Bangladesh (2014) 14,512 0.061 0.080 0.067 0.058 0.057 0.036 0.687 Benin (1996) 5,386 0.200 0.224 0.213 0.212 0.196 0.116 0.722 Benin (2001) 5,691 0.170 0.211 0.183 0.168 0.141 0.109 0.694 Benin (2006) 16,984 0.152 0.169 0.165 0.161 0.143 0.091 0.728 Benin (2012) 12,904 0.084 0.093 0.100 0.087 0.074 0.043 0.723 Bolivia (1998) 9,334 0.117 0.161 0.125 0.119 0.062 0.044 0.574 Bolivia (2004) 10,546 0.103 0.128 0.126 0.103 0.068 0.046 0.688 Bolivia (2008) 10,048 0.080 0.112 0.087 0.074 0.060 0.029 0.597 Brazil (1996) 6,023 0.071 0.113 0.067 0.045 0.037 0.036 0.477 Burkina Faso (1993) 5,514 0.206 0.206 0.253 0.236 0.221 0.157 0.850 Burkina Faso (1999) 5,702 0.230 0.250 0.249 0.251 0.249 0.152 0.751 Burkina Faso (2003) 12,060 0.200 0.201 0.227 0.204 0.208 0.144 0.804 Burkina Faso (2010) 16,759 0.164 0.186 0.186 0.162 0.157 0.110 0.756 Burundi (2011) 6,016 0.137 0.170 0.163 0.152 0.136 0.074 0.761 Cambodia (2000) 12,071 0.131 0.171 0.144 0.120 0.116 0.072 0.646 Cambodia (2011) 7,258 0.081 0.113 0.104 0.084 0.050 0.038 0.633 Cambodia (2014) 8,272 0.060 0.093 0.073 0.051 0.041 0.029 0.611 Cameroon (1991) 3,140 0.149 0.210 0.204 0.146 0.131 0.088 0.771 Cameroon (1998) 4,080 0.145 0.212 0.176 0.145 0.101 0.096 0.662 Cameroon (2004) 7,535 0.157 0.207 0.181 0.155 0.102 0.090 0.645 Cameroon (2011) 10,812 0.133 0.188 0.148 0.126 0.095 0.076 0.676 CAR (1995) 4,429 0.166 0.204 0.181 0.167 0.166 0.093 0.692 Chad (1997) 6,941 0.201 0.173 0.230 0.227 0.223 0.167 0.854 Chad (2004) 6,260 0.201 0.191 0.215 0.231 0.217 0.178 0.822 Chad (2015) 18,985 0.144 0.160 0.160 0.136 0.132 0.135 0.798 Colombia (1990) 4,087 0.041 0.069 0.055 0.034 0.032 0.025 0.754 Colombia (1995) 5,041 0.040 0.053 0.041 0.029 0.042 0.026 0.655 Colombia (2005) 15,630 0.032 0.047 0.032 0.026 0.020 0.021 0.598 Comoros (1996) 2,208 0.116 0.132 0.139 0.108 0.094 0.091 0.715 Comoros (2012) 3,390 0.050 0.051 0.054 0.052 0.055 0.035 0.725 DRC (2005) 4,419 0.134 0.157 0.141 0.137 0.143 0.081 0.745 DRC (2007) 7,971 0.172 0.207 0.195 0.180 0.155 0.107 0.734 DRC (2012) 7,597 0.097 0.105 0.106 0.082 0.066 0.071 0.501 DRC (2014) 15,132 0.125 0.137 0.137 0.124 0.128 0.077 0.717 Coˆte d’Ivoire (1999) 2,757 0.158 0.195 0.172 0.189 0.136 0.110 0.789 Coˆte d’Ivoire (2005) 3,812 0.127 0.149 0.127 0.125 0.115 0.097 0.673 Coˆte d’Ivoire (2012) 7,224 0.140 0.145 0.145 0.170 0.124 0.087 0.762 (Continued) PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 6 / 23 Measure inequality within and between groups of births and to identify left behind populations in LMIC PLOS ONE PLOS ONE Table 1. PLOS ONE (Continued) Country (Survey Year) N U5MR U5MR by Wealth Quintile NPD First Second Third Fourth Fifth Dominican Republic (1999) 3,250 0.070 0.093 0.074 0.071 0.049 0.019 0.575 Dominican Republic (2002) 12,941 0.049 0.071 0.045 0.039 0.039 0.019 0.541 Dominican Republic (2007) 13,945 0.037 0.047 0.037 0.032 0.025 0.028 0.558 Dominican Republic (2013) 4,782 0.042 0.057 0.042 0.032 0.024 0.023 0.505 Egypt (1996) 12,791 0.110 0.158 0.133 0.107 0.070 0.038 0.605 Egypt (2003) 11,850 0.070 0.099 0.079 0.067 0.047 0.036 0.611 Egypt (2008) 11,394 0.039 0.061 0.035 0.035 0.026 0.025 0.592 Egypt (2014) 14,486 0.035 0.051 0.043 0.033 0.028 0.021 0.700 Eswatini (2007) 2,421 0.102 0.118 0.108 0.097 0.102 0.091 0.782 Ethiopia (1997) 12,984 0.141 0.134 0.168 0.153 0.158 0.104 0.743 Ethiopia (2003) 13,218 0.129 0.149 0.132 0.132 0.135 0.084 0.636 Gabon (2001) 3,783 0.093 0.095 0.117 0.099 0.083 0.040 0.685 Gabon (2012) 5,149 0.070 0.082 0.073 0.061 0.047 0.035 0.453 Ghana (1994) 3,281 0.147 0.181 0.188 0.155 0.114 0.078 0.751 Ghana (1999) 3,226 0.126 0.156 0.142 0.126 0.103 0.048 0.565 Ghana (2003) 4,134 0.127 0.155 0.120 0.125 0.112 0.088 0.603 Ghana (2008) 3,258 0.096 0.114 0.093 0.105 0.076 0.068 0.620 Ghana (2014) 6,370 0.084 0.107 0.077 0.067 0.062 0.080 0.570 Guatemala (1999) 7,083 0.078 0.085 0.087 0.081 0.065 0.034 0.637 Guatemala (2015) 11,719 0.041 0.057 0.043 0.038 0.030 0.021 0.623 Guinea (1999) 6,867 0.195 0.235 0.218 0.196 0.182 0.128 0.721 Guinea (2005) 7,807 0.201 0.219 0.230 0.220 0.172 0.125 0.741 Guinea (2012) 8,010 0.143 0.180 0.151 0.152 0.122 0.073 0.684 Guyana (2005) 1,268 0.046 0.030 0.052 0.029 0.038 0.088 0.828 Guyana (2009) 2,464 0.037 0.027 0.042 0.030 0.065 0.036 0.700 Haiti (1995) 3,020 0.157 0.194 0.182 0.147 0.139 0.100 0.722 Haiti (2000) 7,063 0.152 0.185 0.140 0.151 0.137 0.114 0.655 Haiti (2006) 5,907 0.107 0.130 0.115 0.098 0.098 0.067 0.655 Haiti (2012) 6,944 0.096 0.101 0.102 0.102 0.091 0.065 0.691 Honduras (2006) 12,380 0.045 0.055 0.052 0.041 0.029 0.022 0.606 Honduras (2012) 10,065 0.031 0.041 0.027 0.028 0.025 0.020 0.553 India (1993) 65,681 0.113 0.162 0.157 0.115 0.085 0.055 0.723 India (2000) 53,079 0.099 0.146 0.122 0.104 0.070 0.045 0.686 India (2006) 59,240 0.080 0.128 0.099 0.080 0.061 0.037 0.699 Indonesia (1997) 23,155 0.085 0.111 0.101 0.084 0.060 0.031 0.565 Indonesia (2003) 16,049 0.064 0.091 0.068 0.056 0.043 0.027 0.515 Indonesia (2007) 20,592 0.067 0.100 0.072 0.054 0.042 0.034 0.529 Indonesia (2012) 19,788 0.054 0.087 0.057 0.038 0.037 0.019 0.490 Jordan (1990) 9,308 0.046 0.061 0.056 0.043 0.038 0.034 0.796 Jordan (1997) 6,408 0.036 0.046 0.040 0.036 0.026 0.029 0.707 Jordan (2002) 7,098 0.037 0.040 0.041 0.037 0.031 0.029 0.708 Jordan (2009) 13,691 0.029 0.035 0.022 0.028 0.029 0.026 0.611 Jordan (2012) 11,205 0.024 0.029 0.023 0.021 0.024 0.015 0.670 Kazakhstan (1999) 2,651 0.057 0.069 0.062 0.067 0.052 0.038 0.762 Kenya (1993) 6,514 0.097 0.138 0.129 0.078 0.067 0.060 0.681 Kenya (1998) 5,789 0.104 0.140 0.119 0.104 0.076 0.058 0.668 (Continued) PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 7 / 23 Measure inequality within and between groups of births and to identify left behind populations in LMIC PLOS ONE PLOS ONE Table 1. PLOS ONE (Continued) Country (Survey Year) N U5MR U5MR by Wealth Quintile NPD First Second Third Fourth Fifth Kenya (2009) 5,412 0.095 0.103 0.106 0.098 0.074 0.084 0.686 Kenya (2014) 23,924 0.055 0.053 0.066 0.054 0.053 0.044 0.674 Kyrgyzstan (1997) 2,400 0.074 0.094 0.092 0.079 0.051 0.043 0.669 Kyrgyzstan (2012) 3,705 0.036 0.031 0.037 0.048 0.032 0.031 0.799 Lesotho (2005) 3,115 0.093 0.113 0.107 0.090 0.075 0.077 0.746 Lesotho (2010) 3,107 0.087 0.077 0.095 0.098 0.090 0.079 0.737 Lesotho (2014) 3,250 0.100 0.080 0.103 0.117 0.120 0.087 0.791 Liberia (2009) 6,871 0.173 0.195 0.176 0.158 0.173 0.149 0.713 Liberia (2013) 8,220 0.132 0.147 0.131 0.123 0.108 0.126 0.618 Madagascar (1997) 5,960 0.165 0.208 0.186 0.178 0.137 0.098 0.675 Madagascar (2004) 5,268 0.106 0.163 0.142 0.114 0.095 0.058 0.699 Madagascar (2009) 12,686 0.087 0.111 0.098 0.093 0.070 0.045 0.651 Malawi (1992) 4,746 0.231 0.273 0.242 0.259 0.256 0.154 0.799 Malawi (2005) 9,663 0.180 0.216 0.192 0.193 0.167 0.124 0.777 Malawi (2010) 20,677 0.129 0.145 0.136 0.133 0.115 0.110 0.748 Malawi (2016) 16,793 0.079 0.094 0.082 0.088 0.076 0.053 0.756 Mali (1996) 9,960 0.259 0.310 0.292 0.262 0.238 0.175 0.757 Mali (2001) 13,031 0.257 0.264 0.271 0.287 0.271 0.148 0.776 Mali (2006) 15,201 0.222 0.248 0.261 0.229 0.210 0.134 0.773 Mali (2013) 9,249 0.113 0.120 0.140 0.130 0.108 0.063 0.779 Moldova (2005) 1,744 0.033 0.036 0.031 0.044 0.036 0.018 0.789 Morocco (1992) 5,422 0.088 0.110 0.094 0.092 0.074 0.050 0.695 Morocco (2004) 6,493 0.061 0.085 0.069 0.048 0.046 0.027 0.602 Mozambique (1997) 6,834 0.200 0.262 0.213 0.210 0.183 0.120 0.674 Mozambique (2004) 8,942 0.195 0.229 0.222 0.227 0.168 0.115 0.716 Mozambique (2011) 10,379 0.112 0.137 0.112 0.126 0.100 0.093 0.783 Namibia (1992) 3,692 0.109 0.137 0.100 0.103 0.120 0.079 0.718 Namibia (2000) 4,354 0.063 0.073 0.090 0.072 0.058 0.033 0.778 Namibia (2007) 4,668 0.069 0.097 0.078 0.064 0.062 0.032 0.703 Namibia (2013) 4,691 0.058 0.065 0.074 0.060 0.056 0.023 0.745 Nicaragua (1998) 8,665 0.062 0.067 0.070 0.060 0.054 0.041 0.661 Nicaragua (2001) 9,008 0.049 0.063 0.053 0.048 0.036 0.018 0.600 Niger (1998) 7,644 0.306 0.294 0.376 0.356 0.329 0.194 0.823 Niger (2006) 9,820 0.206 0.189 0.237 0.248 0.227 0.151 0.812 Niger (2012) 13,573 0.151 0.153 0.175 0.175 0.162 0.099 0.805 Nigeria (1990) 8,696 0.190 0.247 0.243 0.213 0.165 0.105 0.729 Nigeria (2003) 5,848 0.221 0.246 0.291 0.213 0.201 0.092 0.721 Nigeria (2008) 30,182 0.185 0.224 0.226 0.169 0.137 0.091 0.657 Nigeria (2013) 34,186 0.158 0.204 0.202 0.146 0.109 0.085 0.685 Pakistan (1991) 8,356 0.110 0.109 0.140 0.128 0.110 0.074 0.864 Pakistan (2007) 9,531 0.089 0.112 0.097 0.076 0.085 0.060 0.698 Pakistan (2013) 11,854 0.093 0.122 0.099 0.091 0.082 0.057 0.673 Paraguay (1990) 4,375 0.053 0.069 0.055 0.054 0.045 0.018 0.597 Peru (1992) 9,085 0.112 0.155 0.133 0.083 0.055 0.035 0.553 Peru (1996) 19,554 0.088 0.121 0.097 0.067 0.058 0.026 0.527 Peru (2000) 17,334 0.081 0.112 0.094 0.060 0.037 0.016 0.536 (Continued) PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 8 / 23 Measure inequality within and between groups of births and to identify left behind populations in LMIC https://doi.org/10.1371/journal.pone.0238847.t001 PLOS ONE (Continued) Country (Survey Year) N U5MR U5MR by Wealth Quintile NPD First Second Third Fourth Fifth Peru (2008) 13,739 0.040 0.063 0.047 0.037 0.025 0.019 0.720 Peru (2012) 31,443 0.033 0.046 0.035 0.026 0.020 0.013 0.544 Philippines (1993) 9,340 0.075 0.101 0.088 0.068 0.038 0.052 0.625 Philippines (1998) 8,361 0.065 0.091 0.070 0.052 0.039 0.031 0.530 Philippines (2003) 7,863 0.045 0.073 0.048 0.033 0.020 0.023 0.526 Philippines (2008) 7,480 0.044 0.066 0.043 0.030 0.032 0.024 0.535 Philippines (2013) 8,159 0.033 0.051 0.032 0.025 0.018 0.017 0.485 Rwanda (1992) 6,071 0.174 0.165 0.218 0.155 0.211 0.134 0.795 Rwanda (2005) 9,139 0.202 0.223 0.224 0.200 0.224 0.132 0.744 Rwanda (2008) 4,865 0.149 0.176 0.166 0.159 0.159 0.087 0.824 Rwanda (2015) 8,096 0.071 0.082 0.082 0.077 0.068 0.040 0.731 Sao Tome and Principe (2009) 1,685 0.081 0.087 0.076 0.082 0.106 0.034 0.728 Senegal (1997) 7,311 0.157 0.189 0.192 0.165 0.109 0.076 0.706 Senegal (2005) 10,284 0.162 0.210 0.186 0.158 0.100 0.079 0.677 Senegal (2009) 13,229 0.124 0.154 0.135 0.107 0.063 0.067 0.575 Senegal (2015) 12,606 0.084 0.110 0.089 0.075 0.054 0.046 0.596 Sierra Leone (2008) 6,413 0.179 0.214 0.184 0.163 0.173 0.155 0.739 Sierra Leone (2013) 13,981 0.187 0.206 0.197 0.192 0.179 0.142 0.746 South Africa (1998) 5,564 0.057 0.085 0.073 0.048 0.031 0.022 0.610 Tanzania (1999) 6,715 0.150 0.159 0.167 0.167 0.169 0.098 0.764 Tanzania (2005) 7,200 0.143 0.166 0.158 0.160 0.124 0.101 0.755 Tanzania (2010) 11,262 0.101 0.126 0.110 0.098 0.092 0.071 0.737 Tanzania (2016) 8,745 0.079 0.085 0.081 0.076 0.084 0.062 0.755 Timor-Leste (2010) 9,499 0.089 0.096 0.102 0.095 0.091 0.059 0.758 Togo (1998) 7,211 0.155 0.174 0.181 0.159 0.119 0.102 0.720 Togo (2014) 6,901 0.109 0.131 0.122 0.112 0.084 0.045 0.588 Turkey (1993) 4,998 0.090 0.144 0.095 0.087 0.073 0.030 0.639 Turkey (1998) 4,162 0.064 0.096 0.065 0.058 0.045 0.033 0.615 Turkey (2004) 4,765 0.058 0.087 0.065 0.051 0.034 0.031 0.587 Uganda (1995) 6,244 0.159 0.199 0.183 0.158 0.163 0.114 0.778 Uganda (2001) 5,933 0.154 0.192 0.194 0.170 0.136 0.102 0.784 Uganda (2010) 5,912 0.142 0.168 0.149 0.138 0.134 0.104 0.690 Uganda (2011) 7,852 0.117 0.137 0.137 0.110 0.112 0.080 0.684 Ukraine (2007) 1,494 0.021 0.021 0.015 0.021 0.041 0.011 0.806 Uzbekistan (1996) 2,656 0.054 0.064 0.039 0.054 0.065 0.049 0.776 Vietnam (2002) 4,060 0.039 0.055 0.045 0.031 0.030 0.023 0.643 Zambia (1997) 5,614 0.192 0.214 0.226 0.192 0.169 0.126 0.660 Zambia (2002) 6,027 0.171 0.204 0.188 0.196 0.142 0.084 0.722 Zambia (2007) 5,808 0.147 0.125 0.171 0.172 0.142 0.102 0.821 Zambia (2014) 12,324 0.088 0.109 0.091 0.087 0.069 0.072 0.728 Zimbabwe (1994) 4,622 0.066 0.073 0.084 0.050 0.073 0.045 0.702 Zimbabwe (1999) 3,713 0.078 0.085 0.087 0.081 0.081 0.043 0.697 Zimbabwe (2006) 4,357 0.062 0.064 0.071 0.069 0.055 0.047 0.748 Zimbabwe (2011) 4,374 0.067 0.075 0.075 0.074 0.052 0.057 0.718 Zimbabwe (2015) 5,726 0.093 0.118 0.102 0.103 0.084 0.062 0.726 https://doi.org/10.1371/journal.pone.0238847.t001 PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 9 / 23 PLOS ONE Measure inequality within and between groups of births and to identify left behind populations in LMIC mortality rates are generally higher for the poorest wealth quintiles, reflecting a socioeconomic gradient in mortality. 3.2 Quantifying within and between group variability Fig 1 presents box plots showing the distribution of mortality risk for the last survey of each country. Countries are ordered from the highest median mortality risk (Sierra Leone) to the lowest median mortality risk (Ukraine). As the median mortality risk gets smaller, variance decreases as well. There is considerable overlap in mortality risk across countries. This suggests that country of birth explains only a small fraction of mortality risk and that all countries have some children with very high mortality risk. Fig 2 presents the distribution of mortality risk across countries stratified by wealth quintile. Only the most recent survey is shown, and countries are ranked from highest to lowest median mortality risk, from top left to bottom right. Outliers are not shown and all graphs are pre- sented on the same scale. For all countries and surveys in our sample, there is considerable overlap in mortality risk across socioeconomic groups within countries and this is true irre- spective of a country’s average mortality level. Among higher mortality countries, Sierra Leone and the Central African Republic have clear socioeconomic gradients in mortality risk. Among lower mortality countries, Bolivia, Brazil, Nigeria, and Cameroon have the largest socioeconomic gradients in mortality risk. High mortality countries like Niger and Lesotho exhibit no socioeconomic gradients in mortality, and this is also true for some lower mortality countries, such as Ukraine, Armenia and Jordan. Conclusions from Fig 2 are thus consistent with those from Table 1. Table 2 presents results from our analysis. The first column gives the country and year in which the survey was taken, and first row presenting the results across all surveys combined. Columns two through five show the mean, median, and standard deviation of the mortality risk distribution from our analysis, and the R2 of our ANOVA, which quantifies how much of the variance in mortality risk is explained by wealth quintile. Globally, wealth quintile only explains about 3% of the variability in mortality risk. How- ever, there is substantial country to country heterogeneity. The countries with the highest R2 values are India (23%), Nigeria (17%), Indonesia (14%), and Cameroon (14%). In contrast, Eswatini, Lesotho, Tanzania, Moldova, Sao Tome and Principe, Kyrgyzstan, Uzbekistan, Kenya, Ukraine, and Comoros all have R2 point estimates that are less than 1%. Further there is not a clear relationship between R2 and mean/median mortality risk. PLOS ONE Some countries, such as Egypt, exhibit a consistent decrease in mortality with increasing wealth quintile. In a few countries, mortality increases from the poorest to the second poorest quintile, such as in Burkina Faso (2003). In general, the NPD are typically between 50% and 75%. These results show that there are high risk children in all socioeco- nomic groups. 3.2 Quantifying within and between group variability Using country of birth in the ANOVA gives a posterior mean R2 of 19%. Thus the ANOVA results confirm the find- ings from the boxplots of mortality rates in Figs 1 and 2 which show that while there is sub- stantial country to country heterogeneity, within a given country wealth does not explain much of the variability in mortality risk. Mortality risk distributions have a long right tail and in Table 2 the mean mortality risk is always higher than the median. In every country, there are individuals that face much higher mortality risk than the national average. https://doi.org/10.1371/journal.pone.0238847.g002 our approach is much more efficient in identifying high risk births than targeting the poor. Efficiency gains range from 26% in Mali (1996), to more than 550% in Guyana (2009). Effi- ciency gains are not strongly related to a country’s average mortality rates. PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 3.3 Comparing mortality among highest risk and poorest children Poverty status alone is often used to decide which families will be targeted by health interven- tions. However, high within group variability for socioeconomic groups suggests that targeting PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 10 / 23 PLOS ONE Measure inequality within and between groups of births and to identify left behind populations in LMIC based on a single demographic variable is inefficient because there are high risk births in all Fig 1. Box plots for mortality risk by country and survey. Lines are ±1.5 times interquartile range, boxes are lower to upper quartile, and dark line is the median mortality risk. Outliers are not shown. https://doi.org/10.1371/journal.pone.0238847.g001 Fig 1. Box plots for mortality risk by country and survey. Lines are ±1.5 times interquartile range, boxes are lower to upper quartile, and dark line is the median mortality risk. Outliers are not shown. https://doi.org/10.1371/journal.pone.0238847.g001 based on a single demographic variable is inefficient because there are high risk births in all socioeconomic groups. We formally demonstrate the validity of this hypothesis for the last sur- vey of each country, comparing efficiency gains of targeting the 20% poorest compared to tar- geting the 20% highest risk. Results are presented in Table 3. For all surveys and all countries, PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 11 / 23 PLOS ONE Measure inequality within and between groups of births and to identify left behind populations in LMIC Fig 2. Box plots for mortality risk by wealth quintile, country, and survey. Countries are ranked from lowest to highest mortality. Lines are ±1.5 times interquartile range, boxes are lower to upper quartile, and dark line is the median mortality risk. Outliers are not shown. Fig 2. Box plots for mortality risk by wealth quintile, country, and survey. Countries are ranked from lowest to highest mortality. Lines are ±1.5 times interquartile range, boxes are lower to upper quartile, and dark line is the median mortality risk. Outliers are not shown. https //doi org/10 1371/jo rnal pone 0238847 g002 Fig 2. Box plots for mortality risk by wealth quintile, country, and survey. Countries are ranked from lowest to highest mortality. Lines are ±1.5 times interquartile range, boxes are lower to upper quartile, and dark line is the median mortality risk. Outliers are not shown https://doi.org/10.1371/journal.pone.0238847.g002 PLOS ONE PLOS ONE Table 2. Results from ANOVA of posterior mean of mortality risk on wealth quintiles. Countries are ordered by median mortality risk. Mean, median, variance, and R2 are presented as posterior means and 95% intervals. 3.4 Who are the highest risk children? We define the high risk (low risk) births for a particular country and survey as those in the top 20% (bottom 80%) of all births in terms of mortality risk as estimated by our model. For each of the continuous (categorical) variables, we calculate means of the variable for high and low risk births and the difference (odds ratio). Results are presented in Tables 1-7 in S1 Appendix for the last survey in each country. Higher risk births have younger mothers on average com- pared to lower risk births, but the differences are not substantively important: mothers from low risk group are usually less than a year older than mothers from the high risk group. High risk and low risk groups are also comparable for birth gender. For maternal education, there is often a significant difference between high risk and low risk births, but the difference is not substantively important. There is on average less than a year of additional education for PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 12 / 23 Measure inequality within and between groups of births and to identify left behind populations in LMIC (Continued) PLOS ONE Country Mean Median Variance R2 (Wealth) Overall 6.9% (6.8%, 6.9%) 4.0% (4.0%, 4.1%) 0.7% (0.7%, 0.7%) 3.3% (3.1%, 3.5%) Sierra Leone 2013 15.7% (15.1%, 16.3%) 12.7% (12.0%, 13.3%) 1.3% (1.1%, 1.5%) 2.8% (1.3%, 4.7%) Central African Republic 1995 14.0% (13.0%, 15.1%) 10.8% (9.8%, 12.0%) 1.5% (1.2%, 1.8%) 5.9% (2.7%, 10.3%) Burkina Faso 2010 13.3% (12.8%, 13.8%) 11.4% (10.8%, 11.9%) 0.8% (0.7%, 0.9%) 5.5% (3.2%, 8.3%) Niger 2012 12.3% (11.7%, 12.8%) 9.8% (9.2%, 10.4%) 0.9% (0.8%, 1.1%) 2.9% (1.3%, 4.9%) Nigeria 2013 12.9% (12.6%, 13.3%) 9.9% (9.5%, 10.3%) 1.0% (0.9%, 1.1%) 17.4% (14.9%, 20.0%) Guinea 2012 11.6% (11.0%, 12.4%) 9.1% (8.4%, 9.8%) 1.0% (0.8%, 1.2%) 7.7% (4.5%, 11.6%) Burundi 2011 11.2% (10.4%, 12.0%) 8.5% (7.7%, 9.3%) 1.0% (0.8%, 1.3%) 8.7% (4.9%, 13.1%) Chad 2015 11.6% (11.2%, 12.1%) 8.8% (8.4%, 9.3%) 0.9% (0.8%, 1.0%) 1.2% (0.3%, 2.3%) Coˆte d’Ivoire 2012 11.5% (10.8%, 12.2%) 8.2% (7.5%, 9.0%) 1.1% (0.9%, 1.4%) 1.4% (0.2%, 3.2%) Liberia 2013 10.6% (9.9%, 11.3%) 8.0% (7.4%, 8.6%) 0.9% (0.7%, 1.0%) 1.5% (0.2%, 3.6%) Ethiopia 2003 10.2% (9.8%, 10.7%) 8.0% (7.5%, 8.5%) 0.7% (0.6%, 0.9%) 3.6% (1.7%, 6.0%) Cameroon 2011 10.7% (10.1%, 11.3%) 7.8% (7.2%, 8.4%) 0.9% (0.7%, 1.1%) 13.6% (9.7%, 18.1%) Uganda 2011 9.3% (8.7%, 9.9%) 7.0% (6.4%, 7.7%) 0.7% (0.6%, 0.9%) 3.8% (1.5%, 6.9%) Congo Democratic Republic 2014 9.9% (9.4%, 10.4%) 7.5% (7.0%, 7.9%) 0.7% (0.6%, 0.8%) 2.5% (1.1%, 4.4%) Togo 2014 8.7% (8.1%, 9.3%) 6.1% (5.5%, 6.7%) 0.8% (0.6%, 1.0%) 5.5% (2.7%, 8.8%) Mozambique 2011 8.9% (8.4%, 9.5%) 6.1% (5.6%, 6.6%) 0.8% (0.7%, 1.0%) 2.2% (0.7%, 4.2%) Angola 2011 8.8% (8.2%, 9.5%) 5.4% (4.8%, 6.0%) 1.0% (0.8%, 1.3%) 2.4% (0.6%, 4.8%) Lesotho 2014 8.1% (7.3%, 9.0%) 4.6% (3.8%, 5.4%) 1.2% (0.9%, 1.5%) 0.4% (0.0%, 1.6%) Haiti 2012 7.5% (6.9%, 8.0%) 5.2% (4.7%, 5.8%) 0.6% (0.5%, 0.8%) 0.6% (0.0%, 1.9%) Congo 2012 7.5% (7.0%, 8.1%) 5.3% (4.8%, 5.8%) 0.6% (0.5%, 0.7%) 1.3% (0.1%, 3.1%) Mali 2013 9.1% (8.6%, 9.7%) 5.4% (4.9%, 6.0%) 1.0% (0.9%, 1.2%) 2.6% (1.2%, 4.3%) Eswatini 2007 8.6% (7.6%, 9.6%) 4.0% (3.3%, 4.9%) 1.7% (1.3%, 2.1%) 0.4% (0.0%, 1.5%) Zimbabwe 2015 7.4% (6.8%, 8.0%) 4.5% (4.0%, 5.1%) 0.9% (0.7%, 1.1%) 3.0% (1.2%, 5.4%) Pakistan 2013 7.1% (6.7%, 7.6%) 4.9% (4.5%, 5.4%) 0.5% (0.4%, 0.6%) 5.7% (3.2%, 8.7%) Timor-Leste 2010 6.9% (6.4%, 7.4%) 4.7% (4.2%, 5.1%) 0.5% (0.4%, 0.6%) 1.8% (0.5%, 3.9%) Senegal 2015 6.3% (5.9%, 6.7%) 4.8% (4.4%, 5.2%) 0.3% (0.3%, 0.4%) 7.6% (4.1%, 11.5%) Zambia 2014 6.7% (6.3%, 7.1%) 4.5% (4.1%, 4.9%) 0.5% (0.4%, 0.6%) 2.9% (1.2%, 5.1%) Ghana 2014 6.4% (5.9%, 7.0%) 4.1% (3.7%, 4.7%) 0.5% (0.4%, 0.7%) 2.8% (0.6%, 5.8%) Madagascar 2009 6.7% (6.3%, 7.1%) 4.5% (4.1%, 4.9%) 0.5% (0.4%, 0.6%) 6.0% (3.7%, 8.9%) Tanzania 2016 6.0% (5.5%, 6.4%) 4.1% (3.7%, 4.5%) 0.4% (0.3%, 0.5%) 0.3% (0.0%, 1.3%) Bolivia 2008 6.1% (5.6%, 6.6%) 4.1% (3.6%, 4.5%) 0.5% (0.4%, 0.6%) 11.2% (7.6%, 15.4%) India 2006 5.9% (5.7%, 6.1%) 4.3% (4.2%, 4.5%) 0.3% (0.2%, 0.3%) 22.7% (19.8%, 25.9%) Malawi 2016 5.8% (5.5%, 6.2%) 4.2% (3.9%, 4.5%) 0.3% (0.2%, 0.4%) 3.4% (1.6%, 5.7%) Rwanda 2015 5.3% (4.8%, 5.8%) 3.5% (3.1%, 4.0%) 0.4% (0.3%, 0.5%) 3.0% (0.9%, 6.0%) Benin 2012 6.6% (6.2%, 7.0%) 3.6% (3.2%, 3.9%) 0.8% (0.6%, 0.9%) 2.1% (1.1%, 3.4%) Gabon 2012 5.5% (4.9%, 6.1%) 2.7% (2.3%, 3.2%) 0.7% (0.6%, 0.9%) 2.3% (0.7%, 4.5%) Bangladesh 2014 4.4% (4.1%, 4.7%) 3.2% (2.9%, 3.5%) 0.2% (0.2%, 0.3%) 6.4% (3.4%, 9.8%) Brazil 1996 5.6% (5.1%, 6.1%) 2.7% (2.2%, 3.1%) 0.8% (0.6%, 1.0%) 9.9% (6.6%, 13.6%) Sao Tome and Principe 2009 6.9% (6.0%, 8.0%) 1.9% (1.3%, 2.6%) 1.9% (1.4%, 2.3%) 0.2% (0.0%, 1.2%) Azerbaijan 2006 5.0% (4.3%, 5.8%) 2.0% (1.5%, 2.7%) 0.9% (0.6%, 1.1%) 0.8% (0.0%, 2.4%) Morocco 2004 4.6% (4.1%, 5.1%) 2.4% (2.0%, 2.8%) 0.5% (0.4%, 0.6%) 6.0% (3.0%, 9.5%) South Africa 1998 4.4% (3.9%, 4.9%) 2.1% (1.7%, 2.5%) 0.5% (0.4%, 0.7%) 7.9% (4.7%, 11.8%) Kenya 2014 4.0% (3.7%, 4.2%) 2.6% (2.4%, 2.8%) 0.2% (0.2%, 0.2%) 0.1% (0.0%, 0.5%) Kazakhstan 1999 4.5% (3.8%, 5.3%) 1.8% (1.3%, 2.3%) 0.8% (0.5%, 1.0%) 1.7% (0.2%, 4.0%) Namibia 2013 4.5% (4.0%, 5.0%) 2.0% (1.6%, 2.4%) 0.7% (0.5%, 0.9%) 1.8% (0.5%, 3.8%) Table 2. PLOS ONE PLOS ONE Table 2. (Continued) Country Mean Median Variance R2 (Wealth) Cambodia 2014 4.5% (4.1%, 5.0%) 2.3% (2.0%, 2.6%) 0.5% (0.4%, 0.6%) 8.3% (5.5%, 11.6%) Nicaragua 2001 3.6% (3.2%, 3.9%) 2.1% (1.8%, 2.4%) 0.2% (0.2%, 0.3%) 4.2% (1.9%, 7.4%) Indonesia 2012 4.0% (3.7%, 4.3%) 2.2% (1.9%, 2.4%) 0.3% (0.2%, 0.4%) 13.7% (10.5%, 16.9%) Paraguay 1990 4.1% (3.6%, 4.6%) 1.8% (1.5%, 2.2%) 0.5% (0.4%, 0.7%) 2.0% (0.5%, 4.2%) Turkey 2004 4.5% (4.0%, 5.1%) 1.9% (1.5%, 2.3%) 0.6% (0.5%, 0.8%) 6.1% (3.3%, 9.7%) Comoros 2012 4.1% (3.5%, 4.7%) 1.1% (0.8%, 1.5%) 0.8% (0.6%, 1.0%) 0.1% (0.0%, 0.6%) Uzbekistan 1996 4.4% (3.7%, 5.1%) 1.3% (0.9%, 1.7%) 0.9% (0.7%, 1.2%) 0.1% (0.0%, 0.6%) Guatemala 2015 2.9% (2.6%, 3.2%) 1.7% (1.5%, 1.9%) 0.2% (0.1%, 0.2%) 4.7% (2.4%, 7.5%) Dominican Republic 2013 3.2% (2.8%, 3.7%) 1.3% (1.0%, 1.6%) 0.4% (0.3%, 0.6%) 2.5% (0.8%, 4.9%) Kyrgyzstan 2012 2.8% (2.4%, 3.3%) 0.9% (0.6%, 1.3%) 0.5% (0.3%, 0.6%) 0.1% (0.0%, 0.6%) Vietnam 2002 3.0% (2.5%, 3.4%) 1.0% (0.8%, 1.4%) 0.4% (0.3%, 0.5%) 2.1% (0.4%, 4.6%) Peru 2012 2.3% (2.1%, 2.4%) 1.5% (1.4%, 1.6%) 0.1% (0.1%, 0.1%) 10.5% (7.4%, 14.3%) Colombia 2005 2.2% (2.0%, 2.4%) 1.2% (1.0%, 1.4%) 0.1% (0.1%, 0.2%) 4.1% (2.1%, 6.6%) Egypt 2014 2.5% (2.3%, 2.7%) 1.3% (1.1%, 1.5%) 0.2% (0.2%, 0.3%) 3.5% (1.8%, 5.7%) Honduras 2012 2.2% (2.0%, 2.5%) 1.1% (0.9%, 1.3%) 0.2% (0.1%, 0.3%) 1.3% (0.3%, 2.8%) Philippines 2013 2.5% (2.2%, 2.8%) 0.9% (0.7%, 1.2%) 0.3% (0.2%, 0.4%) 3.8% (2.0%, 5.7%) Guyana 2009 3.1% (2.5%, 3.7%) 0.4% (0.1%, 0.7%) 0.9% (0.7%, 1.3%) 0.8% (0.1%, 1.9%) Jordan 2012 1.7% (1.5%, 1.9%) 0.7% (0.5%, 0.8%) 0.2% (0.1%, 0.2%) 0.5% (0.0%, 1.5%) Albania 2009 2.4% (1.9%, 2.9%) 0.1% (0.0%, 0.2%) 0.9% (0.6%, 1.2%) 1.0% (0.2%, 2.3%) Armenia 2010 2.4% (1.9%, 3.1%) 0.1% (0.0%, 0.3%) 1.0% (0.7%, 1.3%) 0.5% (0.0%, 1.6%) Moldova 2005 2.9% (2.3%, 3.6%) 0.0% (0.0%, 0.2%) 1.2% (0.7%, 1.7%) 0.3% (0.0%, 1.1%) Ukraine 2007 1.8% (1.3%, 2.4%) 0.0% (0.0%, 0.2%) 0.7% (0.3%, 1.1%) 0.1% (0.0%, 0.6%) https://doi.org/10.1371/journal.pone.0238847.t002 mothers from the low risk group. There is also often a statistical, but not substantive difference in birth order. mothers from the low risk group. There is also often a statistical, but not substantive difference in birth order. The most substantial differences between the higher and lower risk groups are for residency (urban/rural), wealth, and previous death of a sibling. High risk births are substantively poorer than the remaining 80% of the population. PLOS ONE In Cambodia, high risk births average at the poorest 32nd percentile of wealth while the low risk births average around the 53rd percentile of wealth. We find similar results for other countries: Bolivia: 32% against 52%; Brazil: 31% against 53%; Peru: 30% against 53%; Nigeria: 32% against 53%. High risk births are disproportionately born to mothers that have already experienced a prior death of another child. The odds ratio is 18.8 (13.1, 26.7) in Benin; 16.3 (10.9, 24.1) in Mali; and 15.4 (11.9, 19.9) in Nigeria. Even for relatively wealthier countries, the odds ratio for another death is high for mothers that have experienced a prior death. The only countries in which a prior death is not a significant risk factor for a subsequent birth are Moldova and Viet- nam. Ukraine seems an exception, but the fractions of the births with a prior death are small, and this makes the odds ratio for Ukraine not very meaningful. PLOS ONE Results from ANOVA of posterior mean of mortality risk on wealth quintiles. Countries are ordered by medi R2 are presented as posterior means and 95% intervals. PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 13 / 23 Measure inequality within and between groups of births and to identify left behind populations in LMIC 4 Discussion In this study we have investigated inequality in under-5 mortality within and between socio- economic groups for a large pool of LMIC. We have made three related contributions to the existing research. First, we show that for all 67 countries in our sample, most of the variability in mortality risk exists within socioeconomic groups, not between groups. Second, we show that within countries the average mortality risk—which is closely related to national averages PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 14 / 23 Measure inequality within and between groups of births and to identify left behind populations in LMIC PLOS ONE Table 3. Efficiency gains by targeting 20% highest risk as estimated from our model versus targeting the poorest 20%. The first column gives the country and year. The second column gives sample size per survey. The third column is under 5 mortality rate in the 20% poorest. The fourth column is the mortality rate in the 20% identi- fied as having the highest mortality risk for each sample with 95% posterior intervals. Efficiency Gain is defined as (HRDeaths—PoorDeaths)/PoorDeaths. CAR is Central African Republic. Table 3. Efficiency gains by targeting 20% highest risk as estimated from our model versus targeting the poorest 20%. The first column gives the country and year. The second column gives sample size per survey. The third column is under 5 mortality rate in the 20% poorest. The fourth column is the mortality rate in the 20% identi- fied as having the highest mortality risk for each sample with 95% posterior intervals. Efficiency Gain is defined as (HRDeaths—PoorDeaths)/PoorDeaths. CAR is Central African Republic Table 3. Efficiency gains by targeting 20% highest risk as estimated from our model versus targeting the poorest 20%. The first column gives the country and year. The second column gives sample size per survey. The third column is under 5 mortality rate in the 20% poorest. The fourth column is the mortality rate in the 20% identi- fied as having the highest mortality risk for each sample with 95% posterior intervals. Efficiency Gain is defined as (HRDeaths—PoorDeaths)/PoorDeaths. CAR is Central Af i R bli fied as having the highest mortality risk for each sample with 95% posterior intervals. Efficiency Gain is defined as (HRDeaths—PoorDeaths)/PoorDeaths. CAR is Central African Republic. (Continued) PLOS ONE Table 3. 4 Discussion Country Year Sample Size Mortality Poor Mortality High Risk Efficiency Gains Albania 2009 2,481 32% 86% (74%, 94%) 168% (132%, 195%) Armenia 2000 2,602 22% 53% (47%, 59%) 144% (116%, 175%) Armenia 2010 1,545 23% 77% (63%, 91%) 230% (170%, 290%) Angola 2011 5,812 24% 44% (41%, 46%) 80% (69%, 89%) Azerbaijan 2006 2,739 27% 49% (42%, 55%) 83% (59%, 107%) Bangladesh 2000 9,061 26% 39% (36%, 40%) 50% (42%, 57%) Bangladesh 2004 7,261 24% 38% (35%, 40%) 55% (44%, 65%) Bangladesh 2007 6,929 25% 43% (40%, 45%) 69% (58%, 79%) Bangladesh 2014 14,512 27% 40% (37%, 42%) 47% (39%, 56%) Burkina Faso 1993 5,514 19% 31% (29%, 33%) 61% (51%, 70%) Burkina Faso 1999 5,702 21% 30% (28%, 31%) 39% (32%, 46%) Burkina Faso 2003 12,060 20% 33% (32%, 34%) 69% (62%, 74%) Burkina Faso 2010 16,759 23% 35% (33%, 36%) 52% (47%, 57%) Benin 1996 5,386 22% 32% (31%, 34%) 47% (39%, 55%) Benin 2001 5,691 26% 34% (32%, 36%) 33% (26%, 41%) Benin 2006 16,984 22% 35% (34%, 36%) 57% (52%, 63%) Benin 2012 12,904 21% 51% (49%, 53%) 139% (129%, 148%) Bolivia 1998 9,334 29% 42% (40%, 45%) 48% (40%, 55%) Bolivia 2004 10,546 25% 42% (40%, 44%) 67% (58%, 75%) Bolivia 2008 10,048 29% 42% (40%, 45%) 46% (37%, 55%) Brazil 1996 6,023 34% 51% (47%, 55%) 50% (40%, 62%) Burundi 2011 6,016 25% 38% (35%, 40%) 54% (45%, 64%) Cambodia 2000 12,071 27% 39% (38%, 41%) 47% (42%, 53%) Cambodia 2011 7,258 28% 47% (43%, 49%) 64% (53%, 74%) Cambodia 2014 8,272 31% 51% (48%, 54%) 67% (56%, 77%) CAR 1995 4,429 25% 36% (34%, 39%) 46% (35%, 54%) Chad 1997 6,941 15% 34% (33%, 36%) 133% (123%, 143%) Chad 2004 6,260 18% 34% (33%, 36%) 93% (83%, 102%) Chad 2015 18,985 22% 39% (38%, 40%) 77% (72%, 82%) Congo 2005 4,419 23% 40% (38%, 42%) 74% (63%, 85%) Congo 2012 7,597 22% 39% (36%, 41%) 74% (63%, 84%) Coˆte d’Ivoire 1999 2,757 25% 42% (39%, 45%) 73% (60%, 84%) Coˆte d’Ivoire 2005 3,812 22% 38% (35%, 41%) 70% (56%, 83%) Coˆte d’Ivoire 2012 7,224 21% 40% (38%, 42%) 94% (85%, 104%) Cameroon 1991 3,140 26% 41% (39%, 44%) 57% (46%, 67%) Cameroon 1998 4,080 30% 43% (40%, 45%) 43% (34%, 51%) Cameroon 2004 7,535 26% 40% (38%, 41%) 52% (46%, 58%) Cameroon 2011 10,812 29% 39% (38%, 41%) 35% (30%, 41%) Colombia 1990 4,087 34% 63% (56%, 70%) 86% (65%, 105%) Colombia 1995 5,041 30% 56% (49%, 62%) 90% (66%, 112%) Colombia 2005 15,630 32% 51% (46%, 57%) 58% (43%, 76%) Comoros 1996 2,208 22% 42% (38%, 46%) 93% (75%, 111%) Comoros 2012 3,390 18% 57% (51%, 62%) 223% (190%, 253%) DRC 2007 7,971 24% 40% (39%, 42%) 70% (64%, 76%) DRC 2014 15,132 22% 39% (37%, 40%) 74% (68%, 81%) (Continued) PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 15 / 23 Measure inequality within and between groups of births and to identify left behind populations in LMIC (Continued) PLOS ONE PLOS ONE Table 3. PLOS ONE (Continued) Country Year Sample Size Mortality Poor Mortality High Risk Efficiency Gains DR 1999 3,250 26% 50% (46%, 55%) 97% (78%, 116%) DR 2002 12,941 32% 49% (45%, 52%) 50% (40%, 61%) DR 2007 13,945 26% 46% (42%, 50%) 77% (61%, 93%) DR 2013 4,782 30% 51% (45%, 56%) 67% (48%, 85%) Egypt 1996 12,791 30% 41% (40%, 43%) 40% (35%, 46%) Egypt 2003 11,850 30% 49% (46%, 52%) 64% (54%, 72%) Egypt 2008 11,394 32% 51% (46%, 55%) 59% (46%, 72%) Egypt 2014 14,486 29% 52% (48%, 56%) 78% (64%, 91%) Eswatini 2007 2,421 22% 47% (42%, 51%) 111% (91%, 129%) Ethiopia 1997 12,984 18% 38% (36%, 39%) 109% (101%, 116%) Ethiopia 2003 13,218 22% 38% (37%, 40%) 79% (71%, 86%) Gabon 2001 3,783 20% 43% (39%, 47%) 111% (93%, 129%) Gabon 2012 5,149 26% 48% (44%, 52%) 84% (70%, 98%) Ghana 1994 3,281 24% 42% (40%, 46%) 77% (65%, 90%) Ghana 1999 3,226 23% 42% (39%, 45%) 84% (69%, 98%) Ghana 2003 4,134 25% 41% (38%, 44%) 63% (52%, 75%) Ghana 2008 3,258 25% 44% (40%, 49%) 81% (64%, 99%) Ghana 2014 6,370 27% 41% (37%, 44%) 53% (40%, 64%) Guinea 1999 6,867 24% 32% (30%, 34%) 32% (26%, 39%) Guinea 2005 7,807 22% 33% (31%, 34%) 48% (42%, 54%) Guinea 2012 8,010 26% 37% (35%, 39%) 45% (38%, 52%) Guatemala 1999 7,083 23% 42% (38%, 45%) 81% (68%, 95%) Guatemala 2015 11,719 28% 46% (43%, 49%) 66% (54%, 79%) Guyana 2005 1,268 14% 86% (76%, 93%) 525% (450%, 575%) Guyana 2009 2,464 10% 68% (59%, 77%) 578% (489%, 667%) Honduras 2006 12,380 23% 43% (39%, 46%) 88% (72%, 102%) Honduras 2012 10,065 26% 50% (44%, 55%) 93% (71%, 113%) Haiti 1995 3,020 24% 40% (37%, 43%) 65% (53%, 75%) Haiti 2000 7,063 23% 36% (34%, 38%) 56% (48%, 64%) Haiti 2006 5,907 26% 41% (39%, 44%) 60% (49%, 70%) Haiti 2012 6,944 20% 40% (37%, 43%) 102% (88%, 116%) India 1993 65,681 29% 41% (41%, 42%) 45% (42%, 47%) India 2000 53,079 30% 39% (38%, 39%) 31% (29%, 33%) India 2006 59,240 32% 43% (42%, 44%) 35% (32%, 38%) Indonesia 1997 23,155 26% 46% (45%, 48%) 79% (73%, 85%) Indonesia 2003 16,049 31% 51% (48%, 53%) 65% (57%, 73%) Indonesia 2007 20,592 34% 49% (48%, 51%) 44% (39%, 49%) Indonesia 2012 19,788 35% 52% (49%, 54%) 46% (39%, 53%) Jordan 1990 9,308 28% 47% (44%, 51%) 67% (55%, 79%) Jordan 1997 6,408 27% 54% (50%, 59%) 100% (84%, 119%) Jordan 2002 7,098 23% 48% (43%, 53%) 108% (88%, 130%) Jordan 2009 13,691 23% 52% (47%, 56%) 123% (104%, 141%) Jordan 2012 11,205 26% 54% (48%, 59%) 106% (84%, 128%) Kenya 1993 6,514 29% 45% (42%, 48%) 55% (46%, 65%) Kenya 1998 5,789 28% 51% (48%, 53%) 83% (73%, 93%) Kenya 2009 5,412 20% 48% (45%, 51%) 136% (120%, 150%) Kenya 2014 23,924 16% 44% (42%, 46%) 179% (166%, 192%) (Continued) PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 16 / 23 Measure inequality within and between groups of births and to identify left behind populations in LMIC (Continued) PLOS ONE PLOS ONE Table 3. https://doi.org/10.1371/journal.pone.0238847.t003 PLOS ONE (Continued) Country Year Sample Size Mortality Poor Mortality High Risk Efficiency Gains Kazakhstan 1999 2,651 25% 50% (43%, 56%) 103% (76%, 127%) Kyrgyzstan 1997 2,400 27% 52% (46%, 57%) 92% (71%, 110%) Kyrgyzstan 2012 3,705 19% 53% (45%, 60%) 184% (141%, 223%) Liberia 2009 6,871 22% 36% (35%, 38%) 65% (57%, 73%) Liberia 2013 8,220 24% 37% (36%, 39%) 59% (51%, 66%) Lesotho 2005 3,115 23% 46% (42%, 51%) 103% (85%, 123%) Lesotho 2010 3,107 17% 45% (41%, 50%) 160% (134%, 185%) Lesotho 2014 3,250 17% 44% (39%, 48%) 158% (133%, 184%) Morocco 1992 5,422 27% 39% (36%, 42%) 46% (35%, 57%) Morocco 2004 6,493 30% 49% (45%, 53%) 63% (50%, 76%) Moldova 2005 1,744 21% 88% (72%, 96%) 317% (244%, 358%) Madagascar 1997 5,960 19% 36% (34%, 38%) 93% (83%, 103%) Madagascar 2004 5,268 30% 45% (42%, 48%) 50% (40%, 59%) Madagascar 2009 12,686 25% 43% (41%, 45%) 73% (64%, 81%) Mali 1996 9,960 24% 30% (29%, 31%) 26% (22%, 30%) Mali 2001 13,031 21% 32% (31%, 33%) 56% (52%, 59%) Mali 2006 15,201 23% 33% (32%, 34%) 45% (41%, 49%) Mali 2013 9,249 21% 47% (45%, 48%) 122% (113%, 131%) Malawi 1992 4,746 21% 34% (33%, 36%) 62% (54%, 69%) Malawi 2005 9,663 24% 34% (32%, 35%) 42% (35%, 48%) Malawi 2010 20,677 22% 34% (33%, 35%) 53% (47%, 58%) Malawi 2016 16,793 24% 41% (39%, 43%) 70% (61%, 77%) Mozambique 1997 6,834 23% 39% (38%, 41%) 72% (65%, 78%) Mozambique 2004 8,942 19% 36% (34%, 37%) 91% (83%, 99%) Mozambique 2011 10,379 24% 42% (40%, 44%) 77% (69%, 85%) Nicaragua 1998 8,665 20% 46% (43%, 49%) 131% (115%, 147%) Nicaragua 2001 9,008 23% 43% (40%, 47%) 87% (72%, 103%) Nigeria 1990 8,696 25% 43% (42%, 45%) 70% (65%, 75%) Nigeria 2003 5,848 22% 36% (34%, 37%) 61% (54%, 67%) Nigeria 2008 30,182 24% 37% (36%, 37%) 55% (52%, 58%) Nigeria 2013 34,186 26% 40% (39%, 41%) 56% (53%, 58%) Niger 1998 7,644 18% 32% (31%, 33%) 81% (75%, 86%) Niger 2006 9,820 17% 35% (33%, 36%) 106% (98%, 114%) Niger 2012 13,573 20% 37% (36%, 39%) 84% (77%, 89%) Namibia 1992 3,692 22% 46% (43%, 49%) 104% (91%, 118%) Namibia 2000 4,354 23% 55% (50%, 59%) 138% (119%, 157%) Namibia 2007 4,668 29% 50% (46%, 54%) 73% (59%, 87%) Namibia 2013 4,691 22% 51% (46%, 56%) 130% (108%, 152%) Pakistan 1991 8,356 20% 46% (45%, 48%) 129% (120%, 138%) Pakistan 2007 9,531 26% 47% (44%, 49%) 78% (69%, 87%) Pakistan 2013 11,854 26% 42% (40%, 44%) 64% (56%, 71%) Peru 1992 9,085 27% 44% (42%, 46%) 66% (58%, 74%) Peru 1996 19,554 28% 42% (40%, 44%) 51% (45%, 58%) Peru 2000 17,334 29% 42% (40%, 44%) 48% (41%, 54%) Peru 2008 13,739 30% 45% (41%, 49%) 48% (35%, 60%) Peru 2012 31,443 32% 44% (41%, 46%) 35% (27%, 42%) (Continued) PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 17 / 23 Measure inequality within and between groups of births and to identify left behind populations in LMIC PLOS ONE (Continued) Country Year Sample Size Mortality Poor Mortality High Risk Efficiency Gains Paraguay 1990 4,375 22% 48% (43%, 53%) 118% (96%, 139%) Philippines 1993 9,340 27% 48% (46%, 51%) 78% (68%, 87%) Philippines 1998 8,361 27% 50% (46%, 53%) 84% (72%, 97%) Philippines 2003 7,863 33% 53% (49%, 58%) 63% (50%, 77%) Philippines 2008 7,480 29% 54% (50%, 58%) 83% (70%, 97%) Philippines 2013 8,159 32% 56% (51%, 61%) 77% (62%, 92%) Rwanda 1992 6,071 18% 37% (35%, 38%) 99% (90%, 109%) Rwanda 2005 9,139 22% 34% (32%, 35%) 52% (46%, 58%) Rwanda 2008 4,865 18% 43% (41%, 45%) 143% (130%, 156%) Rwanda 2015 8,096 24% 41% (38%, 45%) 74% (60%, 88%) Sierra Leone 2008 6,413 23% 40% (38%, 42%) 71% (63%, 78%) Sierra Leone 2013 13,981 22% 36% (35%, 37%) 64% (59%, 69%) Senegal 1997 7,311 24% 34% (32%, 36%) 41% (33%, 49%) Senegal 2005 10,284 26% 36% (34%, 37%) 36% (30%, 42%) Senegal 2009 13,229 24% 36% (34%, 37%) 49% (43%, 55%) Senegal 2015 12,606 27% 38% (36%, 40%) 45% (37%, 52%) Sao Tome and Principe 2009 1,685 19% 54% (48%, 60%) 181% (150%, 212%) Togo 1998 7,211 22% 33% (32%, 35%) 51% (43%, 58%) Togo 2014 6,901 24% 40% (38%, 43%) 70% (60%, 80%) Timor-Leste 2010 9,499 22% 42% (39%, 44%) 87% (76%, 97%) Turkey 1993 4,998 33% 49% (45%, 52%) 47% (38%, 56%) Turkey 1998 4,162 31% 51% (46%, 56%) 64% (48%, 78%) Turkey 2004 4,765 30% 51% (47%, 55%) 68% (54%, 81%) Tanzania 1999 6,715 20% 36% (33%, 37%) 81% (70%, 90%) Tanzania 2005 7,200 23% 36% (34%, 38%) 55% (46%, 63%) Tanzania 2010 11,262 25% 38% (36%, 40%) 54% (47%, 61%) Tanzania 2016 8,745 21% 40% (37%, 42%) 86% (73%, 99%) Ukraine 2007 1,494 19% 87% (65%, 100%) 350% (233%, 417%) Uganda 1995 6,244 25% 35% (33%, 37%) 43% (35%, 51%) Uganda 2001 5,933 25% 36% (34%, 38%) 46% (37%, 55%) Uganda 2010 5,912 24% 36% (34%, 38%) 47% (38%, 56%) Uganda 2011 7,852 24% 37% (35%, 39%) 57% (47%, 65%) Uzbekistan 1996 2,656 24% 55% (50%, 60%) 129% (109%, 153%) Vietnam 2002 4,060 29% 54% (47%, 59%) 83% (61%, 102%) South Africa 1998 5,564 33% 48% (44%, 53%) 45% (31%, 58%) Zambia 1997 5,614 23% 35% (34%, 37%) 52% (45%, 61%) Zambia 2002 6,027 24% 35% (33%, 37%) 46% (39%, 55%) Zambia 2007 5,808 16% 36% (34%, 38%) 122% (108%, 135%) Zambia 2014 12,324 26% 42% (39%, 44%) 61% (51%, 69%) Zimbabwe 1994 4,622 21% 47% (43%, 50%) 122% (103%, 141%) Zimbabwe 1999 3,713 21% 50% (46%, 54%) 135% (115%, 153%) Zimbabwe 2006 4,357 19% 46% (41%, 50%) 137% (112%, 162%) Zimbabwe 2011 4,374 20% 49% (45%, 54%) 146% (124%, 168%) Zimbabwe 2015 5,726 25% 44% (41%, 46%) 72% (61%, 82%) https://doi.org/10.1371/journal.pone.0238847.t003 PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 18 / 23 PLOS ONE Measure inequality within and between groups of births and to identify left behind populations in LMIC of child mortality—is far from the typical (modal) mortality risk experienced by most births. PLOS ONE Third, we show that poverty status alone, while important, is a poor proxy for being at the higher risk of an an early death than the general population. All these findings have important policy implications. In addition, we have developed new methods to analyse inequality in mor- tality risk which have broad applicability. While quantifying inequality in under-5 mortality between socioeconomic groups is impor- tant it misses a larger within-group inequality. In particular, we have shown that for most countries socioeconomic group explains less than 5% of the total variability in mortality. Even in countries where socioeconomic inequality matters the most, socioeconomic group explains very little of the variation in U5MR. For example, socioeconomic status explains 11% of U5MR in Bolivia and 22% in India. This means that there is a large overlap in mortality risk among births from different socioeconomic groups and, as a consequence, there is a large a number of high risk individuals outside that poorest group. In addition, being born to a partic- ular country does not predict your mortality risk very well, which means that between country comparisons also miss most of the variability in mortality risk. In addition of being incomplete, between country comparisons are often done in terms of average level of child mortality. However, we show that countries’ distributions of mortality risk are right skewed because some births experience substantially higher mortality risk than the national averages. These are left behind populations who are largely unnoticed when we only look at average mortality in socioeconomic groups. The typical modal mortality rate in each country is very different from the national averages of child mortality. Thus between- country comparisons using national averages are not comparing typical mortality levels between countries. Finally, most equity based policy strategies that target births are based on a single risk fac- tor, usually poverty status. However, efficiency gains from targeting the 20% highest risk births versus the 20% poorest are substantively important for all countries that we have data for, with efficiency gains ranging from 26% in Mali (1996), to more than 550% in Guyana (2009), likely due to the fact that it is one of the few countries with an apparent decrease in mortality risk with increasing wealth. PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 PLOS ONE Thus we are not suggesting major changes in interven- tions targeting high risk populations. Instead, we are proposing a new methodology that combines information from multiple well known risk factors simultaneously to identify high risk births. Our approach considers interactions among risk factors that are readily available for LMIC via nationally representative health surveys, and frees researchers and policy makers from having to decide which risk factors capture most of the inequality in each country-year. The methods developed in this paper have broader applicability and are flexible enough to be applied to a number of different scenarios. For example, some countries with good vital reg- istration system could use their administrative data instead of surveys. When people wish to implement an intervention in a particular country, our methodology points the way to a more targeted and impactful intervention. Implementers will need to choose variables, and they may choose different predictors than we have chosen, depending on data available and political and medical considerations. This is acceptable and something we consider a necessary part of implementing our methods in practice. Our recommendations are also related to a large body of literature in medicine and public health that develops risk scores for individuals to identify those at risk of some event. These scores have been applied to a variety of outcomes and our results suggest the possible useful- ness of such scores for identification of high risk children. [35] Our approach requires repre- sentative surveys of the population, such as DHS or Multiple Indicator Cluster Surveys (MICS) so that we can rank children by mortality risk based on demographics. Policy makers could use mobile apps, which are now widely used for data collection, to collect and combine information on the children, calculate their risk, and then check whether their score is above or below a pre-determined threshold. We would not suggest a single risk score for the entire world. Rather, we would develop a score for each country, and we would update the score as new data became available. The calculus of the efficiency gains assumes that interventions have the same costs for each birth. In reality, costs need to be adjusted according to local conditions. However, our approach provides a baseline to which any other allocation algorithm should be compared. Every comparison allocation scheme also needs to accommodate costs, not just our allocation scheme. PLOS ONE Although the 20% highest risk births are usually the poorest and from rural areas, as might be expected, including other risk factors and their interactions consider- ably improves the identification of left behind individuals. One previously overlooked characteristic is the importance of having experienced a prior death of a child. [28, 29] This is likely the case because this variable represents several unmea- sured risk factors at the maternal level. However, it is an observable variable and can be the object of policy targeting. And it should be used to do so. We find that this is a particularly important characteristic for Sub-Saharan Africa countries in our sample. For these countries, just targeting mothers that have already experienced the death of a child could be an effective way to reach high risk populations. Taken together these results support the view that measuring national averages of under- 5 mortality is insufficient to identify left behind groups. [5, 30–34] The concerns raised by United Nations General Assembly Resolution 68/261 are real and important, and we have shown that policy makers and international agencies should routinely implement disaggre- gation of inequality measures by several demographic variables simultaneously. [4] However, our findings suggest that monitoring inequality between socioeconomic groups of births may not enable policy makers to accurately identify many left behind children. We recom- mend using nationally representative surveys or administrative data to estimate mortality risk at the individual level to identify left behind populations that can be the target of inter- ventions. We also recommend our methods to properly quantify and monitor high risk populations. 19 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 PLOS ONE Measure inequality within and between groups of births and to identify left behind populations in LMIC Our findings should not be interpreted as recommending against targeting the poor. Pov- erty alone is not the best guide for equity based policies because other risk factors are also important. Poverty status needs to be combined with other available information to identify high risk births. This is important for both low and high mortality countries, because children in need are spread out across socioeconomic groups. Further, since high risk children tend to be poor and from rural areas, most interventions that work for the poorest children will proba- bly work for the highest risk children. Author Contributions Conceptualization: Antonio P. Ramos, Robert E. Weiss. Conceptualization: Antonio P. Ramos, Robert E. Weiss. Data curation: Antonio P. Ramos. Formal analysis: Antonio P. Ramos, Martin J. Flores. Funding acquisition: Antonio P. Ramos. Investigation: Antonio P. Ramos. Methodology: Antonio P. Ramos, Martin J. Flores, Robert E. Weiss. Project administration: Antonio P. Ramos. Software: Martin J. Flores. Supervision: Robert E. Weiss. Writing – original draft: Antonio P. Ramos, Martin J. Flores. Writing – review & editing: Antonio P. Ramos, Martin J. Flores, Robert E. Weiss. Methodology: Antonio P. Ramos, Martin J. Flores, Robert E. Weiss. Project administration: Antonio P. Ramos. Writing – original draft: Antonio P. Ramos, Martin J. Flores. Writing – review & editing: Antonio P. Ramos, Martin J. Flores, Robert E. Weiss. Supporting information S1 Appendix. [36]. (PDF) PLOS ONE For example, targeting the poor is likely easier in urban settings than in rural settings, and this would be a differential cost for the simple “intervene with the poor” intervention. It is possible to incorporate costs; one would multiply estimated probability of mortality times cost, then follow our same procedure to identify a combination of cheapest and most at risk to intervene with, until the budget had been spent. Instead of identifying the 20% most at risk, one would tabulate costs until the allocation funds had been spent. No matter differential costs, combining information from multiple observable risk factors better identifies high risk populations. Having identified higher risk populations, public health officials can then work to bring down costs, and best target at-risk births. Our methodology has not explicitly included the complex sampling design from the DHS. We did this to create a more parsimonious set of methodological innovations. We treated DHS samples as a random sample. However, we have included all variables used to stratify the surveys, which implicitly incorporates some of the sample design in our analysis. Future research should explicitly incorporate survey design. 20 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 PLOS ONE Measure inequality within and between groups of births and to identify left behind populations in LMIC In conclusion, our results show that despite progress toward reducing national averages of under-5 mortality, we still have substantial inequality within groups of births defined by com- monly used stratifiers that measure progress toward SDG’s. Our results suggest that research- ers and policy makers should also quantify inequality in mortality risk within groups of births in addition to between-groups comparisons. Quantifying both between and within group inequality helps us to have an accurate picture of inequality in under-5 mortality and to iden- tify left behind populations that otherwise cannot be easily identified. References 1. Moser KA, Leon DA, Gwatkin DR. How does progress towards the child mortality millennium develop- ment goal affect inequalities between the poorest and least poor? Analysis of Demographic and Health Survey data. British Medical Journal. 2005; 331 (7526). https://doi.org/10.1136/bmj.38659.588125.79 PMID: 16284209 2. Stuckler D, Basu S, Mckee M. Drivers of inequality in Millennium Development Goal progress: a statisti- cal analysis. PLoS Medicine. 2010; 7(3):e1000241. https://doi.org/10.1371/journal.pmed.1000241 PMID: 20209000 3. Mullholand E, Smith L, Carneiro I, Becher H, Lehmann D. Equity and Child Survival Strategies. Bulletin of the World Health Organization. 2008; 86(5). https://doi.org/10.2471/BLT.07.044545 PMID: 18545743 4. Economic and Social Council: Statistical Commision. Report of the Inter-Agency and Expert Group on Sustainable Development Goal Indicators. United Nations; 2016. 5. Chao F, You D, Pedersen J, Hug L, Alkema L. National and regional under-5 mortality rate by economic status for low-income and middle-income countries: a systematic assessment. The Lancet Global Health. 2018; 6:e535–47. https://doi.org/10.1016/S2214-109X(18)30059-7 PMID: 29653627 6. De P, Dhar A. Inequality in child mortality across different states of India: A comparative study. Journal of Child Health Care. 2013; 17(4):397–409. https://doi.org/10.1177/1367493512468359 PMID: 23435164 7. Ramos AP, Weiss RE. Measuring Within and Between Group Inequality in Early-Life Mortality Over Time: A Bayesian Approach with Application to India; 2018. 21 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 PLOS ONE Measure inequality within and between groups of births and to identify left behind populations in LMIC 8. Wagstaff A. Socioeconomic inequalities in child mortality: comparisons across nine developing coun- tries. Bulletin of The World Health Organization. 2000; 78(1):19–29. PMID: 10686730 9. Victora C, Wagstaff A, Schellenberg J, Gwatkin D, Claeson M, Habicht J. Applying an equity lens to child health and mortality: more of the same is not enough. The Lancet. 2003; 362(9379):233–241. https://doi.org/10.1016/S0140-6736(03)13917-7 PMID: 12885488 10. Sastry N. Trends in socioeconomic inequalities in mortality in developing countries: the case of child survival in Sao Paulo, Brazil. Demography. 2004; 41(3):443–464. https://doi.org/10.1353/dem.2004. 0027 PMID: 15461009 11. Brockerhoff M, Hewett P. Inequality of child mortality among ethnic groups in sub-Saharan Africa. Bulle- tin of The World Health Organization. 2000; 78(1). PMID: 10686731 12. Antai D. Regional inequalities in under-5 mortality in Nigeria: a population-based analysis of individual- and community-level determinants. Population Health Metrics. 2011; 9(1):6. https://doi.org/10.1186/ 1478-7954-9-6 PMID: 21388522 13. Jankowska M, Benza M, Weeks JR. Estimating spatial inequalities of urban child mortality. Demo- graphic Research. 2013; 28(2):33–62. https://doi.org/10.4054/DemRes.2013.28.2 PMID: 24478594 14. References Houweling TA, Kunst AE. Socio-economic inequalities in childhood mortality in low and middle income countries: a review of the evidence. British Medical Bulletin. 2010; 93:7–26. https://doi.org/10.1093/ bmb/ldp048 PMID: 20007188 15. Gwatkin D, Bhuiya A, Victora C. Making health systems more equitable. The Lancet. 2004; 364:1273– 80. https://doi.org/10.1016/S0140-6736(04)17145-6 PMID: 15464189 16. Black R, Morris S, Bryce J, Venis S. Where and why are 10 million children dying every year? Commen- tary. The Lancet. 2003; 361:2226–34. https://doi.org/10.1016/S0140-6736(03)13779-8 PMID: 12842379 17. Braveman P, Starfield B, Geiger J. World Health Report 2000: how it removes equity from the agenda for public health monitoring and policy. British Medical Journal. 2001; 323:678–81. https://doi.org/10. 1136/bmj.323.7314.678 PMID: 11566834 18. Bryce J, el Arifeen S, Pariyo G, Lanata C, Gwatkin D, Habicht J. Reducing child mortality: can public health deliver? The Lancet. 2003; 362(9378):159–164. https://doi.org/10.1016/S0140-6736(03)13870- 6 PMID: 12867119 19. Jones G, Steketee R, Black R, Bhutta Z, Morris S. How many child deaths can we prevent this year? The Lancet. 2003; 362:65–71. https://doi.org/10.1016/S0140-6736(03)13811-1 PMID: 12853204 20. Malqvist M, Yuan B, Trygg N, Selling K, Thomsen S. Targeted interventions for improved equity in maternal and child health in low- and middle-income settings: a systematic review and meta-analysis. PLoS ONE. 2013; 8(6). https://doi.org/10.1371/journal.pone.0066453 PMID: 23840474 21. Glassman A, Duran D, Fleisher L, Singer D, Sturke R, Angeles G, et al. Impact of Conditional Cash Transfers on Maternal and Newborn Health. Journal of Health, Population and Nutrition. 2013; 31.4 Suppl 2 (2013):S48–S66. PMID: 24992803 22. Basset L. Can Conditional Cash Transfer Programs Play a Greater Role in Reducing Child Undernutri- tion? World Bank; 2008. 23. Akresh R, de Walque D, Kazianga H. Alternative Cash Transfer Delivery Mechanism: Impacts on Rou- tine Preventive Health Clinic Visits in Burkina Faso. National Bureau of Economic Research; 2015. 17785. https://doi.org/10.3386/w17785 24. Banerjee A, Duflo E, Glennerster R, Kothari D. Improving immunisation coverage in rural India: clus- tered randomised controlled evaluation of immunisation campaigns with and without incentives. British Medical Journal. 2010; 340:c2220. https://doi.org/10.1136/bmj.c2220 PMID: 20478960 25. Ramos AP, Weiss R. Measuring Within and Between Group Inequality in Early-Life Mortality Over Time: A Bayesian Approach with Application to India; 2018. 26. Corsi DJ, Neuman M, Finlay JE, Subramanian SV. Demographic and health surveys: a profile. Interna- tional Journal of Epidemiology. 2012; 41(6):1602–1613. https://doi.org/10.1093/ije/dys184 PMID: 23148108 27. Fabic MMS, Choi YY, Bird SS. A systematic review of Demographic and Health Surveys: data availabil- ity and utilization for research. PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 References Bulletin of The World Health Organization. 2012; 90(8):604–612. https:// doi.org/10.2471/BLT.11.095513 PMID: 22893744 28. Abir T, Agho KE, Page AN, Milton AH, Dibley MJ. Risk factors for under-5 mortality: evidence from Ban- gladesh Demographic and Health Survey, 2004–2011. British Medical Journal. 2015; 5(8). https://doi. org/10.1136/bmjopen-2014-006722 PMID: 26297357 29. Gubhaju BB. The Effect of Previous Child Death on Infant and Child Mortality in Rural Nepal. Studies in Family Planning. 1985; 16(4):231–236. https://doi.org/10.2307/1967085 PMID: 4035724 22 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0238847 October 14, 2020 PLOS ONE Measure inequality within and between groups of births and to identify left behind populations in LMIC 30. Gwatkin D. How much would the poor gain from faster progress towards the Millenium Development Goals for health? The Lancet. 2005; 365:813–17. https://doi.org/10.1016/S0140-6736(05)17992-6 PMID: 15733726 31. UNICEF. Infant and Child Mortality in India: Levels, Trends and Determinants. UNICEF; 2012. 32. Gakidou E, King G. Measuring total health inequality: adding individual variation to group-level differ- ences. International Journal for Equity in Health. 2002; 1(1):3–3. https://doi.org/10.1186/1475-9276-1-3 PMID: 12379153 33. Black R, Cousens S, Johnson H, Lawn J, Rudan I, Bassani D, et al. Global, regional, and national causes of child mortality in 2008: a systematic analysis. The Lancet. 2010; 375:1969–1987. https://doi. org/10.1016/S0140-6736(10)60549-1 PMID: 20466419 34. Ramos AP, Heymann J, Weiss R. Improving Program Targeting to Combat Early-life Mortality by Identi- fying High Risk Births in India. Population Health Metrics. 2018; 16(17). https://doi.org/10.1186/s12963- 018-0172-6 PMID: 30139376 35. Mpimbaza A, Sears D, Sserwanga A, Kigozi R, Rubahika D, Nadler A, et al. Admission Risk Score to Predict Inpatient Pediatric Mortality at Four Public Hospitals in Uganda. PLoS ONE. 2015; 10(7): e0133950. https://doi.org/10.1371/journal.pone.0133950 PMID: 26218274 36. Rutstein SO. The DHS wealth index: approaches for rural and urban areas; 2008. 23 / 23
https://openalex.org/W4285678243	https://periodicos.ufms.br/index.php/orbital/article/download/16204/11003	English	null	Effects of Different Irrigation Solutions on the Chemical Composition of Multi-Root Canals of Radicular Dentine in Upper Premolar Teeth	Orbital	2,022	cc-by	7,047	Graphical abstract Graphical abstract Keywords Root canal irrigation solutions Upper premolar teeth Radicular dentin X-ray fluorescence Major and minor elements Article history Received 17 Fev 2022 Revised 26 Jun 2022 Accepted 27 Jun 2022 Available online 06 Jul 2022 Handling Editor: Marcelo Oliveira a Endodontic and Operative Dentistry Department, College of Dentistry, Damascus University, Damascus, Syrian Arab Republic. b Department of Chemistry, Atomic Energy Commission of Syria (AECS), P. O. Box 6091, Damascus, Syrian Arab Republic. Corresponding author. E-mail: morgana.quteifan@yahoo.com Orbital: Electron. J. Chem. 2022, 14(2), 95-102 Orbital: Electron. J. Chem. 2022, 14(2), 95-102 Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br Effects of Different Irrigation Solutions on the Chemical Composition of Multi-Root Canals of Radicular Dentine in Upper Premolar Teeth Morgana Quteifani Quteifan a and Ali Khuder b The objective of this research was to assess the effects of several irrigation solutions (ISOLs) on two major (P and Ca) and some minor (Fe, Zn, and Sr) elements of multi-root canals of radicular dentin in upper premolar teeth using X-ray fluorescence analysis (XRF). Powder of fifty radicular dentin specimens were prepared and divided into five groups based on treatments with different single and combined ISOLs. These groups were: G1-distilled water (control); G2- sodium hypochlorite (NaOCl, 5.25%); G3- sodium hypochlorite (NaOCl, 5.25%) and ethylendiaminetetraacetic acid (EDTA, 17%); G4- distilled water and a mixture of tetracycline, an acid, and a detergent (MTAD); and G5- sodium hypochlorite (NaOCl, 1.3%) and MTAD. The XRF results confirmed that the ISOL G5 (Ca/P=1.6254) was the most effective one. The ratio-to-ratio correlations revealed the common origins of Fe, Zn, and Sr in root canals of radicular dentin, which was highly interacted with the major elements of P and Ca each. These interactions were insignificantly affected by the used ISOLs. Finally, XRF is highly recommended as a safe, fast, nondestructive, and relatively not expensive technique to analyze the endodontic therapy of teeth. Graphical abstract 1 I t d ti 2.2 XRF measurements The measurements of dentin powders were performed using XRF instrument, which was assembled by the Syrian Atomic Energy Commission. The detector was PGT Si(Li) with an energy resolution of 160 eV at 5.9 keV and the preamplifier was a GAMMA-PGT system 4000 (Amp., ADC and MCA). The XRF instrument was equipped with a 2 kW Mo tube. Three secondary targets, i.e. Ti, Cu, and Mo, were used for X-ray excitations of (P), (Fe), and (Ca, Zn, and Sr) in teeth specimens with the operating conditions of (35 mA and 40 kV), (10 mA and 35 kV), and (20 mA and 35 kV), respectively. The live time was 2000 s for each of the X-ray excitation modes. The elements in the obtained spectra were determined using AXIL- QXAS software package, which was developed in the International Atomic Energy Agency (IAEA) [19]. One primary objective in the root canal treatments is to ensure chemomechanical removal of the microorganisms and organic/inorganic tissue residues from the infected root canals. Studies in the literature report that due to the structure of the root canals, mechanical instrumentation alone could not perform a satisfying cleaning in the canal. In ex vivo and clinical studies, some areas have remained unreached and infected, thus irrigation has played a key role in the achievement of root canal treatments [17, 18]. The accuracy of XRF method was checked by the analysis of SRM 1486 - Bone Meal, NIST standard sample; while, the hydroxyapatite salt (Ca5(PO4)3.OH, GR, Merck, Darmstadt, Germany) was used for estimation of only P accuracy. To establish the reproducibility of the instrumental technique, five different standard samples with masses of 0.200 g each were put in polyethylene cups, having sample area of approximately 0.2826 cm2. The cups were covered from one side with Mylar foils and sequentially exposed to XRF excitations using the previously mentioned operating conditions and modes. In a similar way as for the standard samples, the teeth powders were prepared and measured by XRF. 1. Introduction Teeth are a composite of organic, inorganic, and water fractions in various amounts. The organic components may consist of reacted coagulated proteins, necrotic or viable pulp tissue, odontoblastic processes, and microorganisms [1, 2]. The inorganic phase consists of the unit cell (Ca,X)10(P,C)6(O,OH)26,in which Ca and P are the two main elements in teeth [3]. Any change in the Ca:P ratio may change the original proportion of the organic and inorganic components, which in turn changes different properties of the tooth such as microhardness, permeability, and solubility characteristics of dentin and may also adversely affect the sealing ability and adhesion of dental materials such as resin- a Endodontic and Operative Dentistry Department, College of Dentistry, Damascus University, Damascus, Syrian Arab Republic. b Department of Chemistry, Atomic Energy Commission of Syria (AECS), P. O. Box 6091, Damascus, Syrian Arab Republic. *Corresponding author. E-mail: morgana.quteifan@yahoo.com Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br 95 Orbital: Electron. J. Chem. 2022, 14(2), 95-102 based cements and root canal sealers to dentin [4]. Changes in the Ca:P ratio may be caused by the chelating agents, which are used to improve chemomechanical debridement in the root canal treatment by removing the smear layer created during root canal instrumentation and composed of dentin structure and some nonspecific inorganic contaminants [1, 5]. out using reciprocation files and rotary instrumentation systems [20]. Then the specimens were stored in normal saline solution until used. The crowns of the teeth were removed at the cementoenamel junction using diamond disc (Brasseler, Savannah, GA, USA). Afterward, the digital X Rays of the specimens were performed to ensure that they are free from caries, fractures, and damages [20]. In addition to the previous two major elements, teeth accumulate a variety of minor elements such Sr, Zn, and Fe, which can be ingested by humans via different routes, including mainly food and water sources [6-8]. Strontium and Zn can enrich in teeth by substituting in hydroxyapatite [9, 10]. A negative correlation between Ca and Sr in main parts of teeth, e.g. the enamel and cementum, was obtained, confirming that Ca can be substituted by Sr either as a whole or in fractions for the formation of hydroxyapatite [9]. Zinc is readily acquired by synthetic hydroxyapatite, competing with Ca for positions on the surface of the apatite crystal. 2.2 XRF measurements The aim of this work was to (i) assess the distributions of P and Ca as major elements and Fe, Zn, and Sr as minor elements in root canals of radicular dentine in upper premolar teeth using XRF technique; (ii) study the effect of several single and combined ISOLs on the chemical composition of root canal regimes; and (iii) study the mutual interactions between the previously mentioned major and minor elements in the control and the irrigation treatment groups. 1. Introduction Zinc pretreatment of hydroxyapatite produces a resistance to acid dissolution similar in magnitude to that produced by equivalent molar concentrations of fluoride [11]. Among the minor elements studied in this work, Fe is presented in teeth and its ions (Fe3+) can induce the formation of apatite of high crystallinity and these ions can inhibit caries progression [12]. The teeth were randomly divided into the following five groups (n=10) according to protocol of ISOLs used. Group I- G1, the root canal dentin tissues were irrigated for 20 minutes with 5 mL of distilled water. Group II-G2, the root canal dentin tissues were irrigated for 20 minutes with 5 mL of5.25% NaOCl (Clorox, KSA). Group III-G3, the root canal dentin tissues were irrigated for 19 minutes with 5 mL of NaOCl 5.25%, then for 1 minute with 5 mL of 17% EDTA (META BIOMED CO., LTD Republic of Korea). Group IV-G4, the root canal dentin tissues were irrigated for 15 minutes with 5 mL of distilled water, then for 5 minutes with 5 mL of BioPure MTAD as instructed by the company (Tulsa Dental Specialties Company, OK, USA). Finally, Group V-G5, the root canal dentin tissues were irrigated for 15 minutes with 5 mL of 1.3% NaOCl, then for 5 minutes with 5 mL of BioPure MTAD. Dentin powders were obtained by Gates-Glidden burs after using the ISOLs. The powder of each canal dentin was kept in 5 cm3 sterilized plastic vial for further XRF analysis. Major and minor elements in teeth were analyzed using methods from different techniques, i.e., atomic absorption spectroscopy (AAS) and inductively coupled plasma mass spectroscopy (ICP-MS) were becoming more routine, but they are considered destructive techniques [3,13]. Proton induced X-ray emission (PIXE) [9, 14] and X-ray fluorescence (XRF) [15, 16] were used for multi-elemental non-destructive analysis of teeth samples. XRF showed a very fast, simultaneous, reliable, quantitative, multi-elemental and non-destructive technique with a very good ability to determine major and minor elements in a wide range of concentrations. Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br 3.1 XRF measurements The suitability of XRF method for the determination of the elements in dentin powders was checked by estimation of the accuracy (A), the precision (P), and the limits of detection (LOD). The A values were determined by the percentages of the relative differences between the obtained and the certified values of Ca, Fe, Zn, and Sr in five different subsamples (n = 5) of the bone meal standard reference material (SRM 1486 - Bone Meal, NIST), while the accuracy of P was determined by measurements of five different subsamples (n = 5) of the apatite reference sample. The results in Table 1 showed that there were no significant differences between the obtained elemental concentrations and the certified values (P>0.05) according to Student’s t-test for independent samples. The lower limits of detection (LOD) of the analyzed elements were estimated using the equation (1) [21]: LOD = (3/S)√(Ib/t), (1) (1) Where, S is the sensitivity (net peak counts per second per concentration); Ib is the intensity of the background of the respective peak (counts per second); and t is the counting time (s). Where, S is the sensitivity (net peak counts per second per concentration); Ib is the intensity of the background of the respective peak (counts per second); and t is the counting time (s). The results in Table 1 showed that all studied elements were with concentrations much higher than the obtained LOD values. Table 1. Characterization of element determination in the standard samples using XRF technique with different secondary targets Element Standard Secondary targeta Obtained concentration Certified values (A%)b (P%)c (LOD)d pe P (%) hydroxyapatite Ti 18.1±0.4 18.5 -2.16 ±2.21 0.484 0.085 Ca (%) bone meal SRM Mo 26.4±1.4 26.58 -0.677 ±5.30 0.040 0.826 Fe (µg/g) bone meal SRM Cu 104±5 99 +5.05 ±4.81 3.68 0.291 Zn (µg/g) bone meal SRM Mo 158±13 147 +7.48 ±8.23 6.25 0.255 Sr (µg/g) bone meal SRM Mo 266±15 264 +0.758 ±5.53 1.90 0.773 a The operating conditions were (35 mA and 40 kV) for P; (20 mA and 35 kV) for Ca, Zn, and Sr; and (10 mA and 35 kV) for Fe; collecting time was 2000 s for all used XRF targets each. b the percent accuracy, which is estimated by dividing the difference between the obtained and the certified values by the certified value and multiply the result by 100. Orbital: Electron. J. Chem. 2022, 14(2), 95-102 Orbital: Electron. J. Chem. 2022, 14(2), 95-102 Orbital: Electron. J. Chem. 2022, 14(2), 95-102 The precision of XRF measurements was evaluated in terms of the relative standard deviation (RSD = SD/Ci)×100, where SD is the standard deviation and Ci is the mean of ith element concentrations. The results showed that the precision was better than ± 8.3% for all determined elements. 3.1 XRF measurements c the percent precision, which is estimated by dividing the standard deviation by the obtained concentration and multiplying the result by 100. d the limits of detection. e the probability. Ca (%) bone meal SRM Mo 26.4±1.4 26.58 -0.677 ±5.30 0.040 0.826 Fe (µg/g) bone meal SRM Cu 104±5 99 +5.05 ±4.81 3.68 0.291 Zn (µg/g) bone meal SRM Mo 158±13 147 +7.48 ±8.23 6.25 0.255 Sr (µg/g) bone meal SRM Mo 266±15 264 +0.758 ±5.53 1.90 0.773 a The operating conditions were (35 mA and 40 kV) for P; (20 mA and 35 kV) for Ca, Zn, and Sr; and (10 mA and 35 kV) for Fe; collecting time was 2000 s for all used XRF targets each. b the percent accuracy, which is estimated by dividing the difference between the obtained and the certified values by the certified value and multiply the result by 100. c the percent precision, which is estimated by dividing the standard deviation by the obtained concentration and multiplying the result by 100. d the limits of detection. e the probability. Sr (µg/g) bone meal SRM Mo 266±15 264 +0.758 ±5.53 1.90 0.773 a The operating conditions were (35 mA and 40 kV) for P; (20 mA and 35 kV) for Ca, Zn, and Sr; and (10 mA and 35 kV) for Fe; collecting time was 2000 s for all used XRF targets each. b the percent accuracy, which is estimated by dividing the difference between the obtained and the certified values by the certified value and multiply the result by 100. c the percent precision, which is estimated by dividing the standard deviation by the obtained concentration and multiplying the result by 100. d the limits of detection. e the probability. The operating conditions were (35 mA and 40 kV) for P; (20 mA and 35 kV) for Ca, Zn, and Sr; and (10 mA and 35 kV time was 2000 s for all used XRF targets each. b the percent accuracy, which is estimated by dividing the difference bet and the certified values by the certified value and multiply the result by 100. c the percent precision, which is estimat standard deviation by the obtained concentration and multiplying the result by 100. d the limits of detection. e the proba 3.2 Root canal dentin analysis 434) µg/g. This could be explained as a result of high-protein diets, which characterized many individuals in the country; taking into account that, beef and lamb, as well as nuts, whole grains, and legumes are significant sources of zinc in the diet [23]. According to the previously mentioned reasons, diet and other individual factors could affect the Zn distributions in teeth specimens. Fifty fine and homogenized specimens of the collected root canal dentins were analyzed by the XRF technique with the previously mentioned operating conditions and modes. The XRF results of P, Ca, Fe, Zn, and Sr in the powders of the root canal dentins are shown in Table 2. 3.4.1 Major elements A very well positive linear relationship was obtained between the concentration means of P and Ca in the treatment groups (Fig. 1). The coefficient of determination (R2) was with a value of 0.9996. The results in Fig. 1 also shows a very interesting correlation between the ratios of Ca/P and the concentration means of P with R2 value of 0.9974. It was obviously that, the least value of Ca/P was for the control group (1.5145), accepting those obtained in literature [24]. Distilled water (DW) was used in the control group (G1) as an ISOL, which had only a mechanical flushing action to remove material from the root canal system [25]. According to this application, the instrumented specimens in G1 group were expected to contain significant amounts of P not only in the hydroxyapatite fraction form, but also in the organic and the remnant fraction forms. Depending on latter interpretation, higher P concentrations and thus a minimal ratio of Ca/P was expected to find in the control group compared to the other treatment groups. As a result, the different ratios of Ca/P means were to reflect the different effects of the single and the combined ISOLs on the chemical compositions of the organic and inorganic fractions of the root canal dentin tissues. Overall, Ca/P ratios were increased in the following order: The results showed that P and Ca were the two main elements in the root canals dentin specimens of the control group with mean values of 17.3±0.9% and 26.2±0.8%, respectively. These two elements were normally distributed with relatively narrow ranges of P (16.0-18.6%) and Ca (25.0- 27.1%), accepting the results obtained in literature [3,4,10], where P and Ca were distributed homogenously through the root dentin tissues. The results also showed that, the normal distributions of P and Ca were not influenced by treatments using the ISOLs. In addition, the normality was obtained for the minor elements in the control and the treatment groups, except two instances belonged to Zn in control (G1) and G2 treatment groups. Accordingly, Fe and Sr were normally distributed in the control group with concentration values of 69.0 (28-132) µg/g and 102 (48-154) µg/g, respectively, which were in good agreement with those reported in literature [3,14]. 3.4 ISOLs effects Several workers reported that X-ray fluorescence (XRF) is a useful technique for the analysis of major and minor elements in teeth samples [15,16,22]. In the present work, XRF showed the possibility to determine P and Ca as two major elements and Fe, Zn, and Sr as three minor elements in the root canal dentin powders. Table 2 shows the XRF results of the means, the medians and the ranges of the studied elements in the root canal dentin specimens. 2.3 Statistics This study was approved by the Research Ethics Committee at Damascus University. All patients signed an informed consent to use their teeth in this study. Fifty human teeth samples were collected from the patients who visited the dental clinics at the faculty of dentistry in Damascus University for orthodontic reasons. The teeth were extracted from the first upper molars of the left side of adults aged 20- 35 year. The teeth were put in sodium hypochlorite NaOCl 5.25% for 2 hours [19]. Apical extrusion of debris was carried The normal distributions of data were examined using Shapiro-Wilk test. The significant differences (p = 0.05) between the concentration means of the studied elements in treatment groups (G2-G5) and those in control group (G1) were analyzed with Tukey’s honest significance test and Mann-Whitney U test. Pearson correlation analysis was used to investigate the mutual relations between each pair of the studied elements. Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br 96 Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br 3.4.1 Major elements G1, G2, G3, G4, and G5 are the treatment groups with distilled water (DW), 5.25% NaOCl, 5.25% NaOCl+17% EDTA, DW+MTAD, and 1.3% NaOCl+MTAD ISOLs, respectively. Fig. 2. Relative changes (RC) in P (a) and Ca (b) concentration means in root canal dentin tissues caused by different ISOLs. G1 is the control group; G2, G3, G4, and G5 are the treatment groups; SE and SD are the standard error and standard deviation, respectively. Fig. 1. Relationships of Ca and Ca/P versus P concentration. means. G1, G2, G3, G4, and G5 are the treatment groups with distilled water (DW), 5.25% NaOCl, 5.25% NaOCl+17% EDTA, DW+MTAD, and 1.3% NaOCl+MTAD ISOLs, respectively. Fig. 2 shows the relative changes of P and Ca concentration means with the standard errors (SE) and the standard deviations (SD) in all studied treatment groups. The results shows significant negative changes (p<0.05) in P and Ca each in all treatment groups compared to the control. Thus, in the instance where the NaOCl ISOL in G2 treatment group was used, the significant RC(P) and RC(Ca) (p<0.05) with values of -0.127 and -0.052, respectively, were obtained; indicating significant losses in P and Ca concentrations, which could occur as a result of proteins losses from the organic components found in dentin canal system [4,26]. While, in the instance of MTAD ISOL use in G4 treatment group, the RC(P) and RC(Ca) values were of -0.153 and -0.106, respectively; indicating higher losses of P and Ca compared to those obtained in G2 group. This result confirmed the higher efficacy of MTAD ISOL in comparison to that of NaOCl. In the instance of G3 irrigation treatment group, the use of the combined ISOLs of 5.25% NaOCl+17% EDTA revealed RC(P) value of - 0.115, confirming the negligible effect of 17% EDTA ISOL on P distribution when it was used after NaOCl. This could be explained by the fact that, dentin has a dense collagen network covered by hydroxyapatite; thus, when NaOCl used before EDTA, the hydroxyapatite coating appeared to protect Fig. 2. Relative changes (RC) in P (a) and Ca (b) concentration means in root canal dentin tissues caused by different ISOLs. G1 is the control group; G2, G3, G4, and G5 are the treatment groups; SE and SD are the standard error and standard deviation, respectively. 3.4.1 Major elements On the other hand, Zn concentrations in the control group were positively skewed with mean value of 273 (215- (Ca/P)G1=1.5145<(Ca/P)G3=1.5849<(Ca/P)G2=1.5958<(Ca/ Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br 97 Orbital: Electron. J. Chem. 2022, 14(2), 95-102 Orbital: Electron. J. Chem. 2022, 14(2), 95-102 Orbital: Electron. J. Chem. 2022, 14(2), 95-102 P)G4=1.6100<(Ca/P)G5=1.6254 Where RC(P) and RC(Ca) are the relative changes of P and Ca concentration means, respectively; Gi is the ith group; PG1 and CaG1 are the concentration means of P and Ca in control group (G1), respectively. Where RC(P) and RC(Ca) are the relative changes of P and Ca concentration means, respectively; Gi is the ith group; PG1 and CaG1 are the concentration means of P and Ca in control group (G1), respectively. Table 2. Means, medians and ranges of concentrations of elements in powders of root canals of radicular dentin specimens, following treatment with different test solutions Groupa Descriptive statistics P (%) Ca (%) Fe (µg/g) Zn (µg/g) Sr (µg/g) G1 Meanb 17.3±0.9 26.2±0.8 69.0±39.0 273±71 102±23 Median 17.5 26.3 59.0 249 96.8 Range 16.0-18.6 25.0-27.1 28-132 215-434 74.7-154 G2 Mean 15.1±0.1 24.8±0.8 44.3±16.3 370±172 114±15 Median 15.3 24.8 44.5 306 112 Range 13.0-17.1 23.8-26.3 24.0-77.0 222-714 92.1-138 G3 Mean 15.3±1.1 24.1±1.1 53.4±20.1 369±147 145±31 Median 15.5 24.0 49.0 345 139 Range 13.6-17.2 22.4-26.2 30.0-81.0 163-704 109-207 G4 Mean 14.6±1.4 23.4±1.1 45.3±10.3 312±58 120±25 Median 14.5 23.4 45.0 311 128 Range 12.3-17.3 21.8-25.3 24.0-62.0 207-393 69.4-150 G5 Mean 14.1±1.2 22.8±0.3 57.0±29.9 341±97 144±42 Median 14.3 23.1 43.0 309 134 Range 11.9-15.6 22.4-23.4 34.0-108 239-526 98.7-225 a G1, G2, G3, G4, and G5 are the treatment groups of distilled water (DW), 5.25% NaOCl, 5.25% NaOCl+17% EDTA, distilled water+MTAD, and 1.3% NaOCl+MTAD, respectively. Number of specimens in each group, n=10. b the concentration presented as mean ± standard deviation. Table 2. Means, medians and ranges of concentrations of elements in powders of root canals of radicular dentin spe treatment with different test solutions medians and ranges of concentrations of elements in powders of root canals of radicular dentin specimens, following erent test solutions the collagen fibers, avoiding the direct NaOCl action [27]. Finally, the use of the combined ISOLs of 1.3% NaOCl+MTAD (G5 irrigation treatment group) revealed the greatest negative values of RC(P) (-0.182) and RC(Ca) (-0.114) compared to those of other instances in this work. Fig. 1. Relationships of Ca and Ca/P versus P concentration. means. Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br 3.4.2 Minor elements The significant increase in Zn/P ratio was obtained only for G5 treatment group with a value of (24.3±7.2).10-4. Fig. 3b shows insignificant positive changes in concentration means of Zn in all treatment groups compared to the control. This could be attributed to Zn2+ ion which has the ability to substitute Ca2+ in the apatite lattice [30]. The changes in Zn means were in the following order: RC(Zn)G2≈RC(Zn)G3>RC(Zn)G5>RC(Zn)G4>RC(Zn)G1, ( ) ( ) ( ) ( ) ( ) indicating the higher substitution of Zn2+ for Ca2+ in treatment groups with NaOCl ISOL. In addition, the ratio of Zn/Ca was calculated. This was with a value of (10.4±2.5).10-4 for the control group. Insignificant increases in Zn/Ca ratios were obtained for all treatment groups, except G3, which was estimated with a value of (15.4±6.1).10-4. The ratio of Zn/P was also calculated. This was with a value of (15.8±4.1).10-4 for the control group. The significant increase in Zn/P ratio was obtained only for G5 treatment group with a value of (24.3±7.2).10-4. It is well known that, Sr is one of the component elements of teeth and the Sr2+ ion would strongly replace Ca2+ in the apatite lattice [3,31]. The results in Fig. 3c showed positive changes in Sr concentration means in all irrigation treatment groups, confirming the specific substitution ability of Sr2+ ion for Ca2+ in the apatite lattice. The significant changes in Sr means were obtained by using the ISOLs in G3 and G5 treatment groups. Overall, the changes in Sr means were in the following order: RC(Sr)G3≈RC(Sr)G5>RC(Sr)G4>RC(Sr)G2>RC(Sr)G1, indicating the higher effect of the combined ISOLs on Sr compared to the effect of the single ISOLs. Furthermore, the ratio of Sr/Ca was calculated with a value of (4.95±3.40).10-4 for the control group, accepting those obtained in literature [3]. The previous ratio was insignificantly changed in the irrigation treatment groups. The ratio of Sr/P was also calculated for the control group. This was with a value of (7.42±4.83).10-4. The significant differences were obtained for Sr/P ratios of G3, (11.0±4.7).10-4, and G5, (10.3±3.3).10-4, compared to that ratio of the control group. 3.4.2 Minor elements Fig. 3a shows insignificant negative changes in concentration means of Fe in all treatment groups compared to that of the control. The relative changes in Fe means were in the following order: Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br 98 Orbital: Electron. J. Chem. 2022, 14(2), 95-102 0.75), Fe/Ca-Zn/Fe (r = -0.87), Sr/Ca-Zn/Sr (r = - 0.73), Fe/P- Zn/Fe (r = -0.87), and Sr/P-Zn/Sr (r = -0.73); further indications showed two strong positive correlations: Zn/Ca-Zn/Sr (r = 0.67) and Zn/P-Zn/Sr (r = 0.68). RC(Fe)G2>RC(Fe)G4>RC(Fe)G3>RC(Fe)G5>RC(Fe)G1. RC(Fe)G2>RC(Fe)G4>RC(Fe)G3>RC(Fe)G5>RC(Fe)G1. ( ) ( ) ( ) ( ) ( ) According to the previous results, the single ISOLs of NaOCl and MTAD have more efficiencies on removing Fe from the dentin tissues than the combined ISOLs of NaOCl+EDTA and NaOCl+MTAD. Thus, the efficacies of EDTA as a chelating agent to sequester metal ions such as Fe3+ [28] or MTAD as a demineralizing agent [29] were reduced when they were used after the ISOL of NaOCl. Finally, the results indicated that the ratios of Fe/Ca and Fe/P for the treatment groups (G2-G5) were insignificantly changed compared to those of the control group: (2.63±1.44).10-4 and (3.96±2.10).104, respectively. 67) and Zn/P Zn/Sr (r 0.68). Fig. 3. Relative changes (RC) in Fe (a), Zn (b), and Sr (c) concentrations means in root canal dentin tissues caused by different ISOLs. G1 is the control group; G2, G3, G4, and G5 are the treatment groups; SE and SD are the standard error and standard deviation, respectively. In G2 treatment group, seven very strong positive orrelations were obtained: Fe/Ca-Sr/Ca (r = 0.78), Fe/Ca g p ( ) ( ) , p y Fig. 3b shows insignificant positive changes in concentration means of Zn in all treatment groups compared to the control. This could be attributed to Zn2+ ion which has the ability to substitute Ca2+ in the apatite lattice [30]. The changes in Zn means were in the following order: RC(Zn)G2≈RC(Zn)G3>RC(Zn)G5>RC(Zn)G4>RC(Zn)G1, indicating the higher substitution of Zn2+ for Ca2+ in treatment groups with NaOCl ISOL. In addition, the ratio of Zn/Ca was calculated. This was with a value of (10.4±2.5).10-4 for the control group. Insignificant increases in Zn/Ca ratios were obtained for all treatment groups, except G3, which was estimated with a value of (15.4±6.1).10-4. The ratio of Zn/P was also calculated. This was with a value of (15.8±4.1).10-4 for the control group. 3.4.3 Correlation study The data on element-to-element correlations in the control and the treatment groups were studied. The results showed only one positive strong correlation in the control group: P-Ca (r = 0.65); while, there were two correlations in G2 treatment group: one strong positive correlation (r = 0.69) between P-Zn and one very strong negative correlation (r = -0.76) between Ca-Fe. Whereas, there were no strong correlations between the studied elements in the other treatment groups, i.e. G3, G4, and G5. Fig. 3. Relative changes (RC) in Fe (a), Zn (b), and Sr (c) concentrations means in root canal dentin tissues caused by different ISOLs. G1 is the control group; G2, G3, G4, and G5 are the treatment groups; SE and SD are the standard error and standard deviation, respectively. In G2 treatment group, seven very strong positive correlations were obtained: Fe/Ca-Sr/Ca (r = 0.78), Fe/Ca- Fe/P (r = 0.93), Zn/Ca-Zn/P (r = 0.95), Zn/Ca-Zn/Sr (r = 0.91), Zn/P-Zn/Fe (r = 0.70), Zn/P-Zn/Sr (r = 0.97), and Zn/Fe-Zn/Sr (r = 0.72); while, there were four very strong negative correlations: Ca/P-Fe/Ca (r = -0.77), Ca/P-Zn/Ca (r = -0.77), Ca/P-Sr/Ca (r = -0.78), and Fe/P-Sr/Fe (r = -0.91). In G2 treatment group there was only one strong positive correlation: Zn/Fe-Sr/Fe (r = 0.68). The data on ratio-to-ratio correlations in all treatment groups were also studied. The results in Table 3 showed surprisingly many strong and very strong correlations in the studied groups. There were four very strong positive correlations and a similar number of the very strong negative correlations in the control group (G1): Fe/Ca-Fe/P (r = 1.00), Zn/Ca-Zn/P (r = 0.99), Sr/Ca-Sr/P (r = 1.00), Zn/Fe-Sr/Fe (r = 99 ublished by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br Orbital: Electron. J. Chem. 2022, 14(2), 95-102 Orbital: Electron. J. Chem. 2022, 14(2), 95 102 Table 3. Pearson correlation coefficient matrix of element ratios in treatment and control groups. 3.4.3 Correlation study Groupa Ratio Ca/P Fe/Ca Zn/Ca Sr/Ca Fe/P Zn/P Sr/P Zn/Fe Sr/Fe Zn/Sr G1 Ca/P 1.00 Fe/Ca -0.35 1.00 Zn/Ca 0.31 -0.03 1.00 Sr/Ca -0.40 0.50 -0.09 1.00 Fe/P -0.30 1.00 -0.02 0.45 1.00 Zn/P 0.43 -0.08 0.99 -0.14 -0.07 1.00 Sr/P -0.35 0.48 -0.07 1.00 0.43 -0.12 1.00 Zn/Fe 0.49 -0.87 0.29 -0.25 -0.87 0.35 -0.21 1.00 Sr/Fe 0.19 -0.53 -0.05 0.35 -0.56 -0.02 0.39 0.75 1.00 Zn/Sr 0.34 -0.18 0.67 -0.73 -0.13 0.68 -0.73 0.11 -0.54 1.00 G2 Ca/P 1.00 Fe/Ca -0.77 1.00 Zn/Ca -0.77 0.59 1.00 Sr/Ca -0.78 0.78 0.68 1.00 Fe/P -0.51 0.93 0.31 0.55 1.00 Zn/P -0.58 0.35 0.95 0.53 0.08 1.00 Sr/P 0.10 0.21 -0.02 0.51 0.13 0.03 1.00 Zn/Fe -0.06 -0.41 0.47 -0.01 -0.63 0.70 -0.03 1.00 Sr/Fe 0.32 -0.77 -0.12 -0.23 -0.91 0.07 0.22 0.68 1.00 Zn/Sr -0.57 0.28 0.91 0.36 0.03 0.97 -0.23 0.72 0.05 1.00 G3 Ca/P 1.00 Fe/Ca -0.18 1.00 Zn/Ca -0.21 0.82 1.00 Sr/Ca 0.85 -0.36 -0.46 1.00 Fe/P -0.17 1.00 0.82 -0.35 1.00 Zn/P -0.02 0.77 0.98 -0.33 0.78 1.00 Sr/P 0.90 -0.31 -0.45 0.99 -0.30 -0.31 1.00 Zn/Fe 0.35 -0.63 -0.25 0.26 -0.63 -0.14 0.24 1.00 Sr/Fe 0.79 -0.49 -0.63 0.90 -0.48 -0.51 0.92 0.44 1.00 Zn/Sr -0.40 0.92 0.93 -0.63 0.91 0.86 -0.60 -0.46 -0.73 1.00 G4 Ca/P 1.00 Fe/Ca 0.26 1.00 Zn/Ca 0.14 0.12 1.00 Sr/Ca 0.08 -0.31 0.02 1.00 Fe/P 0.57 0.94 0.16 -0.22 1.00 Zn/P 0.67 0.30 0.83 0.07 0.53 1.00 Sr/P 0.57 -0.12 0.09 0.86 0.12 0.44 1.00 Zn/Fe -0.35 -0.85 0.38 0.40 -0.76 0.09 0.10 1.00 Sr/Fe -0.29 -0.84 -0.06 0.70 -0.78 -0.23 0.42 0.84 1.00 Zn/Sr -0.08 0.28 0.49 -0.85 0.15 0.32 -0.70 -0.08 -0.58 1.00 G5 Ca/P 1.00 Fe/Ca -0.35 1.00 Zn/Ca 0.11 -0.11 1.00 Sr/Ca -0.05 -0.19 -0.26 1.00 Fe/P -0.30 1.00 -0.11 -0.15 1.00 Zn/P 0.39 -0.19 0.95 -0.20 -0.16 1.00 Sr/P 0.20 -0.22 -0.23 0.95 -0.17 -0.08 1.00 Zn/Fe 0.14 -0.67 0.46 -0.15 -0.70 0.43 -0.13 1.00 Sr/Fe 0.18 -0.71 0.03 0.24 -0.74 0.05 0.24 0.80 1.00 Zn/Sr 0.19 0.05 0.80 -0.76 0.04 0.77 -0.69 0.37 -0.15 1.00 a G1, G2, G3, G4, and G5 are the treatment groups of distilled water (DW), 5.25% NaOCl, 5.25% NaOCl+17% EDTA, distilled water+MTAD, and 1.3% NaOCl+MTAD, respectively. Number of specimens in each group, n=10. Higher numbers of the very strong positive and negative correlations were obtained in G3 treatment group. 3.4.3 Correlation study Thus, there were sixteen very strong positive and one very strong negative correlations: Ca/P-Sr/Ca (r = 0.85), Ca/P-Sr/P (r = 0.90), Ca/P- Sr/Fe (r = 0.79), Fe/Ca-Zn/Ca (r = 0.82), Fe/Ca-Fe/P (r = 1.00), Fe/Ca-Zn/P (r = 0.77), Fe/Ca-Zn/Sr (r = 0.92), Zn/Ca-Fe/P (r = 0.82), Zn/Ca-Zn/P (r = 0.98), Zn/Ca-Zn/Sr (r = 0.93), Sr/Ca-Sr/P r = 0.99), Sr/Ca-Sr/Fe (r = 0.90), Fe/P-Zn/P (r = 0.78), Fe/P- Zn/Sr (r = 0.91), Zn/P-Zn/Sr (r = 0.86), Sr/P-Sr/Fe (r = 0.92), and six very strong negative correlations: Fe/Ca-Fe/P (r = 0.94), Zn/Ca-Zn/P (r = 0.83), Sr/Ca-Sr/P (r = 0.86), Sr/Ca-Sr/Fe (r = 0.70), Zn/Fe-Sr/Fe (r = 0.84), Fe/Ca-Zn/Fe (r = -0.85), Fe/Ca-Sr/Fe (r = -0.84), Sr/Ca-Zn/Sr (r = -0.85), Fe/P-Zn/Fe (r = -0.76), Fe/P-Sr/Fe (r = -0.78), and Sr/P-Zn/Sr (r=-0.70). Strong positive and strong negative correlations were not obtained in G4 treatment group. In G5 treatment group, there were six very strong positive and four very strong negative correlations: Fe/Ca Fe/P (r = and six very strong negative correlations: Fe/Ca-Fe/P (r = 0.94), Zn/Ca-Zn/P (r = 0.83), Sr/Ca-Sr/P (r = 0.86), Sr/Ca-Sr/Fe (r = 0.70), Zn/Fe-Sr/Fe (r = 0.84), Fe/Ca-Zn/Fe (r = -0.85), Fe/Ca-Sr/Fe (r = -0.84), Sr/Ca-Zn/Sr (r = -0.85), Fe/P-Zn/Fe (r = -0.76), Fe/P-Sr/Fe (r = -0.78), and Sr/P-Zn/Sr (r=-0.70). Strong positive and strong negative correlations were not obtained in G4 treatment group. and six very strong negative correlations: Fe/Ca-Fe/P (r = 0.94), Zn/Ca-Zn/P (r = 0.83), Sr/Ca-Sr/P (r = 0.86), Sr/Ca-Sr/Fe (r = 0.70), Zn/Fe-Sr/Fe (r = 0.84), Fe/Ca-Zn/Fe (r = -0.85), Fe/Ca-Sr/Fe (r = -0.84), Sr/Ca-Zn/Sr (r = -0.85), Fe/P-Zn/Fe (r = -0.76), Fe/P-Sr/Fe (r = -0.78), and Sr/P-Zn/Sr (r=-0.70). Strong positive and strong negative correlations were not obtained in G4 treatment group. Higher numbers of the very strong positive and negative correlations were obtained in G3 treatment group. Thus, there were sixteen very strong positive and one very strong negative correlations: Ca/P-Sr/Ca (r = 0.85), Ca/P-Sr/P (r = 0.90), Ca/P- Sr/Fe (r = 0.79), Fe/Ca-Zn/Ca (r = 0.82), Fe/Ca-Fe/P (r = 1.00), Fe/Ca-Zn/P (r = 0.77), Fe/Ca-Zn/Sr (r = 0.92), Zn/Ca-Fe/P (r = 0.82), Zn/Ca-Zn/P (r = 0.98), Zn/Ca-Zn/Sr (r = 0.93), Sr/Ca-Sr/P (r = 0.99), Sr/Ca-Sr/Fe (r = 0.90), Fe/P-Zn/P (r = 0.78), Fe/P- Zn/Sr (r = 0.91), Zn/P-Zn/Sr (r = 0.86), Sr/P-Sr/Fe (r = 0.92), and Sr/Fe-Zn/Sr (r = -0.73). 4. Conclusions [8] [8] Wang, Y.-L.; Chang, H.-H.; Chiang, Y.-C.; Lin, C.-H.; Lin, C.-P. J. Formos. Med. Assoc. 2019, 118, 39. [Crossref] In the present work, XRF technique was well suited for the determination of major elements (P and Ca) and minor elements (Fe, Zn, and Sr) in powders of root canal dentins with errors less than 8.3%. The previously mentioned elements, except Zn in G2 and G3 treatment groups, were normally distributed. [9] Graziani, G.; Boi, M.; Bianchi, M. Coatings 2018, 8, 269. [Crossref] [10] Dorozhkin, S. V. Materials 2009, 2, 1975. [Crossref] [11] Adhani, R.; Widodo; Sukmana, B. I.; Suhartono, E. Int. J. Chem. Eng. Appl. 2015, 6, 138. [Crossref] The XRF results showed greater decreases in P means of the treatment groups than those of Ca. The ratios of Ca/P were decreased as follows: Ca/PG5(1.3% NaOCl+MTAD)>Ca/PG4(MTAD)>Ca/PG2(5.25% NaOCl)>Ca/PG3(5.25% NaOCl+17% EDTA)>Ca/PG1(Distilled water). Accordingly, the highest direct action on the root canal dentin was attributed to the combined ISOL of 1.3% NaOCl+MTAD. On the other hand, the ISOLs used in this work showed different effects on the distribution of the minor elements (Fe, Zn, and Sr) in root canal dentin. Accordingly, higher removal of Fe from dentin tissues was investigated by using the single ISOLs of NaOCl or MTAD, while Zn and Sr had, apparently, the ability to replace Ca2+ in the apatite lattice regardless of the type of the used ISOLs. [12] Costa, E. M.; Azevedo, J. A. P. d.; Martins, R. F. M.; Rodrigues, V. P.; Alves, C. M. C.; Ribeiro, C. C. C.; Thomaz, E. B. A. F. Rev. Bras. Ginecol. Obstet. 2017, 39, 094. [Crossref] [13] Rautray, T. R.; Das, S.; Rautray, A. C. Nucl. Instrum. Methods Phys. Res., Sect. B 2010, 268, 2371. [Crossref] [14] Falla-Sotelo, F. O.; Rizzutto, M. A.; Tabacniks, M. H.; Added, N.; Barbosa, M. D. L. Braz. J. Phys. 2005, 35, 761. [Crossref] [15] Christensen, A. M.; Smith, M. A.; Thomas, R. M. J. Forensic Sci. 2012, 57, 47. [Crossref] Study on ratio-to-ratio correlations showed significant mutual interactions between minor-major elements in all studied treatment groups. Moreover, several minor-to-minor element correlations were significantly affected by the type of the ISOLs. This indicated the re-distribution ability of the minor elements during the exposure of the root canal dentin tissues to the action of the ISOL regimes. [16] Uo, M.; Wada, T.; Sugiyama, T. Jpn. Dent. Sci. Rev. 2015, 51, 2. Author Contributions It could be concluded hereby that, the correlation between pairs of the studied elements (element-to-element) in canals of radicular dentin specimens was observed only for a limited number of instances, whereas the higher numbers of the relationships were observed by correlating the ratios of the studied elements by each other (ratio-to-ratio). MQ contributed to the research concept and design, collection and/or assembly of data, data analysis and interpretation and writing the article. AK contributed to the research concept and design, data analysis and interpretation, writing and critical revision of the article. All authors read, revised and approved the final manuscript. The results showed that several important ratio-to-ratio correlations in the control group such (Sr/Ca-Sr/P), (Zn/Fe- Sr/Fe), (Fe/Ca-Zn/Fe), (Sr/Ca-Zn/Sr), (Fe/P-Zn/Fe), (Sr/P- Zn/Sr), (Zn/Ca-Zn/Sr), and (Zn/P-Zn/Sr) were not found in the treatment groups: (G2), (G3), (G2 and G5), (G2 and G4), (G2 and G3), (G2 and G3), (G4), and (G4), respectively. 3.4.3 Correlation study Whereas, there were only two strong negative correlations in G3 treatment group: Fe/Ca- Zn/Fe (r = -0.63) and Sr/Ca-Zn/Sr (r = -0.63). In G5 treatment group, there were six very strong positive and four very strong negative correlations: Fe/Ca-Fe/P (r = 1.00), Zn/Ca-Zn/P (r = 0.95), Zn/Ca-Zn/Sr (r = 0.80), Sr/Ca-Sr/P (r = 0.95), Zn/P-Zn/Sr (r = 0.77), Zn/Fe-Sr/Fe (r = 0.88), Fe/Ca- Sr/Fe (r = -0.71), Sr/Ca-Zn/Sr (r = -0.76), Fe/P-Zn/Fe (r = -0.70), and Fe/P-Sr/Fe (r = -0.74). Only two strong negative In G4 treatment group, there were five very strong positive Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br 100 Orbital: Electron. J. Chem. 2022, 14(2), 95-102 correlations were obtained in G5 treatment group: Fe/Ca- Zn/Fe (r = -0.67) and Sr/P-Zn/Sr (r = -0.69). 4. Conclusions [Crossref] [17] Mohammadi, Z.; Abbott, P. V. Aust. Endod. J. 2009, 35, 131. [Crossref] [18] Nair, P. N. R.; Henry, S.; Cano, V.; Vera, J. Oral Surg. Oral Med. Oral Pathol. Oral Radiol. Endod. 2005, 99, 231. [Crossref] [19] [19] International Atomic Energy Agency Laboratories Seibersdorf, X. R. F. G. S. Quantitative X ray analysis system User's manual and guide to X ray fluorescence technique; International Atomic Energy Agency (IAEA), 2009; p 168. References and Notes [1] Hegde, R.; Bapna, K. Contemp. Clin. Dent. 2016, 7, 216. [Crossref] [2] Prabhakaran, P.; Mariswamy, A. J. Conservative Dent. 2018, 21, 401. [Crossref] The results showed very strong correlations for Fe/Ca- Fe/P and Zn/Ca-Zn/P each, regardless to the ISOL used in treatment process; this well agrees with those reported in literature, where minor elements such Fe and Zn have the ability to incorporate into the hydroxyapatite lattice of the dental tissues [32,33]. The results also showed that Sr could incorporate in the hydroxyapatite lattice, but NaOCl solution was likely to affect the distribution of Sr in the root canal dentin. Strong correlations between the studied minor elements were obtained in all instances, in exception of the combined ISOLs in G3 treatment group, reflecting the possible effect of EDTA solution on the minor elements distribution in root canal system. [3] Liu, H.-Y.; Chao, J.-H.; Chuang, C.-Y.; Chiu, H.-L.; Yang, C.-W.; Sun, Y.-C. Advances in Anthropology 2013, 03, 71. [Crossref] [4] [4] Nogueira, B. M. L.; Pereira, T. I. C.; Pedrinha, V. F.; Rodrigues, P. A. Iran. Endod. J. 2018, 13, 209. [Crossref] [5] Hülsmann, M.; Heckendorff, M.; Lennon, Á. Int. Endod. J. 2003, 36, 810. [Crossref] ] Lynch, R. J. M. Int. Dent. J. 2011, 61, 46. [Crossref] Zhang, F.; Li, Y.; Xun, Z.; Zhang, Q.; Liu, H.; Chen, F. Lett. Appl. Microbiol. 2017, 64, 424. [Crossref] Khuder, A.; Bakir, M. A.; Karjou, J.; Sawan, M. K. J. Radioanal. Nucl. Chem. 2007, 273, 435. [Crossref] Acknowledgments The second author thanks Prof. I. Othman, Director General of AECS, for his encouragements and keen interest in this work. [20] Khuder, A.; Bakir, M. A.; Karjou, J.; Sawan, M. K. J. Radioanal. Nucl. Chem. 2007, 273, 435. [Crossref] Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br 101 Orbital: Electron. J. Chem. 2022, 14(2), 95-102 [29] Wopenka, B.; Pasteris, J. D. Mater. Sci. Eng., C 2005, 25, 131. [Crossref] [21] Oprea, C.; Szalanski, P. J.; Gustova, M. V.; Oprea, I. A.; Buzguta, V. Vacuum 2009, 83, S166. [Crossref] [22] Gutiérrez-Salazar, M. d. P.; Reyes-Gasga, J. J. Mater. Res. 2003, 6, 367. [Crossref] [30] Saxena, V.; Hasan, A.; Pandey, L. M. Mater. Technol. 2018, 33, 79. [Crossref] [23] Gibson, R. S. Adv. Nutr. 2012, 3, 772. [Crossref [31] Sprio, S.; Campodoni, E.; Sandri, M.; Preti, L.; Keppler, T.; Müller, F. A.; Pugno, N. M.; Tampieri, A. Int. J. Mol. Sci. 2018, 19, 3604. [Crossref] [24] Gu, L.-s.; Huang, X.-q.; Griffin, B.; Bergeron, B. R.; Pashley, D. H.; Niu, L.-n.; Tay, F. R. Acta Biomater. 2017, 61, 144. [Crossref] [32] Tonbuloğlu, B.; Aslan, D.; Aydin, H. Eurasian J. Educ. Res. 2016, 1. [Crossref] [25] Zhang, K.; Tay, F. R.; Kim, Y. K.; Mitchell, J. K.; Kim, J. R.; Carrilho, M.; Pashley, D. H.; Ling, J.-q. Dent. Mater. 2010, 26, 514. [Crossref] [33] Uzun, I.; Güler, B.; Özyürek, T.; Tunç, T. Niger. J. Clin. Pract. 2016, 19, 71. [Crossref] [26] Mohammadi, Z.; Shalavi, S.; Jafarzadeh, H. Eur. J. Dent. 2013, 07, S135. [Crossref] How to cite this article How to cite this article Quteifan, M. Q.; Khuder, A. Orbital: Electron. J. Chem. 2022, 14, 95. DOI: http://dx.doi.org/10.17807/orbital.v14i2.16204 [27] De-Deus, G.; Reis, C.; Fidel, S.; Fidel, R.; Paciornik, S. J. Endod. 2007, 33, 1364. [Crossref] [28] Peld, M.; Tõnsuaadu, K.; Bender, V. Environ. Sci. Technol. 2004, 38, 5626. [Crossref] Published by Federal University of Mato Grosso do Sul \| www.orbital.ufms.br 102
https://openalex.org/W2032929264	https://europepmc.org/articles/pmc3567115?pdf=render	English	null	Testing Cost-Benefit Models of Parental Care Evolution Using Lizard Populations Differing in the Expression of Maternal Care	PloS one	2,013	cc-by	7,122	Introduction focal species which vary in the type or intensity of care provided (but in which all populations express care to some degree), or to examine interspecific variation in care strategies. Because parental care is either absent altogether or present in the vast majority of species, our capacity to understand the direct ecological influences of parental care evolution at the population level is limited. Parental care should evolve when the benefits of providing care increase offspring survival above the costs of reduced survival and future reproduction of adults [1–3]. Offspring benefit from parental care through an increased chance of survival, which is relatively easy to measure, but the costs for parents are complex, and thus much more difficult to detect [4]. Parental costs are twofold: (1) reproductive costs, including loss of mating opportu- nities when providing long-term parental care, which can decrease future fecundity or reduce the number of offspring produced during the next reproductive event due to lower energy intake; and (2) survival costs, in which long-term parental care can reduce survival of the parent. Evolutionary theory suggests that parents will adjust their parental care expenditure in relation to the variation in costs to themselves and the benefits to their offspring, so as to maximize fitness [5]. p p p Long-tailed skinks (Eutropis longicaudata) are widely distributed throughout southeast Asia, but only one insular population is known to display maternal care [9]; this presents a unique opportunity for testing how ecological circumstances can contrib- ute to parental care evolution. On Orchid Island (Taiwan), nesting female long-tailed skinks guard their nests during incubation, whereas females in at least 13 other populations abandon the nest immediately after laying eggs [9]. This pattern is driven by significantly higher levels of predation pressure on Orchid Island, necessitating maternal care to ensure reproductive success [9]. We developed cost-benefit models of parental care, which we tested using ecological data from the population which provides maternal care and two other populations lacking maternal care. We focus on the fitness consequences of maternal care, where mothers benefit by increasing egg hatching rates but incur a cost in terms of the survival of the nest-guarding female and reduced opportunity for future reproduction. This extends other models, in which Several mathematical models of parental care have been proposed, most of which involve cost-benefit tradeoffs between the parents and offspring [1–2,5–8]. Abstract * E-mail: wshuang@mail.nmns.edu.tw Testing Cost-Benefit Models of Parental Care Evolution Using Lizard Populations Differing in the Expression of Maternal Care Wen-San Huang1,2, David A. Pike3 1 Department of Zoology, National Museum of Natural Science, Taichung, Taiwan, 2 Department of Life Sciences, National Chung Hsing University, Taichung, Taiwan, 3 School of Marine and Tropical Biology, James Cook University, Townsville, Australia 1 Department of Zoology, National Museum of Natural Science, Taichung, Taiwan, 2 Department of Life Sciences, National Chung Hsing University, Taichung, Taiwan, 3 School of Marine and Tropical Biology, James Cook University, Townsville, Australia February 2013 \| Volume 8 \| Issue 2 \| e54065 Abstract Parents are expected to evolve tactics to care for eggs or offspring when providing such care increases fitness above the costs incurred by this behavior. Costs to the parent include the energetic demands of protecting offspring, delaying future fecundity, and increased risk of predation. We used cost-benefit models to test the ecological conditions favoring the evolution of parental care, using lizard populations that differ in whether or not they express maternal care. We found that predators play an important role in the evolution of maternal care because: (1) evolving maternal care is unlikely when care increases predation pressure on the parents; (2) maternal care cannot evolve under low levels of predation pressure on both parents and offspring; and (3) maternal care evolves only when parents are able to successfully defend offspring from predators without increasing predation risk to themselves. Our studies of one of the only known vertebrate species to exhibit interpopulation differences in the expression of maternal care provide clear support for some of the hypothesized circumstances under which maternal care should evolve (e.g., when nests are in exposed locations, parents are able to defend the eggs from predators, and egg incubation periods are brief), but do not support others (e.g., when nest-sites are scarce, life history strategies are ‘‘risky’’, reproductive frequency is low, and environmental conditions are harsh). We conclude that multiple pathways can lead to the evolution of parental care from a non-caring state, even in a single population of a widespread species. Citation: Huang W-S, Pike DA (2013) Testing Cost-Benefit Models of Parental Care Evolution Using Lizard Populations Differing in the Expression of Maternal Care. PLoS ONE 8(2): e54065. doi:10.1371/journal.pone.0054065 Editor: Ce´dric Sueur, Institut Pluridisciplinaire Hubert Curien, France Received September 21, 2012; Accepted December 5, 2012; Published February 7, 2013 Received September 21, 2012; Accepted December 5, 2012; Published February 7, 2013 Copyright: 2013 Huang, Pike. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Funding was provided by the Kuo Wu Hsiu Luan Culture and Education Foundation and the National Science Council of Taiwan (NSC 99-2621-B-178- 001-MY3). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. Parental Care Evolution in a Lizard parental care evolves solely in response to selection on trade-offs between growth rate and future reproduction, or the relationship between parental care and fecundity [2]. Eventually, a balance between the costs and benefits should lead to parental care evolving from a non-caring ancestor, and if the long-term benefits outweigh the costs, then parental care should be maintained. Our study suggests that natural selection acting directly on the female and her ability to produce offspring exerts a strong influence on the evolution of maternal care. We integrate our findings into existing predictive frameworks that propose mechanisms leading to the evolution of parental care [1], and find support for some hypotheses, but not others. Model overview We used an evolutionary stable strategy (ESS) game theory model [12]. The game theory model predicts that the selective pressures maintaining parental care by both sexes, and desertion by one parent, affects investment by the other parent (Table 2) [12]. We can use the game theory model to define four different evolutionary stable strategies for parental care evolution (ESS): (1) both parents express parental care, and if the costs outweigh the benefits, either parent can abandon care; (2) the female deserts the offspring while the male provides care; (3) the female provides care for the offspring while the male deserts; or (4) both sexes desert the nest resulting in no parental care. Because parental care by male reptiles is virtually nonexistent [1], here we focus on the conditions under which maternal care is likely to evolve (ESS1 and ESS3). The long-tailed skink populations on Green Island and mainland Taiwan are described by evolutionary stable strategy ESS 1: both parents desert the nest, whereas the Orchid Island population is described by ESS 3: the female cares for incubating eggs and the male deserts (Table 2). Because the payoff from these two strategies relates to egg numbers and egg survival, we can calculate Ethics statement This study was approved by the Taiwanese National Museum of Natural Science Animal Care and Use Committee (Protocol Permit NMNSHP02-002). Introduction However, one of the principle constraints with predicting the circumstances under which parental care evolves is that the ecological form of the cost and benefit functions is unknown, making models difficult to test empirically [2]. Consequently, models are normally tested using February 2013 \| Volume 8 \| Issue 2 \| e54065 February 2013 \| Volume 8 \| Issue 2 \| e54065 1 PLOS ONE \| www.plosone.org Parental Care Evolution in a Lizard Study populations and empirical data y p p p From 2001–2010 we collected clutch size data for Eutropis longicaudata on Orchid Island, located 60 km southeast of Taiwan (22u029N, 121u349E); Green Island, located 33 km southeast of Taiwan (22u409N, 121u289E); and mainland Taiwan (Santimen, Pingtung County; 22u429N, 120u389E). At each location long- tailed skinks nest within drainage holes running through a concrete retaining wall located along a mountain road. For full details of study sites and methods, see [9–11]. Incubating eggs fail due to a variety of reasons, including fungal infections, or predation by ants or egg-eating snakes (Oligodon formosanus) [9]. Egg-eating snakes, however, are the major egg predator and the maternal care expressed by Orchid Island long-tailed skinks is aimed at deterring snakes from eating lizard eggs, which are more abundant on this island than our other two study sites [9]. We conducted a manipulative experiment to determine the fitness effects of excluding snake predators (as mother lizards do on Orchid Island) [9]. This entailed gluing mesh over the pipes housing lizard nests; by doing so, we could estimate the fitness effects that maternal care would provide in our two study populations which do not display maternal care (Table 1) [9]. We also conducted these trials on Orchid Island to serve as controls (Table 1) [9]. Table 2. Payoff matrix and four evolutionary stable strategies (ESS) from the game theory model [12]. Female Care Desert Male Care R wP2 WP1 = wP2 WP1 Desert R wP1 WP0 = wP1 (1+p) WP0 (1+p) W: egg without care; w: egg with care; P0: egg survival without parental care; P1: egg survival with uni-parental care; P2: egg survival with bi-parental care; P2. P1. P0; p: male deserts and then obtains another mating opportunity. ESS1: both sexes show care; wP2. WP1, otherwise female deserts; wP2. wP1 (1+p), otherwise male deserts. ESS2: female deserts and male cares; WP1 . wP2, otherwise female cares; WP1 . WP0 (1+P), otherwise male deserts. ESS3: female cares and male deserts; wP1 . WP0, otherwise female deserts; wP1 (1+p) . wP2 otherwise male cares. ESS4: both sexes desert; WP0 . wP1, otherwise the female cares; WP0 (1+p) . WP1, otherwise the male cares. doi:10.1371/journal.pone.0054065.t002 Table 2. Payoff matrix and four evolutionary stable strategies (ESS) from the game theory model [12]. Table 2. Payoff matrix and four evolutionary stable strategies (ESS) from the game theory model [12]. Study populations and empirical data W: egg without care; w: egg with care; P0: egg survival without parental care; P1: egg survival with uni-parental care; P2: egg survival with bi-parental care; P2. P1. P0; p: male deserts and then obtains another mating opportunity. ESS1: both sexes show care; wP2. WP1, otherwise female deserts; wP2. wP1 (1+p), otherwise male deserts. ESS2: female deserts and male cares; WP1 . wP2, otherwise female cares; WP1 . WP0 (1+P), otherwise male deserts. ESS3: female cares and male deserts; wP1 . WP0, otherwise female deserts; wP1 (1+p) . wP2 otherwise male cares. ESS4: both sexes desert; WP0 . wP1, otherwise the female cares; WP0 (1+p) . WP1, otherwise the male cares. doi:10.1371/journal.pone.0054065.t002 Table 1. Predation, competition, and parental care attributes of three long-tailed skink populations. Variable Location Mainland Taiwan Green Island Orchid Island Individual predation Strong Weak Weak Egg predation Weak Weak Strong Parental care No No Yes Clutch size (eggs) 6.7 6.0 6.5 Hatching success with parental care 59% 55% 81% Hatching success without parental care 62% 51% 18% Relative change in hatching success with care 23% 4% 63% Cost-benefit model without maternal care 1.15 0.06 21.83 Cost-benefit model evolving maternal care 0.95 0.30 2.27 Individual predation pressure was assessed using the number of lizard predator species present in each location; intensity of egg predation was estimated using numbers of egg-eating snakes (Oligodon formosanus) [33] encountered during our study; and intra-specific competition was estimated using skink density. Data for hatching success with and without parental care are from [9]. doi:10.1371/journal.pone.0054065.t001 Table 1. Predation, competition, and parental care attributes of three long-tailed skink populations. Individual predation pressure was assessed using the number of lizard predator species present in each location; intensity of egg predation was estimated using numbers of egg-eating snakes (Oligodon formosanus) [33] encountered during our study; and intra-specific competition was estimated using skink density. Data for hatching success with and without parental care are from [9]. doi:10.1371/journal.pone.0054065.t001 February 2013 \| Volume 8 \| Issue 2 \| e54065 PLOS ONE \| www.plosone.org 2 Parental Care Evolution in a Lizard the benefits in terms of numbers of eggs produced and the proportion of those eggs that hatch into baby lizards. study system [13]. Study populations and empirical data Consequently, Cdelay = 23 (in Equation 1, we assumed that a clutch represents 3 points, and the greatest benefit to a female would be a clutch that is not preyed upon by snakes, and thus all eggs hatch. In this case X = 0 and 23+6 (1–0) = 3. If the entire clutch is consumed by snakes, then X = 1 and 23+6 * (1–1) = 23; a female loses at most 3 points by delaying the second clutch). Cpredation and Cantipredator are dependent on the female’s ability to escape lizard predators and to deter egg predators, respectively. These costs sum to at most 7 points because females invest 3 in the clutch. In our study system, female lizards easily deter egg predators (the reptile egg-eating snake Oligodon formosanus, which does not eat lizards; [9,10]), so we assume Cantipredator to be negligible (i.e., 0). Testing the model without parental care We solved the model without parental care (Equation 1) using empirical data (Table 1). The benefits of not providing care outweigh the costs for the mainland and Green Island populations, but not for the Orchid Island population (Table 1). For the Orchid Island population to persist in the absence of maternal care (i.e., to obtain a positive value for the equation) would require either a decrease in Clay or an increase Begg. To obtain a positive value for the equation, X would have to be above 0.46, or, given the current hatching success rate of eggs in the absence of maternal care (18%), females would have to increase their clutch size to .17 eggs (the known maximum clutch size is 13 eggs; [14]). With lower clutch sizes and given the high rate of egg predation in the absence of maternal care, long-tailed skinks would become extirpated from Orchid Island if the eggs were not protected from predation during incubation. Benefits of parental care The benefit of parental care is the successful production of offspring. In long-tailed skinks, the proportion of eggs laid that produce hatchlings is a function of incubation temperature [11] and the probability that the eggs will be located and consumed by predators [10]. Eggs exposed to relatively moderate temperatures have high hatching success ratios [11], which increases the benefit of providing parental care. Similarly, parental care maintains a benefit when females successfully attack and deter egg-eating snake predators [9,10]. Model development The model without maternal care (which is representative of the long-tailed skink populations on mainland Taiwan and Green Island) is: ð1Þ Clay z Begg (1-X) ~ Constant ð1Þ Testing the model evolving maternal care Where Clay is the cost of laying a clutch of eggs by a mother lizard, Begg is the benefit of egg-hatching ratios, and X is the proportion of eggs that fail to hatch. We solved the model evolving maternal care (Equation 2) using empirical data (Table 1). Our experimental exclusion of egg predators, which simulates maternal defense of the nest site, revealed that when eggs are protected from predation, Begg increases from 18% to 81% on Orchid Island, but in the other two populations there is negligible change (,4% change; Table 1 [9]). The benefits of providing care thus outweigh the costs only for the Orchid Island population. The absolute value for the Green Island population is higher for the model evolving maternal care than for the model not evolving maternal care (Table 1), but the difference is so small that when using the empirically-derived hatching success data as input (instead of data from the predator exclusion experiment; Table 1), the values for the models with and without maternal care are equal (i.e., 0.06). Lizards on mainland Taiwan and Green Island should therefore continue burying their eggs in the soil and abandoning them after oviposition (as is common throughout most of the species’ range). We assumed that a mother lizard represents ten points, and that laying a clutch of eggs costs the female 30% of her body mass [10], or three points. If all of the eggs in a clutch hatch, the female benefits ten points; this value can be adjusted based on egg hatching rates. We assumed that a female lays eggs twice during her lifetime, and at least one hatchling from each clutch reaches adulthood. This means that in her lifetime each female will replace herself and one mate (20 lifetime points), thereby maintaining a stable population size. These assumptions can be adjusted based on life history attributes; for example, if a female lays four clutches of eggs in her lifetime, then each clutch would represent five points. The model evolving parental care (which is representative of the long-tailed skink population on Orchid Island) is: Clay z Cdelay z Cpredation z Cantipredator z Begg (1-X) ~ Constant ð2Þ Costs of maternal care The costs of maternal care are fourfold: (1) the energetic cost of producing a clutch; (2) delay of future mating opportunities (when a female remains with the eggs, she delays opportunities for mating and acquiring the energy necessary for producing her next clutch); (3) anti-predator ability (the ability of a lizard to defend the nest from obligate egg predators); and (4) risk of predation (maternal care could reduce survival through increased predator exposure during care). When considering these costs as point values in a model, the total costs cannot exceed the value of the female herself. Discussion Clay z Cdelay z Cpredation z Cantipredator z Begg (1-X) ~ Constant ð2Þ ð2Þ Constant Why the mainland Taiwan population has not evolved maternal care There are two explanations for why the long-tailed skink population on mainland Taiwan has not evolved maternal care. First, on the mainland there is a substantial risk of predation to the parent while guarding the nest because of the diverse array of lizard predators there (e.g., at least 4 snake species; Table 1 [9]). The presence of a higher diversity of lizard predators could increase Crisk. If the mother lizard is preyed upon, then the X value will increase simultaneously because the eggs are vulnerable to predators without maternal care. Secondly, experimental exclusion of vertebrate predators in the mainland population revealed that the benefits of care and desertion are almost equivalent in terms of hatching success (Table 1 [9]). Excluding vertebrate predators did not result in a significant increase in hatching success on mainland Taiwan or Green Island, as compared to a 63% increase for the Orchid Island population (Table 1 [9]). When there is no positive net benefit to maternal care, natural selection will favor nest desertion. Predation is a strong selective force that influences the behavior, morphology and life history of prey species [17–18]. Studies of the effects of predation on prey responses are biased towards studies of life history and morphology. Nevertheless, exposure to predators has been shown to influence prey behavioral changes and evolution (e.g. [18–20]). The current cost-benefit model suggests that both individual predators and reptile egg predators are important to the evolution of maternal care in long-tailed skinks. Direct contact between prey and potential predators is unlikely to induce parental care in smaller species on the island, because they are unable to successfully deter egg-eating snakes. However, when predators are rare (or multiple predators co-occur) it may be difficult to evolve maternal care because the costs of providing care may not outweigh the benefits. For example, the Taiwanese population of skinks is exposed to many different predators, many of which eat lizards. Likewise, on Green Island there are very few predators at all, and consequently maternal care would accrue only minimal benefits (Table 1). Our cost-benefit model predicts that animals will seek further benefits if the environment allows it, regardless of whether the species shows aggression towards egg predators. For example, higher temperatures influence egg incubation time and egg mortality of long-tailed skinks on Orchid Island [11]. Figure 1. A hypothetical cost-benefit model describing the evolution of parental care in long-tailed skinks (Eutropis longicaudata), showing how parental care can evolve based on the relationship between incubation temperature (the benefit is increased hatching success, x-axis) and egg preda- tion pressure from snakes (the cost of laying eggs in populations with varying snake population size, y-axis). Note that these relationships are hypothetical, and provide only one example of how these two variables could influence the evolution of maternal care (similar relationships can be developed between other variables of interest). Parental care has a high likelihood of evolving in instances where female lizards select nest sites that are relatively high (i.e., inside of a retaining wall as compared to in natural habitats) but in which there are large numbers of egg-eating snakes (and thus high risk of any eggs left unattended becoming depredated). This matches the situation on Orchid Island, Taiwan. Maternal care is unlikely to evolve when predation pressure is relatively low, which matches the situation on Green Island and mainland Taiwan. doi:10 1371/journal pone 0054065 g001 the nest for the entire duration of incubation) and allow the hatchlings to have longer activity seasons (and thus reach larger sizes) prior to winter. Thus, these concrete walls are likely to attract female lizards because of the benefits to incubating eggs, regardless of whether that female has the ability to deter predators or even provide maternal care at all. The Taiwanese and Green Island populations are consistent with this prediction in that they have not evolved maternal care, but do lay their eggs inside the concrete walls. the nest for the entire duration of incubation) and allow the hatchlings to have longer activity seasons (and thus reach larger sizes) prior to winter. Thus, these concrete walls are likely to attract female lizards because of the benefits to incubating eggs, regardless of whether that female has the ability to deter predators or even provide maternal care at all. The Taiwanese and Green Island populations are consistent with this prediction in that they have not evolved maternal care, but do lay their eggs inside the concrete walls. Constant A major objective in behavioral and evolutionary ecology is to understand how animals make decisions in complex environments, and the strategies adopted by animals under realistic conditions [15]. Our cost-benefit model used an optimization framework [12,15,16] to examine the evolution of maternal care in complex environments, and revealed four major insights: (1) maternal care can evolve from a non-parental caring ancestor if the benefits outweigh the costs; (2) in species lacking anti-predator behaviors, the benefits of maternal care can increase with ecological or environmental factors; (3) the benefits and costs of maternal care Where Cdelay is the cost to a female in terms of delaying reproduction while guarding eggs, Cpredation is the risk of the female provisioning care being preyed upon, and Cantipredator is the cost the female incurs when using antipredator behaviors. If a female develops maternal care, then the costs relative to not providing care include the cost of future reproduction, the risk of predation, and the cost of deterring predators. Delaying future reproduction may be the most important cost for females in our February 2013 \| Volume 8 \| Issue 2 \| e54065 PLOS ONE \| www.plosone.org February 2013 \| Volume 8 \| Issue 2 \| e54065 3 Parental Care Evolution in a Lizard Figure 1. A hypothetical cost-benefit model describing the evolution of parental care in long-tailed skinks (Eutropis longicaudata), showing how parental care can evolve based on the relationship between incubation temperature (the benefit is increased hatching success, x-axis) and egg preda- tion pressure from snakes (the cost of laying eggs in populations with varying snake population size, y-axis). Note that these relationships are hypothetical, and provide only one example of how these two variables could influence the evolution of maternal care (similar relationships can be developed between other variables of interest). Parental care has a high likelihood of evolving in instances where female lizards select nest sites that are relatively high (i.e., inside of a retaining wall as compared to in natural habitats) but in which there are large numbers of egg-eating snakes (and thus high risk of any eggs left unattended becoming depredated). This matches the situation on Orchid Island, Taiwan. Maternal care is unlikely to evolve when predation pressure is relatively low, which matches the situation on Green Island and mainland Taiwan. Constant doi:10.1371/journal.pone.0054065.g001 increase when individuals actively defend their nests against predation; and (4) predators play important roles in the evolution of maternal care. The evolution of parental care represents a considerable evolutionary challenge: to change behavior in a manner that interacts positively with offspring. Nonetheless, phylogenetic studies from a wide range of taxa all suggest that this is the pathway by which parental care evolves (reviewed by [2]). Although the costs and benefits of parental care are many, we currently know little of the form of the relationships between parental investment and offspring fitness. It seems likely that this relationship could involve one or more threshold effects (e.g., see Fig. 1 for one putative relationship between selecting nest sites with appropriate incubation temperatures and predation pressure). However, there are many problems associated with measuring the costs of reproduction to the parents. Individual variation in phenotypic or genetic quality may obscure the effects of parental expenditure on subsequent fitness. Furthermore, experimental manipulation of parental expenditure is needed, even though it is difficult to measure reproductive costs and offspring benefits using either laboratory of field experiments [2]. Theoretical models of parental investment necessarily incorporate assumptions about the form of cost and benefit functions. Because it is impossible to deal with all of these functions adequately, few predictions concerning the form or extent of parental investment in practice have a basis grounded in ecology. For example, we do not currently know whether parents should generally invest more heavily in offspring of superior quality because they are more likely to survive, or inferior ones because parental expenditure will have a greater influence on their fitness. Because we know nothing about whether environmental factors affect the parents’ decisions, those uncer- tainties lead us to be skeptical of predictions based only on theoretical principles. This may be a problem in our model as well, but our goal was to use biologically meaningful parameters to understand how these might relate to parental care evolution. Characteristics of the external environment Our model predicts that parental care will evolve more commonly when reproductive output is low. This implies that short-lived species are more likely to evolve parental care than long-lived species. Short-lived species often have lower reproductive outputs than long-lived species, and thus could be more likely to pursue ‘‘risky’’ reproductive strategies than long-lived ones [1,22]. Parental care is a ‘‘risky’’ behavior because it can render the parents vulnerable to predators. In lizards, parental care is more common in late-maturing or large species [22,23]. Tinkle (1969) suggested that parental care generally should be found in long- lived iteroparous species, especially those with a short, annual breeding season. However, the long-tailed skink, an iteroparous species with a long breeding season and large body size relative to other sympatric lizards, does not support this prediction. This inconsistency might correlate with a lizard’s ability to defend against predators. For example, Sphenomorphus incognitus, a lizard with a small body size, cannot defend its nest from Oligodon formosanus [24] because its head width (8.1–10.9 mm; [25]) is smaller than the chest diameter of O. formosanus (11.5–14.5 mm; [26]), but the head width of Eutropis females (13.3–18.3 mm; [14]) is larger than the chest diameter of the snake; thus they can defend nests from intruding snakes. Parental Care Evolution in a Lizard Parental Care Evolution in a Lizard Although long-tailed skinks bury their eggs beneath rocks in natural habitats, the eggs of females nesting inside the retaining wall remain exposed during incubation. Females nesting in populations expressing or not expressing maternal care both nest in these retaining walls, suggesting that egg exposure in and of itself does not always lead to maternal care. Characteristics of the external environment 1. Harsh and unpredictable environments. This hypoth- esis suggests that parental care will most often be found in environments in which resources for adults are limited at the time of egg guarding [21]. For example, if resources are scarce prior to or during the period of parental care, then the parents might increase the rate of food searching while guarding the nest, which could increase the risk of predation to the parent or offspring, which would reduce the benefits of providing maternal care [21]. However, if food resources are abundant at the time of nesting, an alternative hypothesis is that parents would not necessarily tradeoff egg guarding with food searching. This would lessen the time spent away from the nest and provide the female with more energy with which to protect the nest. Parental care might strongly benefit egg survivorship in these instances. In our current study, for instance, long-tailed skinks breed from February to September when the food resources are abundant [14], in contrast to the harsh and unpredictable environment hypothesis, but in support of the abundant food resource hypothesis. Although skinks occasionally leave the nest to forage in the forest near the concrete wall [9], these excursions may be infrequent enough or of sufficiently short duration that foraging does not increase the risk of predation to the mother or eggs. Addressing this hypothesis using more detailed behavioral observations will help clarify this. 5. Ability of parents to defend eggs from predators. If a major benefit of parental care is to deter potential egg-predators, then parental care should evolve most often in species in which a parent is physically capable of deterring egg predators [1,3,30]. For example, parental care is common in the largest snakes (Pythoninae), and those with venom (Viperidae, .50% of species and Elapidae, 41% of species). Likewise, the few reports of parental care in the Colubridae (7.5%) are in species with relatively large body sizes (e.g., the genera Elaphe, Farancia, Ptyas) or belonging to the minority of venomous species within the family (e.g., Psammophylax, Rhabdophis; [1]). The ubiquity of parental care in the large and formidable crocodilians is consistent with this prediction. Likewise, due to their relatively large body size, long- tailed skinks appear to be the only lizard species on Orchid Island large enough to successfully deter egg-eating snakes [24]. 2. Selection for ‘‘risky’’ life history strategies. The cost and benefit model in a hypothesis-testing framework Existing hypotheses about the mechanisms behind the evolution of parental care [1] can be integrated into our findings, which results in eight hypotheses spanning: (1) attributes of the external environment, including nesting sites; and (2) reproductive attri- butes of the species, both of which must result in the parental action increasing fitness. We address each of these hypotheses in turn. 4. Suitability of habitat for clutch attendance. Parental care is predicted to be more common when adults occupy the same habitat in which eggs are laid. This is because the parents may be less susceptible to predation inside their own territory, which also provides vital resources. This hypothesis is difficult to test in lizards because the males are more likely to be territorial, although females are more likely to show parental care [1]. Long- tailed skinks do not support this prediction because females generally spend the period outside the nesting season in natural habitats, but move onto the retaining wall during the nesting season [10]. A related hypothesis concerns the abundance of nest- sites available. Parental care might evolve when nest sites are scarce relative to the number of nesting females, because the arrival of new females can destroy existing nests. This scenario has been described in sea turtles [28] and island populations of iguanine lizards [29]. However, at our study areas there are nearly 1,200 identical potential nest-sites available, but we have never observed more than 50 nesting females at any given time. Because the number of suitable nesting sites vastly outnumbers the number of breeding females, our data do not support this prediction. Why the mainland Taiwan population has not evolved maternal care Eggs incubated in the warmer concrete wall have higher hatching success (88%) than natural nests (67%), and hatch 7 days earlier [11]. In warmer environments, shorter incubation times could reduce the duration of maternal care (when the female remains at February 2013 \| Volume 8 \| Issue 2 \| e54065 PLOS ONE \| www.plosone.org 4 Reproductive attributes 6. Parental care is more likely to evolve in species with low reproductive frequency. Parental care may be more likely to evolve in species that produce only a single clutch of large eggs per reproductive season, as opposed to species that produce small and frequent clutches. This is because the cost of caring for a single clutch of eggs is most likely smaller than the cost of caring for several clutches over a much longer duration. Data from several studies seemingly support this prediction [23,31]. However, because multiple clutching is more common in tropical species than in temperate species, parental care should evolve more often in temperate rather than tropical species. This is not the case for reptiles, in which parental care is most common in tropical species (e.g., Iguaninae, Pythoninae, Alligatorinae, Crocodylinae [1]). The costs for long-tailed skinks to guard their eggs are minimal, and similar for females with both large and small clutches [13]. Some long-tailed skinks reproduce twice within a single breeding season, and guard both clutches, which does not support this prediction. 3. Exposed vs hidden nest sites. Tinkle (1969) reviewed parental care of lizards and suggested that lizards laying their eggs in well-hidden nests might be less exposed to dangers than a female that does not accompany her eggs. Likewise, eggs incubating in exposed areas, rather than buried or hidden, may be easier for predators to locate. Hence, parental care should evolve more often in species that do not bury their eggs [1,27]. 7. Brief incubation periods. Parental care should evolve when incubation periods are short because females investing less time engaged in parental care will incur fewer costs. A comparison of the incubation periods of egg-attending vs non-attending species February 2013 \| Volume 8 \| Issue 2 \| e54065 PLOS ONE \| www.plosone.org 5 Parental Care Evolution in a Lizard ectotherms than in ectotherms. This could help explain why almost all birds and mammals show parental care. supports this hypothesis [21]. However, such interspecific com- parisons could be confounded if large species that tend to have large eggs (which need longer incubation periods) are dispropor- tionately represented in the non-attending species. Our field data support hypothesis; long-tailed skink eggs incubate more quickly on Orchid Island than on the mainland [13]. References 1. Shine R (1988) Parental care in reptiles, Pages 275–330 in Gans C, Huey RB, eds. Biology of the Reptilia Vol.16. Alan R. Liss. New York. 17. Lima SL, Dill LM (1990) Behavioral decisions made under the risk of predation: a review and prospectus. Canad J Zool 68: 619–640. 18. Chivers DP, Kiesecker JM, Marco A, DeVito J, Anderson MT, et al. (2001) Predator-induced life history changes in amphibians: egg predation induces hatching. Oikos 92: 135–142. 2. Clutton-Brock TH (1991) The evolution of parental care: Monographs in behavior and ecology. Princeton University Press. 3. Greene HW, May PG, Hardy SDL, Sciturro JM, Farrell TM (2002) Parental behavior by vipers, Pages 179–205 in Schuett GW, Hoggren M, Douglas ME, Greene HW, eds. Biology of the Vipers. Eagle Mountain Publ. Eagle Mountain, UT. g 19. Crowl TA, Covich AP (1990) Predator-induced life history shifts in a freshwater snail. Science 247: 949–951. 20. Ball SL, Baker RL (1996) Predator-induced life history changes: antipredator behavior costs or facultative life history shifts. Ecology 77: 1116–1124. 4. Reguera P, Gomendio M (1999) Predation costs associated with parental care in the golden egg bug Phyllomorpha laciniata (Heteroptera: Coreidae). Behav Ecol 10: 541–544. 21. Tinkle DW, Gibbons JW (1977) The distribution and evolution of viviparity in reptiles. Misc. Pub. Museum Zool. Univ. Mich. 154: 1–54. 21. Tinkle DW, Gibbons JW (1977) The distribution and evolutio reptiles. Misc. Pub. Museum Zool. Univ. Mich. 154: 1–54. 22. Williams GC (1966) Adaptation and natural selection. Princeton University Press, Princeton, NJ. 5. Winkler DW (1987) A general model for parental care. Am Nat 130: 526–543. 6. Reynolds JD, Gross MR (1990) Costs and benefits of female mate choice: is there a lek paradox? Am Nat 136: 230–243. 23. Tinkle DW (1969) The concept of reproductive effort and its relation to the evolution of life histories of lizards. Am Nat 103: 501–516. 7. Rosenheim JA (1999) Characterizing the cost of oviposition in insects: a dynamic model. Evol Ecol 13: 141–165. 24. Tseng HY, Huang WS (2012) Sphenomorphus incognitus (Brown forest skink). Parental care. Herpetol Rev 43: 141. 8. Besnard A, Gimenez O, Lebreton JD (2002) A model for the evolution of creching in gulls. Evol Ecol 16: 489–503. 25. Huang WS (2010) Foraging behaviors of two sympatric ant species in response to lizard eggs. Zoology 113: 85–90. 9. Acknowledgments We thank C.H. Chang and several assistants for help in the field. This study was approved by the Taiwanese National Museum of Natural Science Animal Care and Use Committee (Protocol Permit NMNSHP02- 002). Parental action enhancing fitness 8. Parental care is more likely to evolve when it increases egg hatching ratios. According to our model, the higher the hatching ratio (in terms of benefit, Begg) the greater the opportunity for parental care evolution. For example, long-tailed skink eggs laid inside concrete walls have higher hatching success than those in natural nests [11]. Exposing the eggs to higher incubation temperatures might play an important role in the evolution of parental care. This prediction may help explain why tropical species more commonly display parental care than temperate species [1]. Eggs in tropical environments are exposed to higher temperatures, and in many species this results in higher hatching success, which would bring the benefits in line with our model. Further, our model might predict that endotherms would show parental care more commonly than ectotherms, because most of the temperatures experienced by embryos are higher in Reproductive attributes However, the incubation period may not be a major factor in the evolution of maternal care in this species because females in populations with and without maternal care nest in two habitat types: some females bury eggs beneath the soil (where they are not visible to predators) and some females lay eggs inside of a retaining wall (where they are visible to predators; [32]). Furthermore, the duration of maternal care does not always last the entire incubation period, but can vary due to the frequency of attempted egg predation by snakes [9]. Hence, this prediction may be overly simplistic. ectotherms than in ectotherms. This could help explain why almost all birds and mammals show parental care. Author Contributions Conceived and designed the experiments: WSH. Performed the experi- ments: WSH. Analyzed the data: WSH DAP. Contributed reagents/ materials/analysis tools: WSH DAP. Wrote the paper: WSH DAP. Conceived and designed the experiments: WSH. Performed the experi- ments: WSH. Analyzed the data: WSH DAP. Contributed reagents/ materials/analysis tools: WSH DAP. Wrote the paper: WSH DAP. Conclusions We know so little about the mechanisms leading to the evolution of parental care that it is difficult to predict how parents should invest in their offspring. Our general cost-benefit framework can be used to examine how a relatively large number of biologically- relevant variables simultaneously influence quantitative estimates of parental care. This approach revealed that the evolution of maternal care in long-tailed skinks is related to at least five different factors: the energetic investment in a clutch, the delay of future fecundity, the risk of predation to the parent, the ability to defend the eggs from predation, and the increase in hatching success provided by maternal care. The relative magnitudes of the cost and benefits of these variables are important in determining the evolution of parental care. Consequently, multiple pathways can lead to the evolution of parental care from a non-caring state, even in a single population of a widespread species. References Huang WS, Lin SM, Dubey S, Pike DA (In press) Predation drives interpopulation differences in parental care expression. J Anim Ecol in press. gg gy 26. Huang WS (2004) Reptile ecology and the evolution of 26. Huang WS (2004) Reptile ecology and the evolution of parental care on a tropical Asian island. Ph. D thesis, Cornell University, NY. g ( ) p gy tropical Asian island. Ph. D thesis, Cornell University, NY. 10. Huang WS (2006a) Parental care in the long-tailed skink, Mabuya longicaudata on a tropical Asian island. Anim Behav 72: 791–795. 27. Noble GK, Mason ER (1933) Experiments on the brooding habits of the lizards Eumeces and Ophisaurus. Am Mus Novit 619: 1–29. 11. Huang WS, Pike DA (2011) Climate change impacts on fitness depend on nesting habitat in lizards. Funct Ecol 25: 1125–1136. 28. Girondot M, Tucker AD, Rivalan P, Godfrey MH, Chevalier J (2002) Density- dependent nest destruction and population fluctuations of Guianan leatherback turtles. Anim Conserv 5: 75–84. 12. Maynard-Smith J (1977) Parental investment: A prospective analysis. Anim Behav 25: 1–9. Rand AS (1968) A nesting aggregation of iguanas. Copeia 1968: 552– 29. Rand AS (1968) A nesting aggregation of iguanas. Cop 13. Huang WS (2007) Costs of egg caring in the skink, Mabuya longicaudata. Ecol Res 22: 659–664. 30. Fraipont MD, Clobert J, Barbault R (1996) The evolution of oviparity with egg guarding and viviparity in lizards and snakes: a phylogenetic analysis. Evolution 50: 391–400. 14. Huang WS (2006b) Ecological characteristics of the skink, Mabuya longicaudata, on a tropical East Asian island. Copeia 2006: 293–300. 31. Bull JJ, Shine R (1979) Iteroparous animals that skip opportunities for reproduction. Am Nat 114: 296–316. 15. Schmitz OJ, Cohon JL, Rothley KD, Beckerman AP (1998) Reconciling variability and optimal behaviour using multiple criteria in optimization models. Evol Ecol 12: 73–94. 32. Huang WS (1994) Report on egg clutch size of long-tailed skink, Mabuya longicaudata from Taiwan. J Taiw Muse 47: 45–47. 16. Stearns SC (1992) The evolution of life histories. Oxford University Press, Oxford UK. 33. Huang WS, Greene HW, Chang TJ, Shine R (2011) Territorial behaviour in Taiwanese kukrisnakes (Oligodon formosanus). Proc Natl Acad Sci 108: 7455–7459. February 2013 \| Volume 8 \| Issue 2 \| e54065 PLOS ONE \| www.plosone.org February 2013 \| Volume 8 \| Issue 2 \| e54065 6
https://openalex.org/W4308485430	https://www.nature.com/articles/s41598-022-23415-2.pdf	English	null	Acute anxiety during the COVID-19 pandemic was associated with higher levels of everyday altruism	Scientific reports	2,022	cc-by	10,982	Joana B. Vieira 1,2, Stephen Pierzchajlo3, Simon Jangard2, Abigail A. Marsh4 & Andreas Olsson2 Joana B. Vieira 1,2, Stephen Pierzchajlo3, Simon Jangard2, Abigail A. Marsh4 & Andreas Olsson2 Prior laboratory research has suggested that humans may become more prosocial in stressful or threatening situations, but it is unknown whether the link between prosociality and defense generalizes to real-life. Here, we examined the association between defensive responses to a real- world threat (the COVID-19 pandemic) and everyday altruism. Four independent samples of 150 (N = 600) US residents were recruited online at 4 different timepoints, and self-report measures of perceived COVID-19 threat, defensive emotions (e.g., stress and anxiety), and everyday altruism were collected. Our operationalization of defensive emotions was inspired by the threat imminence framework, an ecological model of how humans and animals respond to varying levels of threat. We found that perceived COVID-19 threat was associated with higher levels of everyday altruism (assessed by the Self-report Altruism scale). Importantly, there was a robust association between experiencing acute anxiety and high physiological arousal during the pandemic (responses typically characteristic of higher perceived threat imminence), and propensity to engage in everyday altruism. Non-significant or negative associations were found with less acute defensive responses like stress. These findings support a real-life relation between defensive and altruistic motivation in humans, which may be modulated by perceived threat imminence. The beginning of 2020 was met with an unprecedented global challenge—the novel coronavirus outbreak. A defining feature of the COVID-19 pandemic, especially its earlier stages, was a widespread feeling of being under an invisible threat, a feature likely to profoundly impact psychological functioning and behaviour. Indeed, a wealth of laboratory research has demonstrated how defensive processes triggered by stressful or threatening situations affect not only psychological health1 and decision-making2, but also prosocial behaviour3. Still, little is known about how ecological threats of varying proximity impact behaviour outside the laboratory setting. The pandemic presented a unique opportunity to answer this question. Here, we examined the association between changing COVID-19 threat and prosocial behaviour, particularly everyday altruism.h g g p p y y y There is a long held popular view that human nature is inherently self-serving, and presumably more likely to reveal itself in challenging contexts or when resources are scarce4. These characteristics are applicable to early stages of a pandemic, wherein the fear of infection co-occurred with that of losing one’s job and/or access to essential goods like medication, food, or even toilet paper. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Scientific Reports \| (2022) 12:18619 Acute anxiety during the COVID‑19 pandemic was associated with higher levels of everyday altruism OPEN Joana B. Vieira 1,2, Stephen Pierzchajlo3, Simon Jangard2, Abigail A. Marsh4 & Andreas Olsson2 Joana B. Vieira 1,2, Stephen Pierzchajlo3, Simon Jangard2, Abigail A. Marsh4 & Andreas Olsson2 However, the notion of a fundamentally selfish human nature is called into question by the ubiquitous nature of everyday altruism in modern societies5,6, and by instances of extraordinary altruism in highly risky (e.g., heroic rescues; https://www.carnegiehero.org/) or costly scenarios (e.g., non-directed organ donation)7. Importantly, it has been suggested that challenging contexts may in fact promote rather than hinder altruistic motivation3,8. Previous studies have shown that inducing acute social stress through paradigms like the Trier Social Stress Test increases prosocial behaviour in subsequent economic exchanges9 and hypothetical moral decisions10, as well as the ability to empathize with others’ pain11. Other studies, however, indicated the link between stress and prosocial behaviour is more nuanced, as it may be shaped by additional individual and situational variables12–14. 1Department of Psychology, Faculty of Health and Life Sciences, University of Exeter, Exeter, UK. 2Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden. 3Department of Psychology, Stockholm University, Stockholm, Sweden. 4Department of Psychology, Georgetown University, Washington, DC, USA. email: j.m.b.vieira@exeter.ac.uk \| https://doi.org/10.1038/s41598-022-23415-2 Scientific Reports \| (2022) 12:18619 Scientific Reports \| (2022) 12:18619 www.nature.com/scientificreports/ One such variable could be the imminence of danger, and the way it is subjectively perceived by individuals. It has been proposed that, in humans and other species, defensive responses vary based on the spatiotemporal proximity of the threat, along a so-called threat imminence continuum15. Distal threats predominantly trigger vigilance, risk assessment and more flexible escape strategies, whereas imminent threats activate more stereotyped avoidance responses (e.g., freezing and fight-or-flight)15–17. The terms pre-encounter, post-encounter, and circa- strike have been used to describe different defensive contexts along the imminence continuum, and research in both animals and humans has begun to elucidate the neural circuitries implicated in adaptively responding to each context (see17 for a review). Importantly, it has been proposed that human behaviours, cognitions and emo- tions that characterize states like anxiety, fear and panic may be understood as contiguous transitions in response to the perceived imminence of a threat17–19. For example, intermittent anticipatory anxiety accompanied by cogni- tive control and reappraisal strategies are typical in response to an unpredictable threat that is not immediately present but may appear at any point (pre-encounter). By contrast, encounter anxiety may result from a threat that is already present but has consequences that are still somewhat unpredictable (post-encounter), and acute fear or panic may arise from a threat that is present and predictable, for which immediate avoidance is needed (circa-strike). Joana B. Vieira 1,2, Stephen Pierzchajlo3, Simon Jangard2, Abigail A. Marsh4 & Andreas Olsson2 An important point is that defensive responses are known to be variable between individuals20–22, and it is thus reasonable to expect some degree of inter-individual variability in this defensive continuum, i.e., a threat may be perceived as more imminent by some than others, giving rise to varying defensive behaviours to the same external event.hff There is laboratory evidence that defensive states along the imminence continuum may have different effects on prosocial behaviour. Specifically, we have recently demonstrated that healthy individuals were more likely to help a co-participant avoid aversive electrical shocks (at the risk of also being shocked) when helping decisions were made immediately before the shock delivery (imminent threat) than in the beginning of the trial (distal threat)23. Further, responses made during imminent relative to distal threats were faster, and accompanied by increased heart rate. These findings suggested that acute defensive states triggered in situations of imminent danger not only enable self-preservation responses, but may also promote motivation to defend/help others24. It is, however, unknown whether the link between threat imminence and altruistic motivation would persist in a larger spatiotemporal scale and in a real-life context. The COVID-19 pandemic offered the opportunity to exam- ine this question. In the United States, the first confirmed case of COVID-19 was reported on January 21st, and the government declared the outbreak a Public Health Emergency on February 3rd. On March 11th, the World Health Organization officially classified the outbreak as a global pandemic. By then, other countries like China, Italy and Iran had already recorded a rampant number of cases. This sequence of unfolding events presumably contributed to a perception of increased threat imminence, as individuals witnessed the negative impact of the virus in other countries, and anticipated similar consequences in their own country. y Our main goal here was to determine the association between COVID-19 threat and everyday altruism, defined as voluntary actions that benefit others at a variable personal cost (e.g., money, time, effort). Based on our findings that increased threat imminence favors helping decisions23, and evidence that acute stress promotes prosociality3,9–11, we hypothesized that increased COVID-19 threat over time would be accompanied by an increase in self-reported altruistic behaviours at the population level. To test this, we performed four independent data collections (cross-sectional design), corresponding to one-week apart time points during the COVID-19 pandemic (March–April), during which number of confirmed cases and COVID-19 fatalities were objectively increasing. Joana B. Vieira 1,2, Stephen Pierzchajlo3, Simon Jangard2, Abigail A. Marsh4 & Andreas Olsson2 At each time point, self-report measures of everyday altruism, perceived COVID-19 threat, and experienced defensive emotional states were collected (see “Methods”). We then examined whether perceived COVID-19 threat and everyday altruism increased over those 4 weeks. Given the expected inter-individual variability in responding to threat imminence, we also hypothesized that individuals experiencing more acute defensive states (presumably linked with perception of higher threat imminence) would report higher engage- ment in everyday altruism. This prediction was tested by modelling everyday altruism as a function of different defensive emotions, while accounting for week-by-week variation. Our measure of everyday altruism was the Self-report Altruism Scale26, which assesses the frequency with which participants have engaged in different eve- ryday altruistic behaviours in their lifetime. These ratings are believed to reflect both the frequency of performed acts, and the individual’s endorsement of helping others, thus being considered both a measure of self-reported behaviour and dispositional altruism25. This is an important point to keep in mind when interpreting our results, as they can reflect both behavioural and trait-level associations between defence and altruism. Results Ch Changes in everyday altruism and perceived COVID‑19 threat over time. The total number of confirmed COVID-19 cases increased over the 4 weeks of data collection, but we found no evidence that per- ceived COVID-19 threat increased over time (F(3, 596) = 1.82, p = 0.14; Fig. 1). Regarding levels of everyday altruism assessed by the SRA, although the ANOVA suggested changes across the four weeks (F(3, 596 = 3.04, p = 0.03), follow-up pairwise comparisons were not statistically significant (all ps > 0.15).fi i Because some of the items of the SRA describe actions that may be difficult to perform during a lockdown (e.g., helping a stranger with car troubles), we repeated the ANOVA using as dependent variable a subset of SRA items referring to donations, given that the ability to carry out those behaviours should be less affected. When using the SRA-donations score, results suggested changes over the 4 weeks (F(3, 596) = 3.69, p = 0.012). Pairwise comparisons revealed the only significant difference was a decrease between Week 1 and 2 (p = 0.012). if Altogether, our results provided limited evidence about how population levels of overall altruism changed over the 4 weeks of data collection, but we found some indication that donation behaviour may have decreased from the first to the second week. These results could have been impacted by the non-changing threat perception, and/or limitations in the SRA’s ability to capture behavioural changes in that time period. https://doi.org/10.1038/s41598-022-23415-2 Scientific Reports \| (2022) 12:18619 \| www.nature.com/scientificreports/ Figure 1. Total number of confirmed cases averaged across states, over the 4 weeks of data collection (top left); Perceived COVID-19 threat over the 4 weeks of data collection (top right ); Everyday altruism over the 4 weeks of data collection (SRA total and donations only score, bottom panel). The dashed line connects the mean across samples (i.e. weeks). Note that the distributions shown here correspond to independent samples collected on each week. Figure 1. Total number of confirmed cases averaged across states, over the 4 weeks of data collection (top left); Perceived COVID-19 threat over the 4 weeks of data collection (top right ); Everyday altruism over the 4 weeks of data collection (SRA total and donations only score, bottom panel). The dashed line connects the mean across samples (i.e. weeks). Note that the distributions shown here correspond to independent samples collected on each week. Results Ch Modelling everyday altruism (SRA) as a function of perceived threat and defensive emo‑ tions. To examine whether experiencing different defensive emotional states during the pandemic was asso- ciated with altruism at the individual level, we adopted a Generalized Linear Mixed Models (GLMMs) approach. This allowed us to account for variation in everyday altruism that was explained by our variables of interest (fixed effects), as well as by random sampling of timepoint (random effect) (see “Methods” and Supplementary material for a detail description of all the modelling checks and steps).iif p g p We first examined how everyday altruism (SRA) varied as a function of our key fixed effects of interest Model 1), namely: a) Perceived COVID-19 threat (assessed by the COVID-19 Risk Perception Scale, RP26);i y p b) Stress, defined as feelings of uncontrollability and unpredictability, which are characteristic in response to situations of lower threat imminence (assessed by Perceived Stress Scale-10, PSS-1027); y p b) Stress, defined as feelings of uncontrollability and unpredictability, which are characteristic in response to situations of lower threat imminence (assessed by Perceived Stress Scale-10, PSS-1027);i ( y ) c) Anxiety, defined as a state of high autonomic arousal, acute anxiety and panic, consistent with emotional responses to threats perceived as predictable and imminent (assessed by the Depression Anxiety Stress Scales, DASS-2128). y c) Anxiety, defined as a state of high autonomic arousal, acute anxiety and panic, consistent with emotional responses to threats perceived as predictable and imminent (assessed by the Depression Anxiety Stress Scales, DASS-2128). Although not the focus of our hypotheses, Depression (assessed by the DASS-21) was also included as a fixed effect, since it is thought to have important associations with stress and anxiety-related responses29. Results showed only perceived COVID-19 threat and anxiety were significantly associated with increased everyday altruism (Table 1, Fig. 2A). When modelling these variables with only random intercept per week, we found additionally that stress and depression were negatively associated with altruism (see Supplementary material— Addressing model singularity, Table m1.1). We next examined the association between other demographic variables and everyday altruism, focusing on age, gender, socio-economic status and employment. Rather than introducing all these additional fixed effects in the main model, we instead first modelled them separately, with the goal of identifying those significantly associated with altruism. This strategy helped limit the complexity of the main model and avoid convergence problems (see Supplementary material for results of the model with all predictors, fixed effects of interest plus additional demographics, which did not converge but showed results in the same direction and significance as those reported here). Age was found to be positively associated with altruism, and employment was margin- ally negatively associated (Supplementary material). We thus added these variables to the main model (Model 2). Results showed that age, perceived COVID-10 threat, and anxiety were uniquely associated with increased Scientific Reports \| (2022) 12:18619 \| https://doi.org/10.1038/s41598-022-23415-2 www.nature.com/scientificreports/ Table 1. Model 1 estimates (DV: Everyday altruism, SRA). Estimate Std. a) Perceived COVID-19 threat (assessed by the COVID-19 Risk Perception Scale, RP26);i Error df t value Pr(>\|t\|) CI_lower CI_upper (Intercept) − 0.006 0.070 3.084 − 0.091 0.933 − 0.143 0.131 Perceived COVID-19 threat 0.169 0.050 5.565 3.377 0.017 0.071 0.267 Stress − 0.146 0.085 3.282 − 1.719 0.176 − 0.312 0.020 Anxiety 0.305 0.055 9.893 5.526 0.000* 0.197 0.414 Depression − 0.172 0.088 3.570 − 1.945 0.132 − 0.345 0.001 Table 1. Model 1 estimates (DV: Everyday altruism, SRA). Table 1. Model 1 estimates (DV: Everyday altruism, SRA). Figure 2. (A) Beta values and SEs for Model 1. (B) Beta values and SEs for Model 2. Dependent variable in both models is everyday altruism (SRA). Figure 2. (A) Beta values and SEs for Model 1. (B) Beta values and SEs for Model 2. Dependent variable in both models is everyday altruism (SRA). everyday altruism (Table 2, Figs. 2B and 3). When using a simpler model with only a random intercept by week, in addition to these effects, we found depression, stress and unemployment were negatively associated with altruism (Supplementary material Table m3.3). everyday altruism (Table 2, Figs. 2B and 3). When using a simpler model with only a random intercept by week, in addition to these effects, we found depression, stress and unemployment were negatively associated with altruism (Supplementary material Table m3.3). y Overall, results showed higher perceived COVID-19 threat, and higher acute anxiety were robustly associ- ated with increased everyday altruism, with some evidence emerging for a negative association with stress. This pattern of results can be interpreted as higher threat perception and acute defence during a crisis being linked with increased engagement in altruism, or alternatively that those higher in altruistic tendencies are more likely to experience threat and acute anxiety in response to a crisis. Modelling other indicators of altruism during the pandemic. We next examined whether the observed associations between everyday altruism, perceived COVID-19 threat, and anxiety would hold when using an indicator that specified altruistic action during the pandemic, namely the reported frequency of altru- istic behaviours towards strangers or charity in the last few weeks. Results showed that anxiety and age were positively associated with reporting altruistic behaviours towards a stranger or charity during the COVID-19 pandemic (Table 3). https://doi.org/10.1038/s41598-022-23415-2 Scientific Reports \| (2022) 12:18619 \| www.nature.com/scientificreports/ Table 2. Model 2 estimates (DV: Everyday altruism, SRA). Reference class = Employed. Estimate Std. Table 3. Model 3 estimates (DV: Help stranger). 1 Reference class = Employed. a) Perceived COVID-19 threat (assessed by the COVID-19 Risk Perception Scale, RP26);i Error df t value Pr( >\|t\|) CI_lower CI_upper (Intercept) 0.119 0.070 3.580 1.694 0.174 − 0.019 0.256 Perceived COVID-19 threat 0.156 0.051 4.674 3.090 0.030 0.057 0.256 Stress − 0.084 0.063 5.228 − 1.327 0.239 − 0.208 0.040 Anxiety 0.296 0.052 9.196 5.641 0.000* 0.193 0.399 Depression − 0.132 0.069 5.115 − 1.915 0.112 − 0.268 0.003 Age 0.314 0.042 577.963 7.498 0.000* 0.232 0.396 Employment* = Student − 0.120 0.119 8.789 − 1.009 0.340 − 0.352 0.113 Employment = Unemployed − 0.416 0.149 3.231 − 2.786 0.063 − 0.708 − 0.123 Table 2. Model 2 estimates (DV: Everyday altruism, SRA). Reference class = Employed. Figure 3. Model 2 predictions of everyday altruism (SRA) as a function of Anxiety (top left), Perceived COVID-19 threat (top right), Age (bottom left) and Stress (bottom right), depicting estimated random slopes per week. Figure 3. Model 2 predictions of everyday altruism (SRA) as a function of Anxiety (top left), Perceived COVID-19 threat (top right), Age (bottom left) and Stress (bottom right), depicting estimated random slope per week. Table 3. Model 3 estimates (DV: Help stranger). 1 Reference class = Employed. Estimate Std. error df t value Pr( >\|t\|) CI_lower CI_upper (Intercept) 0.040 0.081 3.994 0.493 0.648 − 0.120 0.200 Perceived COVID-19 threat 0.047 0.068 2.601 0.686 0.549 − 0.087 0.180 Stress − 0.050 0.132 1.170 − 0.380 0.761 − 0.309 0.209 Anxiety 0.267 0.079 3.389 3.380 0.036 0.112 0.422 Depression − 0.137 0.155 1.068 − 0.883 0.532 − 0.441 0.167 Age 0.230 0.067 291.869 3.419 0.001* 0.098 0.362 Employment (1 = Student)1 0.099 0.162 16.485 0.611 0.549 − 0.219 0.417 Employmen (2 = Unemployed) − 0.130 0.171 1.585 − 0.759 0.544 − 0.465 0.205 Table 3. Model 3 estimates (DV: Help stranger). 1 Reference class = Employed. https://doi.org/10.1038/s41598-022-23415-2 Scientific Reports \| (2022) 12:18619 \| www.nature.com/scientificreports/ Table 4. Model 4 estimates (DV: Donations). 1 Reference class = Employed. Estimate Std. a) Perceived COVID-19 threat (assessed by the COVID-19 Risk Perception Scale, RP26);i error df t value Pr( >\|t\|) CI_lower CI_upper (Intercept) 0.109 0.072 3.366 1.520 0.216 − 0.032 0.250 Perceived COVID-19 threat 0.129 0.051 4.868 2.534 0.054 0.029 0.228 Anxiety 0.210 0.061 5.125 3.471 0.017* 0.092 0.329 Depression − 0.158 0.063 20.345 − 2.510 0.021* − 0.281 − 0.035 Stress − 0.027 0.067 5.632 − 0.398 0.705 − 0.157 0.104 Age 0.282 0.055 3.992 5.115 0.007* 0.174 0.391 Employment (1 = Student)1 − 0.039 0.125 6.831 − 0.313 0.764 − 0.285 0.207 Employment (2 = Unemployed) − 0.443 0.149 3.336 − 2.975 0.051 − 0.735 − 0.151 Table 4. Model 4 estimates (DV: Donations). 1 Reference class = Employed. Table 4. Model 4 estimates (DV: Donations). 1 Reference class = Employed. Figure 4. Zero-order correlations between indices of prosocial behaviour and defensive responses. The numbers in each cell correspond to the Pearson r coefficients. Cells with a white background correspond to correlations that did not survive correction for multiple comparisons. Covid19_th: Perceived COVID-19 threat; PSS: Perceived Stress Scale; Depression: DASS-21 Depression scale; Stress: DASS-21 Stress scale; Anxiety: DASS- 21 Anxiety scale; SRA: Self-report Altruism; PBI: Prosocial Behavioural Intentions scale; stranger: Frequency of altruistic acts towards stranger; friend: Frequency of altruistic acts towards a friend; relative: Frequency of altruistic acts towards a relative; ingroup: Frequency of altruistic acts towards an ingroup member; outgroup: Frequency of altruistic acts towards an outgroup member; acquaint: Frequency of altruistic acts towards an acquaintance. Figure 4. Zero-order correlations between indices of prosocial behaviour and defensive responses. The numbers in each cell correspond to the Pearson r coefficients. Cells with a white background correspond to correlations that did not survive correction for multiple comparisons. Covid19_th: Perceived COVID-19 threat; PSS: Perceived Stress Scale; Depression: DASS-21 Depression scale; Stress: DASS-21 Stress scale; Anxiety: DASS- 21 Anxiety scale; SRA: Self-report Altruism; PBI: Prosocial Behavioural Intentions scale; stranger: Frequency of altruistic acts towards stranger; friend: Frequency of altruistic acts towards a friend; relative: Frequency of altruistic acts towards a relative; ingroup: Frequency of altruistic acts towards an ingroup member; outgroup: Frequency of altruistic acts towards an outgroup member; acquaint: Frequency of altruistic acts towards an acquaintance. We also replaced the SRA total score with that referring to donations only in the full model (Model 4). Results showed that anxiety and age remained significantly associated with increased donation behaviours. Additionally, Depression was negatively associated with those behaviours (Table 4). Table 4. Model 4 estimates (DV: Donations). 1 Reference class = Employed. Discussionh This study investigated the association between everyday altruistic behaviour, changing level of COVID-19 threat over time (objective threat imminence), and type of defensive emotions experienced by individuals during this period (perceived threat imminence). It has been proposed that emotional states like stress, anxiety, fear and panic can be understood as reflecting contiguous defensive responses to varying levels of perceived threat17,19. Based on observations that defensive responses to increased threat imminence23 and acute stress3 may facilitate prosocial action in experimental settings, we hypothesized that experiencing more acute defensive emotions during the pandemic would be associated with higher levels of self-reported altruism. Our results supported this hypothesis, showing that individuals reporting higher perceived COVID-19 threat and experiencing higher acute anxiety (conceptualized as high physiological arousal and feelings of being close to panic), also reported higher everyday altruism. These associations persisted while controlling for additional variables such as age, and employment status. The association between everyday altruism and acute anxiety was particularly strong and consistent both across, and within each week (when each week was analysed separately, acute anxiety emerged as a robust predictor of everyday altruism—see Supplementary material), suggesting that acute defensive states may be particularly predictive of altruistic outcomes.i y p y p Our findings are consistent with previous reports that acute social stressors promote prosocial behaviour. For instance, individuals submitted to the Trier Social Stress Test, a paradigm that increases salivary cortisol and heart rate in experimental studies, have been shown to subsequently display increased levels of trustwor- thiness and sharing9,30, empathy for others’ pain11, and self-reported emotional empathy31. Of note, despite the different terminology, “acute stress” in these experimental studies has a physiological manifestation and is more closely related to our operationalization of acute anxiety. That acute social stressors promote affiliative behaviour has often been interpreted with reference to the “tend-and-befriend” hypothesis by Taylor32,33. This hypothesis proposes an oxytocin-mediated biobehavioural system that, on the one hand, fosters affiliative behaviours that maximize species survival (tending), and, on the other, reduces stress through the establishing of social bonds (befriending). The “tend-and-befriend” proposal thus highlights the role of oxytocin in promoting social bonds and, as a result, in dampening defensive responses in social stressful situations. However, animal and human research have demonstrated that oxytocin plays a more general role in regulating individual defensive behaviour in non-social contexts as well34. Discussionh For instance, oxytocinergic neurons in the central amygdala are implicated in allowing transitions between freezing and active defense (fight-or-flight) in response to imminent danger35,36. Moreover, in rodents, the oxytocin-mediated ability to inhibit freezing in favor of active threat coping behaviours is necessary to allow females to engage in offspring defense37. In light of these data, one possibility is that acute defensive states enable prosocial action regardless of the social nature of the stressor, by recruiting active coping mechanisms also implicated in dealing with first-hand threats. This is consistent with recent experimental data showing that the induction of fight-or-flight states by increasing threat imminence promotes costly helping deci- sions in humans23. It is also consistent with previous theoretical accounts suggesting acute physiological stress (versus diffuse or chronic stress) promotes caregiving specifically in situations of immediate need8,38. While open questions remain regarding the underlying mechanisms, our results demonstrate a positive relation between acute defensive states and prosocial behaviour, which generalizes beyond time bounded laboratory experiments to real-life threatening contexts that extend over time. I i l f d i h i ifi i i i (d di h h d l g Interestingly, we found either a non-significant or negative association (depending on how the models were defined) between altruism and stress, characterized here by diffuse anxiety and feelings of uncontrollability and unpredictability. One possible interpretation for this finding, is that the association between altruism and threat perception is modulated by perceived threat imminence. i.e., altruism is differently associated with defensive emotions typical of higher perceived threat imminence (acute anxiety), and those eventually elicited when threat imminence is perceived to be lower (stress). It should be noted however, that our questionnaire measures do not directly assess perceived threat imminence. Rather, our measures of stress mainly tapped onto feelings of control- lability and unpredictability. Therefore, our interpretation, and the potentially dissociable association between altruism and perceived threat imminence, warrants further examination in more controlled experiments (e.g.,23). Additionally, because the (negative) relation between stress and altruism was not as robust as the (positive) relation with acute anxiety, replication of the stress effect in future well-powered studies is especially warranted. a) Perceived COVID-19 threat (assessed by the COVID-19 Risk Perception Scale, RP26);i We also replaced the SRA total score with that referring to donations only in the full model (Model 4). Results showed that anxiety and age remained significantly associated with increased donation behaviours. Additionally, Depression was negatively associated with those behaviours (Table 4). Finally, we re-ran the full model using the Prosocial Behavioral Intentions (PBI) score as dependent measure. We found no evidence of a significant association between prosocial intentions and any of our predictors (Sup- plementary material). Of note, this measure assesses the willingness to engage in future prosocial behaviour rather than self-reported performed behaviours. Also, visual inspection of the PBI data revealed skewed distribu- tions, with the majority of scores concentrated in the upper end of the scale. This suggests participants generally rated their behavioural intentions as very highly prosocial (maybe due to social desirability), and variability in responses was low (see Supplementary material). This limitation of the measure could have contributed for the disparity in findings between prosocial intentions and the other altruistic measures. Associations among measures of prosociality and defensive emotions. At an exploratory level, we examined the associations between everyday altruism, other measures of prosocial behaviour, and measures assessing defensive emotional states (anxiety and stress). Correlation analyses (corrected for multiple compari- sons, all reported ps < 0.00065) revealed moderate positive correlations between everyday altruism and other indicators of prosocial behaviour, including prosocial behavioural intentions, and frequency of altruistic behav- iours towards friends, acquaintances, strangers, ingroup and outgroup members (Fig. 4). Perceived COVID-19 Scientific Reports \| (2022) 12:18619 \| https://doi.org/10.1038/s41598-022-23415-2 www.nature.com/scientificreports/ threat was positively associated with state measures of anxiety and stress. In line with the main analysis, everyday altruism was positively correlated with perceived COVID-19 threat and anxiety. There results point to a coherent network of associations between measures assessing defensive and prosocial behaviour, respectively, and further support a unique link between altruistic behaviour and acute defensive responses to the pandemic. threat was positively associated with state measures of anxiety and stress. In line with the main analysis, everyday altruism was positively correlated with perceived COVID-19 threat and anxiety. There results point to a coherent network of associations between measures assessing defensive and prosocial behaviour, respectively, and further support a unique link between altruistic behaviour and acute defensive responses to the pandemic. www.nature.com/scientificreports/ systems responsible for defensive responding and those implicated in empathy and care. Future research would benefit from considering additional trait level moderators of interest in the association between defense and altruism, such as individual differences in threat sensitivity and dispositional empathy. f y p p y We had also hypothesized that increased COVID-19 threat over time (i.e., increasing number of cases and fatalities) would result in increasing indices of everyday altruism at the population level. Our results for this hypothesis were inconclusive. Although donations decreased from Week 1 to Week 2, no clear indications of increase or decrease in altruism emerged over the 4 weeks of data collection. Critically, our inability to detect meaningful alterations in altruism may have been due to the fact that, despite the increasing number of cases and fatalities, subjective perception of COVID-19 threat did not significantly change over those 4 weeks. There are several potential, not mutually exclusive reasons for the disparity between objective and subjectively perceived threat. For one, the use of different samples each week may have compromised our power to capture changes over time. Another likely reason is that our study had insufficient temporal resolution to detect changes in sub- jective threat perception that accompanied the course of the pandemic. It is possible that individuals perceived the COVID-19 situation to be substantially more severe prior to the first data collection on March 24th, and that perception became relatively stable over the following weeks. Indeed, in the early stages of the pandemic in the United States, information about the coronavirus and its transmission was still limited and unclear. The percep- tion of danger in these early stages was therefore presumably higher, as suggested by the outages of toilet paper and other goods at supermarkets (https://www.mckinsey.com/industries/consumer-packaged-goods/our-insig hts/us-food-supply-chain-disruptions-and-implications-from-covid-19), and increase in gun sales across the country (https://www.theguardian.com/us-news/2021/may/31/us-gun-sales-rise-pandemic). Some research sup- ports this interpretation, with data suggesting that more substantial changes in COVID-19 risk perception may have occurred prior to the beginning of our study39,40. Another possibility is that objective indices of COVID-19 threat, like number of cases and deaths, are not the only contributing factor for subjective threat perception. For instance, knowledge about the measures that are being implemented to control the virus spread may modulate how severe individuals perceive the situation to be. www.nature.com/scientificreports/ In that regard, the start of the lockdown in the US during the data collection period may have mitigated the perceived threat. Accounting for the state of residence of the participants could also have improved the sensitivity of the analysis to detect changes in threat perception, since the evolution of the pandemic was quite heterogeneous across different locations. Our dataset had an uneven number of cases per US state in each week, substantially limiting statistical power of any spatially-based analyses. However, even accounting for the number of COVID-19 cases specifically in the participants’ state of resident, we did not find indication that it was associated with perceived COVID-19 threat (r = 0.006, p = 0.88). In sum, given that we cannot discount the influence of between-sample variation in our data, that our time resolution may have been insufficient, and that perceived threat did not change significantly over the 4 weeks, our study was inconclusive to assess a potential link between changing threat imminence over time and population levels of altruistic behaviour. Future research on population-level effects would benefit from including information regarding location, as well as additional demographic aspects, such as having children or not, and health status of the participants.ih p p As a final note, it is worth pointing out the strong association that emerged between age and altruism. This association has been previously documented, e.g.,41–46, including in relation to situations of acute stress47. The positive association between age and altruism is likely explained by a combination of biological and sociocultural factors, which are out of the scope of the present study. Of note, we did not find indication that this association was driven by other demographical aspects (such as employment or socio-economic level), which likely covary with age but were controlled for in our analysis. The inclusion of additional variables, such as having children and health status, could help to further clarify the apparent positive impact of age on everyday altruism. It should also be noted that the strength of the age effect we found could, to some extent, be driven by how the SRA scale is formulated, i.e., in measuring accumulated behaviour over time, older participants might by default report more engagement in altruistic behaviours.i g g Some specificities of our study should be acknowledged. Discussionh yf y One aspect to consider is that our main indicator of altruism, the Self-report Altruism Scale (SRA), does not specify a time window for the occurrence of behaviour, and thus altruistic behaviours reported by participants may have included actions not necessarily undertaken during the pandemic. We have accounted for this in the analysis by testing our model with an index of altruistic behaviours performed during the pandemic, having confirmed the same pattern of results with acute anxiety. In addition, the SRA includes actions that could be dif- ficult to perform during the lockdown. Our analysis also accounted for this aspect, demonstrating results in the same direction with acute anxiety while using only SRA items related to donation behaviours, which although still hindered to some extent by the lockdown context (e.g., donating blood)could more easily be undertaken (e.g., charity donations). Despite these limitations (which we have accounted for in the analysis), the SRA has the advantage of being the only validated and well-established measure of self-reported altruistic behaviour in our study. Crucially, even if understood as trait measure, our findings with the SRA suggest that altruistic dispositions are associated with the propensity to experience acute anxiety (but not stress) during a real-life crisis. Such trait- level association has important theoretical implications, since it further supports a link between neurobiological Scientific Reports \| (2022) 12:18619 \| https://doi.org/10.1038/s41598-022-23415-2 www.nature.com/scientificreports/ Methods Participants. Participants. 602 participants were recruited through the Prolific participant pool (Prolific.co). The sample size was decided based on resource availability (largest N possible with available resources), and not a formal power calculation. To mitigate the lack of a power calculation, we performed simulation-based power analysis after completing the study, fixing the sample size to the one we used to determine our power to detect effect sizes of different magnitudes. This procedure provides an estimate of how much power our study had, without relying on the obtained effects sizes48,49. This analysis suggested that our sample size and design would have between 88 and 94% power to detect a 0.3 effect size (d) (details on this analysis and power curve in Supplementary mate- rial).l Participant inclusion criteria were Nationality and residence in the United States, and fluency in English. Data were collected at 4 timepoints, over a 4-week period: 24-03-2020 (Week 1, n = 150), 31-03-2020 (Week 2, n = 150), 07-04-2020 (Week 3, n = 150), and 14-04-2020 (Week 4, n = 151). Data collection took place at approximately 20:30 (CEST) each week (i.e., the study was set to active on Prolific at that time). The collection of 4 samples at different timepoints (rather than N = 600 at one timepoint) provided a way of examining the robustness of the findings across independent samples, which combined with the statistical approach adopted (Generalized Lin- ear Mixed Effects Models, see below) has been suggested to improve the generalizability of research findings50. f gg p g yi g Participants completed an online survey consisting of several questionnaire measures, in addition to demo- graphical information, namely age, gender, employment status, state of residence, socio-economic status and education level. The full survey (available here: https://osf.io/c8zhn/) took about 15 min to complete. One entry was excluded because the data appeared to be a duplicate (i.e., two entries had identical responses in open-ended fields for participant comments and questions, and all demographics except age were identical). The data seemed otherwise valid, and thus only the first entry was retained. Another participant was excluded due to incomplete information regarding state of residence, yielding a total of 600 participants in the final sample (Mage = 31.16, SDage = 10.70, 56.1% female, 2.0% other; Fig. 5). All participants provided informed consent prior to participating. www.nature.com/scientificreports/ At the conceptual level, it is important to note that, while our operationalization of different defensive states was informed by the threat imminence framework, it does not entirely match how threat imminence has been previously manipulated in experimental studies. Specifi- cally, experimental studies have typically manipulated features of threatening stimuli (e.g., spatial proximity), thereby changing actual threat imminence. Conversely, in our study, we took two approaches: first, we inferred threat imminence indirectly based on the number of COVID-19 cases; and second, we quantified individual defensive responses that are thought to reflect variation in perceived threat imminence. This difference in actual versus perceived threat should be taken into account when interpreting our results, although it is noteworthy that our results were consistent with a recent experimental study that directly manipulated threat imminence23.f A related limitation refers to the self-report measures used to index different defensive states. While these measures (DASS-21 scales and Perceived Stress Scale) tapped into dimensions closely related to varying threat imminence (e.g., predictability/controllability, versus high arousal and feelings of panic), they did not provide a direct measure of perceived threat imminence. Moreover, our ability to capture extreme levels of stress and anxiety may have been limited by the fact that high scorers in those measures (e.g., individuals with small chil- dren at home, or caring for ill relatives) may be less likely to participate in an online study. Finally, it is worth acknowledging the difficulty in measuring altruistic behaviour during the pandemic due to, among other reasons, reliance on self-report, and behavioural restrictions imposed by lockdown. Scientific Reports \| (2022) 12:18619 \| https://doi.org/10.1038/s41598-022-23415-2 www.nature.com/scientificreports/ Figure 5. Demographic information for the sample collected in each week. l ll d h k Figure 5. Demographic information for the sample collected in each week. www.nature.com/scientificreports/ Objective COVID-19 threat was operationalized by the number of total confirmed cases per state of residence. Statistical analysis. Changes in everyday altruism and perceived COVID‑19 threat over time. Our first question was whether increased COVID-19 threat over time was accompanied by higher frequency of self- reported altruistic behaviours at the population level. To test our hypothesis, we performed one-way ANOVAs, using perceived COVID-19 threat and everyday altruism as dependent variables. Our main indicator of eve- ryday altruism was the SRA score. Because some of the items of the SRA describe actions that may be difficult to perform during a lockdown (e.g., helping a stranger with care troubles), we repeated the ANOVA using as dependent variable a subset of SRA items referring to donations, given that the ability to carry out those be- haviours should be less affected (items: 4. I have given money to a charity, 5. I have given money to a stranger who needed it (or asked me for it), 6. I have donated goods or clothes to a charity, 7. I have done volunteer work for a charity, 8. I have donated blood; Cronbach’s alpha for SRA donations = 0.7). In each ANOVA, we assessed whether Perceived COVID-19 threat and indices of everyday altruism (SRA and SRA-donations) varied over the 4 weeks of data collection. Modelling everyday altruism (SRA) as a function of perceived COVID‑19 threat and defensive emotions. Our second question was whether experiencing different defensive emotional states during the pandemic was associ- ated with self-reported altruism at the individual level. To assess this question, we adopted a Generalized Linear Mixed Models (GLMMs) approach. GLMMs explicitly model variance in a hierarchical fashion, in keeping with the hierarchical structure of datasets in which different observations belong to different groups or were collected at different timepoints (e.g., week). Thus, this approach allowed us to account for variation in everyday altruism (SRA; dependent variable) that was explained by our variables of interest (fixed effects), as well as by random sampling of timepoint (random effect). Initially, both state of residence and week were included as random effects. However, given (a) the limited and unbalanced number of observations from some US states, (b) the high number of random effect estimates necessary to model state as a random effect, and c) the negligible vari- ance explained by the random effect of state, we opted to exclude state from the model. www.nature.com/scientificreports/ Perceived COVID-19 threat was operation- alized through the COVID-19 Risk Perception Scale (RP) developed by Wise and collaborators26 (M = 45.27, range = [10, 70]), which measures self-reported feelings and thoughts about the COVID-19 pandemic, including perceived risk of infection and economic repercussions (e.g., “How likely do you think you are to catch the virus”, 1 = min, 7 = max). Defensive emotional responses were assessed through the Perceived Stress Scale-10 (PSS-1027) and the Depression Anxiety Stress Scales (DASS-2128). The PSS-10 (M = 20.73, range = 3, 40]) includes 10 items assess- ing experienced stress, specifically the degree to which individuals perceived their lives to be unpredictable, uncontrollable, and overloaded in the past month (e.g., “How often have you been upset because something that happened unexpectedly?”, 1 = never, 5 = very often).h y yt The DASS-21 includes 21 items assessing states of stress, anxiety and depression over the past week, using a 1 (Did not apply to me at all), to 4 (applied to me very much or most of the time) scale. The Stress scale (M = 15.29, range = [0, 42]) measures difficulty relaxing, unspecific nervousness, and being easily upset and impatient (e.g., “I found it hard to wind down). The Anxiety scale (M = 8.23, range = [0, 40]), on the other hand, assesses states of acute anxiety, characterized by high autonomic arousal and skeletal muscle effects (e.g., “I was aware of dry- ness of my mouth”, “I felt close to panic”), presumably more characteristic of higher perceived threat imminence (post-encounter and circa-strike contexts). Finally, the Depression scale (M = 14.2, range = [0, 42]) assessed feeling of dysphoria, hopelessness, devaluation of life, self-deprecation, lack of interest, anhedonia, and inertia (e.g., “I couldn’t seem to experience any positive feeling at all”). Although not the focus of the present study, the depression scale was included to account for shared variance with stress and anxiety. p y All questionnaires were used at all time points, with the exception of the items assessing frequency of altruistic behaviours towards different targets, which were only measured on Weeks 3 and 4.h f The COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University (https://github.com/CSSEGISandData/COVID-19/) was used as a resource for obtaining objective COVID-19 data during the time window of the study, including the number of cases and fatalities per state in the United States. www.nature.com/scientificreports/ another person, without any benefit for yourself. Examples would be donating money, clothes or food, caring for someone who was unwell, assisting someone in a task they were unable to complete on their own, etc. If any of these or similar behaviours occurred in the last few weeks, how many of those behaviours were directed at…”. Subsequently, participants rated the amount of such behaviours that were directed at a relative (Help relative), a friend (Help friend), an acquaintance (Help acquaintance), a stranger or a charity (Help stranger), a person having a lot in common with (Help ingroup), and a person having nothing or very little in common with (Help outgroup; 1 = none, 4 = all). The idea to include these measures arose after the study had been initiated, and therefore these additional items are only available for Weeks 3 and 4 (n = 300). another person, without any benefit for yourself. Examples would be donating money, clothes or food, caring for someone who was unwell, assisting someone in a task they were unable to complete on their own, etc. If any of these or similar behaviours occurred in the last few weeks, how many of those behaviours were directed at…”. Subsequently, participants rated the amount of such behaviours that were directed at a relative (Help relative), a friend (Help friend), an acquaintance (Help acquaintance), a stranger or a charity (Help stranger), a person having a lot in common with (Help ingroup), and a person having nothing or very little in common with (Help outgroup; 1 = none, 4 = all). The idea to include these measures arose after the study had been initiated, and therefore these additional items are only available for Weeks 3 and 4 (n = 300). y Additionally, to assess prosocial intentions (rather than performed behaviour), we used the Prosocial Behav- ioural Intentions Scale52, wherein individuals rate how likely they would be to engage in 4 types of prosocial actions (“Comfort someone I know after they experience a hardship”, “Help a stranger find something they lost, like their key or a pet”, “Help care for a sick friend or relative”, “Assist a stranger with a small task (e.g., help carry groceries, watch their things while they use the restroom”). Perceived COVID‑19 threat and defensive emotional responses. Methods Participants. age g p p p p p p g All methods and experimental protocols were carried out in accordance with relevant guidelines and regula- tions, and approved by the Karolinska Institutet (Sweden) and the Etikprövningsmyndigheten (Ethics review authority). Questionnaire measures. Everyday altruism. Everyday altruism was measured through the widely used Self-Report Altruism Scale51 (M = 50.83, SD = 11.44, range = [24, 98]; SRA normative data from51 mean = 55.40 and SD = 10.57), in which participants rate the frequency (1 = never, 5 = very often) with which they engaged in 20 different everyday altruistic behaviours (e.g., “I have given money to a charity”, “I have offered my seat on a bus or train to a stranger who was standing”). These ratings are believed to reflect both the frequency of per- formed altruistic acts, and the individual’s endorsement of helping others in everyday situations51.h The SRA does not specify a timeframe for the occurrence of altruistic actions, thereby including both acts performed in the weeks leading up to the moment of data collection (i.e., during the pandemic), as well as prior in the participant’s life. To obtain more detailed information about self-reported altruistic actions specifically during the timeframe of the study, we created 5 additional items in which participants rated the frequency of altruistic actions towards different targets: “In the last few weeks, think of anything you may have done to help or benefit Scientific Reports \| (2022) 12:18619 \| https://doi.org/10.1038/s41598-022-23415-2 www.nature.com/scientificreports/ www.nature.com/scientificreports/ www.nature.com/scientificreports/ Key fixed effects of interest included in Model 1 were: (1) perceived COVID-19 threat (COVID-19 RP scale); (2) stress (PSS), defined as feelings of uncontrollability and unpredictability, which are more characteristic in response to situations of lower threat imminence; and, (3) anxiety (DASS-21 Anxiety), defined as a state of high autonomic arousal, acute anxiety and panic, consistent with emotional responses to threats perceived as predictable and imminent. Of note, the PSS score was chosen over the DASS-21 Stress scale to index stress due to the stronger correlation between DASS-21 Stress and Anxiety (Fig. 4). Importantly, we verified that includ- ing DASS-21 Stress instead of PSS in the models did not change the direction of results (see Supplementary material). Although not the focus of our hypotheses, Depression was also included as a fixed effect, since it is thought to have important associations with stress and anxiety-related responses29. Including Depression in our models was critical to account for shared variance with stress and anxiety, especially given we hypothesized the opposite association between altruism and acute defensive emotions (positive) to that previously described for Depression (negative)53,54. Multicollinearity diagnostics using the Kappa statistic and Variance Inflation Factor (VIF) did not suggest problematic collinearity when stress, anxiety and depression were included in the same models (all kappas < 3.66; all VIFs < 5). Also, including Depression did not affect the direction or significance of the main findings (Supplementary material).if In addition to the fixed effects of interest, we modelled demographic variables with a presumed association with altruistic motivation, namely age, gender, socio-economic status and employment (Model 2). Briefly, (1) age has been shown before to be positively associated with prosocial tendencies55,56; (2) gender may modulate prosocial behaviour6,57, particularly in stressful situations32; and (3) socio-economic and employment status could compromise one’s ability to carry out some types of altruistic acts (e.g., money donations), and affect subjective wellbeing, which has been previously linked with altruism7. Modelling other indicators of altruism during the pandemic. Because our main measure of every- day altruism (SRA) does not specify a timeframe for the reported behaviours, we performed additional GLMM using the Help stranger item as dependent variable (Model 3). This item assessed the frequency of altruistic acts towards a stranger or charity in the last few weeks. www.nature.com/scientificreports/ This measure was chosen because: (1) it specifies a timeframe consistent with the pandemic; (2) in assessing altruism towards a stranger/charity, it was deemed a “purer” meas- ure of altruism than the items assessing frequency of altruistic actions towards relatives, friends or acquaint- ances, which can be inherently rewarding, or reciprocity-based.fi y g p y Another limitation of using the SRA is that some of its items describe actions that may be difficult to perform during a lockdown (e.g., helping a stranger with care troubles). To minimize this potential concern, we also repeated the GLMM analysis using the SRA-donations as dependent variable (Model 4). p y g p ( ) Finally, we repeated the analysis using the Prosocial Behavioral Intentions scale (PBI).hf The description of the different models tested is presented together with the respective results in the next section, to improve readability. Also, for the sake of conciseness, not all steps of the modelling were included in the main text. All modelling steps and results, code, model assumptions, as well as an analysis assessing the robustness of the effects within each week are available in Supplementary material. Correlations among measures of prosociality and defensive emotions. Finally, at an exploratory level, we used zero-order Pearson correlations to examine associations between self-reported altruism, other measures of prosocial behaviour (Prosocial Behavioral Intentions Scale, items assessing altruistic behaviours towards dif- ferent targets), and measures assessing defensive emotional states (anxiety and stress). These analyses were per- formed across all weeks (N = 600), except for the items assessing altruism towards different targets, for which only two Week 2 and 3 were available (N = 300). Correlation analyses between these items were Sidak-corrected for multiple comparisons. p p All analyses were implemented in R, and the GLMMs performed using lme4 package58 with standardized variables. The significance level was set at α < 0.05. In line with open science practices, all analytical decisions and steps are detailed in Supplementary material, and all data, study materials and code are publicly available on the OSF project page (https://osf.io/c8zhn/). Data availability All data, code and study materials are publicly available on OSF (https://osf.io/c8zhn/). All data, code and study materials are publicly available on OSF (https://osf.io/c8zhn Received: 26 April 2022; Accepted: 31 October 2022 Received: 26 April 2022; Accepted: 31 October 2022 h gf g y 9. von Dawans, B., Ditzen, B., Trueg, A., Fischbacher, U. & Heinrichs, M. Effects of acute stress on social behavior in women. Psy‑ choneuroendocrinology 99, 137–144 (2019). ( ) 8. Preston, S. D. The origins of altruism in offspring care. Psychol. Bull. 139, 1305–1341 (2013).f www.nature.com/scientificreports/ Week was modelled as a random effect, with random intercept and slope per each fixed effect (details in Supplementary material). https://doi.org/10.1038/s41598-022-23415-2 Scientific Reports \| (2022) 12:18619 \| References 1. Atrooz, F., Liu, H. & Salim, S. Stress, psychiatric disorders, molecular targets, and more. Prog. Mol. Biol. Transl. Sci. 167, 77–10 (2019). 1. Atrooz, F., Liu, H. & Salim, S. Stress, psychiatric disorders, molecular targets, and more. Prog. Mol. Biol. Transl. Sci. 167, 77–1 (2019). ( ) 2. Starcke, K. & Brand, M. Decision making under stress: A selective review. Neurosci. Biobehav. Rev. 36, 1228–1248 (2012). 3. Buchanan, T. W. & Preston, S. D. Stress leads to prosocial action in immediate need situations. Front. Behav. Neurosci. 8, (2014). 4. Feinberg, J. Psychological egoism. in Ethical Theory: An Anthology (Wiley, 2014).h 3. Buchanan, T. W. & Preston, S. D. Stress leads to prosocial action in immediate need situations. Front. Behav. Neurosci. 8, (2014 4. Feinberg, J. Psychological egoism. in Ethical Theory: An Anthology (Wiley, 2014).h 3. Buchanan, T. W. & Preston, S. D. Stress leads to prosocial action in immediate need situations. Front. Behav. Neurosci. 8, (2014). 4. Feinberg, J. Psychological egoism. in Ethical Theory: An Anthology (Wiley, 2014).h p 4. Feinberg, J. Psychological egoism. in Ethical Theory: An Anthology (Wiley, 2014).h g y g gh y gy y 5. Mattis, J. S. et al. The social production of altruism: Motivations for caring action in a low-income urban community. Am. J munity Psychol. 43, 71–84 (2009). y y ( ) 6. Sisco, M. R. & Weber, E. U. Examining charitable giving in real-world online donations. Nat. Commun. 10, 3968 (2019). h l h h l d ff b ll b d d l h y y 6. Sisco, M. R. & Weber, E. U. Examining charitable giving in real-world online donations. Nat. Commun. 10, 3968 (2019). 7. Brethel-Haurwitz, K. M. & Marsh, A. A. Geographical differences in subjective well-being predict extraordinary altruism. Psychol. y y 6. Sisco, M. R. & Weber, E. U. Examining charitable giving in real-world online donations. Nat. Commun. 10, 3968 (2019). 7 B th l H it K M & M h A A G hi l diff i bj ti ll b i di t t di lt i P h l 6. Sisco, M. R. & Weber, E. U. Examining charitable giving in real-world online donations. Nat. Commun. 10, 3968 (2019). 7. Brethel-Haurwitz, K. M. & Marsh, A. A. Geographical differences in subjective well-being predict extraordinary altruism. Psychol. Sci. 25, 762–771 (2014). 6. Sisco, M. R. & Weber, E. U. www.nature.com/scientificreports/ www.nature.com/scientificreports/ 0. Singer, N. et al. Acute psychosocial stress and everyday moral decision-making in young healthy men: The impact of cortisol Horm. Behav. 93, 72–81 (2017). 1. Tomova, L. et al. Increased neural responses to empathy for pain might explain how acute stress increases prosociality. Soc. Cogn Affect. Neurosci. 12, 401–408 (2017).ih ff 2. Buchanan, T. W., Bagley, S. L., Stansfield, R. B. & Preston, S. D. The empathic, physiological resonance of stress. Soc. Neurosci. 7 191–201 (2012). 3. Starcke, K., Polzer, C., Wolf, O. T. & Brand, M. Does stress alter everyday moral decision-making?. Psychoneuroendocrinology 36 210–219 (2011).lh 4. Zhang, Q., Ma, J. & Nater, U. M. How cortisol reactivity influences prosocial decision-making: The moderating role of sex and empathic concern. Front. Hum. Neurosci. 13, 415 (2019). p 15. Fanselow, M. S. & Lester, L. A functional behavioristic approach to aversively motivated behavior: predatory imminence as a determinant of the topography of defensive behavior. in Evolution and Learning, Evol. 185–212 (1988). 6. Blanchard, R. J. & Blanchard, D. C. Anti-predator defense as models of animal fear and anxiety. in Fear and Defence 89–108 (Har- wood Academic Publishers, 1990). 17. Mobbs, D., Headley, D. B., Ding, W. & Dayan, P. Space, time, and fear: survival computations along defensive circuits. Trends Cogn. Sci. 24, 228–241 (2020). d h d d f h f h h l h h l ( ) 8. Hamm, A. O. Fear, anxiety, and their disorders from the perspective of psychophysiology. Psychophysiology 57, e13474 (2020). 9. McNaughton, N. & Corr, P. J. A two-dimensional neuropsychology of defense: Fear/anxiety and defensive distance. Neurosci Biobehav. Rev. 28, 285–305 (2004). 0. Ehlers, M. R., Nold, J., Kuhn, M., Klingelhöfer-Jens, M. & Lonsdorf, T. B. Revisiting potential associations between brain morphol- ogy, fear acquisition and extinction through new data and a literature review. Sci. Rep. 10, 19894 (2020). 1. Perkins, A. M., Cooper, A., Abdelall, M., Smillie, L. D. & Corr, P. J. Personality and defensive reactions: Fear, trait anxiety, and threat magnification. J. Pers. 78, 1071–1090 (2010).f gi 22. Perkins, A. M. & Corr, P. J. Reactions to threat and personality: Psychometric differentiation of intensity and direction dimensions of human defensive behaviour. Behav. Brain Res. 169, 21–28 (2006).l 23. Vieira, J. B., Schellhaas, S., Enström, E. & Olsson, A. Help or flight? Increased threat imminence promotes defensive helping in humans. Proc. Biol. Sci. 287, 20201473 (2020). 4. Vieira, J. B. www.nature.com/scientificreports/ The basolateral amygdala is essential for rapid escape: A human and rodent study. Cell 175, 723-735.e16 (2018). e, P. et al. Midbrain circuits for defensive behaviour. Nature 534, 206–212 (2016). gh yg p p y ( ) 6. Tovote, P. et al. Midbrain circuits for defensive behaviour. Nature 534, 206–212 (2016). 36. Tovote, P. et al. Midbrain circuits for defensive behaviour. Nature 534, 206–212 (2016). 7. Rickenbacher, E., Perry, R. E., Sullivan, R. M. & Moita, M. A. Freezing suppression by oxytocin in central amygdala allows alternate defensive behaviours and mother-pup interactions. Elife 6, e24080 (2017).hh f 8. Preston, S. D. The evolution and neurobiology of heroism. in The Handbook of Heroism and Heroic Leadership (Taylor & Francis Routledge, 2016).f g ) 39. Ayers, J. D. et al. How is the COVID-19 pandemic affecting cooperation? (2020). 10.31234/osf.io/pk6jy. 39. Ayers, J. D. et al. How is the COVID-19 pandemic affecting cooperation? (2020). 10.31234/osf.io/pk6jy. 40 Wise T Zbozinek T D Michelini G Hagan C C & mobbs dean Changes in risk perception and protective behavior during f 40. Wise, T., Zbozinek, T. D., Michelini, G., Hagan, C. C. & mobbs, dean. Changes in risk perception and protective behavior duri first week of the COVID-19 pandemic in the United States. (2020). 10.31234/osf.io/dz428. i 1. Beadle, J. N. & de la Vega, C. E. Impact of aging on empathy: Review of psychological and neural mechanisms. Front. Psychiatry 10, 331 (2019). 2. Cutler, J. et al. Ageing disrupts reinforcement learning whilst learning to help others is preserved. bioRxiv 2020.12.02.407718 (2020). https://doi.org/10.1101/2020.12.02.407718.fi 43. Freund, A. M. & Blanchard-Fields, F. Age-related differences in altruism across adulthood: Making personal financial gain versus contributing to the public good. Dev. Psychol. 50, 1125–1136 (2014). k d l l f ff h d f ( ) g g y 44. Lockwood, P. et al. Ageing increases prosocial motivation for effort. https://doi.org/10.31234/osf.io/8c5ra (2020).f 44. Lockwood, P. et al. Ageing increases prosocial motivation for effort. https://doi.o f 45. Rosi, A., Nola, M., Lecce, S. & Cavallini, E. Prosocial behavior in aging: Which factors can explain age-related differences in social- economic decision making?. Int. Psychogeriatr. 31, 1747–1757 (2019). 6. Cutler, J., Nitschke, J. P., Lamm, C. & Lockwood, P. L. Older adults across the globe exhibit increased prosocial behavior but also greater in-group preferences. Nat. Aging 1, 880–888 (2021). greater in-group preferences. Nat. Aging 1, 880–888 (2021) 7. Sparrow, E. P., Armstrong, B. References Examining charitable giving in real-world online donations. Nat. Commun. 10, 3968 (2019). 7. Brethel-Haurwitz, K. M. & Marsh, A. A. Geographical differences in subjective well-being predict extraordinary altruism. Psychol. 6. Sisco, M. R. & Weber, E. U. Examining charitable giving in real-world online donations. Nat. Commun. 10, 3968 (2019). 7. Brethel-Haurwitz, K. M. & Marsh, A. A. Geographical differences in subjective well-being predict extraordinary altruism. Psychol. Sci 25 762–771 (2014) 8. Preston, S. D. The origins of altruism in offspring care. Psychol. Bull. 139, 1305–1341 (2013).f https://doi.org/10.1038/s41598-022-23415-2 Scientific Reports \| (2022) 12:18619 \| www.nature.com/scientificreports/ & Olsson, A. Neural defensive circuits underlie helping under threat in humans. eLife. 11, e78162. https://doi.org/10 7554/eLife.78162 (2022).h 25. Philippe Rushton, J., Chrisjohn, R. D. & Cynthia Fekken, G. The altruistic personality and the self-report altruism scale. Personal. Individ. Differ. 2, 293–302 (1981). f 26. Wise, T., Zbozinek, T. D., Michelini, G., Hagan, C. C. & Mobbs, D. Changes in risk perception and self-reported protective behav iour during the first week of the COVID-19 pandemic in the United States. R. Soc. Open Sci. 7, 200742 (2020). 26. Wise, T., Zbozinek, T. D., Michelini, G., Hagan, C. C. & Mobbs, D. Changes in risk perception and self reported protective behav iour during the first week of the COVID-19 pandemic in the United States. R. Soc. Open Sci. 7, 200742 (2020). i 27. Cohen, S., Kamarck, T. & Mermelstein, R. A global measure of perceived stress. J. Health Soc. Behav. 24, 385–396 (1983). g p 28. Lovibond, S. & Lovibond, P. Manual for the Depression Anxiety Stress Scales (Psychology Foundation, 1995).h tical relationship between anxiety and depression. Am. J. Psychiatr 29. Kalin, N. H. The critical relationship between anxiety and depression. Am. J. Psychiatry 177, 365–367 (2020).h h y y y 30. von Dawans, B., Fischbacher, U., Kirschbaum, C., Fehr, E. & Heinrichs, M. The social dimension of stress reactivity: Acute increases prosocial behavior in humans. Psychol. Sci. 23, 651–660 (2012).t y 31. Wolf, O. T. et al. Enhanced emotional empathy after psychosocial stress in young healthy men. Stress 18, 631–637 (2015).il O. T. et al. Enhanced emotional empathy after psychosocial stress i p yt p y y g y 32. Taylor, S. E. et al. Biobehavioral responses to stress in females: Tend-and-befriend, not fight-or-flight. Psychol. Rev. 107, 41 (2000).fi ( ) 33. Taylor, S. E. Tend and befriend: Biobehavioral bases of affiliation under stress. Curr. Dir. Psychol. Sci. (2016). fi 4. Neumann, I. D. & Landgraf, R. Balance of brain oxytocin and vasopressin: Implications for anxiety, depression, and social behaviors Trends Neurosci. 35, 649–659 (2012).h 5. Terburg, D. et al. The basolateral amygdala is essential for rapid escape: A human and rodent study. Cell 175, 723-735.e16 (2018) 35. Terburg, D. et al. The basolateral amygdala is essential for rapid escape: A human and rodent study. Cell 175, 723-735.e16 (2018). 36 Tovote P et al Midbrain circuits for defensive behaviour Nature 534 206 212 (2016) urg, D. et al. www.nature.com/scientificreports/ www.nature.com/scientificreports/ 57. Willer, R., Wimer, C. & Owens, L. A. What drives the gender gap in charitable giving? Lower empathy leads men to give less to poverty relief. Soc. Sci. Res. 52, 83–98 (2015). 58. Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1–48 (2015). 57. Willer, R., Wimer, C. & Owens, L. A. What drives the gender gap in charitable giving? Lower empathy leads men to give less to poverty relief. Soc. Sci. Res. 52, 83–98 (2015).ft p y 8. Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1–48 (2015). Author contributions B.V. designed the study, analysed data, and drafted the manuscript; S.R.P. analysed data; S.J. created study materials; A.A.M. and A.O. edited the manuscript. Funding g Open access funding provided by Karolinska Institute. This research was supported by the Knut and Alice Wal- lenberg Foundation (KAW 495 2014.0237), and a Consolidator Grant (2018-00877) from the Swedish Research Council (Vetenskapsrådet) to Andreas Olsson. www.nature.com/scientificreports/ A., Fiocco, A. J. & Spaniol, J. Acute stress and altruism in younger and older adults. Psychoneuroen docrinology 100, 10–17 (2019).f gy 48. DeBruine, L. & Barr, D. J. Understanding mixed effects models through data simulation. (2019). https://doi.org/10.31234/osf.io/ xp5cy. p y 49. Green, P. & MacLeod, C. J. SIMR: An R package for power analysis of generalized linear mixed models by simulation. Methods Ecol. Evol. 7, 493–498 (2016).h 50. Yarkoni, T. The Generalizability Crisis. (2019). https://doi.org/10.31234/osf.io/jqw35.h h y p g jq 51. Rushton, J. P., Chrisjohn, F. R. D. & Fekken, G. C. The Altruist Ic Personal Ity and the Self-Report Altruism Scale*. (1981). l T l Th d l f l b h l l h 51. Rushton, J. P., Chrisjohn, F. R. D. & Fekken, G. C. The Altruist Ic Personal Ity and the Self-Rep 52. Baumsteiger, R. & Siegel, J. T. Measuring prosociality: The development of a prosocial behavioral intentions scale. J. Pers. Assess. 101, 305–314 (2019).h 53. Frick, A., Thinnes, I., Hofmann, S. G., Windmann, S. & Stangier, U. Reduced social connectedness and compassion toward close others in patients with chronic depression compared to a non-clinical sample. Front. Psychiatry 12, 608607 (2021). 4. Kupferberg, A., Bicks, L. & Hasler, G. Social functioning in major depressive disorder. Neurosci. Biobehav. Rev. 69, 313–332 (2016) 54. Kupferberg, A., Bicks, L. & Hasler, G. Social functioning in major depressive disorder. Neurosci. Biobehav. Rev. 69, 313–332 (2016). 55. Matsumoto, Y., Yamagishi, T., Li, Y. & Kiyonari, T. Prosocial behavior increases with age across five economic games. PLoS ONE 11, e0158671 (2016). 6. Sze, J. A., Gyurak, A., Goodkind, M. S. & Levenson, R. W. Greater emotional empathy and prosocial behavior in late life. Emot Wash. DC 12, 1129–1140 (2012). https://doi.org/10.1038/s41598-022-23415-2 Scientific Reports \| (2022) 12:18619 \| Competing interests h p g The authors declare no competing interests. Additional informationh Additional information Supplementary Information The online version contains supplementary material available at https://doi.org/ 10.1038/s41598-022-23415-2. Supplementary Information The online version contains supplementary material available at https://doi.org/ 10.1038/s41598-022-23415-2. Correspondence and requests for materials should be addressed to J.B.V. Correspondence and requests for materials should be addressed to J.B.V. Reprints and permissions information is available at www.nature.com/reprints. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. © Crown 2022 https://doi.org/10.1038/s41598-022-23415-2 Scientific Reports \| (2022) 12:18619 \|
https://openalex.org/W4253041945	https://europepmc.org/articles/pmc3986083?pdf=render	English	null	Correction: Exocytosis of ATP From Astrocytes Modulates Phasic and Tonic Inhibition in the Neocortex	PLoS biology	2,014	cc-by	882	The PLOS Biology Staff The authors noticed an error in Panel 2A of Figure 2, and they have now provided a corrected version here. Figure 2. Ca2+-dependent release of ATP can be evoked in the cortical astrocytes of the wild-type but not of the dn-SNARE mice. Release of ATP from cortical astrocytes of wild-type (A, B) and dn- SNARE mice (C) was detected using the sniffer cells as described in Figure 1. (A) Elevation of cytosolic Ca2+ level was elicited in the astrocytes by UV uncaging and by rapid application of the agonist of PAR-1 metabotropic receptor TFLLR (10 mM) or the agonist of glutamate ionotropic receptor NMDA (20 mM). (B) Inhibition of vacuolar H-ATPase in the astrocytes with Bafilomycin A1 (1 mM for 2 h) dramatically decreased both the amplitude and frequency of phasic currents. (C) Elevation of the Ca2+ level in any of the dn-SNARE astrocytes did not lead to activation of phasic purinergic currents in the sniffer cell. Inlays in (A–C) show examples of individual phasic currents recorded at moments indicated; scale bars are 50 ms and 10 pA. (D) The amplitude and decay time distributions of purinergic currents recorded in the HEK293-P2X2 cells after stimulation of the astrocytes; data were pooled for number of experiments indicated in (E). The grey dotted line shows the best fit of quantal model to the distribution of UV-activated currents. (E) The pooled data (mean 6 SD for indicated numbers of experiments) on net release of ATP were assessed as total charge transferred by spontaneous currents in the sniffer cell. The effects of bafylomicin and dn-SNARE expression on net charge transferred by purinergic currents were statistically significant at p = 0.005 (two- population t test). doi:10 1371/journal pbio 1001747 g002 Figure 2. Ca2+-dependent release of ATP can be evoked in the cortical astrocytes of the wild-type but not of the dn-SNARE mice. Release of ATP from cortical astrocytes of wild-type (A, B) and dn- SNARE mice (C) was detected using the sniffer cells as described in Figure 1. (A) Elevation of cytosolic Ca2+ level was elicited in the astrocytes by UV uncaging and by rapid application of the agonist of Figure 2. Ca2+-dependent release of ATP can be evoked in the cortical astrocytes of the wild-type but not of the dn-SNARE mice. Correction Correction: Exocytosis of ATP From Astrocytes Modulates Phasic and Tonic Inhibition in the Neocortex Correction Correction: Exocytosis of ATP From Astrocytes Modulates Phasic and Tonic Inhibition in the Neocortex Correction Correction: Exocytosis of ATP From Astrocytes Modulates Phasic and Tonic Inhibition in the Neocortex The PLOS Biology Staff Release of ATP from cortical astrocytes of wild-type (A, B) and dn- SNARE mice (C) was detected using the sniffer cells as described in Figure 1. (A) Elevation of cytosolic Ca2+ level was elicited in the astrocytes by UV uncaging and by rapid application of the agonist of PAR-1 metabotropic receptor TFLLR (10 mM) or the agonist of glutamate ionotropic receptor NMDA (20 mM). (B) Inhibition of vacuolar H-ATPase in the astrocytes with Bafilomycin A1 (1 mM for 2 h) dramatically decreased both the amplitude and frequency of phasic currents. (C) Elevation of the Ca2+ level in any of the dn-SNARE astrocytes did not lead to activation of phasic purinergic currents in the sniffer cell. Inlays in (A–C) show examples of individual phasic currents recorded at moments indicated; scale bars are 50 ms and 10 pA. (D) The amplitude and decay time distributions of purinergic currents recorded in the HEK293-P2X2 cells after stimulation of the astrocytes; data were pooled for number of experiments indicated in (E). The grey dotted line shows the best fit of quantal model to the distribution of UV-activated currents. (E) The pooled data (mean 6 SD for indicated numbers of experiments) on net release of ATP were assessed as total charge transferred by spontaneous currents in the sniffer cell. The effects of bafylomicin and dn-SNARE expression on net charge transferred by purinergic currents were statistically significant at p = 0.005 (two- population t test). Figure 2. Ca2+-dependent release of ATP can be evoked in the cortical astrocytes of the wild-type but not of the dn-SNARE mice. Release of ATP from cortical astrocytes of wild-type (A, B) and dn- ff Citation: The PLOS Biology Staff (2014) Correction: Exocytosis of ATP From Astrocytes Modulates Phasic and Tonic Inhibition in the Neocortex. PLoS Biol 12(4): e1001857. doi:10.1371/journal.pbio.1001857 p p doi:10.1371/journal.pbio.1001747.g002 p p doi:10.1371/journal.pbio.1001747.g002 1. Lalo U, Palygin O, Rasooli-Nejad S, Andrew J, Haydon PG, et al. (2014) Exocytosis of ATP From Astrocytes Modulates Phasic and Tonic Inhibition in the Neocortex. PLoS Biol 12(1): e1001747. doi:10.1371/journal.pbio.1001747 Correction: Exocytosis of ATP From Astrocytes Modulates Phasic and Tonic Inhibition in the Neocortex The PLOS Biology Staff Citation: The PLOS Biology Staff (2014) Correction: Exocytosis of ATP From Astrocytes Modulates Phasic and Tonic Inhibition in the Neocortex. PLoS Biol 12(4): e1001857. doi:10.1371/journal.pbio.1001857 Published April 14, 2014 Published April 14, 2014 Copyright: 2014 The PLOS Biology Staff. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. PLOS Biology \| www.plosbiology.org 1 April 2014 \| Volume 12 \| Issue 4 \| e1001857 April 2014 \| Volume 12 \| Issue 4 \| e1001857
https://openalex.org/W4254447024	https://journals.abcjournal.aosis.co.za/index.php/abc/article/download/104/104	English	null	Asteraceae	Bothalia	2,013	cc-by	2,146	Bothalia 43,2 (2013) Bothalia 43,2 (2013) 227 SENECIO PSEUDOLONGIFOLIUS, A NEW NAME FOR THE MISAPPLIED S. LINIFOLIUS FIGURE 1.—Lectotype of Senecio linifolius L. (G00360056). Senecio linifolius L. (1759) was based on a South African collection in the herbarium of J. Burman (Wijnands 1983). Linnaeus (1763) later published the replacement name S. longifolius L. for the species to avoid homonymy with his new combination S. linifo- lius (L.) L. (1763) for a Mediterranean species based on Solidago linifolia L. (1753). These two nomenclatural steps transgress two articles in the International Code of Nomenclature (McNeill et al. 2012): the replacement S. longifolius L. (1763) is superfluous and illegitimate (Art. 52), and the combination Senecio linifolius (L.) L. (1763) is a heterotypic later homonym of S. linifolius L. (1759) and therefore nomenclaturally illegitimate (Art. 53). The name S. linifolius L. (1759) is therefore cor- rectly applied to the South African species. For the Med- iterranean species, the earliest available name appears to be Senecio malacitanus Huter (1905). The illegitimate Senecio longifolius L. (1763) has, however, been widely taken up and used in the taxo- nomic literature (e.g. Bergius 1767; Willdenow 1803; Candolle 1838; Schultz 1844; Harvey 1865; Juel 1918; Bond & Goldblatt 1984) and has also made the transi- tion to the popular literature (Batten & Bokelman 1966). It is only recently that the name S. linifolius has been applied to the South African taxon (Manning & Gold- blatt 2012). As currently understood, S. linifolius is a small shrub from the Eastern Cape, with the leaves clus- tered apically on basally leafless stems and abruptly dif- ferentiated from the inflorescence bracts. The terminal synflorescences are distinctly pedunculate, loose cymes with the capitula subtended by four or five supplemen- tary bracts (bracteoles) and containing five ligulate flo- rets each. FIGURE 1.—Lectotype of Senecio linifolius L. (G00360056). Senecio linifolius L. (1759) was lectotypified by Wijnands (1983) against a specimen in the Burman Herbarium at the Conservatoire et Jardin botaniques de la Ville de Genève (G) (Figure 1). This specimen com- prises a single flowering branch. The leaves are scattered along the entire length of the stem and grade gradually into the inflorescence bracts. The capitula are subtended by numerous supplementary bracts (bracteoles) and each contains more than five ligulate florets. In none of these respects does it match the current usage of S. linifolius L., which appears to have been established by Harvey (1865), and whose description for the Flora capensis matches the current concept exactly. SENECIO PSEUDOLONGIFOLIUS, A NEW NAME FOR THE MISAPPLIED S. LINIFOLIUS Harvey (1865) also included the name Cineraria filifolia Thunb. (1800) as a synonym of S. longifolius L. The collection under this respects does it match the current usage of S. linifolius L., which appears to have been established by Harvey (1865), and whose description for the Flora capensis matches the current concept exactly. Harvey (1865) also included the name Cineraria filifolia Thunb. (1800) as a synonym of S. longifolius L. The collection under this 228 Bothalia 43,2 (2013) name in Thunberg’s herbarium (UPS-THUNB19915) is consistent with this interpretation of the name (Figure 2). FIGURE 2.—Lectotype of Cineraria filifolia [Senecio pseudolongifo- lius] (Thunberg s.n. UPS-THUNB19915). It is evident, therefore, that the name S. linifolius L. has been incorrectly applied since Harvey’s (1865) time. From the Burman lectotype, it appears to us to closely resemble S. burchellii DC., S. inaequidens DC. or an allied species, but this group is in need of a comprehen- sive taxonomic revision and its precise identity remains to be established. As an aside, the name S. odontophyllus Wallich ex C.Jeffrey (1992), published by Jeffrey (1992) as a replacement name for the Indian species until then known under the illegitimate later homonym S. linifolius (Wallich ex DC.) C.B.Clarke (1876) [non S. linifolius L. (1759)], was mistakenly applied to S. linifolius L. by Goldblatt & Manning (2000) but later corrected (Gold- blatt & Manning 2010). What, then, is the correct name for the Eastern Cape taxon currently treated as S. linifolius? The transfer to Senecio of the only available name, Cineraria filifolia Thunb. (1800), is unfortunately blocked by the exist- ence of S. filifolius P.J Bergius (1767). C.H. Schultz (1844) included the manuscript name S. pseudolongi- folius as a synonym of S. longifolius in his Compositae Kraussianae, and this name remains invalidly published (MacNeill et al. 2012: Art. 34.1). The relevant mate- rial to which this name applies is a collection made by C.F. Krauss during April and June 1839 near Uitenhage in Eastern Cape under the number 552. We have located two duplicates of this collection, one at TUB, compris- ing four fragments, and the other at MO comprising a single fragment. Krauss’s collection is consistent with the current interpretation of S. linifolius in both mor- phology and locality, making it clear that Schultz inter- preted the species as currently understood. We therefore validate the name S. pseudolongifolius as a replacement name for Cineraria filifolia Thunb. (1800). SENECIO PSEUDOLONGIFOLIUS, A NEW NAME FOR THE MISAPPLIED S. LINIFOLIUS FIGURE 2.—Lectotype of Cineraria filifolia [Senecio pseudolongifo- lius] (Thunberg s.n. UPS-THUNB19915). Senecio pseudolongifolius Sch.Bip. ex J.Calvo, nom. nov. pro Cineraria filifolia Thunb., Prodr. Pl. Cap.: 154 (1800) [non Senecio filifolius P.J.Bergius (1767)]. Type: South Africa, Cap. bonae Spei, without precise locality or date, C.P. Thunberg s.n. (UPS-THUNB19915, lecto., designated here). Senecio linifolius [as ‘linifolia’] L., Systema naturae, ed. 10, 2: 1215 (1759). S. longifolius L.: 1222 (1763), nom. illegit. superfl. Type: ‘Habitat ad Cap. b. spei.’, herb. Burmann [G00360056, lecto., designated by Wijnands: 83 (1983)]. Senecio linifolius [as ‘linifolia’] L., Systema naturae, ed. 10, 2: 1215 (1759). S. longifolius L.: 1222 (1763), nom. illegit. superfl. Type: ‘Habitat ad Cap. b. spei.’, herb. Burmann [G00360056, lecto., designated by Wijnands: 83 (1983)]. Senecio pseudolongifolius Sch.Bip. in Flora 27(2): 699 (1844), nom. inval. pro syn. S. longifolius L. Senecio pseudolongifolius Sch.Bip. in Flora 27(2): 699 (1844), nom. inval. pro syn. S. longifolius L. Senecio malacitanus Huter in Oesterreichische bota- nische Zeitschrift; gemeinütziges Organ für Botanik 55: 402 (1905). Type: Herb. Huter. Senecio malacitanus Huter in Oesterreichische bota- nische Zeitschrift; gemeinütziges Organ für Botanik 55: 402 (1905). Type: Herb. Huter. Senecio longifolius sensu Harv. (1865), non L. (1759). Senecio odontophyllus sensu Goldblatt & Man- ning (2000), non [Wallich ex] C.Jeffrey (1992). Senecio odontophyllus sensu Goldblatt & Man- ning (2000), non [Wallich ex] C.Jeffrey (1992). Solidago linifolia L.: 881 (1753). Senecio lini- folius (L.) L.: 1220 (1763), hom. illegit. [non Senecio linifolius L. (1759)]. Type: ‘Iacobaea Linifolio His- panica et Ital.’, illustration in Boccone: t. 49 (1697) [lecto., designated by Wijnands: 83 (1983)]. ACKNOWLEDGEMENTS We are grateful to Cornelia Dilger-Endrulat (TUB), Mats Hjertson (UPS), and the staff of the Herbarium at the Conservatoire et Jardin botaniques de la Ville de Genève for kindly sending photographs of type mate- rial. Special thanks to Joan Pedrol who recorded the first naturalized population of S. pseudolongifolius in the Iberian Peninsula, which led to our interest on this mat- ter. This work was financed by the Flora iberica project (CGL2011-28613-C03-01). Senecio odontophyllus Wallich [Aster odontophyllus Wallich, nom. nud. in Cat. no. 3285 comp. no. 395] ex C.Jeffrey in Kew Bulletin 47(1): 95 (1992), as a replace- ment name for S. linifolius (Wallich ex DC.) C.B.Clarke: 202 (1876), non S. linifolius L. (1759). Doronicum lini- folius Wallich [Aster linifolius Wallich, nom. nud in herb.] ex DC.: 322 (1838). Type: India, ‘Khasia’, Herb. Wallich. 229 Bothalia 43,2 (2013) ** Compton Herbarium, South African National Biodiversity Institute, Private Bag X7, 7735 Claremont, Cape Town; Research Centre for Plant Growth and Development, School of Life Sciences, University of KwaZulu-Natal, Pietermaritzburg, Private Bag X01, Scottsville 3209, South Africa. * Real Jardín Botánico-CSIC, Plaza de Murillo 2, E-28014 Madrid, Spain. E-mail: calvocasas@gmail.com. REFERENCES ** Compton Herbarium, South African National Biodiversity Institute, Private Bag X7, 7735 Claremont, Cape Town; Research Centre for Plant Growth and Development, School of Life Sciences, University of KwaZulu-Natal, Pietermaritzburg, Private Bag X01, Scottsville 3209, South Africa. JUEL, H.O. 1918. Plantae Thunbergianae. Akademiska bokhandeln, Uppsala & Otto Harrassowitz, Leipzig. JEFFREY, C. 1992. The tribe Senecioneae (Compositae) in the Mas- carene Islands with an annotated world check-list of the genera of the tribe. Kew Bulletin 47,1: 95. Ethesia tanquana Mart.-Azorín & M.B.Crespo (Mar- tínez-Azorín & Crespo 2012), Stellarioides chartacea Mart.-Azorín et al. (Martínez-Azorín et al. 2013a) and Trimelopter craibii Mart.-Azorín et al. (Martínez-Azorín et al. 2013b) (Hyacinthaceae: Ornithogaloideae) are recently described novelties from southern Africa. They were named following the generic classification of the subfamily proposed by Martínez-Azorín et al. (2011), one of two recent classifications for the subfamily deri- ved from substantively the same molecular phylogeny. The alternative classification adopted by Manning et al. (2009) is the preferred option among southern African botanists, where it has been adopted by the major her- baria in the country (e.g. BOL, NBG, PRE). We provide combinations for the three new taxa in this classification to permit their integration into collections and literature. Ornithogalum recurvum Oberm. (1971) was treated as a synonym of O. stapffii Schinz until resurrected by Martínez-Azorín & Crespo (in press). We provide the necessary combination in Albuca. We also provide the new combination Albuca comosa (Welw. ex Baker) J.C.Manning & Goldblatt, based on Urginea comosa Welw. ex Baker (Baker 1874), as the correct name for the species known until now as either Ornithogalum pulchrum Schinz or Albuca pulchra (Schinz) J.C.Manning & Goldblatt. The transfer of this basionym to Ornithogalum was precluded by the exis- tence of the earlier Ornithogalum comosum L. (1753), but we overlooked its availability in Albuca until alerted to this fact by Martínez-Azorín & Crespo (2013). Finally, we recommend that all future descriptions of species and any new combinations in Ornithogaloideae include an optional combination in the alternative clas- sification (i.e. Manning et al. 2009 or Martínez-Azorín et al. 2011) to obviate the necessity for separate publica- tion of these combinations, and to facilitate their direct integration into existing collections and databases. Albuca chartacea (Mart.-Azorín et al.) J.C.Manning & Goldblatt, comb. nov. [subg. Urophyllon (Salisb.) J.C.Manning & Goldblatt]. Stellarioides chartacea Mart.-Azorín et al. in Phytotaxa 85: 2 (2013). Albuca craibii (Mart.-Azorín et al.) J.C.Manning & Goldblatt, comb. REFERENCES LINNAEUS, C. 1753. Species plantarum. Salvius, Stockholm. LINNAEUS, C. 1759. Systema naturae, ed. 10. Salvius, Stockho BATTEN, A. & BOKELMANN, H. 1966. Wild flowers of the Eastern Cape Province. Books of Africa, Cape Town. LINNAEUS, C. 1763. Species plantarum, ed. 2. Salvius, Stockholm. MANNING, J.C. & GOLDBLATT, P. 2012. Plants of the Greater Cape Floristic Region 1: the Core Cape Flora. Strelitzia 29. South African National Biodiversity Institute, Pretoria. p p BERGIUS, P.J. 1767. Descriptiones plantarum ex Capite Bonae Spei. Salvius, Stockholm. , BOCCONE, P. 1697. Museo di piante rare della Sicilia, Malta, Corsica, Italia, Piemonte, e Germania. Per Io: Baptista Zuccato, Venetia. MCNEILL, J., BARRIE, F.R., BUCK, W.R., DEMOULIN, H.M., GREUTER, W., HAWKSWORTH, D.L., HERENDEEN, P.S., KNAPP, S., MARHOLD, K., PRADO, J., PRUD’HOMME van REINE, W.F., SMITH, G.F., WIERSEMA, J.H. & TURLAND, N.J. 2012. International Code of Nomenclature for algae, fungi, and plants (Melbourne Code). Regnum Vegetabile 154. Koeltz Scientific Books, Königstein. Italia, Piemonte, e Germania. Per Io: Baptista Zuccato, Vene BOND, P. & GOLDBLATT, P. 1984. Plants of the Cape Flora: a des- criptive catalogue. Journal of South African Botany Supple- mentary Volume 13. National Botanic Gardens of South Africa, Kirstenbosch. CANDOLLE, de, A.P. 1838. Prodromus systematis naturalis regni vegetabilis 6. Sumptibus Sociorum Treuttel et Würtz, Paris. i g SCHULTZ, C.H. 1844. Compositae Kraussianae. Flora 27,2: 667–702. CLARKE, C.B. 1876. Compositae indica. Thacker, Spink & Co., Cal- cutta; Thacker, Viking & Co., Bombay; W. Thacker & Co., Lon- don. THUNBERG, C.P. 1800. Prodromus plantarum capensium, pars poste- rior. J.F. Edman, Uppsala. WIJNANDS, O. 1983. The botany of the Commelins. A.A. Balkema, Rotterdam. GOLDBLATT, P. & MANNING, J.C. 2000. Cape plants: a conspectus of the Cape flora of South Africa. Strelitzia 9. National Botanical Institute, Cape Town & Missouri Botanical Garden, St. Louis. WILLDENOW. 1803. Species plantarum, ed. 4. 3(3). G.C. Nauk, Ber- lin. p GOLDBLATT, P. & MANNING, J.C. 2010. New synonyms and a new name in Asteraceae: Senecioneae from the southern African win- ter rainfall region. Bothalia 40,1: 37–46. J. CALVO, J. MANNING, F. MUÑOZ GARMENDIA & C. AEDO* HARVEY, W.H. 1865. Compositae. In W.H. Harvey & O.W. Sonder (eds.), Flora capensis 3: 44–530. Hodges, Smith & Co., Dublin; I.C. Juta, Cape Town. * Real Jardín Botánico-CSIC, Plaza de Murillo 2, E-28014 Madrid, Spain. E-mail: calvocasas@gmail.com. p HUTER, R. 1905. Herbar-Studien. Oesterreichische botanische Zeitschrift; gemeinütziges Organ für Botanik 55,10: 400–406. REFERENCES nov. [subg. Monarchos (U.Müll.- Doblies & D.Müll.-Doblies) J.C.Manning & Goldblatt]. Trimelopter craibii Mart.-Azorín et al. in Phytotaxa 87: 52 (2013). Albuca comosa (Welw. ex Baker) J.C.Manning & Goldblatt, comb. nov. [subg. Namibiogalum (U.Müll.- Doblies & D.Müll.-Doblies) J.C.Manning & Goldblatt]. Urginea comosa Welw. ex Baker in Transactions of the Linnaean Society London, Botany 1: 247 (1878). HYACINTHACEAE FIVE NEW COMBINATIONS IN ORNITHOGALOIDEAE IN SOUTHERN AFRICA AND A RECOMMENDATION FOR OPTIONAL COMBINATIONS IN THE SUBFAMILY
https://openalex.org/W3032305800	https://pure.eur.nl/ws/files/48371397/Repub_130950_O-A.pdf	English	null	Metabolomics, machine learning and immunohistochemistry to predict succinate dehydrogenase mutational status in phaeochromocytomas and paragangliomas	Journal of pathology	2,020	cc-by	8,364	Journal of Pathology J Pathol 2020; 251: 378–387 Published online 1 July 2020 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/path.5472 Journal of Pathology J Pathol 2020; 251: 378–387 Published online 1 July 2020 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/path.5472 ORIGINAL PAPER ORIGINAL PAPER Metabolomics, machine learning and immunohistochemistry to predict succinate dehydrogenase mutational status in phaeochromocytomas and paragangliomas E-mail: susan.richter@uniklinikum-dresden.de Metabolomics, machine learning and immunohistochemistry to predict succinate dehydrogenase mutational status in phaeochromocytomas and paragangliomas Paal W Wallace1 , Catleen Conrad1, Sascha Brückmann2, Ying Pang3, Eduardo Caleiras4, Masanori Murakami5, Esther Korpershoek6, Zhengping Zhuang7, Elena Rapizzi8, Matthias Kroiss9, Volker Gudziol10,11, Henri JLM Timmers12, Massimo Mannelli8, Jens Pietzsch14,15, Felix Beuschlein5,16, Karel Pacak3, Mercedes Robledo17, Barbara Klink18,19, Mirko Peitzsch1, Anthony J Gill20,21,22, Arthur S Tischler23, Ronald R de Krijger24,25, Thomas Papathomas26, Daniela Aust27, Graeme Eisenhofer1,13 and Susan Richter1* 1 Institute of Clinical Chemistry and Laboratory Medicine, University Hospital Carl Gustav Carus, Medical Faculty Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany 2 Institute of Pathology, University Hospital Carl Gustav Carus, Medical Faculty Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany 3 Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD, USA 4 Histopathology Core Unit, Spanish National Cancer Research Centre (CNIO), Calle de Melchor Fernández Almagro, Madrid, Spain 5 Medizinische Klinik and Poliklinik IV Ludwig Maximilians Universität München Munich Germany 2 Institute of Pathology, University Hospital Carl Gustav Carus, Medical Faculty Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany 3 Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD, USA 4 Histopathology Core Unit, Spanish National Cancer Research Centre (CNIO), Calle de Melchor Fernández Almagro, Madrid, Spain 5 6 Department of Pathology, Erasmus MC-University Medical Center Rotterdam, Rotterdam, The Netherlan 6 Department of Pathology, Erasmus MC-University Medical Center Rotterdam, Rotterdam, The Netherlands 7 Surgical Neurology Branch, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesd 7 Surgical Neurology Branch, National Institute of Neurological Disorders and Stroke, National Institutes of Health 8 Department of Experimental and Clinical Medicine, University of Florence, Florence, Italy 8 Department of Experimental and Clinical Medicine, University of Florence, Florence, Italy tal and Clinical Medicine, University of Florence, Florence, Italy 9 Department of Internal Medicine, Division of Endocrinology, University Hospital, University o 9 Department of Internal Medicine, Division of Endocrinology, University Hospital, University of Würzburg, Würzburg, Germany 0 10 Klinik für Hals-Nasen-Ohrenheilkunde, Kopf- und Hals-Chirurgie, Plastische Op Lehrkrankenhaus der Technischen Universität Dresden, Dresden, Germany f pf g Lehrkrankenhaus der Technischen Universität Dresden, Dresden, Germany Lehrkrankenhaus der Technischen Universität Dresden, Dresden, Germany Departments of Otorhinolaryngology, University Hospital Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany 11 Departments of Otorhinolaryngology, University Hospital Carl Gustav Carus, Technische Universität Dresd 12 ents of Otorhinolaryngology, University Hospital Carl Gustav Caru 12 Department of Internal Medicine, Radboud University Medical Centre, Nijmegen, The Netherlands 12 Department of Internal Medicine, Radboud University Medical Centre, Nijmegen, The Netherlands 13 Department of Medicine III, University Hospital Dresden, Dresden, Germany 13 Department of Medicine III, University Hospital Dresden, Dresden, Germany nt of Medicine III, University Hospital Dresden, Dresden, Germany 13 Department of Medicine III, University Hospital Dresden, Dresden, Germany 14 14 Department of Radiopharmaceutical and Chemical Biology, Institute of Radiopharmaceutical Cancer Research, Helmholtz-Zentrum Dresden- Rossendorf, Dresden, Germany 15 15 Faculty of Chemistry and Food Chemistry, School of Science, Technische Universität Dresden, Dresden, Germany 16 15 Faculty of Chemistry and Food Chemistry, School of Science, Technische Universität Dresden, Dresden, Germany 16 16 Department for Endocrinology, Diabetology and Clinical Nutrition, UniversitätsSpital Zürich, Zurich, Switzerland 16 Department for Endocrinology, Diabetology and Clinical Nutrition, UniversitätsSpital Zürich, Zurich, Switzerland 17 Hereditary Endocrine Cancer Group, CNIO, Madrid, Spain and Centro de Investigación Biomédica en Red de Enfermedades Raras (CIBERER), Madrid, Spain 18 Institute for Clinical Genetics, Medical Faculty Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany 18 Institute for Clinical Genetics, Medical Faculty Carl Gustav Carus, Technische Universität Dresden, Dresden, Ger 19 Department of Genetics, Laboratoire National de Santé, Dudelange, Luxembourg 19 Department of Genetics, Laboratoire National de Santé, Dudelange, Luxembourg 20 Royal North Shore Hospital, Cancer Diagnosis and Pathology Group, Kolling Institute of Medical Research, Sydney, Australia 21 Royal North Shore Hospital, Cancer Diagnosis and Pathology Group, Kolling Institute of Medical Research, Sydney, Australia 21 School of Medicine, University of Sydney, Sydney, Australia 21 School of Medicine, University of Sydney, Sydney, Australia 22 21 School of Medicine, University of Sydney, Sydney, Australia 22 NSW Health Pathology, Department of Anatomical Pathology, Royal North Shore Hospital, St Leonards, Australia 23 23 Department of Pathology and Laboratory Medicine, Tufts University School of Medicine, Boston, MA, USA 24 Department of Pathology, University Medical Center Utrecht, Utrecht, The Netherlands 26 Institute of Metabolism and Systems Research, University of Birmingham, Edgbaston, Birmingham, UK 27 26 Institute of Metabolism and Systems Research, University of Birmingham, Edgbaston, Birmingham, UK 27 Correspondence to: S Richter, Institute of Clinical Chemistry and Laboratory Medicine, University Hospital Carl Gustav Carus, Fetscherstraße 74, 01307 Dresden, Germany. Introduction Another method to assess SDH functionality is based on measurements of Krebs cycle metabolites by liquid chromatography–mass spectrometry (LC–MS/MS), the same instrument also now used for biochemical diagno- sis of PPGL [11,12]. Metabolite proﬁling assesses the functionality of SDH directly at the catalytic level by measuring the precursor succinate and the product fuma- rate. The ratio of these two metabolites, the tissue succi- nate to fumarate ratio (SFR), can predict SDHx mutations with a high sensitivity and speciﬁcity in PPGL and is also now being applied to other tumours [11,13]. Mitochondrial enzymes, such as succinate dehydroge- nase (SDH), or complex II of the respiratory chain, play a central role in energy homeostasis within the cell. The complex is made up of four subunits (SDHA, SDHB, SDHC, and SDHD) and assembly is assisted by several factors, including SDHAF2. Mutations of genes encod- ing these proteins can result in phaeochromocytomas and paragangliomas (PPGLs), gastrointestinal stromal tumours, renal cell carcinomas, and pituitary adenomas [1]. The mutations occur almost exclusively in the germ- line, leaving the patient and potentially family members at lifelong risk for disease. Mutations in SDHB in partic- ular predispose to metastatic PPGL and are associated with increased mortality [2,3]. Endocrine guidelines therefore advise that genetic testing should be offered to all patients with PPGL [4]. So far, there has been no formal comparison of the two methods for predicting SDHx mutational status. The nature of the techniques is different, one involving func- tional assessment of enzyme activity versus histological information about the presence of protein, but both offer- ing complementary potential. Based on the ability of machine learning to recognise patterns in data in a way the human mind is not trained [14,15], we also investi- gated whether such an approach can improve predictions from metabolite data beyond the currently used SFR [11,12]. Since one of the disadvantages of SDHB-IHC relates to the subjective nature of image interpretation, we further assessed whether local pathologists from dif- ferent institutions scored slides differently from investi- gators experienced in the method (referred to as experts). An overall goal of the study was to explore how SDHB-IHC and metabolite proﬁling might be use- ful for streamlining diagnostic procedures for patients and their families suffering from PPGLs due to SDH impairment. Predicting succinate dehydrogenase mutations in PPGLs 379 machine learning algorithms to metabolite proﬁles improved predictive ability over that of the SFR, in particular for hard-to-interpret cases of head and neck paragangliomas (AUC 0.9821 versus 0.9613, p = 0.044). Importantly, the combination of metabolite proﬁling with SDHB-IHC has complementary utility, as SDHB-IHC correctly classiﬁed all but one of the false negatives from metabolite proﬁling strategies, while metabolite proﬁling correctly classiﬁed all but one of the false negatives/positives from SDHB-IHC. From 186 tumours with conﬁrmed status of SDHx variant pathogenicity, the combination of the two methods resulted in 185 correct predictions, highlighting the beneﬁts of both strategies for patient management. © 2020 The Authors The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. Keywords: mass spectrometry; succinate to fumarate ratio; multi-observer; Krebs cycle metabolites; linear discriminant analysis; LC–MS/MS; diagnostics; variants of unknown signiﬁcance; metabolite proﬁling; prediction models Received 6 December 2019; Revised 28 March 2020; Accepted 16 May 2020 No conﬂicts of interest were declared. Received 6 December 2019; Revised 28 March 2020; Accepted 16 May 2020 No conﬂicts of interest were declared. Introduction With advances in gene sequencing techniques and decreasing costs, genetic testing is becoming more prac- tical and widely available, but this has also led to new challenges [5]. Variants of unknown signiﬁcance, where the functional impact of the mutation has not been estab- lished, are increasingly troublesome. Also, even with advanced genetic testing, some functional variants or mutations may be missed. This includes intronic variants and epimutations, as well as mutations in other genes impacting mitochondrial energy metabolism [6–8]. Such problems can be addressed by methods assessing the functionality of involved proteins, thereby allowing classiﬁcation of variants with uncertain mutational status. For gene variants affecting SDH, the routinely applied method is immunohistochemistry for SDHB (SDHB-IHC), where staining intensity of tumoural cells is compared with that of non-tumoural cells as internal control [9]. Positively stained cells show a granular pattern, whereas negative tissue at most has a weak diffuse cytoplasmic blush. Importantly, the protein is also degraded when subunits other than SDHB are lost [10]. © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc.org Abstract Phaeochromocytomas and paragangliomas (PPGLs) are rare neuroendocrine tumours with a hereditary background in over one-third of patients. Mutations in succinate dehydrogenase (SDH) genes increase the risk for PPGLs and several other tumours. Mutations in subunit B (SDHB) in particular are a risk factor for metastatic disease, further highlight- ing the importance of identifying SDHx mutations for patient management. Genetic variants of unknown signiﬁ- cance, where implications for the patient and family members are unclear, are a problem for interpretation. For such cases, reliable methods for evaluating protein functionality are required. Immunohistochemistry for SDHB (SDHB-IHC) is the method of choice but does not assess functionality at the enzymatic level. Liquid chromatogra- phy–mass spectrometry-based measurements of metabolite precursors and products of enzymatic reactions provide an alternative method. Here, we compare SDHB-IHC with metabolite proﬁling in 189 tumours from 187 PPGL patients. Besides evaluating succinate:fumarate ratios (SFRs), machine learning algorithms were developed to estab- lish predictive models for interpreting metabolite data. Metabolite proﬁling showed higher diagnostic speciﬁcity compared to SDHB-IHC (99.2% versus 92.5%, p = 0.021), whereas sensitivity was comparable. Application of © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Patient cohorts and tumour procurement Patient cohorts and tumour procurement Tumour collections were approved under Intramural Review Board protocols with informed consent signed at each participating centre. A total of 397 patients with 401 different tumours were included in this study (sup- plementary material, Table S1). The present report © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc.org J Pathol 2020; 251: 378–387 www.thejournalofpathology.com PW Wallace et al 380 builds on our previously reported data on metabolite pro- ﬁles in 391 of these patients [12] by additional compari- sons to immunohistopathological data and introduction of machine learning for interpretation of metabolite pro- ﬁles. Tumour material was collected as freshly frozen (FF) and/or formalin-ﬁxed and parafﬁn-embedded (FFPE) samples. Patients were divided into two separate cohorts (Table 1 and supplementary material, Figure S1): cohort 1 with 187 patients and 189 PPGLs included results for matched metabolite proﬁling and SDHB- IHC, whereas for cohort 2 (210 patients with 212 PPGLs) data from metabolite proﬁles, without availabil- ity of SDHB-IHC, were included and used for training the machine learning algorithms. staining; as inconclusive, when both patterns were pre- sent; or as non-informative, when tissue or staining arte- facts were observed. For 50 samples, the SDHB-IHC interpretations were as established previously, from combined interpretations of seven expert pathologists [20]; for these cases, results were rated as inconclusive when fewer than ﬁve pathologists agreed. All patholo- gists were blinded to the genetic status. Interpretation of immunohistochemistry by experienced pathologists In a subset of cohort 1, termed subcohort 1b (see supple- mentary material, Figure S1), SDHB-IHC slides from the local centres were scanned and high-resolution images were provided to a panel of three experts in SDHB-IHC. Experts rated the staining according to the four categories described above. Genetic characteristics Genetic testing, accomplished as previously described [12], yielded ﬁndings of germline or somatic variants in 18 genes in 49.1% (195/397) of patients (for simplic- ity, only 11 are displayed in Table 1). Eleven patients had variants of unknown signiﬁcance in SDH genes, classiﬁed as variants of unknown signiﬁcance according to the standards and guidelines of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology [16]. In silico prediction of mutation signiﬁcance was performed on variants of unknown signiﬁcance using Mutation Taster [17], Poly- phen [18], and SIFT [19]. Metabolite measurements Seven carboxylic acids of the Krebs cycle (succinate, fumarate, malate, citrate, isocitrate, cis-aconitate, α-ketoglutarate), 2-hydroxyglutarate, pyruvate, and lac- tate were measured in methanol extracts of FF or FFPE tissue by LC–MS/MS as detailed in supplementary material, Supplementary materials and methods and with resulting data provided in supplementary material, Table S2. The cut-off for succinate:fumarate-based interpretation was 97.7, as previously established [11]. Machine learning-assisted interpretation of metabolite data FFPE tissue was sectioned and stained for SDHB using rabbit polyclonal anti-SDHB (HPA002868, 1:400 dilu- tion; Sigma-Aldrich, St Louis, MO, USA) according to the local procedures of six different centres (Dresden, Bethesda, Madrid, Florence, Nijmegen, and Rotterdam) [9]. For 23% of tumours (44 samples), a tissue microar- ray was constructed with three cores of 1.0 mm per sam- ple. Local pathologists evaluated SDHB staining in one slide per tumour and gave the results in four categories: as positive, for the typical granular staining pattern; as negative, for completely negative or weak diffuse Tissue metabolite concentrations (ng/mg tissue) were normalised to natural logarithmic values. These and their ratios were used for formulating predictive models. To establish the need for batch corrections (according to measurement dates), a principal component analysis was generated with the normalised metabolite values of SDHx-mutated or SDHx-wild-type patients (supplemen- tary material, Figure S2). There was a clear distinction between the groups and none of the 27 different batches showed any bias towards SDHx-mutated or SDHx-wild type, allowing formulation of models without the need for batch correction. Table 1. Diagnostic performance of SDHB-IHC compared to LC-MS/ MS based measurements of succinate:fumarate [SFR] in cohort 1. SDHB-IHC SFR p-value Sensitivity [%] 85.2 [46/54] 88.1 [52/59] 0.774 Speciﬁcity [%] 92.5 [111/120] 99.2 [126/127] 0.021 Accuracy [%] 90.2 [157/174] 95.7 [178/186] AUC 0.88 [0.82,0.91] 0.96 [0.89,0.98] 0.048 SFR FP TN FN TP SDHB-IHC FP [n = 9] 0 9 – – FN [n = 8] – – 1 7 inc. [n = 12] 0 7 1 4 Inconclusive samples (12) not included; inc. = inconclusive IHC results, FP: false positive; FN: false negative; TP: true positives; TN: true negatives. Sensitivity, speciﬁcity and accuracy are given as percentages with absolute numbers in brackets. AUC is given as a ratio with CI in brackets. © 2020 The Authors. The Journal of Pathology published by John Wiley & So on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc Table 1. Diagnostic performance of SDHB-IHC compared to LC-MS/ MS based measurements of succinate:fumarate [SFR] in cohort 1. Table 1. Diagnostic performance of SDHB-IHC compared to LC-MS/ MS based measurements of succinate:fumarate [SFR] in cohort 1. © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc.org Machine learning-assisted interpretation of metabolite data SDHB-IHC SFR p-value Sensitivity [%] 85.2 [46/54] 88.1 [52/59] 0.774 Speciﬁcity [%] 92.5 [111/120] 99.2 [126/127] 0.021 Accuracy [%] 90.2 [157/174] 95.7 [178/186] AUC 0.88 [0.82,0.91] 0.96 [0.89,0.98] 0.048 SFR FP TN FN TP SDHB-IHC FP [n = 9] 0 9 – – FN [n = 8] – – 1 7 inc. [n = 12] 0 7 1 4 Feature selection for models was performed using either logarithmically transformed values or ratios of all metabolites against each other. The results of genetic testing were used to separate patients into the categories SDHx-mutated or SDHx-wild type. The ‘LDA MATLAB’ function (MATLAB; MathWorks, Natick, MA, USA) with application of a cross validation was used to train the algorithm and generate models based on linear discriminant analysis (LDA) [21]. Patients from cohort 2 (excluding SDHx variants of unknown signiﬁcance and FFPE only tissue) were used to develop the models and were randomly divided into training and internal validation sets in ratios ranging from 50/50 to 90/10 in steps of 10%. This randomisation and model generation was performed ten times and an average SDHB-IHC* SFR p-value Sensitivity [%] 85.2 [46/54] 88.1 [52/59] 0.774 Speciﬁcity [%] 92.5 [111/120] 99.2 [126/127] 0.021 Accuracy [%] 90.2 [157/174] 95.7 [178/186] AUC 0.88 [0.82,0.91] 0.96 [0.89,0.98] 0.048 Inconclusive samples (12) not included; inc. = inconclusive IHC results, FP: false positive; FN: false negative; TP: true positives; TN: true negatives. Sensitivity, speciﬁcity and accuracy are given as percentages with absolute numbers in brackets. AUC is given as a ratio with CI in brackets. Inconclusive samples (12) not included; inc. = inconclusive IHC results, FP: false positive; FN: false negative; TP: true positives; TN: true negatives. Sensitivity, speciﬁcity and accuracy are given as percentages with absolute numbers in brackets. AUC is given as a ratio with CI in brackets. J Pathol 2020; 251: 378–387 www.thejournalofpathology.com Predicting succinate dehydrogenase mutations in PPGLs 381 positives) and missed SDH impairment in eight cases (false negatives) (Table 1). In 12 tumours, the results were deemed inconclusive according to heterogeneous staining patterns or observer disagreements (the latter applied to ﬁve cases with SDHB-IHC results taken from a previously published study). SFR-based metabolite proﬁling correctly predicted all nine false positives and seven of eight false negatives, and all but one inconclu- sive case were predicted correctly by the SFR. Machine learning-assisted interpretation of metabolite data Diagnos- tic speciﬁcity (p = 0.021) and AUC (p = 0.048) were higher for SFR-based metabolite proﬁling than for SDHB-IHC. model was generated. In total, this resulted in ﬁve LDA models for absolute values and ﬁve models for metabo- lite ratios. The predictive models were applied to the PPGLs of cohort 1 to calculate the likelihood of SDHx mutations (external validation). Performance scores for the different models on the training, internal validation, and external validation sets were calculated (supplemen- tary material, Figure S3) and the two best models selected according to performance (supplementary mate- rial, Figure S4). Selected models are referred to as LDA A, using absolute values, and LDA B for metabolite ratios. Since the models were built on values only from FF tissues, ten samples with metabolite values for FF and FFPE tissue were used to assess the suitability of LDA B for PPGLs stored as FFPE tissues. Results Since model generation of LDA B was performed exclusively on data from freshly frozen (FF) samples, we determined if the model could be applied to both FF and FFPE tissues. For this purpose, LDA B was applied to ten tumours where both FF and FFPE tissues were available. LDA B correctly predicted the © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc.org Calculation of diagnostic performance Two different methods for machine learning-assisted interpretation of metabolite proﬁles were compared with the SFR (Table 2). Model LDA A uses four metabolites and is restricted to measurements from frozen tissue, where weight normalisation is possible. LDA B requires input of ratios of ten metabolites and is applicable to both FF and FFPE specimens (supplementary material, Table S3; supplementary material, Table S4 to produce the ratios). Both models are provided as supplementary material, MATLAB File S1 and MATLAB File S2 (for MATLAB Model LDA A and B, respectively). Based on confusion matrices of genetically determined versus predicted mutational status for SDHx, diagnostic performance was assessed from estimates of sensitivity, speciﬁcity, accuracy, and precis sensitivity = TP/(TP + FN), speciﬁcity = TN/(TN + FP), sensitivity = TP/(TP + FN), y speciﬁcity = TN/(TN + FP), speciﬁcity = TN/(TN + FP), p y ( ) accuracy = (TN + TP)/(TN + TP + FN + FP), y ( ) ( ) precision/positive predictive value = TP/(TP + FP), precision/positive predictive value = TP/(TP + FP), F1-Score = 2 * [(precision * sensitivity)/(precision + sensitivity)], where TP represents true positives, FN false negatives, TN true negatives, and FP false positives. , p y) Our metabolite proﬁling-based models (LDA A, LDA B, and SFR) were applied to 186 tumours (cohort 1). In the case of LDA B, only 185 tumours of the 186 could be used because measurement of one metabolite failed in one tumour. Comparisons of metabolite proﬁling predic- tions showed that the performance of LDA A was improved over that of the SFR (p = 0.044), while there were no signiﬁcant differences between LDA B and SFR or LDA A and LDA B (Figure 1 and Table 3). SFR-based interpretation of metabolite proﬁles resulted in seven false negatives, whereas LDA B produced six and LDA A four false negatives (Table 2). The LDA models calculate probabilities for the likelihood of an SDHx mutation based on the metabolite inputs; for the samples differently rated in LDA A and LDA B, the probabilities were 57% and 99% for LDA A, and 42% and 32% for LDA B, respectively (supplementary mate- rial, Table S5). Statistics Receiver operating characteristics (ROC) curves were pro- duced using logistic regression. Areas under ROC curves (AUCs) of the SDHB-IHC, SFR, LDA A, and LDA B were compared using the Model Comparison tool in JMP Pro (version 14; SAS, Cary, NC, USA). Logistic regression models for combinations of SDHB-IHC with any of the other models were produced in JMP and these combined models’ AUCs were compared with those from using only SFR, LDA Model A or LDA Model B. As there were different numbers of tumours available for the different models, the AUC comparisons were performed with all available tumours (186 versus 185 versus 186 ver- sus 174) and with equal numbers (173 versus 173 versus 173 versus 173). Comparisons of sensitivity and speciﬁc- ity between different predictive methods in the same patient group utilised McNemar’s test for matched pairs data, while comparisons of predictive methods in different patient groups utilised Fisher’s exact test. Differences were considered signiﬁcant for P values below 0.05. It should also be noted that of the eight tumours where SFR had an erroneous prediction, six were head and neck paragangliomas with all seven false negatives car- rying an SDHB, SDHC or SDHD mutation. When LDA A and LDA B were applied to these samples, LDA A correctly predicted three of the head and neck paragangliomas, while LDA B correctly predicted two of them (supplementary material, Table S5). Machine learning-assisted interpretation of metabolite proﬁling Calculation of diagnostic performance Diagnostic performance of SDHB-IHC and metabolite proﬁling by SFR In turn, SDHB-IHC correctly identiﬁed all but one of the false-negative cases from metabolite proﬁling (sup- plementary material, Table S5). With the available num- ber of samples, statistical signiﬁcance was not reached (Table 3). As it was observed that paragangliomas of the head and neck region were classiﬁed more often as false neg- atives with metabolite proﬁling, we compared the diag- nostic sensitivity and speciﬁcity between phaeochromocytomas, paragangliomas of the thorax or abdomen, and head and neck paragangliomas (Table 4). Although statistical signiﬁcance was not reached, it was apparent that the sensitivity was lowest for head and neck paragangliomas with all four methods, SDHB- IHC, SFR, LDA A, and LDA B. Speciﬁcity, on the other hand, showed no regional bias with metabolite-based methods, but was slightly lower for all paragangliomas compared with phaeochromocytomas using SDHB-IHC. mutational status of all ten FFPE and nine of ten FF tis- sues (supplementary material, Figure S5). Combining metabolite proﬁling and SDHB-IHC for best possible predictions on SDHx mutational status We observed a trend towards improved predictions when SDHB-IHC results were combined with the results from metabolite proﬁling compared with the latter alone, supporting complementary utility of SDHB-IHC and LC–MS/MS approaches (Figure 1). Metabolite proﬁling complemented SDHB-IHC by correctly predicting the SDHx mutational status of all but one case where Combining metabolite proﬁling and SDHB-IHC for best possible predictions on SDHx mutational status SDHB-IHC interpretations by local pathologists and a panel of experienced experts To address the question of whether expertise in SDHB- IHC inﬂuences the interpretation of SDHB-IHC results, we utilised a subset of samples in which interpretations Figure 1. ROC curve comparisons for SDHB-IHC, metabolite proﬁling, and their combination. (A) ROC curves of SFR, SDHB-IHC, and their combination. (B) ROC curves of LDA A, SDHB-IHC, and their combination. (C) ROC curves of LDA B, SDHB-IHC, and their combination. Figure 1. ROC curve comparisons for SDHB-IHC, metabolite proﬁling, and their combination. (A) ROC curves of SFR, SDHB-IHC, and their combination. (B) ROC curves of LDA A, SDHB-IHC, and their combination. (C) ROC curves of LDA B, SDHB-IHC, and their combination. Table 4. Comparison of predictive methods in the different subcategories of tumours. Diagnostic performance of SDHB-IHC and metabolite proﬁling by SFR Among a total of 186 PPGLs (cohort 1 excluding three variants of unknown signiﬁcance), SDHB-IHC incor- rectly predicted the SDH status in nine cases (false J Pathol 2020; 251: 378–387 www.thejournalofpathology.com J Pathol 2020; 251: 378–387 www.thejournalofpathology.com J Pathol 2020; 251: 378–387 www.thejournalofpathology.com 382 PW Wallace et al Table 2. Comparison of different predictive models for SDHx impairment based on metabolite proﬁling. LDA-Model A LDA-Model B SFR Model Parameters Tissue amount normalized values (ng/mg) of: Ratios of Metabolites: Ratios of metabolites: • Succinate • Succinate • Succinate • Citrate • Fumarate • Fumarate • Malate • Citrate • Pyruvate • Malate • Pyruvate • Cis-aconitate • Isocitrate • Lactate • 2-hydroxyglutarate • α-ketoglutarate Tissue Type FF FF + FFPE FF + FFPE Confusion Matrix Predicted Predicted Predicted n=186 NO YES n=185 NO YES n=186 NO YES Actual NO 126 1 Actual NO 126 1 Actual NO 126 1 YES 4 55 YES 6 52 YES 7 52 Diagnostic Performance p-value p-value p-value Sensitivity 93.2% 0.344 Sensitivity 89.7% 0.774 Sensitivity 88.1% 0.774 Speciﬁcity 99.2% 0.022 Speciﬁcity 99.2% 0.021 Speciﬁcity 99.2% 0.021 Accuracy 97.3% Accuracy 96.2% Accuracy 95.7% AUC 0.982 0.004 AUC 0.977 0.012 AUC 0.961 0.048 SDHB-IHC SDHB-IHC SDHB-IHC FP TN FN TP inc FP TN FN TP inc FP TN FN TP inc FP(1) 0 1 – – 0 FP(1) 0 1 – – 0 FP(1) 0 1 – – 0 FN(4) – – 1 3 0 FN(6) – – 1 5 0 FN(7) – – 1 5 1 FP: false positive; TN: True Negative; FN: False Negative; TP: True Positive; inc: inconclusive results in IHC. Journal of Pathology published by John Wiley & Sons Ltd ociety of Great Britain and Ireland. www.pathsoc.org www © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc.org Predicting succinate dehydrogenase mutations in PPGLs 383 Table 3. Statistical comparisons of AUC differences. Predictor versus Predictor AUC Difference p-value* IHC SFR −0.070 0.048 IHC LDA A −0.092 0.004 IHC LDA B −0.086 0.012 SFR LDA A −0.023 0.044 SFR LDA B −0.016 0.533 LDA A LDA B 0.006 0.775 SFR SFR + IHC −0.031 0.113 LDA B LDA B + IHC −0.021 0.235 LDA A LDA A +IHC −0.014 0.301 p<0.05 considered signiﬁcant. SDHB-IHC incorrectly predicted the mutational status. © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc.org Discussion from local pathologists were compared with those of experts (see supplementary material, Figure S1). Accu- racy of predictions of SDHx mutations according to SDHB-IHC indicated no signiﬁcant differences between local pathologists and experts, suggesting that SDHB- IHC does not require specialised training (Table 5). In 71% of cases, all pathologists agreed and when at least two experts on the panel agreed on the prediction (100 of 105 cases), they in turn agreed with the local patholo- gist in 83 of these cases (79%) (supplementary material, Figure S6). Agreement was higher in non-SDHx cases (86%) than in SDHx-mutated cases (70%) (Table 5). This study establishes for the ﬁrst time that SDHB-IHC and metabolite proﬁling provide complementary diag- nostic tools for the prediction of SDH impairment in PPGL tumour tissue. Moreover, we show that diagnostic performance of metabolite proﬁling can be improved by machine learning-assisted interpretation of metabolite data and that there is a trend towards further improve- ment by inclusion of ﬁndings from SDHB-IHC. We therefore propose an approach that combines metabolite proﬁling and SDHB-IHC to better facilitate identifying or excluding SDH impairment when tumour material is available, particularly for selected patients in whom there is a suspicion of the presence of SDH mutation and where genetic testing yields equivocal or negative results or is unavailable. First, the high speciﬁcity of metabolite proﬁling (99%) translates to high positive predictive value of a positive result, strongly indicating a mutation in an SDHx gene. If the genetic change iden- tiﬁed is a variant of unknown signiﬁcance, the predictive models will, together with in silico prediction tools, aid in determining whether the variant is pathogenic and whether the patient and affected family members require life-long surveillance. If metabolite proﬁling predicts no SDH impairment, then SDHB-IHC provides utility to exclude false negatives by metabolite proﬁling. Apply- ing this approach to cohort 1 (excluding three variants of unknown signiﬁcance), we would have correctly pre- dicted impairments of SDH in 185 of the 186 PPGLs, providing an advantage over either method alone. Amongst the samples with full agreement between all pathologists, there were six cases (out of 75) with incor- rect predictions. These six cases comprised ﬁve head and neck paragangliomas with SDHx mutations and one adrenal PPGL without an SDHx mutation, indicating again that head and neck paragangliomas are the most challenging specimens to interpret (supplementary material, Table S5). Diagnostic performance of SDHB-IHC and metabolite proﬁling by SFR PHEO PGL HNP HNP versus (PHEO + PGL) n = 112 n = 30 n = 44 p-value IHC Sensitivity (%) 88.8 [8/9] 90.9 [20/22] 78.3 [18/23] 0.264 Speciﬁcity (%) 93.8 [91/97] 87.5 [7/8] 86.7 [13/15] 0.313 inconclusive (n) 6 0 6 SFR Sensitivity (%) 100 [9/9] 95.5 [21/22] 78.6 [22/28] 0.130 Speciﬁcity (%) 99 [102/103] 100 [8/8] 100 [16/16] 1.000 LDA A Sensitivity (%) 100 [9/9] 95.5 [21/22] 89.3 [25/28] 0.337 Speciﬁcity (%) 99 [102/103] 100 [8/8] 100 [16/16] 1.000 LDA B Sensitivity (%) 100 [8/8] 95.5 [21/22] 82.1 [23/28] 0.097 Speciﬁcity (%) 99 [102/103] 100 [8/8] 100 [16/16] 1.000 calculated using Fisher’s exact test. HNP = Head and neck paraganglioma, PGL = paraganglioma, PHEO = Phaeochromocytomas. Table 4. Comparison of predictive methods in the different subcategories of tumours. © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc.org J Pathol 2020; 251: 378–387 www.thejournalofpathology.com PW Wallace et al 384 Table 5. SDHB-IHC interpretations of a panel of experienced researchers versus local pathologists in cohort 1b. IHC was available, but did not support the metabolite- based interpretations. LDA Model B predicted SDH impairment in two further PPGLs, a splice site variant in SDHA (NM_004168.3:c.457-1G>A) and an indel var- iant in SDHC (NM_003001.3:c.256_257insTTT, p. (Gly86delinsValCys). For the former, the same variant was found in a second unrelated patient, where SDHB- IHC showed negative staining, supporting the classiﬁca- tion as ‘likely pathogenic’. The missense variant of SDHA, NM_004168.3:c.1772C>T, p.(Ala591Val), was predicted to have no functional impact based on metab- olite proﬁling and SDHB-IHC as interpreted by a local pathologist. Experts, however, all agreed on a negative staining pattern for SDHB. Two out of three in silico protein prediction tools rated the variant as ‘disease causing’ or ‘possibly damaging’. researchers versus local pathologists in cohort 1b. Local Panel# p-value Sensitivity [%] 80.0 [16/20] 65.0 [13/20] 0.250 Speciﬁcity [%] 92.4 [61/66] 98.5 [65/66] 0.125 Accuracy [%] 89.5 [77/86] 90.7 [78/86] # Inconclusive/non- informative cases 9 11 Agreement amongst panel Agreement of panel [>2/3] with local interpretations 3 2 0 non-SDHx (n = 73) 62 9 2 63 [86.3%] SDHx (n = 30) 21 6 3 19 [70.0%] SDHx VUS (n = 2) 2 0 0 1 [50.0%] Total (n = 105) 85 15 5 83 [79.0%] *Inconclusive cases were removed from the analysis. Diagnostic performance of SDHB-IHC and metabolite proﬁling by SFR #For the panel of experienced researchers, interpretations where at least two out of three researchers agreed, were used. VUS, Variant of Unknown Signiﬁcance. Discussion Local pathologists rated nine slides as inconclusive, whereas experts rated ﬁve slides as inconclusive and six as non-informative. Only one inconclusive case overlapped between the two groups, indicating some variable subjectivity of interpretations. Non-informative cases arose due to technical problems with scanned images or staining of slides, such as related to uneven staining or high background (example images in supplementary material, Figure S7). © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc.org J Pathol 2020; 251: 378–387 www.thejournalofpathology.com Re-evaluation of variants of unknown signiﬁcance in SDHx SDHB-IHC requires a simple setup and can be easily incorporated into pathology workﬂows. Since our pre- liminary evidence suggests that interpretation does not require expert review, the technique is readily adoptable anywhere. LC–MS/MS, on the other hand, requires spe- cialised instrumentation and expertise, but is becoming more and more available in clinical laboratories where the instruments are used for many routine diagnostic tests. While SDHB-IHC assesses presence of the pro- tein, LC–MS/MS-based metabolite proﬁling provides Amongst the tumours evaluated, there were 11 patients (three in cohort 1 and eight in cohort 2) with a variant of unknown signiﬁcance in one of the SDHx genes. Both LDA models and the SFR were applied to the metabolite proﬁles and compared with in silico predictions for pro- tein changes (supplementary material, Table S6). Both LDA models agreed in all four cases where the SFR pre- dicted SDH impairment. In two of these cases, SDHB- Predicting succinate dehydrogenase mutations in PPGLs 385 information about functionality of the succinate complex through measurements of precursor and product metabo- lites. There is also the added beneﬁt of measuring a panel of metabolites to identify impaired function of other enzymes, such as fumarate hydratase and isocitrate dehydrogenase [12,22]. Metabolite measurements also address some of the limitations of SDHB-IHC: in partic- ular, there is no subjective bias of interpretation and there is always a numerical result rather than inconclu- sive interpretations. straightforward since it only requires tissue concentra- tions of measured metabolites or ratios of metabolites (to calculate the matrices needed for input of this data into the LDA models, see supplementary material, Table S4). Results are provided as per cent probabilities of the tumour harbouring an SDHx mutation. The interpretation of ten metabolites is thereby converted into a simpliﬁed single output variable to guide clinical decision-making. With ongoing data collection for PPGLs and other tumour entities, machine learning-assisted data interpre- tation can be used to further stratify patients according to mutational background or other clinically relevant fea- tures as new models can be generated as more data become available. This approach was also suggested in the context of steroid metabolomics for the diagnosis of adrenal cortical tumours [28]. On the other hand, there are limitations of metabolite proﬁling. Generating cut-offs or machine learning models requires large numbers of samples and whether such data are transferrable among laboratories (i.e. method harmonisation) using LC–MS/MS is not yet established. © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc.org Re-evaluation of variants of unknown signiﬁcance in SDHx Our study also highlights the beneﬁt of interdisciplinary connections between physi- cians, pathologists, clinical chemists, geneticists, and data scientists. By working together, advanced genotypic strat- iﬁcation of PPGLs can be expected to better facilitate tar- geted therapies with increased efﬁcacy and improved patient outcomes [30–33]. Limitations of the present study are that SDHB-IHC data were not available for all patients and that immunohisto- chemistry was performed in different centres to varying standards and quality. The latter may partly explain the lower level of agreement among pathologists for interpret- ing SDHB-IHC than found in a previous study [20]. Despite thislimitation,diagnostic sensitivity andspeciﬁcity for SDHB-IHC were in the range of former reports [9,20]. Although machine learning was only applied to metab- olite data and not to images from SDHB-IHC, we could show that there is potential for this approach to improve diagnostic and prognostic workﬂows, especially when data complexity is high. Our study also highlights the beneﬁt of interdisciplinary connections between physi- cians, pathologists, clinical chemists, geneticists, and data scientists. By working together, advanced genotypic strat- iﬁcation of PPGLs can be expected to better facilitate tar- geted therapies with increased efﬁcacy and improved patient outcomes [30–33]. In this study, we used pattern recognition and multidi- mensional strategies from the ﬁeld of artiﬁcial intelli- gence for analysing metabolite data to gain information beyond simple ratios such as the SFR. In this way, machine learning offered improved diagnostic sensitiv- ity. This was especially useful for identifying functional impairment of SDH in head and neck paragangliomas, for which false negatives can be a problem when relying on the SFR [11]. We advise that the tumour content of the input material for head and neck paragangliomas be evaluated carefully and that further available methods to test protein status, such as immunohistochemistry for SDHB, SDHA or SDHD, be used [9,26,27]. Re-evaluation of variants of unknown signiﬁcance in SDHx Another limitation of metabolite proﬁling is tissue selection. False negatives can occur due to excessive amounts of non-tumour tissue in the sample. We suspect that this is also the reason for different pre- dictions (probabilities for SDH impairment) produced by LDA A and LDA B, since depending on the type of stromal contaminant, metabolite levels will differ. One possible solution, requiring interdisciplinary connec- tions between anatomic and chemical diagnostic labora- tories, is to assess tumour content ﬁrst by haematoxylin and eosin staining, perform macro-dissection of tumour areas, and use this material for metabolite extraction. A challenge of next-generation panel sequencing in genetic testing is the interpretation of variants of unknown signiﬁcance. In our combined cohorts, a total of 11 vari- ants of unknown signiﬁcance in SDHx genes were identi- ﬁed. From those, four had elevated SFR indicating loss of functionality; however, in silico predictions and SDHB- IHC showed varying agreement. Such discrepancies between SDHB-IHC and in silico predictions were reported previously [29]. In two more tumours, where SFR failed to indicate SDH impairment, the new machine learning-based LDA B predicted loss of SDH functional- ity. At least in one of these cases, SDHA NM_004168.3: c.457-1G>A, evidence suggests a true mutation. Inter-observer variability of SDHB-IHC interpreta- tions could be addressed by applying deep learning to establish a pipeline for automated image interpretation, as has been done for immunohistochemistry directed to other purposes [14,23]. Machine learning was also dem- onstrated to be suitable for cancer diagnosis on whole- slide images [24]. Nevertheless, not only histology but also biomarker interpretation and analysis of omics data (transcriptomics, proteomics, and metabolomics) can beneﬁt from machine learning approaches [15]. A recent example is the identiﬁcation of PPGL-speciﬁc long intergenic noncoding RNAs and their use for molecular subtyping of PPGL patients [25]. Limitations of the present study are that SDHB-IHC data were not available for all patients and that immunohisto- chemistry was performed in different centres to varying standards and quality. The latter may partly explain the lower level of agreement among pathologists for interpret- ing SDHB-IHC than found in a previous study [20]. Despite thislimitation,diagnostic sensitivity andspeciﬁcity for SDHB-IHC were in the range of former reports [9,20]. Although machine learning was only applied to metab- olite data and not to images from SDHB-IHC, we could show that there is potential for this approach to improve diagnostic and prognostic workﬂows, especially when data complexity is high. References 20. Papathomas TG, Oudijk L, Persu A, et al. SDHB/SDHA immunohis- tochemistry in pheochromocytomas and paragangliomas: a multicen- ter interobserver variation analysis using virtual microscopy: a Multinational Study of the European Network for the Study of Adre- nal Tumors (ENS@T). Mod Pathol 2015; 28: 807–821. 1. Gill AJ. Succinate dehydrogenase (SDH)-deﬁcient neoplasia. Histopathology 2018; 72: 106–116. 2. Amar L, Baudin E, Burnichon N, et al. Succinate dehydrogenase B gene mutations predict survival in patients with malignant pheochro- mocytomas or paragangliomas. J Clin Endoncrinol Metab 2007; 92: 3822–3828. 21. Fischer RA. The use of multiple measurements in taxonomic prob- lems. Ann Eugen 1936; 7: 179–188. 22. Dwight T, Kim E, Novos T, et al. Metabolomics in the diagnosis of pheochromocytoma and paraganglioma. Horm Metab Res 2019; 51: 443–450. 3. Rijken JA, van Hulsteijn LT, Dekkers OM, et al. Increased mortality in SDHB but not in SDHD pathogenic variant carriers. Cancers (Basel) 2019; 11: e103. 23. Wang S, Yang DM, Rong R, et al. Pathology image analysis using segmentation deep learning algorithms. Am J Pathol 2019; 189: 1686–1698. 4. Lenders JW, Duh QY, Eisenhofer G, et al. Pheochromocytoma and paraganglioma: an Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab 2014; 99: 1915–1942. 24. Zhang Z, Chen P, McGough M, et al. Pathologist-level interpretable whole-slide cancer diagnosis with deep learning. Nat Mach Intell 2019; 1: 236–245. 5. Toledo RA, Burnichon N, Cascon A, et al. Consensus statement on next-generation-sequencing-based diagnostic testing of hereditary phaeochromocytomas and paragangliomas. Nat Rev Endocrinol 2017; 13: 233–247. 25. Ghosal S, Das S, Pang Y, et al. Long intergenic non-coding RNA pro- ﬁles of pheochromocytoma and paraganglioma: a novel prognostic biomarker. Int J Cancer 2020; 146: 2326–2335. 6. Curras-Freixes M, Pineiro-Yanez E, Montero-Conde C, et al. Pheo- Seq: a targeted next-generation sequencing assay for pheochromocy- toma and paraganglioma diagnostics. J Mol Diagn 2017; 19: 575–588. 26. Korpershoek E, Favier J, Gaal J, et al. SDHA immunohistochemistry detects germline SDHA gene mutations in apparently sporadic para- gangliomas and pheochromocytomas. J Clin Endocrinol Metab 2011; 96: E1472–E1476. 7. Richter S, Klink B, Nacke B, et al. Epigenetic mutation of the succi- nate dehydrogenase C promoter in a patient with two paragangliomas. J Clin Endocrinol Metab 2016; 101: 359–363. 27. Menara M, Oudijk L, Badoual C, et al. SDHD immunohistochemis- try: a new tool to validate SDHx mutations in pheochromocytoma/par- aganglioma. J Clin Endocrinol Metab 2015; 100: E287–E291. 8. PW Wallace et al 386 Paradifference Foundation, and the Intramural Research Program of the NIH, NICHD. Paradifference Foundation, and the Intramural Research Program of the NIH, NICHD. 14. Abels E, Pantanowitz L, Aeffner F, et al. Computational pathology deﬁnitions, best practices, and recommendations for regulatory guid- ance: a white paper from the Digital Pathology Association. J Pathol 2019; 249: 286–294. 15. Grapov D, Fahrmann J, Wanichthanarak K, et al. Rise of deep learn- ing for genomic, proteomic, and metabolomic data integration in pre- cision medicine. OMICS 2018; 22: 630–636. Author contributions statement CC, YP, EC, EK, ER, MUM, and SB carried out experi- ments. SB, AT, RRK, EK, ZZ, and AJG interpreted IHC slides. KP, FB, MR, MAM, MK, VG, HJLMT, and DA provided specimens. MP, GE, PWW, SR, KB, MR, SB, DA, and GE analysed and interpreted data and performed statistical analysis. SR, TP, GE, PWW, and JP designed and conceived the experiments. SR, GE, PWW, and CC drafted the manuscript. All the authors were involved in writing the paper and approved the ﬁnal version. 16. Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommenda- tion of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med 2015; 17: 405–424. 17. Schwarz JM, Cooper DN, Schuelke M, et al. MutationTaster2: muta- tion prediction for the deep-sequencing age. Nat Methods 2014; 11: 361–362. 18. Ramensky V, Bork P, Sunyaev S. Human non-synonymous SNPs: server and survey. Nucleic Acids Res 2002; 30: 3894–3900. 19. Sim NL, Kumar P, Hu J, et al. SIFT web server: predicting effects of amino acid substitutions on proteins. Nucleic Acids Res 2012; 40: W452–W457. Acknowledgements This study was funded by the Deutsche Forschungsge- meinschaft (DFG, German Research Foundation) Pro- jektnummer: 314061271 – TRR 205; RI 2684/1-1; KL 2541/2-1, the AES PI17/01796, co-ﬁnanced by Fondo Europeo de Desarrollo Regional (FEDER), the Euro- pean Union Seventh Framework Programme (FP7/ 2007-2013) under grant agreement No 259735, the While the predictive models generated in this study were targeted towards identifying SDHx mutations, other models could be generated based on measurements of metabolites in the same panel to predict mutations and functional deﬁciencies impacting other enzymes. Appli- cation of the generated models (supplied as supplemen- tary material, MATLAB Files S1 and S2) is relatively J Pathol 2020; 251: 378–387 www.thejournalofpathology.com © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc.org References Remacha L, Comino-Mendez I, Richter S, et al. Targeted exome sequencing of Krebs cycle genes reveals candidate cancer-predispos- ing mutations in pheochromocytomas and paragangliomas. Clin Cancer Res 2017; 23: 6315–6324. 28. Eisenhofer G, Durán C, Chavakis T, et al. Steroid metabolomics: machine learning and multidimensional diagnostics for adrenal corti- cal tumors, hyperplasias, and related disorders. Curr Opin Endocrinol Metab Res 2019; 8: 40–49. 9. van Nederveen FH, Gaal J, Favier J, et al. An immunohistochemical procedure to detect patients with paraganglioma and phaeochromocy- toma with germline SDHB, SDHC, or SDHD gene mutations: a retro- spective and prospective analysis. Lancet Oncol 2009; 10: 764–771. 29. Evenepoel L, Papathomas TG, Krol N, et al. Toward an improved def- inition of the genetic and tumor spectrum associated with SDH germ- line mutations. Genet Med 2015; 17: 610–620. 10. Gill AJ. Succinate dehydrogenase (SDH) and mitochondrial driven neoplasia. Pathology 2012; 44: 285–292. 30. Pang Y, Liu Y, Pacak K, et al. Pheochromocytomas and paraganglio- mas: from genetic diversity to targeted therapies. Cancers (Basel) 2019; 11: e436. 11. Richter S, Peitzsch M, Rapizzi E, et al. Krebs cycle metabolite proﬁl- ing for identiﬁcation and stratiﬁcation of pheochromocytomas/para- gangliomas due to succinate dehydrogenase deﬁciency. J Clin Endocrinol Metab 2014; 99: 3903–3911. 31. Pang Y, Lu Y, Caisova V, et al. Targeting NAD+/PARP DNA repair pathway as a novel therapeutic approach to SDHB-mutated cluster I pheochromocytoma and paraganglioma. Clin Cancer Res 2018; 24: 3423–3432. 12. Richter S, Gieldon L, Pang Y, et al. Metabolome-guided genomics to identify pathogenic variants in isocitrate dehydrogenase, fumarate hydratase, and succinate dehydrogenase genes in pheochromocytoma and paraganglioma. Genet Med 2019; 21: 705–717. 32. Pacak K, Taieb D. Pheochromocytoma (PHEO) and paraganglioma (PGL). Cancers (Basel) 2019; 11: e1391. 13. Kim E, Wright MJ, Sioson L, et al. Utility of the succinate:fumarate ratio for assessing SDH dysfunction in different tumor types. Mol Genet Metab Rep 2016; 10: 45–49. 33. Liu Y, Pang Y, Caisova V, et al. Targeting NRF2-governed glutathi- one synthesis for SDHB-mutated pheochromocytoma and paragan- glioma. Cancers (Basel) 2020; 12: e280. J Pathol 2020; 251: 378–387 www.thejournalofpathology.com J Pathol 2020; 251: 378–387 www.thejournalofpathology.com Predicting succinate dehydrogenase mutations in PPGLs 387 SUPPLEMENTARY MATERIAL ONLINE Supplementary materials and methods Figure S1. Flow diagram of patient and tumour sample numbers used in this study Figure S2. J Pathol 2020; 251: 378–387 www.thejournalofpathology.com © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc.org References PCA plot of all tumours showing grouping of SDHx mutations (black triangles) and SDHx wild types (open circles) based on the normalised values of all ten measured metabolites Figure S3. Performance of the models generated based on different ratios of how to separate cohort 2 into learning set and validation set Figure S4. Performance of the two chosen models LDA A and LDA B Figure S5. Comparison of predictions generated with LDA Model B from FFPE and FF tissue extracted metabolites Figure S6. Distribution of true negatives (TN), true positives (TP), false positives (FP), and false negatives (FN) in cases of agreement between the panel of researchers and local pathologists Figure S7. Example images for SDHB-IHC scored by local pathologists and experts in the ﬁeld of SDHB-IHC Table S1. Characterisation of patient/tumour cohorts Table S2. Tumour identiﬁers, clinical data, IHC results, and metabolite concentrations Table S3. Input ratios for LDA model B in the order they need to be supplied in MATLAB Table S4. Excel ﬁle to produce matrices for use in LDA models Table S5. Results of all predictive methods where the tumour is misclassiﬁed by at least one of the methods Table S6. Interpretation of metabolite proﬁles for SDHx VUS based on predictive models generated with linear discriminant analysis (LDA) Table S7. Concentrations of pre-calibrators Pre-Cal 1–8 Table S8. Concentrations of internal standards in the Internal Standard Mix (IS-Mix) Table S9. Concentrations of calibrators Cal 0–8 Table S10. Assay precision estimated using two quality control (QC) samples Table S11. Multiple reaction monitoring transitions, fragmentation parameters,m and quantiﬁers MATLAB File S1. Matlab_Model_LDA_A.mat – a MATLAB model for use with absolute metabolite concentrations MATLAB File S2. Matlab_Model_LDA_B.mat – a MATLAB model for use with relative metabolite concentrations Predicting succinate dehydrogenase mutations in PPGLs 387 SUPPLEMENTARY MATERIAL ONLINE © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland. www.pathsoc.org
https://openalex.org/W2002458125	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0107997&type=printable	English	null	The Hemodynamically-Regulated Vascular Microenvironment Promotes Migration of the Steroidogenic Tissue during Its Interaction with Chromaffin Cells in the Zebrafish Embryo	PloS one	2,014	cc-by	14,874	Abstract This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Copyright: 2014 Chou et al. This is an open-access article distributed under the terms of the Creative Commons Attributi unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All relevant data are within the paper and its Supporting Information files. nding: CWC, YLZ, ZYJ, and YWL were supported by Ministry of Science and Technology (http://web1.most.gov.tw/) grants (101-2313-B 9-002-MY3 and 102-2321-B-400-018). The funder had no role in study design, data collection and analysis, decision to publish, or prepara and YWL were supported by Ministry of Science and Technology (http://web1.most.gov.tw/) grants (101-2313-B-029-001, 102-2628-B- 21-B-400-018). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript Competing Interests: The authors have declared that no competing interests exist. * Email: dlslys@thu.edu.tw Chih-Wei Chou, You-Lin Zhuo, Zhe-Yu Jiang, Yi-Wen Liu* Chih-Wei Chou, You-Lin Zhuo, Zhe-Yu Jiang, Yi-Wen Liu* Department of Life Science, Tunghai University, Taichung, Taiwan Department of Life Science, Tunghai University, Taichung, Taiwan Department of Life Science, Tunghai University, Taichung, Taiwan Abstract Background: While the endothelium-organ interaction is critical for regulating cellular behaviors during development and disease, the role of blood flow in these processes is only partially understood. The dorsal aorta performs paracrine functions for the timely migration and differentiation of the sympatho-adrenal system. However, it is unclear how the adrenal cortex and medulla achieve and maintain specific integration and whether hemodynamic forces play a role. Methodology and Principal Findings: In this study, the possible modulation of steroidogenic and chromaffin cell integration by blood flow was investigated in the teleostean counterpart of the adrenal gland, the interrenal gland, in the zebrafish (Danio rerio). Steroidogenic tissue migration and angiogenesis were suppressed by genetic or pharmacologic inhibition of blood flow, and enhanced by acceleration of blood flow upon norepinephrine treatment. Repressed steroidogenic tissue migration and angiogenesis due to flow deficiency were recoverable following restoration of flow. The regulation of interrenal morphogenesis by blood flow was found to be mediated through the vascular microenvironment and the Fibronectin-phosphorylated Focal Adhesion Kinase (Fn-pFak) signaling. Moreover, the knockdown of kru¨ppel-like factor 2a (klf2a) or matrix metalloproteinase 2 (mmp2), two genes regulated by the hemodynamic force, phenocopied the defects in migration, angiogenesis, the vascular microenvironment, and pFak signaling of the steroidogenic tissue observed in flow-deficient embryos, indicating a direct requirement of mechanotransduction in these processes. Interestingly, epithelial-type steroidogenic cells assumed a mesenchymal-like character and downregulated b-Catenin at cell-cell junctions during interaction with chromaffin cells, which was reversed by inhibiting blood flow or Fn-pFak signaling. Blood flow obstruction also affected the migration of chromaffin cells, but not through mechanosensitive or Fn-pFak dependent mechanisms. Conclusions and Significance: These results demonstrate that hemodynamically regulated Fn-pFak signaling promotes the migration of steroidogenic cells, ensuring their interaction with chromaffin cells along both sides of the midline during interrenal gland development. Citation: Chou C-W, Zhuo Y-L, Jiang Z-Y, Liu Y-W (2014) The Hemodynamically-Regulated Vascular Microenvironment Promotes Migration of the Steroidogenic Tissue during Its Interaction with Chromaffin Cells in the Zebrafish Embryo. PLoS ONE 9(9): e107997. doi:10.1371/journal.pone.0107997 Editor: Sheng-Ping Lucinda Hwang, Institute of Cellular and Organismic Biology, Taiwan Editor: Sheng-Ping Lucinda Hwang, Institute of Cellular and Organismic Biology, Taiwan Received April 10, 2014; Accepted August 24, 2014; Published September 23, 2014 Copyright: 2014 Chou et al. September 2014 \| Volume 9 \| Issue 9 \| e107997 Introduction However, the existence of shared paracrine factors does not explain why chromaffin cells colonize the adrenal cortex rather than non-adrenal regions surrounding the DA, and additional molecular and cellular factors could participate in the integration of steroidogenic and chromaf- fin cells. It was hypothesized that in addition to instructing the migration and differentiation of chromaffin cells, the vasculature near the adrenal gland also specifies the behavior of adrenocortical cells, thereby promoting cortex-medulla amalgamation. This possibility was investigated in the present study in zebrafish, an established model for exploring the development and diseases of the cardiovascular and endocrine systems. through the endothelium to influence closely associated interrenal cells. In this study, the possible role of blood flow for the integration of steroidogenic and chromaffin cells was examined in the zebrafish interrenal gland, by using genetic and pharmacological approaches to abolish blood flow in the embryo. The vascular structure and associated ECM microenvironment in the interrenal region were examined for changes in the architecture of the developing interrenal tissue. The modulation of interrenal morphogenesis by blood flow through mechanotransduction was investigated by knocking down the mechanosensitive proteins Kru¨ppel-like factor 2a (Klf2a) and Matrix metalloproteinase (Mmp)2. In addition, we demonstrated that steroidogenic cells undergo an epithelial-to-mesenchymal transition (EMT)-like change during organ assembly, which was correlated with a reduction in epithelial and a rise in mesenchymal markers. During EMT, which occurs at many critical steps during embryonic development, cell-cell contacts and polarity are lost and the cytoskeleton is extensively remodeled [35]. The present findings underscore the role of hemodynamics in regulating Fn-phosphor- ylated Focal adhesion kinase (pFak) signaling in the developing interrenal tissue, which in turn induces an EMT-like transforma- tion in steroidogenic cells. Thus, in addition to the known chemoattractive function for chromaffin cells, the axial vasculature regulates the migration of steroidogenic cells through hemody- namically regulated signaling. The teleostean interrenal gland is functionally equivalent to the adrenal gland in mammals, with steroidogenic and chromaffin cell populations arising from conserved molecular programs [16,17,18]. The integration of these two cell types occurs between 1.5 and 3 days post-fertilization (dpf), which is immediately followed by de novo cortisol synthesis in response to stress [18,19,20]. Within the same temporal window, the interrenal vessel (IRV) is patterned along with a vessel-derived, Fibronectin (Fn)-enriched microenvironment [21], which is essential for IRV growth, steroidogenic tissue morphogenesis, and positioning the interrenal organ. Introduction the crucial role of hemodynamics in the morphogenesis of heart, kidney, and brain vasculature [3,4,5,6]. Moreover, it is possible to study the role of hemodynamics in establishing the architecture of endocrine tissues in the zebrafish embryo, since the specification and differentiation of a variety of endocrine cells proceed even in the complete absence of vasculature [7,8,9,10,11]. Although blood vessels have long been known to respond to hemodynamic forces through mechanotransduction, only recently have researchers begun to understand the influence of hemody- namics on organogenesis through modulation of cellular behav- iors, the extracellular matrix (ECM) microenvironment, as well as cell signaling events [1]. The early zebrafish embryo does not rely on blood circulation to transport oxygen [2], making it an excellent in vivo model for studying the effect of blood flow on development. Various genetic and pharmacological approaches have been developed in the zebrafish model, which have revealed How the adrenal cortex and medulla—arising because of distinct cell fate decisions in physically separated precursor cells— assemble to form the adrenal gland remains incompletely understood. The adrenal cortex is comprised of steroidogenic cells differentiated from the intermediate mesoderm, while the medulla contains chromaffin cells that originate from the neural September 2014 \| Volume 9 \| Issue 9 \| e107997 September 2014 \| Volume 9 \| Issue 9 \| e107997 PLOS ONE \| www.plosone.org 1 Hemodynamic Regulation of Interrenal Gland Assembly crest and are subsequently segregated from the sympatho-adrenal lineage [12]. Mice deficient in the transcription factor steroido- genic factor-1 (SF-1, NR5A1) lack an adrenal cortex, but exhibit normal differentiation of chromaffin cells, half of which are present in the suprarenal region, arguing against a role for the adrenal cortex in attracting chromaffin cells [13]. However, ectopic adrenocortical cells in the mouse thorax, induced through the transgenic overexpression of SF-1, are capable of recruiting sympatho-adrenal progenitors [14]. These findings suggestive of an undefined role of the adrenal cortex have been clarified by the recent demonstration of the dorsal aorta (DA) as a morphogenetic center that instructs the specification and segregation of the sympatho-adrenal lineage in the chick embryo [15]. The DA and the adrenal cortex both secrete Neuregulin 1, which attracts chromaffin cells to the suprarenal region. Zebrafish Husbandry Zebrafish (Danio rerio) were reared according to standard protocols [36]. Embryos were obtained from natural crosses of wild-type or transgenic fish, and staged as previously described [37]. The following lines were used: Tg(wt1b: GFP)(line 1) [38] (a gift from Christoph Englert, Fritz-Lipmann Institute, Jena, Germany); Tg(ff1bEx2: GFP) [39] (a gift from Dr. Woon-Khiong Chan, National University of Singapore); and Tg(kdrl: EGFP)s843 [40] (a gift of Didier Stainier, University of California, San Francisco, CA, USA). Previous studies have shown that Klf2a and MMP2 are hemodynamically regulated: KLF2 is a transcription factor activated in cultured endothelial cells by fluid shear stress from laminar flow [22,23], and the endothelial expression of mouse Klf2 and its ortholog klf2a in zebrafish reflects an increase in fluid- generated forces, while a loss of function leads to defective smooth muscle tone [24]. MMPs are known to mediate ECM remodeling and enable reshaping of tissues through peptidase activity [25]. In the zebrafish embryo, mmp2 is expressed in the endothelium of developing axial vasculature in a flow-dependent manner [5]; and in rats and cultured cells, MMP2 activity in glomerular mesangial cells is induced by stretch [26,27] and regulated by cyclic strains in the endothelium resulting from turbulent flow, which modulates the migration of vascular smooth muscle cells [28,29]. Moreover, MMP2 cleaves a variety of ECM molecules, including type IV collagen, vitronectin, and fibronectin [30,31]. Abundant RNA transcripts of both klf2a and mmp2 are restrictively localized at the axial vasculature at around Prim-25 stage (36 hpf) during zebrafish development [5,32], which is temporally correlated with the initiation of interrenal medial extension and angiogenesis. Furthermore, the nascent DA in the zebrafish does not recruit mural cells until 3 dpf [33], and differentiated vascular smooth muscle cells appear only after 7 dpf [34]; it was therefore hypothesized that hemodynamic forces could be transduced Introduction Nevertheless, little is known about how the Fn- enriched interrenal microenvironment is regulated and the cellular mechanisms governing morphogenetic movements during inte- gration. Ethics Statement All of the zebrafish-use protocols in this research were reviewed and approved by the Institutional Animal Care and Use Committee of Tunghai University (IRB Approval NO. 101–12). September 2014 \| Volume 9 \| Issue 9 \| e107997 Blood Flow is Required for the Morphogenesis of Kidney Glomerulus and Interrenal Tissue For IHC experiments, Tg(ff1bEx2: GFP) and Tg(kdrl: EGFP)s843 embryos were fixed and embedded in 4% NuSieve GTG low-melting agarose (Lonza), cut into 100- mm sections with a VT1000M vibratome (Leica), and permeabilized with phos- phate-buffered saline (PBS) containing 1% Triton X-100 before incubation with rabbit anti-human Fn (Sigma), mouse anti-human pFak (pY397) (BD Transduction Laboratories), mouse anti- chicken b-Catenin (Sigma), mouse anti-pig Vimentin (V9) (Abcam), and rabbit anti-zebrafish E-cadherin (Cdh1) (GeneTex) antibodies at 1:200, 1:100, 1:50, 1:200 and 1:200 dilutions, respectively. Dylight 594- and 650-conjugated anti-rabbit or anti- mouse IgG (abcam) were used as secondary antibodies at 1:200 dilution. Images were captured with an LSM510 confocal microscope with version 3.5 software (Zeiss). The hemodynamic force drives the assembly of zebrafish kidney glomeruli at the midline [5]. Since the kidney and interrenal gland develop in parallel [16], with the DA acting as the source of angiogenesis in both organs [21,47], the role of blood flow in the morphogenesis of the interrenal gland was assessed. Kidney and interrenal tissue morphology was visualized by staining for 3b-Hsd enzymatic activity in Tg(wt1b: GFP) embryos, in which GFP is expressed in the glomerular podocytes, pronephric tubules, and proximal pronephric ducts [38,48], as well as in the exocrine pancreas due to a possible position effect of transgene insertion. The MO against tnnt2a, a gene essential for sarcomere assembly and heart contractility [42], was injected into Tg(wt1b: GFP) embryos, and 100% of the tnnt2a morphants (n = 56) displayed completely abolished heartbeat and blood flow. Consistent with the previous study [5], bilateral kidney glomeruli in control embryos assembled at the midline by 54 hpf, but failed to fuse in tnnt2a morphants (Figure 1A). The steroidogenic tissue in blood flow-deficient embryos grew as a round, tightly packed cell aggregate, without extending protrusions as in the case of controls. The extent of migration was quantified by measuring the distance between the tip of medially extending steroidogenic tissue and the midline, and although there was no difference at 54 hpf, the distance was decreased in morphants relative to control embryos at 77 hpf (Figure 1B). The inhibition of medial extension was not due to growth arrest of interrenal tissue, since organ size—as assessed by densitometric analysis of 3b-Hsd activity staining—was similar in tnnt2aMO-injected and control embryos at all stages examined (Figure 1C). Medial Extension of Steroidogenic Tissue during Interrenal Organ Assembly is Regulated by Blood Interrenal Organ Assembly is Regulated by Blood Flow The pharmacological agent 2,3-BDM, which affects heart rate without affecting cell viability, has previously been used to evaluate the role of blood flow in zebrafish organogenesis [5,6,51]. To rule out the possibility that defective medial extension of the interrenal tissue in tnnt2a morphants was due to an early effect on blood vessel morphogenesis, blood flow was inhibited in embryos by application of 2,3-BDM from 1.5 dpf, by which interrenal medial extension and organ assembly are initiated (Figure 2). The 2,3- BDM treatment on zebrafish embryos leads to decreased myofibrillar ATPase and myocardial force in a dose-dependent manner, which affects heart rate at a concentration as low as 2 mM and is sufficient to eliminate blood flow at 6 mM [46]. Consistent with the previous study, a concentration of 6 mM produced a 53% reduction in heart rate and a cessation of blood flow (Video S1) compared to control embryos (Video S2). A lower Microinjection of Antisense Morpholino Oligonucleotides (MOs) Microinjection of Antisense Morpholino Oligonucleotides (MOs) The MO for cardiac troponin T2a (tnnt2aMO) (59-CAT GTT TGC TCT GAT CTG ACA CGC A-39) [42], along with klf2aMO (59-GGA CCT GTC CAG TTC ATC CTT CCA C-39) [43], and mmp2MO (59-GGG AGC TTA GTA AAC ACA AAC CTG T-39) [44] were synthesized by Genetools LLC and diluted in 16Danieau solution, before injection into one- to two-cell stage embryos using a Nanoject (Drummond Scientific Company) at dosages of 1.0, 1.2, and 1.2 pmole per embryo, respectively. 3b-Hydroxysteroid Dehydrogenase (3b-Hsd) Staining, In Situ Hybridization (ISH), Immunohistochemistry (IHC), Densitometry and Imaging Embryos used for histological analysis were treated with 0.03% phenylthiourea (Sigma) from 12 h post-fertilization (hpf) onwards to inhibit pigmentation. The 3b-Hsd activity staining, ISH [9], and IHC [41] were performed with modifications according to previously published methods. To delineate the morphology of steroidogenic interrenal tissue, histochemical staining for 3b-Hsd enzymatic activity was per- formed on whole embryos, and Nomarski images were captured using a BX51 microscope (Olympus). For whole-mount ISH, digoxigenin (DIG)- and fluorescein- labeled antisense riboprobes were synthesized from linearized plasmids of dopamine b hydroxylase (dbh) and ff1b (nr5a1a) genes, respectively; the probes were detected with alkaline phosphatase- conjugated anti-DIG or -fluorescein antibody (Roche), and visualized with 5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium (Promega) or Fast Red (Roche). Stained embryos were PLOS ONE \| www.plosone.org September 2014 \| Volume 9 \| Issue 9 \| e107997 September 2014 \| Volume 9 \| Issue 9 \| e107997 2 Hemodynamic Regulation of Interrenal Gland Assembly Hemodynamic Regulation of Interrenal Gland Assembly flat-mounted and photographed under an Axioplan II microscope (Zeiss). flat-mounted and photographed under an Axioplan II microscope (Zeiss). Blood Flow is Required for the Morphogenesis of Kidney Glomerulus and Interrenal Tissue To further evaluate the effect of tissue growth on interrenal morphogenetic movement, embryos were treated with camptothecin, which blocks cell proliferation in zebrafish embryos [45], from 48 to 57 hpf. Camptothecin-treated embryos showed an 18% reduction in interrenal tissue size compared to controls. Notably, the formation of protrusions was unaffected by camptothecin treatment (Figure 1D), suggesting that cell growth is not the major determinant for interrenal medial extension. Taken together, these results indicate that blood flow is required for the morphogenetic movement of both kidney and interrenal tissues. Based on the early defects in the ventral DA caused by tnnt2a knockdown [49,50], experiments were per- formed to establish whether interrenal morphogenetic movement during the temporal window of organ assembly is specifically subject to regulation by blood flow. For the quantification of 3b-Hsd activity, images of deyolked embryos in each group were taken with identical illumination and magnification using Axioskop 2 Plus microscope equipped with AxioVision 3.0 software (Carl Zeiss). Signal area and density were measured using Image Gauge Program, version 4.0 (Fuji Photo Film). Videos of embryos oriented with the anterior toward the left were taken by using an SMZ1500 microscope equipped with an AM4023X Dino-Eye eyepiece camera (Nikon). Pharmacological Treatment Camptothecin (Sigma # C9911) treatment was performed according to a previously described method [45] with modifica- tions. The compound (60 mM in 0.1% dimethyl sulfoxide [DMSO]) was applied to 48 hpf embryos, which were harvested at 57 hpf for the 3b-Hsd activity assay. The treatment of embryos with 2,3-butanedione 2-monoxime (2,3-BDM; Sigma #B0753) was as described in an earlier report [46], except that dechor- ionated embryos were immersed in various concentrations of 2,3- BDM starting from 1.5 dpf. Norepinephrine treatment was performed by treating dechorionated embryos with 0.01, 0.1 or 1 mM norepinephrine (Sigma A7257) freshly prepared in egg water. L-NAME treatment was performed by treating dechor- ionated embryos with freshly prepared 100 mM Nv-nitro-L- arginine methyl ester (L-NAME) (#N5751, Sigma) in egg water with 0.1% DMSO at 36 hpf. For L-arginyl-L-glycyl-L-aspartic acid (RGD; #G1269, Sigma) treatment, the peptide was reconstituted to 1 mM in filter-sterilized egg water and applied to dechorionated embryos at a final concentration of 100 mM at 26 hpf; embryos were collected at 2.5 dpf and fixed for histological assays. September 2014 \| Volume 9 \| Issue 9 \| e107997 Statistical Analysis All quantitative data are expressed as the mean 6 standard error of the mean. Data were evaluated by analysis of variance (ANOVA), followed by Duncan’s new multiple range test (Duncan’s multiple test) or Student’s t test. P,0.05 was considered statistically significant. PLOS ONE \| www.plosone.org September 2014 \| Volume 9 \| Issue 9 \| e107997 3 Hemodynamic Regulation of Interrenal Gland Assembly Figure 1. Morphology of pronephros and interrenal tissue in the absence of blood flow. (A) The interrenal steroidogenic tissue positive for 3b-Hsd activity forms an extension that protrudes toward the midline by 54 hpf (cell protrusions marked by red arrows), while kidney glomeruli delineated by wt1b: GFP expression (G, yellow arrowheads) assemble at the midline. Morphogenetic movements of kidney glomeruli and steroidogenic tissues are defective in the tnnt2a morphant. All panels show dorsal views of representative embryos. (B) Quantification of effects of tnnt2aMO injection on interrenal migration. The distance between the midline and the migrating tip of steroidogenic tissue was designated as positive if the migrating tip had not reached the midline, and negative if the tip had migrated across the midline. (C) Relative density of steroidogenic tissue, as assessed by 3b-Hsd activity staining in the ventral surface, in tnnt2a morphants compared to wild-type controls. The number of embryos in each group is indicated in parentheses in (B) and (C). Histograms with different letters above them are significantly different (ANOVA and Duncan’s multiple test, P,0.05). (D) Effect of camptothecin treatment from 48 to 57 hpf on 3b-Hsd activity in steroidogenic cells. A, anterior; P, posterior; L, left; R, right. Broken yellow lines indicate position of the midline. Abbreviations: glomerulus (G). Scale bar, 50 mm. doi:10.1371/journal.pone.0107997.g001 Hemodynamic Regulation of Interrenal Gland Assembly Figure 1. Morphology of pronephros and interrenal tissue in the absence of blood flow. (A) The interrenal steroidogenic tissue positive for 3b-Hsd activity forms an extension that protrudes toward the midline by 54 hpf (cell protrusions marked by red arrows), while kidney glomeruli delineated by wt1b: GFP expression (G, yellow arrowheads) assemble at the midline. Morphogenetic movements of kidney glomeruli and steroidogenic tissues are defective in the tnnt2a morphant. All panels show dorsal views of representative embryos. (B) Quantification of effects of tnnt2aMO injection on interrenal migration. September 2014 \| Volume 9 \| Issue 9 \| e107997 Statistical Analysis The distance between the midline and the migrating tip of steroidogenic tissue was designated as positive if the migrating tip had not reached the midline, and negative if the tip had migrated across the midline. (C) Relative density of steroidogenic tissue, as assessed by 3b-Hsd activity staining in the ventral surface, in tnnt2a morphants compared to wild-type controls. The number of embryos in each group is indicated in parentheses in (B) and (C). Histograms with different letters above them are significantly different (ANOVA and Duncan’s multiple test, P,0.05). (D) Effect of camptothecin treatment from 48 to 57 hpf on 3b-Hsd activity in steroidogenic cells. A, anterior; P, posterior; L, left; R, right. Broken yellow lines indicate position of the midline. Abbreviations: glomerulus (G). Scale bar, 50 mm. doi:10.1371/journal.pone.0107997.g001 concentration of 2,3-BDM (2 mM) caused a 26% decrease in heart rate and a visibly weakened blood flow (Video S3). To examine the morphology of the DA adjacent to the kidney and interrenal regions, 2,3-BDM was applied to Tg(kdrl: EGFP)s843 embryos that express GFP in the developing blood vascular structure [40], which were then harvested at 2.5 dpf, when the migration of interrenal cells can be clearly observed [19]. Treatment with 2 or 6 mM 2,3-BDM suppressed interrenal medial extension across the midline (Figure 2B, B’’, C, C’’), with similar effects observed at both concentrations (Figure 2D). However, the interrenal tissue had more protrusions at 2 than at 6 mM 2,3-BDM (Figure 2B, C), suggesting that the effect of 2,3- BDM was dose-dependent. In contrast, the morphology of the DA and the pronephric glomerulus was unperturbed by 2,3-BDM treatment. Thus, the inhibition of medial extension of the interrenal tissue caused by loss of blood flow was not due to a general defect in the DA. Conversely, as the heart rate was accelerated by norepinephrine treatment from 33 hpf, a significant enhancement of interrenal tissue extension was detected at 2 dpf (Figure 2E-E’’, F-F’’, G-G’’, H). Norepinephrine accelerated the heart rate of developing embryos in a dose-dependent manner (Figure S1). 0.01, 0.1 and 1 mM of norepinephrine treatments on embryos at 33 hpf led to a 23%, 29% and 41% increase of heart rate, respectively. Hemodynamic Regulation of Interrenal Gland Assembly designated as positive if the migrating tip had not reached the midline, and negative if the tip had migrated across the midline. P,0.05; N.S., not significant (Student’s t-test). (I) Suppressing effect of interrenal cell migration by 2,3-BDM at 6 mM from 1.5 dpf was reversible at 3 dpf, as the 2,3- BDM applied from 1.5 dpf was washed out at 2.5 dpf for restoring blood flow. The interrenal tissue in recovered embryos extended across the midline at 3 dpf and displayed a migration distance not significantly different from that in control embryos at 2.5 dpf. Histograms with different letters above them are significantly different (ANOVA and Duncan’s multiple test, P,0.05). A, anterior; P, posterior; L, left; R, right. Broken white lines indicate position of the midline. Abbreviations: glomerulus (G), posterior cardinal vein (PCV). Scale bar, 50 mm. doi:10.1371/journal.pone.0107997.g002 tnnt2a morphant [21]; accordingly, the present results indicated that the blood flow was not required for the sprouting of the IRV from the DA, but may play a role in its extension. Furthermore, the interrenal tissue in 2,3-BDM-treated embryos (Figure 3B–B’, C–C’) and tnnt2a morphants [21] only interacted with the tip of the IRV but not the DA, while the extending interrenal tissue in the control embryo remained closely associated with both the ventral DA and the IRV (Figure 3A–A’). phenotype of interrenal tissue, as verified from the extending protrusions (red arrows in Figure 2F, G), which was consistent with the results of the quantification of interrenal tissue extension (Figure 2H). However, no significant difference in steroidogenic tissue extension could be detected between embryos treated with 0.1 or 1 mM of norepinephrine. Norepinephrine at 0.01 mM also enhanced steroidogenic tissue extension (4.362.0 mm, n = 24) compared to control embryos, while no significant statistical difference was found among norepinephrine treatments at 0.01, 0.1 and 1 mM. Our results thus indicated that a moderate elevation of heart rate by 23% was sufficient to promote migration of steroidogenic interrenal cells, although further increase of heart rate by treating with higher concentrations of norepinephrine did not lead to a dose-dependent enhancing effect on interrenal medial extension. To test whether accelerated blood flow could promote extension of the IRV, embryos were treated with norepinephrine at 0.1 or 1 mM at 33 hpf and harvested at 2 dpf for analysis (Figure 3E-G, E’-G’). Hemodynamic Regulation of Interrenal Gland Assembly It was found that norepinephrine at both concentrations significantly increased length of the IRV (Figure 3H), yet no difference in the IRV growth was observed between 0.1 and 1 mM of norepinephrine treatments. The promoting effects of norepinephrine treatments on extension of the IRV were therefore highly correlated with those on interrenal medial extension (Figure 2H). Similar to the case of interrenal medial extension in Figure 2I, restoring blood flow by 2,3-BDM washout after the treatment from 1.5 to 2.5 dpf led to a recovery of IRV growth at 3 dpf, with the IRV length in 3-dpf recovered embryos not significantly different from that in 2.5-dpf control embryos (Figure 3I, S3). To confirm the relationship between blood flow and steroido- genic cell migration, we further tested whether interrenal medial extension repressed by 2,3-BDM treatment could be recovered by restoring the blood flow (Figure 2I, Figure S2). Steroidogenic tissue migrated across the midline by 2.5 dpf (Figure S2A, A’’) and continued to extend and form a bilobed organ structure by 3 dpf (Figure S2C, C’’). Extension of steroidogenic tissue was arrested as 2,3-BDM at 6 mM was applied to embryos from 1.5 dpf onwards (Figure S2B, B’’, D, D’’). The steroidogenic tissue extension in 2,3- BDM-treated embryos was recovered at 3 dpf as 2,3-BDM was washed out at 2.5 dpf (Figure S2E, E’’). The interrenal tissue in 2,3-BDM-treated embryos appeared to be located further away from the midline at 3 dpf than at 2.5 dpf (Figure 2I), possibly due to continuous growth of peri-interrenal structures from 2.5 to 3 dpf. It is interesting to note that there was no significant difference of migration distance between control embryos at 2.5 dpf and recovered embryos at 3 dpf (Figure 2I), implying that the inhibited interrenal medial extension by 2,3-BDM treatment from 1.5 to 2.5 dpf could be rescued by resuming blood flow for 12 hours. Taken together, results in Figure 2 demonstrated that pharmacologic repression and acceleration of heart rates are well correlated with the extent of interrenal medial extension. Furthermore, the inhibited steroidogenic tissue extension caused by arrested blood flow was recoverable following restoration of blood flow, providing strong evidence that migratory activity of interrenal steroidogenic cells is indeed modulated by blood flow. Our results from pharmacologic inhibition and acceleration of blood flow therefore strongly support that blood flow regulates both interrenal medial extension and IRV growth during interrenal organ assembly. Blood Flow Regulates IRV Extension during Interrenal Organ Assembly Blood Flow Regulates IRV Extension during Interrenal Organ Assembly Hemodynamic Regulation of Interrenal Gland Assembly While the processes of interrenal medial extension and IRV growth occur synchronously during development [21], they are both influenced by blood flow in a highly correlated manner. Therefore, it leads to the hypothesis that there might be a common flow-regulated molecular and cellular mechanism by which both interrenal medial extension and IRV angiogenesis are regulated. The initiation of IRV angiogen- esis requires deposition of Fn at the ventral DA near the interrenal tissue, and the accumulation of this protein in the local microenvironment supports IRV extension [21]. While the interrenal tissue is closely associated with both the DA and the IRV, vessel-derived Fn functions at the tissue-vessel interface and thus modulates the migration of steroidogenic cells [19]. There- fore, it is possible that the vascular microenvironment established during IRV angiogenesis is regulated by blood flow, which in turn modulates the migration of steroidogenic cells. Statistical Analysis Compared to the control embryo (Figure 2E), both 0.1 and 1 mM of norepinephrine treatements led to a more evident migratory September 2014 \| Volume 9 \| Issue 9 \| e107997 PLOS ONE \| www.plosone.org 4 Hemodynamic Regulation of Interrenal Gland Assembly ects of 2,3-BDM and norepinephrine on morphogenetic movements of interrenal tissue. For repression of blo 3 embryos were treated with (A–A’’) vehicle (control), or 2,3-BDM at a concentration of (B–B’’) 2 mM or (C–C’’) 6 mM from 1 the medial extension of steroidogenic cells was observed in 2,3-BDM-treated embryos at 2.5 dpf, while the morphology of th glomerulus (yellow arrowheads) appeared unperturbed. Protrusions (red arrow) formed at the lower concentration; the ph re at the higher concentration. For acceleration of blood flow, Tg(kdrl: GFP)s843 embryos were treated with (E-E’’) vehicle (co e at a concentration of (F–F’’) 0.1 mM or (G-G’’) 1 mM from 33 hpf. An enhancement of interrenal medial extension, as evid of protrusions, was observed in norepinephrine-treated embryos at 2 dpf. The effects of 2,3-BDM and norepinephrine treatm ation were quantified in (D) and (H), respectively. The distance between the midline and migrating tip of steroidogenic ti Figure 2. Effects of 2,3-BDM and norepinephrine on morphogenetic movements of interrenal tissue. For repression of blood flow, Tg(kdrl: GFP)s843 embryos were treated with (A–A’’) vehicle (control), or 2,3-BDM at a concentration of (B–B’’) 2 mM or (C–C’’) 6 mM from 1.5 dpf. A suppression of the medial extension of steroidogenic cells was observed in 2,3-BDM-treated embryos at 2.5 dpf, while the morphology of the DA and the pronephric glomerulus (yellow arrowheads) appeared unperturbed. Protrusions (red arrow) formed at the lower concentration; the phenotype was more severe at the higher concentration. For acceleration of blood flow, Tg(kdrl: GFP)s843 embryos were treated with (E-E’’) vehicle (control), or norepinephrine at a concentration of (F–F’’) 0.1 mM or (G-G’’) 1 mM from 33 hpf. An enhancement of interrenal medial extension, as evidenced by the formation of protrusions, was observed in norepinephrine-treated embryos at 2 dpf. The effects of 2,3-BDM and norepinephrine treatments on interrenal migration were quantified in (D) and (H), respectively. The distance between the midline and migrating tip of steroidogenic tissue was n morphogenetic movements of interrenal tissue. September 2014 \| Volume 9 \| Issue 9 \| e107997 PLOS ONE \| www.plosone.org 5 Blood Flow is Required for Patterning the Fn-rich Microenvironment and pFak Distribution in the Interrenal Region The IRV length in recovered embryos at 3 dpf was not significantly different from that in control embryos at 2.5 dpf. Histograms with different letters above them are significantly different (ANOVA and Duncan’s multiple test, P,0.05). D, dorsal; V, ventral; L, left; R, right. Abbreviations: interrenal tissue (IR), notochord (NC), somite (S). Scale bar, 25 mm. doi:10.1371/journal.pone.0107997.g003 y g y Figure 3. Effects of 2,3-BDM and norepinephrine on IRV formation. For repression of blood flow, Tg(kdrl: GFP)s843 embryos were treated with (A–A’) vehicle (control), or 2,3-BDM at a concentration of (B–B’) 2 mM or (C–C’) 6 mM from 1.5 dpf, and harvested at 2.5 dpf. For acceleration of blood flow, Tg(kdrl: GFP)s843 embryos were treated with (E-E’) vehicle (control), or norepinephrine at a concentration of (F–F’) 0.1 mM or (G-G’) 1 mM from 33 hpf, and harvested at 2 dpf. Transverse sections of harvested embryos were subject to analysis of 3b-Hsd activity (black) and GFP expression (green). IRV lengths of 2,3-BDM- or norepinephrine-treated embryos were quantified in (D) and (H), respectively; which were verified from confocal Z-stacks covering the full range of IRV growth, and measurements were made from single focal planes displaying the maximal range of ventrally extending IRV. P,0.05, **P,0.0005, N.S., not significant (Student’s t-test). (I) Repressing effect of 2,3-BDM (6 mM) on IRV growth was reversible at 3 dpf, as the 2,3-BDM applied from 1.5 dpf was washed out at 2.5 dpf. The IRV length in recovered embryos at 3 dpf was not significantly different from that in control embryos at 2.5 dpf. Histograms with different letters above them are significantly different (ANOVA and Duncan’s multiple test, P,0.05). D, dorsal; V, ventral; L, left; R, right. Abbreviations: interrenal tissue (IR), notochord (NC), somite (S). Scale bar, 25 mm. doi:10.1371/journal.pone.0107997.g003 mediated signaling. In contrast, pFak distribution was readily detected at the somites, gut tube, and swim bladder in tnnt2a morphants (Figure 4B, B’) and 2,3-BDM-treated embryos (Figure 4C, C’). These results indicate that the interrenal microenvironment established during IRV angiogenesis is per- turbed by reduced blood flow, resulting in the suppression of Fn- pFak signaling and steroidogenic cell migration. abnormally distributed, raising the possibility that in the absence of blood flow, the polymerization of Fn into fibrils was disrupted. Blood Flow is Required for Patterning the Fn-rich Microenvironment and pFak Distribution in the Interrenal Region g y Interrenal medial extension temporally coincides with IRV angiogenesis, which is promoted by the IRV-associated vascular microenvironment [21], and therefore the effect of reduced blood flow on IRV angiogenesis was examined from 1.5 to 2.5 dpf (Figure 3). IRV growth was initiated normally when blood flow was inhibited starting at 1.5 dpf. However, in embryos treated with 2 or 6 mM 2,3-BDM, IRV lengths were reduced and the vessels reached but did not extend ventrally through the interrenal tissue, with a more severe phenotype observed at the higher concentra- tion (Figure 3B’–C’, D), indicating a dose-dependent effect of 2,3- BDM on IRV extension. Our previous study showed that IRV directionality, but not initiation of angiogenesis, is perturbed in the Possible perturbations in the interrenal microenvironment of blood flow-deficient embryos were assessed using the Tg(ff1bEx2: GFP) line, in which the GFP expression recapitulates the endogenous expression of ff1b [52]—the teleostean ortholog of mammalian SF-1—and marks the ontogeny of steroidogenic interrenal tissue [16,17]. Embryos were examined for expression of Fn and pFak (Figure 4). The elimination of blood flow by tnnt2aMO injection (Figure 4B–B’’) or 6 mM 2,3-BDM treatment (Figure 4C–C’’) did not diminish Fn accumulation in the interrenal microenvironment (Figure 4D). However, Fn was PLOS ONE \| www.plosone.org September 2014 \| Volume 9 \| Issue 9 \| e107997 6 Hemodynamic Regulation of Interrenal Gland Assembly Figure 3. Effects of 2,3-BDM and norepinephrine on IRV formation. For repression of blood flow, Tg(kdrl: GFP)s843 embryos were treated with (A–A’) vehicle (control), or 2,3-BDM at a concentration of (B–B’) 2 mM or (C–C’) 6 mM from 1.5 dpf, and harvested at 2.5 dpf. For acceleration of blood flow, Tg(kdrl: GFP)s843 embryos were treated with (E-E’) vehicle (control), or norepinephrine at a concentration of (F–F’) 0.1 mM or (G-G’) 1 mM from 33 hpf, and harvested at 2 dpf. Transverse sections of harvested embryos were subject to analysis of 3b-Hsd activity (black) and GFP expression (green). IRV lengths of 2,3-BDM- or norepinephrine-treated embryos were quantified in (D) and (H), respectively; which were verified from confocal Z-stacks covering the full range of IRV growth, and measurements were made from single focal planes displaying the maximal range of ventrally extending IRV. P,0.05, *P,0.0005, N.S., not significant (Student’s t-test). (I) Repressing effect of 2,3-BDM (6 mM) on IRV growth was reversible at 3 dpf, as the 2,3-BDM applied from 1.5 dpf was washed out at 2.5 dpf. Blood Flow is Required for Patterning the Fn-rich Microenvironment and pFak Distribution in the Interrenal Region Transverse sections of Tg(ff1bEx2: GFP) embryos were (A–A’’) uninjected (control), (B–B’’) injected with tnnt2aMO, or (C–C’’) treated with 6 mM 2,3- BDM from 1.5 dpf. Embryos were harvested at 2.5 dpf and assayed for expression of GFP (green), Fn (red), and pFak (blue in A–C and A’–C’; white in A’’–C’’). Images are single confocal planes showing the maximal transverse dimension of ff1bGFP-expressing steroidogenic tissue of a representative embryo, with magnified views shown in (A’’–C’’). (D) Fluorescence intensity of Fn in the DA and IRV selected as regions of interest (ROI; white lines in A’–C’) normalized to the size of the ROI. (E) Total fluorescence intensities of pFak within the steroidogenic tissue (ROI marked by orange lines in A’’– C’’) were normalized to the size of the cluster. The difference between the treatment and the control groups was analyzed by Student’s t-test. P, 0.001, N.S., not significant. D, dorsal; V, ventral; L, left; R, right. Abbreviations: arbitrary units (A.U.), interrenal tissue (IR), notochord (NC), somite (S), swim bladder (SB). Scale bar, 25 mm. doi:10.1371/journal.pone.0107997.g004 Figure 4. Effect of blood flow inhibition on the ECM microenvironment and pFak distribution in interrenal steroidogenic tissue. Transverse sections of Tg(ff1bEx2: GFP) embryos were (A–A’’) uninjected (control), (B–B’’) injected with tnnt2aMO, or (C–C’’) treated with 6 mM 2,3- BDM from 1.5 dpf. Embryos were harvested at 2.5 dpf and assayed for expression of GFP (green), Fn (red), and pFak (blue in A–C and A’–C’; white in A’’–C’’). Images are single confocal planes showing the maximal transverse dimension of ff1bGFP-expressing steroidogenic tissue of a representative embryo, with magnified views shown in (A’’–C’’). (D) Fluorescence intensity of Fn in the DA and IRV selected as regions of interest (ROI; white lines in A’–C’) normalized to the size of the ROI. (E) Total fluorescence intensities of pFak within the steroidogenic tissue (ROI marked by orange lines in A’’– C’’) were normalized to the size of the cluster. The difference between the treatment and the control groups was analyzed by Student’s t-test. P, 0.001, N.S., not significant. D, dorsal; V, ventral; L, left; R, right. Abbreviations: arbitrary units (A.U.), interrenal tissue (IR), notochord (NC), somite (S), swim bladder (SB). Scale bar, 25 mm. doi:10.1371/journal.pone.0107997.g004 Blood Flow is Required for Patterning the Fn-rich Microenvironment and pFak Distribution in the Interrenal Region To verify whether aberrant Fn deposition perturbed signaling events within the interrenal tissue, the localization of pFak—a downstream effector of Fn-Integrin signaling and an indicator of the dynamic reorganization of focal adhesions during cell migration—was examined (Figure 4A–C’’). Indeed, pFak level within the interrenal tissue was significantly reduced compared to control embryos (Figure 4E), suggesting a disruption of Integrin- September 2014 \| Volume 9 \| Issue 9 \| e107997 PLOS ONE \| www.plosone.org 7 PLOS ONE \| www.plosone.org Hemodynamic Regulation of Interrenal Gland Assembly e 4. Effect of blood flow inhibition on the ECM microenvironment and pFak distribution in interrenal steroidogenic tissue. verse sections of Tg(ff1bEx2: GFP) embryos were (A–A’’) uninjected (control), (B–B’’) injected with tnnt2aMO, or (C–C’’) treated with 6 mM 2,3- rom 1.5 dpf. Embryos were harvested at 2.5 dpf and assayed for expression of GFP (green), Fn (red), and pFak (blue in A–C and A’–C’; white in ). Images are single confocal planes showing the maximal transverse dimension of ff1bGFP-expressing steroidogenic tissue of a representative yo, with magnified views shown in (A’’–C’’). (D) Fluorescence intensity of Fn in the DA and IRV selected as regions of interest (ROI; white lines in normalized to the size of the ROI. (E) Total fluorescence intensities of pFak within the steroidogenic tissue (ROI marked by orange lines in A’’– ere normalized to the size of the cluster. The difference between the treatment and the control groups was analyzed by Student’s t-test. P, N S not significant D dorsal; V ventral; L left; R right Abbreviations: arbitrary units (A U ) interrenal tissue (IR) notochord (NC) somite (S) igure 4. Effect of blood flow inhibition on the ECM microenvironment and pFak distribution in interrenal steroidogenic tissue. ransverse sections of Tg(ff1bEx2: GFP) embryos were (A–A’’) uninjected (control), (B–B’’) injected with tnnt2aMO, or (C–C’’) treated with 6 mM 2,3- DM from 1 5 dpf Embryos were harvested at 2 5 dpf and assayed for expression of GFP (green) Fn (red) and pFak (blue in A C and A’ C’; white in flow inhibition on the ECM microenvironment and pFak distribution in interrenal steroidogenic tissue Figure 4. Effect of blood flow inhibition on the ECM microenvironment and pFak distribution in inte Figure 4. Effect of blood flow inhibition on the ECM microenvironment and pFak distribution in interrenal steroidogenic tissue. September 2014 \| Volume 9 \| Issue 9 \| e107997 Klf2a and MMP2 are Required for Migration, Angiogenesis, and Fn-pFak signaling in the interrenal tissue The number of embryos in each group is indicated in parentheses. The extent of interrenal medial extension of control 2.5-dpf embryos in panel 5D was not statistically different from those in Figure 2D and 2I. Fn and pFak expression in the interrenal region was examined in (E–E’’’’) uninjected (control), and (F–F’’’’) klf2aMO- and (G–G’’’’) mmp2MO-injected embryos by IHC. Images show transverse sections of a representative embryo from each treatment group. (H) Quantification of the effects of klf2aMO and mmp2MO on IRV growth, Fn level in the vicinity of the DA and IRV (ROI marked by broken lines in E’’–G’’), and pFak level in the steroidogenic tissue (ROI marked by orange lines in E’’’’–G’’’’). The number of embryos in each group is indicated in parentheses. Fluorescence intensities of Fn and pFak were normalized to their respective ROI sizes. The difference between the treatment and the control groups was analyzed by Student’s t-test. P, 0.05, **P,0.001, N.S., not significant. A, anterior; P, posterior; L, left; R, right; D, dorsal; V, ventral. Abbreviations: notochord (NC). Scale bar, 25 mm. doi:10.1371/journal.pone.0107997.g005 Figure 5. Suppression of interrenal tissue migration in klf2a and mmp2 morphants. Dorsal view of interrenal steroidogenic tissue (IR, red arrows) as detected by 3b-Hsd activity staining, with adjacent vasculature marked by GFP expression. Tg(kdrl: GFP)s843 embryos were (A, A’) uninjected (control), or injected with (B, B’) klf2aMO or (C, C’) mmp2MO. (D) Quantification of the effects of MO-mediated gene knockdown on interrenal migration. The distance between the midline and the migrating tip of steroidogenic tissue was designated as positive if the migrating tip had not reached the midline and negative if it had crossed the midline. The number of embryos in each group is indicated in parentheses. The extent of interrenal medial extension of control 2.5-dpf embryos in panel 5D was not statistically different from those in Figure 2D and 2I. Fn and pFak expression in the interrenal region was examined in (E–E’’’’) uninjected (control), and (F–F’’’’) klf2aMO- and (G–G’’’’) mmp2MO-injected embryos by IHC. Images show transverse sections of a representative embryo from each treatment group. (H) Quantification of the effects of klf2aMO and mmp2MO on IRV growth, Fn level in the vicinity of the DA and IRV (ROI marked by broken lines in E’’–G’’), and pFak level in the steroidogenic tissue (ROI marked by orange lines in E’’’’–G’’’’). The number of embryos in each group is indicated in parentheses. Klf2a and MMP2 are Required for Migration, Angiogenesis, and Fn-pFak signaling in the interrenal tissue observed at 2.5 dpf, and circulation was unaffected, making it possible to evaluate the specific effect of hemodynamic forces on interrenal development at this stage. Consistent with the previous finding that primary vascular structures are not perturbed in the klf2a morphant [24], the axial vasculature was grossly normal in the peri-interrenal region at 2.5 dpf (Figure 5B’). However, the migration of steroidogenic cells was inhibited (Figure 5B, D), and IRV growth and directionality were perturbed (Figure 5F, H). In contrast to the tnnt2a morphant and 2,3-BDM-treated embryos in which the steroidogenic tissue and DA were not closely associated (Figure 3) [21], the interrenal tissue was proximal to the DA in the To confirm whether hemodynamics, and not circulating factors, account for the effect of blood flow on steroidogenic tissue migration and angiogenesis, the role of Klf2a and MMP2 in interrenal morphogenesis was evaluated (Figure 5). While Klf2a deficiency leads to heart failure at 3 dpf [24], as evidenced by pericardial edema and venous pooling of blood around the yolk sac, a heart rate similar to that of control embryos is observed at 2 dpf [3]. In the present study, only mild cardiac edema was September 2014 \| Volume 9 \| Issue 9 \| e107997 PLOS ONE \| www.plosone.org 8 Hemodynamic Regulation of Interrenal Gland Assembly Figure 5. Suppression of interrenal tissue migration in klf2a and mmp2 morphants. Dorsal view of interrenal steroidogenic tissue (IR, red arrows) as detected by 3b-Hsd activity staining, with adjacent vasculature marked by GFP expression. Tg(kdrl: GFP)s843 embryos were (A, A’) uninjected (control), or injected with (B, B’) klf2aMO or (C, C’) mmp2MO. (D) Quantification of the effects of MO-mediated gene knockdown on interrenal migration. The distance between the midline and the migrating tip of steroidogenic tissue was designated as positive if the migrating tip had not Figure 5. Suppression of interrenal tissue migration in klf2a and mmp2 morphants. Dorsal view of interrenal steroidogenic tissue (IR, red arrows) as detected by 3b-Hsd activity staining, with adjacent vasculature marked by GFP expression. Tg(kdrl: GFP)s843 embryos were (A, A’) uninjected (control), or injected with (B, B’) klf2aMO or (C, C’) mmp2MO. (D) Quantification of the effects of MO-mediated gene knockdown on interrenal migration. The distance between the midline and the migrating tip of steroidogenic tissue was designated as positive if the migrating tip had not reached the midline and negative if it had crossed the midline. Hemodynamic Regulation of Interrenal Gland Assembly klf2a morphant. Despite these phenotypic differences between blood flow-deficient and klf2a morphant embryos, both types of embryos showed abnormal Fn distribution and reduced pFak level in the interrenal area (Figure 5F’–F’’’’, H). The RGD peptide, an antagonist of Fn, was applied to Tg(ff1bEx2: GFP) embryos at a concentration of 100 mM starting from 26 hpf, when circulation is initiated; in these embryos, junctional b-Catenin distribution was significantly higher than in controls at 2.5 dpf (Figure 6J–J’, M). RGD effectively reduced pFak level in the interrenal area (Figure S5), thus suggesting that the inhibition of pFak signaling was responsible for the observed suppression of EMT-like changes in steroidogenic cells. Moreover, b-Catenin accumulation at cell-cell junctions was evident in both klf2a and mmp2 morphants at 2.5 dpf (Figure 6K–K’, L–L’, M) as compared to the control embryo. Although blood flow regulates the generation of hematopoietic stem cells from the DA by a klf2a-nitric oxide (NO) pathway [49,50], steroidogenic interrenal tissue migration and angiogenesis were NO-independent (Figure S4). Treatment with the endothelial NO synthase inhibitor L-NAME from 1.5 to 2.5 dpf had no effect on interrenal tissue migration and angiogenesis, indicating that blood flow and Klf2a do not regulate interrenal morphogenesis through activation of NO signaling. Consistent with the results of b-Catenin expression, immuno- histochemial analysis of E-cadherin, a central component of cell- cell adhesion junction which is required for the formation of epithelia [56,57], detected a clear epithelial phenotype of the interrenal tissue at 2 dpf (Figure S6A, A’). At 2.5 dpf, the E- cadherin expression was reduced at the interrenal tissue where a migratory phenotype was manifested (Figure S6B, B’). In contrast, junctional E-cadherin distribution at the interrenal tissue was not reduced in 2.5 dpf embryos where interrenal medial migration was suppressed by a disruption of either circulation (Figure S6C, C’) or pFak-mediated signaling (Figure S6D, D’), or hemodynamics- regulated molecules (Figure S6E, E’, F, F’). It was noted that RGD-treated embryos and mmp2 morphants at 2.5 dpf displayed a higher expression level of junctional E-cadherin than control embryos at 2 dpf did (Figure S6G), with the underlying mechanism remaining unclear. Nevertheless, the immunohistochemistry results of b-Catenin and E-cadherin both indicated a clear reduction of epithelial nature at the interrenal tissue from 2 to 2.5 dpf. Medial Extension of the Interrenal Tissue Involves EMT-like Changes in Steroidogenic Cell Morphology that are Hemodynamically Regulated and pFak-Dependent y y g p p The interaction of cancerous epithelial cells with Fn in vitro promotes EMT [53,54]. The morphology of the interrenal steroidogenic tissue and the associated Fn-enriched microenviron- ment suggested that an EMT-like phenotypic change could occur during interrenal medial extension. An examination of interrenal tissue morphology by high resolution Nomarski microscopy revealed that while steroidogenic cells positive for 3b-Hsd activity formed a cluster to the right of the midline by 48 hpf (Figure 6A, B), protrusions were detected at 48 hpf that became more evident by 60 hpf (Figure 6C) and continuously spread across the midline until a bilobed structure was formed by 84 hpf (Figure 6E). During the morphogenetic movement, interrenal steroidogenic cells became more loosely associated with each other and demonstrated a mesenchymal-like phenotype with cell surface protrusions (Figure 6D, E). These morphological features likely reflected an EMT-like process during organ assembly. To confirm whether interrenal steroidogenic cells undergo a transformation from epithelial to mesenchymal phenotypes, the expression of b-Catenin—a marker for adherent junctions in epithelial cells [55]—was examined. At 2 dpf, b-Catenin was clearly detected at cell-cell junctions of interrenal steroidogenic cells visible by GFP expression in ff1bEx2: GFP embryos (Figure 6F–F’), while at 2.5 dpf, b-Catenin was markedly reduced within the steroidogenic tissue cluster (Figure 6G–G’), reflecting the adoption of a mesenchymal-like character. Hemodynamic Regulation of Interrenal Gland Assembly Moreover, mesenchymal phenotype of the interrenal tissue at 2.5 dpf correlated well with a rise of Vimentin expression in the ff1b-expressing steroidogenic cells (Figure S7A, A’, B, B’). Vimentin, a type III intermediate filament protein and a widely used mesenchymal marker, plays a predominant role for inducing changes in cell shape, adhesion and motility during the EMT [58,59]. In contrast to the reduction of b-Catenin and E-cadherin during interrenal medial extension, the Vimentin expresison is significantly increased from 2 to 2.5 dpf, and this accumulation of Vimentin was not detected in embryos deficient in either blood flow (Figure S7C, C’) or pFak signaling (Figure S6D, D’), or hemodynamic transducers (Figure S7E, E’, F, F’). Therefore, an inverse correlation between epithelial and mesenchymal markers was observed during interrenal medial expression. Taken together, these results indicate that blood flow, through mechanotransduc- tion and Fn-pFak signaling, promotes EMT-like changes in steroidogenic cells during interrenal organ assembly. g g g Embryos injected with 1.2 pmole mmp2MO had no major morphological abnormalities except for a kinked tail and mild (2%) reduction in heartbeat that did not cause any visible changes in blood flow. Nevertheless, morphants showed defects in migration, angiogenesis, and Fn-pFak signaling in the interrenal tissue (Figure 5C, D, G, H). This indicated that MMP2 activity may participate in the blood flow-regulated interrenal microenviron- ment. As observed upon klf2a knockdown, the steroidogenic tissue remained associated with both the DA and the IRV, suggesting that neither Klf2a nor MMP2 was essential for this association, which likely depends on other hemodynamically regulated molecules. Klf2a and MMP2 are Required for Migration, Angiogenesis, and Fn-pFak signaling in the interrenal tissue Fluorescence intensities of Fn and pFak were normalized to their respective ROI sizes. The difference between the treatment and the control groups was analyzed by Student’s t-test. P, 0.05, **P,0.001, N.S., not significant. A, anterior; P, posterior; L, left; R, right; D, dorsal; V, ventral. Abbreviations: notochord (NC). Scale bar, 25 mm. doi:10.1371/journal.pone.0107997.g005 Figure 5. Suppression of interrenal tissue migration in klf2a and mmp2 morphants. Dorsal view of interrenal steroidogenic tissue (IR, red arrows) as detected by 3b-Hsd activity staining, with adjacent vasculature marked by GFP expression. Tg(kdrl: GFP)s843 embryos were (A, A’) uninjected (control), or injected with (B, B’) klf2aMO or (C, C’) mmp2MO. (D) Quantification of the effects of MO-mediated gene knockdown on interrenal migration. The distance between the midline and the migrating tip of steroidogenic tissue was designated as positive if the migrating tip had not reached the midline and negative if it had crossed the midline. The number of embryos in each group is indicated in parentheses. The extent of interrenal medial extension of control 2.5-dpf embryos in panel 5D was not statistically different from those in Figure 2D and 2I. Fn and pFak expression in the interrenal region was examined in (E–E’’’’) uninjected (control), and (F–F’’’’) klf2aMO- and (G–G’’’’) mmp2MO-injected embryos by IHC. Images show transverse sections of a representative embryo from each treatment group. (H) Quantification of the effects of klf2aMO and mmp2MO on IRV growth, Fn level in the vicinity of the DA and IRV (ROI marked by broken lines in E’’–G’’), and pFak level in the steroidogenic tissue (ROI marked by orange lines in E’’’’–G’’’’). The number of embryos in each group is indicated in parentheses. Fluorescence intensities of Fn and pFak were normalized to their respective ROI sizes. The difference between the treatment and the control groups was analyzed by Student’s t-test. P, 0.05, **P,0.001, N.S., not significant. A, anterior; P, posterior; L, left; R, right; D, dorsal; V, ventral. Abbreviations: notochord (NC). Scale bar, 25 mm. doi:10.1371/journal.pone.0107997.g005 September 2014 \| Volume 9 \| Issue 9 \| e107997 PLOS ONE \| www.plosone.org 9 Hemodynamic Regulation of Interrenal Gland Assembly September 2014 \| Volume 9 \| Issue 9 \| e107997 Migration of Differentiated Chromaffin Cells Requires Blood Flow but is Independent of Mechanotransduction and Fn-Mediated Signaling To determine whether blood flow also regulates the develop- ment of the chromaffin cell lineage, ISH was performed to detect transcript expression of ff1b and dbh, markers for steroidogenic and chromaffin cell lineages, respectively (Figure 7). Consistent with the findings of our earlier study [19], the integration of the two cell populations was detected as early as 36 hpf (Figure 7A–A’), and was not affected in tnnt2a morphants (Figure 7B–B’), while interrenal medial migration and organ assembly were observed at 56 hpf and 3 dpf (Figure 7C–C’, E–E’), respectively. Since the medial extension of steroidogenic tissue was inhibited in tnnt2a morphants, chromaffin cells that reached the interrenal region remained closely associated with steroidogenic cells, and were located to the right of the midline (Figure 7D–D’, F–F’). While the convergence of differentiated chromaffin cells colonizing Since an EMT-like change occurred in steroidogenic cells during interrenal medial extension, the roles of hemodynamic forces and Fn-pFak signaling in this process was assessed. The accumulation of b-Catenin was observed in the interrenal tissue of blood flow-deficient embryos generated by tnnt2aMO microin- jection (Figure 6H–H’) or 6 mM 2,3-BDM treatment (Figure 6I–I’), providing evidence that blood flow induces an EMT-like change in steroidogenic cells. Since blood flow regulates interrenal medial migration via Fn-pFak signaling (Figure 4), Fn signaling was inhibited without perturbing blood flow to determine whether the EMT-like change in steroidogenic cells could be repressed. PLOS ONE \| www.plosone.org September 2014 \| Volume 9 \| Issue 9 \| e107997 10 Hemodynamic Regulation of Interrenal Gland Assembly Figure 6. Steroidogenic cells are induced to undergo an EMT-like change by hemodynamic forces and pFak signaling. (A–E) Ventral view of the midtrunk from 34 to 84 hpf; steroidogenic cells become loosely associated and develop protrusions at the cell surface. The accumulation of b-Catenin at cell-cell junctions in the steroidogenic tissue can be seen in cross sections of Tg(ff1bEx2: GFP) embryos at (F, F’) 2 dpf, but not at (G, G’) 2.5 dpf. The decrease in junctional b-Catenin was not observed in (H, H’) tnnt2a morphants, (I, I’) 2,3-BDM- or (J, J’) RGD-treated embryos, or (K, K’) klf2a or (L, L’) mmp2 morphants. Sections are shown of a representative embryo from each treatment group. (M) Fluorescence intensity of b-Catenin in ff1bGFP-expressing steroidogenic tissue (ROI marked by orange lines) is normalized to the size of the cluster, with the number of embryos indicated in parentheses. Migration of Differentiated Chromaffin Cells Requires Blood Flow but is Independent of Mechanotransduction and Fn-Mediated Signaling The difference between 2-dpf control group and any of the other groups was analyzed by Student’s t-test. P,0.05, N.S., not significant. A, anterior; P, posterior; L, left; R, right; D, dorsal; V, ventral. Broken yellow lines indicate position of the midline. Abbreviations: notochord (NC). Scale bar, 25 mm. doi:10.1371/journal.pone.0107997.g006 Figure 6. Steroidogenic cells are induced to undergo an EMT-like change by hemodynamic forces and pFak signaling. (A–E) Ventral view of the midtrunk from 34 to 84 hpf; steroidogenic cells become loosely associated and develop protrusions at the cell surface. The accumulation of b-Catenin at cell-cell junctions in the steroidogenic tissue can be seen in cross sections of Tg(ff1bEx2: GFP) embryos at (F, F’) 2 dpf, but not at (G, G’) 2.5 dpf. The decrease in junctional b-Catenin was not observed in (H, H’) tnnt2a morphants, (I, I’) 2,3-BDM- or (J, J’) RGD-treated embryos, or (K, K’) klf2a or (L, L’) mmp2 morphants. Sections are shown of a representative embryo from each treatment group. (M) Fluorescence intensity of b-Catenin in ff1bGFP-expressing steroidogenic tissue (ROI marked by orange lines) is normalized to the size of the cluster, with the number of embryos indicated in parentheses. The difference between 2-dpf control group and any of the other groups was analyzed by Student’s t-test. P,0.05, N.S., not significant. A, anterior; P, posterior; L, left; R, right; D, dorsal; V, ventral. Broken yellow lines indicate position of the midline. Abbreviations: notochord (NC). Scale bar, 25 mm. doi:10.1371/journal.pone.0107997.g006 the migration of steroidogenic and chromaffin cells are differen- tially modulated by blood flow. the interrenal organ was completed by 3 dpf in wild-type embryos (Figure 7E’), clusters of differentiated chromaffin cells were located outside the interrenal region in tnnt2a morphants (Figure 7F–F’), apparently due to the unsuccessful migration of chromaffin cells, indicating that although they are still capable of interacting with steroidogenic cells, their migration is defective in blood flow- deficient embryos, leading to an incomplete assembly of the interrenal organ. September 2014 \| Volume 9 \| Issue 9 \| e107997 Discussion The results of this study indicate that in addition to supplying steroids and maintaining tissue homeostasis, blood flow ensures maximal interaction between steroidogenic and chromaffin cells in teleosts (Figure 8). During interrenal organ assembly, hemody- namic forces pattern the vascular microenvironment and regulate the morphology of steroidogenic cells and the associated angio- genic endothelium. The data presented here illustrate a mecha- nism by which an EMT-like process and tissue-tissue interactions can be modulated by blood flow. A disruption of the vascular microenvironment or mechanotransduction perturbs steroidogen- ic tissue morphogenesis but not chromaffin cell migration, Interestingly, while blood flow was required, mechanotransduc- tion and Fn-pFak signaling were dispensable for chromaffin cell migration. Embryos injected with klf2aMO or mmp2MO, or treated with 100 mM RGD starting from 26 hpf, had defective medial migration of ff1b-expressing cells but normal convergence of chromaffin cells at the midline (Figure 7G–I, G’–I’), resulting in only partial integration of the two cell lineages. This implies that PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org September 2014 \| Volume 9 \| Issue 9 \| e107997 11 Hemodynamic Regulation of Interrenal Gland Assembly Figure 7. Interaction between interrenal steroidogenic and chromaffin cells in tnnt2a, klf2a, and mmp2 morphants and RGD-treated embryos during interrenal gland assembly. Double ISH assays showing colocalization of ff1b (red) and dbh (black) transcripts in uninjected control embryos and tnnt2a morphants at (A, A’; B, B’) 36 hpf (n = 3 and 6, respectively), (C, C’; D, D’) 56 hpf (n = 3 and 5, respectively), and (E, E’; F, F’) 3 dpf (n = 17 and 15, respectively), and in (G, G’) klf2a (n = 9) and (H, H’) mmp2 (n = 8) morphants and (I, I’) RGD-treated embryos (n = 18) at 3 dpf. Ventral flat mount views are shown for representative embryos in each group, oriented with anterior at the top. Yellow arrowheads indicate chromaffin cell clusters that failed to converge at the interrenal area in tnnt2a morphants. Broken white lines indicate position of the midline. (J) Schematic representation of various phenotypic defects associated with interrenal organ assembly. Panels show ventral views of wild-type, mutant, morphant, and drug-treated embryos at 3 dpf, oriented with anterior at the top. Phenotypes depicted for cloche (clo) and fn1 mutants are based on previous reports [9,21]. Abbreviations: notochord (NC), the third somite (3S), the fourth somite (4S). doi:10.1371/journal.pone.0107997.g007 Figure 7. Discussion Interaction between interrenal steroidogenic and chromaffin cells in tnnt2a, klf2a, and mmp2 morphants and RGD-treated embryos during interrenal gland assembly. Double ISH assays showing colocalization of ff1b (red) and dbh (black) transcripts in uninjected control embryos and tnnt2a morphants at (A, A’; B, B’) 36 hpf (n = 3 and 6, respectively), (C, C’; D, D’) 56 hpf (n = 3 and 5, respectively), and (E, E’; F, F’) 3 dpf (n = 17 and 15, respectively), and in (G, G’) klf2a (n = 9) and (H, H’) mmp2 (n = 8) morphants and (I, I’) RGD-treated embryos (n = 18) at 3 dpf. Ventral flat mount views are shown for representative embryos in each group, oriented with anterior at the top. Yellow arrowheads indicate chromaffin cell clusters that failed to converge at the interrenal area in tnnt2a morphants. Broken white lines indicate position of the midline. (J) Schematic representation of various phenotypic defects associated with interrenal organ assembly. Panels show ventral views of wild-type, mutant, morphant, and drug-treated embryos at 3 dpf, oriented with anterior at the top. Phenotypes depicted for cloche (clo) and fn1 mutants are based on previous reports [9,21]. Abbreviations: notochord (NC), the third somite (3S), the fourth somite (4S). doi:10.1371/journal.pone.0107997.g007 The phenotype of the RGD-treated embryo—that is, defective migration of steroidogenic but not chromaffin cells—was different from that of the fn1 mutant, in which the migration of both cell types was compromised (Figs. 7 and 8). Fn regulates the fusion of bilateral cardiac primordia, and is therefore essential for the development of myocardial epithelia [41]; thus, fn1 mutants have impaired cardiac function and consequently, reduced blood flow, which may cause the aberrant migration of chromaffin cells. Nevertheless, these cell clusters in fn1 mutants were more dispersed than in tnnt2a morphants (Figure 7J), suggesting that the presence of Fn ensures that the microenvironment stimulates chromaffin cell migration prior to the onset of circulation. Trunk neural crest cells, from which the sympathochromaffin lineage is derived, migrate along the medial surface of each somite but not the somite boundary where Fn accumulates [60,61]. However, disrupted somite formation in the fn1 mutant leads to the uncoupling of slow- and fast-twitch muscle fibers, and hence a indicating that blood flow regulates these processes through different pathways. September 2014 \| Volume 9 \| Issue 9 \| e107997 Discussion It therefore remains to be explored whether Wnt signals are directly or indirectly regulated by hemodynamic forces and thereby guide cell migration during interrenal organ assembly. EMT is initiated by transforming, fibroblast, epithelial, and hepatocyte growth factors as well as the oncogene Harvey rat sarcoma. During this process, there is a downregulation of epithelial and concomitant upregulation of mesenchymal markers, while cells assume a proliferative and migratory character [35]. EMT is similar to the endothelial-to-mesenchymal transition (EnMT), a critical step in vertebrate heart development in which endothelial endocardial cells give rise to heart cushion cells that form the mesenchymal portion of septa and valves [63]. In various vertebrate models including zebrafish, alterations in hemodynamic forces during cardiogenesis have profound effects on cardiac septation and valvulogenesis [3,4,46,64], suggesting that EnMT is regulated by hemodynamics. The present study underscores a novel role for hemodynamics in the EMT-like behavior of vessel- associated endocrine cells via modulation of the vascular microenvironment and Fn-pFak signaling. g g g y It will be of interest to explore whether other morphogens participate in the interrenal organogenesis. Pregnenolone, a steroid produced from cholesterol by the steroidogenic enzyme Cyp11a1, promotes cell migration by activating CLIP-170 and stabilizing microtubules [69,70]. The function of pregnenolone for cell movement has been well studied in the zebrafish gastrulation which is well before the onset of blood flow as well as interrenal organogenesis. While cyp11a1 is expressed at the zebrafish interrenal tissue by 1 dpf [18], it remains unclear whether pregnelolone would also form a morphogen gradient during the interrenal morphogenetic movement. On the other hand, the molecular and cellular mechanisms by which blood flow regulates chromaffin cell migration remain to be explored. In the blood flow-deficient zebrafish embryo, the expression of endothelial CXCR4a is upregulated during collateral formation [71]. CXCR4a is the G protein-coupled receptor for the chemokine stromal cell-derived factor 1 (SDF-1). SDF-1, along with Bone Morphogenetic Proteins (BMPs) and Neuregulin 1 of the epidermal growth factor family, are the three groups of paracrine factors that participate in the generation of sympatho-adrenal progenitors from the neural crest [12,72,73,74]. BMPs are produced by the DA and are critical for the production of SDF- 1 and Neuregulin 1 in the vicinity of the DA, while SDF-1 and Neuregulin 1 function as chemoattractants for the migration of neural crest cells [15]. Discussion The obstruction of blood flow in the zebrafish embryo produced a phenotype similar to that of the fn1 mutant [19] (Figure 7), providing evidence that the interrenal Fn-enriched microenviron- ment is regulated by hemodynamic forces (Figure 4). However, while blood flow-deficient embryos have similar defects in steroidogenic and chromaffin cell migration, 20%–30% of fn1 mutants exhibit a more severe phenotype, where the bilateral fusion of early interrenal tissues is unsuccessful (Figure 7). The variable expressivity of this phenotype in fn1 mutants could be due to a more profound effect of Fn deficiency on early development prior to the onset of blood flow [41,60]. Indeed, early bilateral interrenal tissues arise in close association with pre-vascular angioblasts, and their fusion occurs in parallel with the assembly of axial vasculature independently of the initiation of blood flow [9,52]. PLOS ONE \| www.plosone.org September 2014 \| Volume 9 \| Issue 9 \| e107997 12 Hemodynamic Regulation of Interrenal Gland Assembly Figure 8. Schematic representation of interrenal steroidogenic tissue (IR) and chromaffin cell integration that takes place in the vicinity of the DA and extending IRV. An Fn-enriched microenvironment promotes medial extension of the steroidogenic tissue and culminates in steroidogenic-chromaffin interactions on both sides of the midline (left); various defects in extension lead to incomplete assembly of the interrenal organ (middle and right). D, dorsal; V, ventral; L, left; R, right. Abbreviations: notochord (NC). Somite (S). doi:10.1371/journal.pone.0107997.g008 Figure 8. Schematic representation of interrenal steroidogenic tissue (IR) and chromaffin cell integration that takes place in the vicinity of the DA and extending IRV. An Fn-enriched microenvironment promotes medial extension of the steroidogenic tissue and culminates in steroidogenic-chromaffin interactions on both sides of the midline (left); various defects in extension lead to incomplete assembly of the interrenal organ (middle and right). D, dorsal; V, ventral; L, left; R, right. Abbreviations: notochord (NC). Somite (S). doi:10.1371/journal.pone.0107997.g008 disorganized myofibril pattern [62], which is another factor that could compromise chromaffin cell migration. In contrast, the RGD-treatment in this study was initiated at 26 hpf without evident perturbation of cardiac flow and somite morphology, which might explain why chomaffin cells migrate normally in RGD-treated embryos. [67], and is also expressed in the outermost region of the adrenal cortex, where it may play a role in migration events that segregate adrenal and gonadal lineages during early development [68]. Supporting Information Figure S1 Effect of norepinephrine on the heart beat of zebrafish. Norepinephrine treatments resulted in a dose- dependent increase of heart rate at 33 hpf. The difference between groups treated with various concentrations of norepi- nephrine was analyzed by Student’s t-test. P,0.05, **P, 0.0005. Figure S1 Effect of norepinephrine on the heart beat of zebrafish. Norepinephrine treatments resulted in a dose- dependent increase of heart rate at 33 hpf. The difference between groups treated with various concentrations of norepi- nephrine was analyzed by Student’s t-test. P,0.05, **P, 0.0005. Figure S2 Interrenal cell migration suppressed by 2,3- BDM was recovered following the removal of 2,3-BDM. The interrenl tissue stained by 3b-Hsd activity assay in the control Tg(kdrl: GFP)s843 embryo continued to extend across the midline from 2.5 dpf (A-A’’) to 3 dpf (C-C’’), while migration of interrenal cells was repressed by 2,3-BDM treatment at 6 mM from 1.5 to 2.5 dpf (B-B’’) or 3 dpf (D-D’’). Migration of interrenal cells was recovered at 3 dpf as 2,3-BDM was washed out at 2.5 dpf (E-E’’). Protrusions of extending interrenal tissues (red arrows) were detected in control as well as recovered embryos. Broken white lines indicate position of the midline. Abbreviations: glomerulus (G), posterior cardinal vein (PCV). Scale bar, 50 mm. (TIF) Figure S2 Interrenal cell migration suppressed by 2,3- BDM was recovered following the removal of 2,3-BDM. The interrenl tissue stained by 3b-Hsd activity assay in the control Tg(kdrl: GFP)s843 embryo continued to extend across the midline from 2.5 dpf (A-A’’) to 3 dpf (C-C’’), while migration of interrenal cells was repressed by 2,3-BDM treatment at 6 mM from 1.5 to 2.5 dpf (B-B’’) or 3 dpf (D-D’’). Migration of interrenal cells was recovered at 3 dpf as 2,3-BDM was washed out at 2.5 dpf (E-E’’). Protrusions of extending interrenal tissues (red arrows) were detected in control as well as recovered embryos. Broken white lines indicate position of the midline. Abbreviations: glomerulus (G), posterior cardinal vein (PCV). Scale bar, 50 mm. (TIF) Figure S7 Steroidogenic cells display a rise of Vimentin expression which is induced by hemodynamic forces and pFak signaling. Vimentin in the steroidogenic tissue (marked by green fluorescence) of Tg(ff1bEx2: GFP) embryos was increased from 2 dpf (A-A’) to 2.5 dpf (B-B’). The increase in Vimentin was not observed in (C, C’) tnnt2a morphants, (D, D’) RGD-treated embryos, or (E, E’) klf2a or (F, F’) mmp2 morphants. 3. Vermot J, Forouhar AS, Liebling M, Wu D, Plummer D, et al. (2009) Reversing blood flows act through klf2a to ensure normal valvulogenesis in the developing heart. PLoS Biol 7: e1000246. Author Contributions Conceived and designed the experiments: YWL. Performed the experi- ments: CWC YLZ ZYJ YWL. Analyzed the data: CWC YLZ ZYJ YWL. Wrote the paper: YWL. Acknowledgments The authors would like to thank Dr. Christoph Englert, Dr. Woon-Khiong Chan, and Prof. Didier Stainier for generously providing the zebrafish strains; Ru Shiangli, Jamie Lin, Yang Huang, Yusyuan Tian, Kuan-Chieh Wang and Hsin-Yu Hou for excellent technical assistance; Dr. Yi-Ching Lin for helpful advice on statistical analysis; and the Taiwan Zebrafish Core Facility, Zebrafish Core in Academia Sinica, and Zebrafish Core Facility at NTHU-NHRI (ZeTH) for assistance with fish culturing. Figure S5 Effect of RGD treatment on pFAK distribu- tion in the interrenal region. Transverse sections of Tg(kdrl: GFP)s843 embryos either untreated (control; n = 6) or treated with 100 mM RGD peptide (n = 10) from 26 hpf and harvested at 2.5 dpf for evaluation of 3b-Hsd activity and pFak level by IHC. Sections are shown of a representative embryo from each group, oriented with the dorsal side at the top. Abbreviations: interrenal tissue (IR), notochord (NC), somite (S). The authors would like to thank Dr. Christoph Englert, Dr. Woon-Khiong Chan, and Prof. Didier Stainier for generously providing the zebrafish strains; Ru Shiangli, Jamie Lin, Yang Huang, Yusyuan Tian, Kuan-Chieh Wang and Hsin-Yu Hou for excellent technical assistance; Dr. Yi-Ching Lin for helpful advice on statistical analysis; and the Taiwan Zebrafish Core Facility, Zebrafish Core in Academia Sinica, and Zebrafish Core Facility at NTHU-NHRI (ZeTH) for assistance with fish culturing. Discussion The difference between 2-dpf control group and any of the other groups was analyzed by Student’s t-test. P,0.05, P,0.005, N.S., not significant. D, dorsal; V, ventral; L, left; R, right. Abbreviations: notochord (NC). Scale bar, 25 mm. (TIF) Supporting Information Sections are shown of a representative embryo from each treatment group. (G) Fluorescence intensity of Vimentin in ff1bGFP-expressing steroidogenic tissue (ROI marked by orange lines) is normalized to the size of the cluster, with the number of embryos indicated in parentheses. The difference between 2-dpf control group and any of the other groups was analyzed by Student’s t-test. P,0.005, N.S., not significant. D, dorsal; V, ventral; L, left; R, right. Abbreviations: notochord (NC). Scale bar, 25 mm. (TIF) Figure S3 Effects of blood flow inhibition on IRV growth was reversible following the removal of 2,3-BDM. The IRV in the control Tg(kdrl: GFP)s843 embryo continued to extend from 2.5 dpf (A, A’) to 3 dpf (C, C’), while the IRV growth was repressed by 2,3-BDM treatment at 6 mM from 1.5 to 2.5 dpf (B, B’) or 3 dpf (D, D’). Extension of IRV was recovered at 3 dpf as 2,3-BDM was washed out at 2.5 dpf (E, E’). The interrenal tissue (IR) was detected by 3b-Hsd acitivity assay. D, dorsal; V, ventral; L, left; R, right. Broken yellow lines indicate position of the midline. Abbreviations: notochord (NC), somite (S). Scale bar, 50 mm. (TIF) Video S1 Complete blockage of blood flow in the posterior cardinal vein of a 36-hpf embryo treated with 6 mM 2,3-BDM. (WMV) Figure S4 Effect of L-NAME on morphogenetic move- ments of interrenal steroidogenic tissue and IRV formation. (A–A’’) Tg(kdrl: GFP)s843 embryos treated with 100 mM L-NAME from 1.5 dpf onwards had interrenal steroido- genic tissue (IR) morphology (orange arrows) and (B–B’’) IRV length similar to control embryos (Figure 3A, A’, D) at 2.5 dpf (n = 8). The activity of endothelial NO synthase was inhibited by L- NAME at concentrations higher than 10 mM [50]. D, dorsal; V, ventral; L, left; R, right. Abbreviations: posterior cardinal vein (PCV). Scale bar, 50 mm. (TIF) Video S2 Normal blood flow in the posterior cardinal vein of a wild-type 36-hpf embryo. (WMV) Video S3 Partial reduction in blood flow in the poste- rior cardinal vein of a 36-hpf embryo treated with 2 mM 2,3-BDM. 2. Pelster B, Burggren WW (1996) Disruption of hemoglobin oxygen transport does not impact oxygen-dependent physiological processes in developing embryos of zebra fish (Danio rerio). Circ Res 79: 358–362. Discussion It will therefore be intriguing to examine whether a loss of blood flow would influence the activity of SDF-1 through upregulating CXCR4a expression. In addition to being an epithelial cell marker, b-Catenin also transduces canonical Wnt signals and is implicated in cell proliferation [65]. Cytoplasmic b-Catenin is stabilized upon activation of Wnt signaling, leading to its translocation to the nucleus, where it interacts with T-cell factors (TCFs) to stimulate the transcription of target genes. There was no obvious enrichment of nuclear b-Catenin detected in steroidogenic cells during interrenal organ assembly (Figure 6), and it is possible that canonical Wnt signaling is not involved in zebrafish interrenal development. Nevertheless, Wnt4 can inhibit b-Catenin/TCF signaling by redirecting nuclear b-Catenin to the membrane [66]. During mammalian sex differentiation, Wnt4 inhibits the migra- tion of endothelial and steroidogenic cells into the female gonad September 2014 \| Volume 9 \| Issue 9 \| e107997 September 2014 \| Volume 9 \| Issue 9 \| e107997 PLOS ONE \| www.plosone.org 13 Hemodynamic Regulation of Interrenal Gland Assembly Hemodynamic Regulation of Interrenal Gland Assembly forces and pFak signaling. E-cadherin in the fluorescent steroidogenic tissue of Tg(ff1bEx2: GFP) embryos was decreased from 2 dpf (A-A’) to 2.5 dpf (B-B’). The decrease in E-cadherin was not observed in (C, C’) tnnt2a morphants, (D, D’) RGD-treated embryos, or (E, E’) klf2a or (F, F’) mmp2 morphants. Sections are shown of a representative embryo from each treatment group. (F) Fluorescence intensity of E-cadherin in ff1bGFP-expressing steroidogenic tissue is normalized to the size of the cluster, with the number of embryos indicated in parentheses. The difference between 2-dpf control group and any of the other groups was analyzed by Student’s t-test. P,0.05, *P,0.005, N.S., not significant. D, dorsal; V, ventral; L, left; R, right. Abbreviations: notochord (NC). Scale bar, 25 mm. (TIF) forces and pFak signaling. E-cadherin in the fluorescent steroidogenic tissue of Tg(ff1bEx2: GFP) embryos was decreased from 2 dpf (A-A’) to 2.5 dpf (B-B’). The decrease in E-cadherin was not observed in (C, C’) tnnt2a morphants, (D, D’) RGD-treated embryos, or (E, E’) klf2a or (F, F’) mmp2 morphants. Sections are shown of a representative embryo from each treatment group. (F) Fluorescence intensity of E-cadherin in ff1bGFP-expressing steroidogenic tissue is normalized to the size of the cluster, with the number of embryos indicated in parentheses. 1. Culver JC, Dickinson ME (2010) The effects of hemodynamic force on embryonic development. Microcirculation 17: 164–178. 1. Culver JC, Dickinson ME (2010) The effects of hemodynamic force on embryonic development. Microcirculation 17: 164–178. Hemodynamic Regulation of Interrenal Gland Assembly (2006) Klf2 is an essential regulator of vascular hemodynamic forces in vivo. Dev Cell 11: 845–857. g 55. Savagner P (2001) Leaving the neighborhood: molecular mechanisms involved during epithelial-mesenchymal transition. Bioessays 23: 912–923. 25. Page-McCaw A, Ewald AJ, Werb Z (2007) Matrix metalloproteinases and the regulation of tissue remodelling. Nat Rev Mol Cell Biol 8: 221–233. 26. Yasuda T, Kondo S, Homma T, Harris RC (1996) Regulation of extracellular matrix by mechanical stress in rat glomerular mesangial cells. J Clin Invest 98: 1991–2000. 56. Cano A, Perez-Moreno MA, Rodrigo I, Locascio A, Blanco MJ, et al. (2000) The transcription factor snail controls epithelial-mesenchymal transitions by repressing E-cadherin expression. Nat Cell Biol 2: 76–83. 27. Singhal PC, Sagar S, Garg P (1996) Simulated glomerular pressure modulates mesangial cell 72 kDa metalloproteinase activity. Connect Tissue Res 33: 257– 263. 57. Moreno-Bueno G, Cubillo E, Sarrio D, Peinado H, Rodriguez-Pinilla SM, et al. (2006) Genetic profiling of epithelial cells expressing E-cadherin repressors reveals a distinct role for Snail, Slug, and E47 factors in epithelial-mesenchymal transition. Cancer Res 66: 9543–9556. 28. Cummins PM, von Offenberg Sweeney N, Killeen MT, Birney YA, Redmond EM, et al. (2007) Cyclic strain-mediated matrix metalloproteinase regulation within the vascular endothelium: a force to be reckoned with. Am J Physiol Heart Circ Physiol. pp. H28–42. 58. Mendez MG, Kojima S, Goldman RD (2010) Vimentin induces changes in cell shape, motility, and adhesion during the epithelial to mesenchymal transition. Faseb J 24: 1838–1851. 59. Vuoriluoto K, Haugen H, Kiviluoto S, Mpindi JP, Nevo J, et al. (2011) Vimentin regulates EMT induction by Slug and oncogenic H-Ras and migration by governing Axl expression in breast cancer. Oncogene 30: 1436–1448. 29. von Offenberg Sweeney N, Cummins PM, Birney YA, Cullen JP, Redmond EM, et al. (2004) Cyclic strain-mediated regulation of endothelial matrix metalloproteinase-2 expression and activity. Cardiovasc Res 63: 625–634. 30. Xu J, Rodriguez D, Petitclerc E, Kim JJ, Hangai M, et al. (2001) Proteolytic exposure of a cryptic site within collagen type IV is required for angiogenesis and tumor growth in vivo. J Cell Biol 154: 1069–1079. 60. Koshida S, Kishimoto Y, Ustumi H, Shimizu T, Furutani-Seiki M, et al. (2005) Integrinalpha5-dependent fibronectin accumulation for maintenance of somite boundaries in zebrafish embryos. Dev Cell 8: 587–598. y 61. Honjo Y, Eisen JS (2005) Slow muscle regulates the pattern of trunk neural crest migration in zebrafish. Development 132: 4461–4470. 31. Hemodynamic Regulation of Interrenal Gland Assembly 4. Hove JR, Koster RW, Forouhar AS, Acevedo-Bolton G, Fraser SE, et al. (2003) Intracardiac fluid forces are an essential epigenetic factor for embryonic cardiogenesis. Nature 421: 172–177. 35. Acloque H, Adams MS, Fishwick K, Bronner-Fraser M, Nieto MA (2009) Epithelial-mesenchymal transitions: the importance of changing cell state in development and disease. J Clin Invest 119: 1438–1449. p J 36. Westerfield M (2000) The Zebrafish Book: Guide for the Laboratory Use of Zebrafish (Danio rerio). Eugene, OR.: Univ. of Oregon Press. g 5. Serluca FC, Drummond IA, Fishman MC (2002) Endothelial signaling in kidney morphogenesis: a role for hemodynamic forces. Curr Biol 12: 492–497. 37. Kimmel CB, Ballard WW, Kimmel SR, Ullmann B, Schilling TF (1995) Stages of embryonic development of the zebrafish. Dev Dyn 203: 253–310. 6. Banjo T, Grajcarek J, Yoshino D, Osada H, Miyasaka KY, et al. (2013) Haemodynamically dependent valvulogenesis of zebrafish heart is mediated by flow-dependent expression of miR-21. Nat Commun 4: 1978. of embryonic development of the zebrafish. Dev Dyn 203: 253–310 flow-dependent expression of miR-21. Nat Commun 4: 1978. 38. Perner B, Englert C, Bollig F (2007) The Wilms tumor genes wt1a and wt1b control different steps during formation of the zebrafish pronephros. Dev Biol 309: 87–96. p p 7. Field HA, Dong PD, Beis D, Stainier DY (2003) Formation of the digestive system in zebrafish. II. Pancreas morphogenesis. Dev Biol 261: 197–208. 7. Field HA, Dong PD, Beis D, Stainier DY (2003) Formation 39. Quek SI (2009) Molecular characterization of the zebrafish ff1b gene. Singapore: National University of Singapore. y p g 8. Field HA, Ober EA, Roeser T, Stainier DY (2003) Formation of the digestive system in zebrafish. I. Liver morphogenesis. Dev Biol 253: 279–290. . Field HA, Ober EA, Roeser T, Stainier DY (2003) Formation of t system in zebrafish. I. Liver morphogenesis. Dev Biol 253: 279–290 40. Jin SW, Beis D, Mitchell T, Chen JN, Stainier DY (2005) Cellular and molecular analyses of vascular tube and lumen formation in zebrafish. Development 132: 5199–5209. 9. Liu YW, Guo L (2006) Endothelium is required for the promotion of interrenal morphogenetic movement during early zebrafish development. Dev Biol 297: 44–58. 41. Trinh LA, Stainier DY (2004) Fibronectin regulates epithelial organization during myocardial migration in zebrafish. Dev Cell 6: 371–382. 10. Opitz R, Maquet E, Huisken J, Antonica F, Trubiroha A, et al. Hemodynamic Regulation of Interrenal Gland Assembly (2012) Transgenic zebrafish illuminate the dynamics of thyroid morphogenesis and its relationship to cardiovascular development. Dev Biol 372: 203–216. 42. Sehnert AJ, Huq A, Weinstein BM, Walker C, Fishman M, et al. (2002) Cardiac troponin T is essential in sarcomere assembly and cardiac contractility. Nat Genet 31: 106–110. 11. Alt B, Elsalini OA, Schrumpf P, Haufs N, Lawson ND, et al. (2006) Arteries define the position of the thyroid gland during its developmental relocalisation. Development 133: 3797–3804. 43. Nicoli S, Standley C, Walker P, Hurlstone A, Fogarty KE, et al. (2010) MicroRNA-mediated integration of haemodynamics and Vegf signalling during angiogenesis. Nature 464: 1196–1200. 12. Huber K (2006) The sympathoadrenal cell lineage: specification, diversification, and new perspectives. Dev Biol 298: 335–343. g g 44. Detry B, Erpicum C, Paupert J, Blacher S, Maillard C, et al. (2012) Matrix metalloproteinase-2 governs lymphatic vessel formation as an interstitial collagenase. Blood 119: 5048–5056. 13. Gut P, Huber K, Lohr J, Bruhl B, Oberle S, et al. (2005) Lack of an adrenal cortex in Sf1 mutant mice is compatible with the generation and differentiation of chromaffin cells. Development 132: 4611–4619. g 45. Vasilyev A, Liu Y, Mudumana S, Mangos S, Lam PY, et al. (2009) Collective cell migration drives morphogenesis of the kidney nephron. PLoS Biol 7: e9. 14. Zubair M, Oka S, Parker KL, Morohashi K (2009) Transgenic expression of Ad4BP/SF-1 in fetal adrenal progenitor cells leads to ectopic adrenal formation. Mol Endocrinol 23: 1657–1667. 46. Bartman T, Walsh EC, Wen KK, McKane M, Ren J, et al. (2004) Early myocardial function affects endocardial cushion development in zebrafish. PLoS Biol 2: E129. 15. Saito D, Takase Y, Murai H, Takahashi Y (2012) The dorsal aorta initiates a molecular cascade that instructs sympatho-adrenal specification. Science 336: 1578–1581. 47. Drummond IA, Majumdar A, Hentschel H, Elger M, Solnica-Krezel L, et al. (1998) Early development of the zebrafish pronephros and analysis of mutations affecting pronephric function. Development 125: 4655–4667. 16. Hsu HJ, Lin G, Chung BC (2003) Parallel early development of zebrafish interrenal glands and pronephros: differential control by wt1 and ff1b. Development 130: 2107–2116. 48. Bollig F, Perner B, Besenbeck B, Kothe S, Ebert C, et al. (2009) A highly conserved retinoic acid responsive element controls wt1a expression in the zebrafish pronephros. Development 136: 2883–2892. 17. Hemodynamic Regulation of Interrenal Gland Assembly Chai C, Liu YW, Chan WK (2003) Ff1b is required for the development of steroidogenic component of the zebrafish interrenal organ. Dev Biol 260: 226– 244. 49. Wang L, Zhang P, Wei Y, Gao Y, Patient R, et al. (2011) A blood flow- dependent klf2a-NO signaling cascade is required for stabilization of hematopoietic stem cell programming in zebrafish embryos. Blood 118: 4102– 4110. 18. To TT, Hahner S, Nica G, Rohr KB, Hammerschmidt M, et al. (2007) Pituitary-interrenal interaction in zebrafish interrenal organ development. Mol Endocrinol 21: 472–485. 50. North TE, Goessling W, Peeters M, Li P, Ceol C, et al. (2009) Hematopoietic stem cell development is dependent on blood flow. Cell 137: 736–748. 19. Chou CW, Chiu CH, Liu YW (2013) Fibronectin mediates correct positioning of the interrenal organ in zebrafish. Dev Dyn 242: 432–443. 20. Alsop D, Vijayan MM (2009) Molecular programming of the corticosteroid stress axis during zebrafish development. Comp Biochem Physiol A Mol Integr Physiol 153: 49–54. 51. Watson O, Novodvorsky P, Gray C, Rothman AM, Lawrie A, et al. (2013) Blood flow suppresses vascular Notch signalling via dll4 and is required for angiogenesis in response to hypoxic signalling. Cardiovasc Res 100: 252–261. y 21. Chiu CH, Chou CW, Takada S, Liu YW (2012) Development and fibronectin signaling requirements of the zebrafish interrenal vessel. PLoS One 7: e43040. 52. Chou CW, Hsu HC, Quek SI, Chan WK, Liu YW (2010) Arterial and venous vessels are required for modulating developmental relocalization and laterality of the interrenal tissue in zebrafish. Dev Dyn 239: 1995–2004. 22. Parmar KM, Larman HB, Dai G, Zhang Y, Wang ET, et al. (2006) Integration of flow-dependent endothelial phenotypes by Kruppel-like factor 2. J Clin Invest 116: 49–58. 53. Park J, Schwarzbauer JE (2013) Mammary epithelial cell interactions with fibronectin stimulate epithelial-mesenchymal transition. Oncogene: doi: 10.1038/onc.2013.1118. [Epub ahead of print]. 23. Dekker RJ, van Thienen JV, Rohlena J, de Jager SC, Elderkamp YW, et al. (2005) Endothelial KLF2 links local arterial shear stress levels to the expression of vascular tone-regulating genes. Am J Pathol 167: 609–618. 54. Sun XJ, Fa PP, Cui ZW, Xia Y, Sun L, et al. (2013) The EDA-containing cellular fibronectin induces epithelial mesenchymal transition in lung cancer cells through integrin alpha9beta1-mediated activation of PI3-K/Akt and Erk1/2. Carcinogenesis 35: 184–191. 24. Lee JS, Yu Q, Shin JT, Sebzda E, Bertozzi C, et al. References 3. Vermot J, Forouhar AS, Liebling M, Wu D, Plummer D, et al. (2009) Reversing blood flows act through klf2a to ensure normal valvulogenesis in the developing heart. PLoS Biol 7: e1000246. September 2014 \| Volume 9 \| Issue 9 \| e107997 PLOS ONE \| www.plosone.org 14 72. Kasemeier-Kulesa JC, McLennan R, Romine MH, Kulesa PM, Lefcort F (2010) CXCR4 controls ventral migration of sympathetic precursor cells. J Neurosci 30: 13078–13088. Hemodynamic Regulation of Interrenal Gland Assembly Kenny HA, Kaur S, Coussens LM, Lengyel E (2008) The initial steps of ovarian cancer cell metastasis are mediated by MMP-2 cleavage of vitronectin and fibronectin. J Clin Invest 118: 1367–1379. 62. Snow CJ, Peterson MT, Khalil A, Henry CA (2008) Muscle development is disrupted in zebrafish embryos deficient for fibronectin. Dev Dyn 237: 2542– 2553. J 32. Corti P, Young S, Chen CY, Patrick MJ, Rochon ER, et al. (2011) Interaction between alk1 and blood flow in the development of arteriovenous malforma- tions. Development 138: 1573–1582. 63. Kovacic JC, Mercader N, Torres M, Boehm M, Fuster V (2012) Epithelial-to- mesenchymal and endothelial-to-mesenchymal transition: from cardiovascular development to disease. Circulation 125: 1795–1808. p 33. Santoro MM, Pesce G, Stainier DY (2009) Characterization of vascular mural cells during zebrafish development. Mech Dev 126: 638–649. 64. Granados-Riveron JT, Brook JD (2012) The impact of mechanical forces in heart morphogenesis. Circ Cardiovasc Genet 5: 132–142. 34. Miano JM, Georger MA, Rich A, De Mesy Bentley KL (2006) Ultrastructure of zebrafish dorsal aortic cells. Zebrafish 3: 455–463. September 2014 \| Volume 9 \| Issue 9 \| e107997 September 2014 \| Volume 9 \| Issue 9 \| e107997 PLOS ONE \| www.plosone.org 15 74. Shah NM, Groves AK, Anderson DJ (1996) Alternative neural crest cell fates are instructively promoted by TGFbeta superfamily members. Cell 85: 331–343. 73. Britsch S, Li L, Kirchhoff S, Theuring F, Brinkmann V, et al. (1998) The ErbB2 and ErbB3 receptors and their ligand, neuregulin-1, are essential for development of the sympathetic nervous system. Genes Dev 12: 1825–1836. 71. Packham IM, Gray C, Heath PR, Hellewell PG, Ingham PW, et al. (2009) Microarray profiling reveals CXCR4a is downregulated by blood flow in vivo and mediates collateral formation in zebrafish embryos. Physiol Genomics 38: 319–327. 70. Hsu HJ, Liang MR, Chen CT, Chung BC (2006) Pregnenolone stabilizes microtubules and promotes zebrafish embryonic cell movement. Nature 439: 480–483. Hemodynamic Regulation of Interrenal Gland Assembly Hemodynamic Regulation of Interrenal Gland Assembly Hemodynamic Regulation of Interrenal Gland Assembly 65. Klaus A, Birchmeier W (2008) Wnt signalling and its impact on development and cancer. Nat Rev Cancer 8: 387–398. 70. Hsu HJ, Liang MR, Chen CT, Chung BC (2006) Pregnenolone stabilizes microtubules and promotes zebrafish embryonic cell movement. Nature 439: 480–483. 66. Bernard P, Fleming A, Lacombe A, Harley VR, Vilain E (2008) Wnt4 inhibits beta-catenin/TCF signalling by redirecting beta-catenin to the cell membrane. Biol Cell 100: 167–177. 71. Packham IM, Gray C, Heath PR, Hellewell PG, Ingham PW, et al. (2009) Microarray profiling reveals CXCR4a is downregulated by blood flow in vivo and mediates collateral formation in zebrafish embryos. Physiol Genomics 38: 319–327. 67. Jeays-Ward K, Hoyle C, Brennan J, Dandonneau M, Alldus G, et al. (2003) Endothelial and steroidogenic cell migration are regulated by WNT4 in the developing mammalian gonad. Development 130: 3663–3670. 72. Kasemeier-Kulesa JC, McLennan R, Romine MH, Kulesa PM, Lefcort F (2010) CXCR4 controls ventral migration of sympathetic precursor cells. J Neurosci 30: 13078–13088. developing mammalian gonad. Development 130: 3663–3670. 68. Heikkila M, Peltoketo H, Leppaluoto J, Ilves M, Vuolteenaho O, et al. (2002) Wnt-4 deficiency alters mouse adrenal cortex function, reducing aldosterone production. Endocrinology 143: 4358–4365. 73. Britsch S, Li L, Kirchhoff S, Theuring F, Brinkmann V, et al. (1998) The ErbB2 and ErbB3 receptors and their ligand, neuregulin-1, are essential for development of the sympathetic nervous system. Genes Dev 12: 1825–1836. 69. Weng JH, Liang MR, Chen CH, Tong SK, Huang TC, et al. (2013) Pregnenolone activates CLIP-170 to promote microtubule growth and cell migration. Nat Chem Biol 9: 636–642. 74. Shah NM, Groves AK, Anderson DJ (1996) Alternative neural crest cell fates are instructively promoted by TGFbeta superfamily members. Cell 85: 331–343. PLOS ONE \| www.plosone.org September 2014 \| Volume 9 \| Issue 9 \| e107997 PLOS ONE \| www.plosone.org 16
https://openalex.org/W4253922980	https://hal.archives-ouvertes.fr/hal-03711297/file/amt-14-1593-2021.pdf	English	null	Two-dimensional and multi-channel feature detection algorithm for the CALIPSO lidar measurements	null	2,020	cc-by	12,927	To cite this version: Thibault Vaillant de Guélis, Mark Vaughan, David Winker, Zhaoyan Liu. Two-dimensional and multi- channel feature detection algorithm for the CALIPSO lidar measurements. Atmospheric Measurement Techniques, 2021, 14 (2), pp.1593-1613. 10.5194/amt-14-1593-2021. hal-03711297 Two-dimensional and multi-channel feature detection algorithm for the CALIPSO lidar measurements Thibault Vaillant de Guélis1,2, Mark A. Vaughan3, David M. Winker3, and Zhaoyan Liu3 1NASA Postdoctoral Program Fellow, NASA, Langley Research Center, Hampton, VA 23681, USA 2Science Systems and Applications, Inc., Hampton, VA 23666, USA 3NASA Langley Research Center, Hampton, VA 23681, USA orrespondence: Thibault Vaillant de Guélis (thibault.vaillantdeguelis@outlook.com) Received: 14 September 2020 – Discussion started: 29 September 2020 Revised: 22 December 2020 – Accepted: 29 December 2020 – Published: 26 February 2021 Received: 14 September 2020 – Discussion started: 29 September 2020 Revised: 22 December 2020 – Accepted: 29 December 2020 – Published: 26 February 2021 Abstract. In this paper, we describe a new two-dimensional and multi-channel feature detection algorithm (2D-McDA) and demonstrate its application to lidar backscatter measure- ments from the Cloud-Aerosol Lidar and Infrared Pathﬁnder Satellite Observations (CALIPSO) mission. Unlike previous layer detection schemes, this context-sensitive feature ﬁnder algorithm is applied to a 2-D lidar “scene”, i.e., to the im- age formed by many successive lidar proﬁles. Features are identiﬁed when an extended and contiguous 2-D region of enhanced backscatter signal rises signiﬁcantly above the ex- pected “clear air” value. Using an iterated 2-D feature de- tection algorithm dramatically improves the ﬁne details of feature shapes and can accurately identify previously un- detected layers (e.g., subvisible cirrus) that are very thin vertically but horizontally persistent. Because the algorithm looks for contiguous 2-D patterns using successively lower detection thresholds, it reports strongly scattering features separately from weakly scattering features, thus potentially offering improved discrimination of juxtaposed cloud and aerosol layers. Moreover, the 2-D detection algorithm uses the backscatter signals from all available channels: 532 nm parallel, 532 nm perpendicular and 1064 nm total. Since the backscatter from some aerosol or cloud particle types can be more pronounced in one channel than another, simulta- neously assessing the signals from all channels greatly im- proves the layer detection. For example, ice particles in sub- visible cirrus strongly depolarize the lidar signal and, con- sequently, are easier to detect in the 532 nm perpendicular channel. Use of the 1064 nm channel greatly improves the detection of dense smoke layers, because smoke extinction at 532 nm is much larger than at 1064 nm, and hence the range- dependent reduction in lidar signals due to attenuation oc- curs much faster at 532 nm than at 1064 nm. Moreover, the photomultiplier tubes used at 532 nm are known to generate artifacts in an extended area below highly reﬂective liquid clouds, introducing false detections that artiﬁcially lower the apparent cloud base altitude, i.e., the cloud base when the cloud is transparent or the level of complete attenuation of the lidar signal when it is opaque. Two-dimensional and multi-channel feature detection algorithm for the CALIPSO lidar measurements Thibault Vaillant de Guélis1,2, Mark A. Vaughan3, David M. Winker3, and Zhaoyan Liu3 1NASA Postdoctoral Program Fellow, NASA, Langley Research Center, Hampton, VA 23681, USA 2Science Systems and Applications, Inc., Hampton, VA 23666, USA 3NASA Langley Research Center, Hampton, VA 23681, USA By adding the informa- tion available in the 1064 nm channel, this new algorithm can better identify the true apparent cloud base altitudes of such clouds. 1 Introduction The Cloud-Aerosol Lidar and Infrared Pathﬁnder Satellite Observation (CALIPSO) mission (Winker et al., 2010) has provided direct measurements of cloud and aerosol vertical distributions with a very high vertical resolution since 2006. A key component of these measurements is made by the ac- tive remote sensing instrument CALIOP (Cloud-Aerosol Li- dar with Orthogonal Polarization), a two-wavelength (532 and 1064 nm) polarization-sensitive elastic backscatter lidar. The knowledge of the cloud and aerosol vertical distribu- tions and their properties is critical in assessing the planet’s radiation budget (e.g., Shonk and Hogan, 2010), in eval- uating the atmospheric radiative heating rate (e.g., Huang et al., 2009) and for advancing our understanding of cloud– climate feedback cycles that occur as the climate warms (e.g., Tsushima et al., 2006). The Cloud-Aerosol Lidar and Infrared Pathﬁnder Satellite Observation (CALIPSO) mission (Winker et al., 2010) has provided direct measurements of cloud and aerosol vertical distributions with a very high vertical resolution since 2006. A key component of these measurements is made by the ac- tive remote sensing instrument CALIOP (Cloud-Aerosol Li- dar with Orthogonal Polarization), a two-wavelength (532 and 1064 nm) polarization-sensitive elastic backscatter lidar. The knowledge of the cloud and aerosol vertical distribu- tions and their properties is critical in assessing the planet’s radiation budget (e.g., Shonk and Hogan, 2010), in eval- uating the atmospheric radiative heating rate (e.g., Huang et al., 2009) and for advancing our understanding of cloud– climate feedback cycles that occur as the climate warms (e.g., Tsushima et al., 2006). HAL Id: hal-03711297 https://hal.science/hal-03711297v1 Submitted on 1 Jul 2022 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Atmos. Meas. Tech., 14, 1593–1613, 2021 https://doi.org/10.5194/amt-14-1593-2021 © Author(s) 2021. This work is distributed under the Creative Commons Attribution 4.0 License. 2 Threshold-based feature detection Atmospheric lidars measure attenuated signal backscattered by molecules (m) and particles (p): To locate cloud and aerosol layers within lidar backscat- ter proﬁles, two main approaches are generally employed: the slope-based method, which looks for zero crossings in the ﬁrst derivative of the raw signal (e.g., Pal et al., 1992) and threshold-based methods, which search for regions ex- ceeding some expectation of the maximum signal value that could be measured in “clear air” (e.g., Winker and Vaughan, 1994; Clothiaux et al., 1998; Campbell et al., 2008). Some studies use a combination of both methods (e.g., Wang and Sassen, 2001; Lewis et al., 2016). A few others adopt a third method: the wavelet analysis (e.g., Davis et al., 2000; Brooks, 2003). Because these layer detection algorithms are applied successively to individual 1-D proﬁles (either single shot or averaged), we deﬁne them collectively as “proﬁle- based processes”. We also deﬁne a second, more comprehen- sive class of methods as “scene processes”. Scene processes can take advantage of the contextual information provided by a continuous series of proﬁle measurements by searching for cloud and aerosol patterns in the two-dimensional (2-D) image formed by successive lidar proﬁles. While edge de- tection techniques based on 2-D gradient search routines are not well suited for spatial analysis of lidar data (Vaughan et al., 2005), methods based on sliding window operations have been shown to greatly improve the feature shape detec- tion (e.g., Hagihara et al., 2010; van Zadelhoff et al., 2011; Herzfeld et al., 2014). β′(r) = βm(r) + βp(r) Tm(r)2TO3(r)2Tp(r)2, (1) (1) β′(r) = βm(r) + βp(r) Tm(r)2TO3(r)2Tp(r)2, where βm(r) and βp(r) are the volume backscatter coefﬁ- cients for molecules and particulates, and Tm(r)2, TO3(r)2 and Tp(r)2 are, respectively, the two-way transmittances for molecules, ozone and particles, and r is the range from the satellite altitude. If there are no particles in the atmosphere, Eq. (1) reduces to the molecular attenuated backscatter coef- ﬁcient: β′ m(r) = βm(r)Tm(r)2TO3(r)2. (2) (2) A feature, i.e., a cloud or an aerosol layer, appears as an extended and contiguous region of enhanced attenuated backscatter signal that rises signiﬁcantly above the expected clear-sky (molecules only) value. However, not all signals that exceed the expected values of \ β′m(r) necessarily indicate the presence of features; instead, such excursions are often caused by noise. T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 1594 The critically important ﬁrst step in retrieving the spatial and optical properties of clouds and aerosols is to determine where these “features” are located in the vertical, curtain- like images (altitude vs. satellite track) of the backscattered lidar signals (Fig. 1). The CALIPSO feature detection al- gorithms were ﬁrst developed for ground-based observa- tions and then adapted for space-based analyses using LITE measurements and CALIPSO simulations. These algorithms, which were conceived more than 25 years ago (e.g., Winker and Vaughan, 1994), at a time when computational power was considerably lower than what is now available, are in- voked sequentially on single, one-dimensional (1-D) lidar signal proﬁles (possibly generated from averaging data from several consecutive laser pulses). Moreover, in order to min- imize the computational load, the current CALIPSO algo- rithm is only applied to the 532 nm total signal (Vaughan et al., 2009). Section 2 presents a reﬁned method for determining fea- ture detection thresholds, which are a critically important component of the detection algorithm. Section 3 presents the detection algorithm. The detection of the Earth’s surface is described ﬁrst as it is performed ﬁrst and separately from the cloud and aerosol detection. This has been shown to have many practical advantages. Then, the cloud and aerosol de- tection algorithm is described. Finally, the detections from each channel are merged into a composite feature detection mask. Section 4 shows how this new algorithm improves the feature detection compared to the CALIPSO version 4 verti- cal feature mask (VFM). https://doi.org/10.5194/amt-14-1593-2021 2 Threshold-based feature detection \ 1β′m(r) is esti- mated using its proportional relation with the square root of \ β′m(r) (e.g., Liu and Sugimoto, 2002), called the “noise scale factor” (NSF): where 1β′ b(r) is the standard deviation due to background noise (range independent1), \ 1β′m(r) is the expected standard deviation due to the shot noise (range dependent) in the ex- pected clear sky, and N(r) is the number of bins averaged on- board. The fcorr(r) term is a correction function which takes into account the partial vertical correlation in samples due to the limited electronic bandwidth and the shifting and rebin- ning that can occur in the altitude registration phase of the level 1A processing (details in Appendix A). The number of shot noise standard deviations considered in the thresh- old is quantiﬁed by the factor k, which can be tuned accord- ing to the degree of sensitivity needed to avoid false detec- tions. \ β′m(r) is derived from modeled proﬁles of molecular and ozone number densities. 1β′ b(r) is derived from the on- board computation of the rms of the background signal in the high-altitude background region (HABR) between 65 and 80 km for each shot (Hostetler et al., 2006). \ 1β′m(r) is esti- mated using its proportional relation with the square root of \ β′m(r) (e.g., Liu and Sugimoto, 2002), called the “noise scale factor” (NSF): Figure 2. Range-dependent threshold (red) applied to a single-shot lidar signal proﬁle (blue) in clear sky during nighttime. The esti- mated molecular signal is shown in black. Jumps in the lidar signal and threshold at −0.5, 8.2 and 20.2 km reveal the change of onboard averaging resolution. The NSF is evaluated from the solar background signal during daytime for the 532 nm parallel and perpendicu- lar channels (Hostetler et al., 2006; Liu et al., 2006). At 1064 nm, CALIOP uses an avalanche photodiode (APD) de- tector rather than the photomultiplier tubes (PMTs) that are used for the 532 nm channels. Because the APD dark noise overwhelms the 1064 nm shot noise, only the background noise is considered at 1064 nm. Figure 2 shows the range-dependent threshold (red) com- puted from Eq. (3) with k = 2 applied to the 532 nm par- allel lidar signal (blue) for a clear-sky case study during nighttime. Note the noise due to the quantum nature of pho- tons (shot noise) in this ﬁgure. 2 Threshold-based feature detection To distinguish features from the ambient (but noisy) clear-sky signals, a ﬁrst step is to determine a threshold above which signals can be conﬁdently attributed to enhanced scattering arising from clouds or aerosols. We construct this threshold by ﬁrst calculating the expected molecular attenuated backscatter, \ β′m(r), to which we add k times the expected noise-induced standard deviation of the molecular signal. The resulting range-dependent threshold is the sum of \ β′m(r) and, based on error propagation theory (e.g., Bevington and Robinson, 2003), k times the root mean square (rms) of the standard deviations due to both range- independent and range-dependent noise sources. Here, we propose a new 2-D and multi-channel feature de- tection algorithm (2D-McDA). This “context-sensitive” fea- ture ﬁnder algorithm is then applied to a 2-D lidar sig- nal scene, i.e., to the image formed by many successive li- dar proﬁles. Moreover, the 2-D detection algorithm uses the backscatter signals from all available channels: the 532 nm co-polarized (or parallel) signal, the 532 nm cross-polarized (or perpendicular) signal and the 1064 nm signal. Since the backscatter from some aerosol or cloud particle types can be more pronounced in one channel than another, simultane- ously assessing the signals from all channels is expected to greatly improve the layer detection. In constructing thresholds to be applied to CALIOP data, one must take into account the onboard signal averaging that is applied to the backscatter measurements. Because the CALIPSO satellite has limited telemetry bandwidth, the backscatter data are averaged both vertically and horizontally before the data are downlinked from the satellite, with in- creasing amounts of averaging applied to data acquired at https://doi.org/10.5194/amt-14-1593-2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 1595 Figure 1. Curtain of attenuated backscatter signal measured by CALIOP in the 532 nm parallel channel during nighttime observations on 31 August 2018, 21:46:37 UTC (start point), over the Arabian Sea. backscatter signal measured by CALIOP in the 532 nm parallel channel during nighttime observations on (start point), over the Arabian Sea. Figure 1. Curtain of attenuated backscatter signal measured by CALIOP in the 532 nm parallel channel during nighttime observations on 31 August 2018, 21:46:37 UTC (start point), over the Arabian Sea. Figure 2. Range-dependent threshold (red) applied to a single-shot lidar signal proﬁle (blue) in clear sky during nighttime. The esti- mated molecular signal is shown in black. 1The background noise is range- independent in the digitizer- reading domain P . However, it then depends on r when transformed to the β′ domain. 2 Threshold-based feature detection Jumps in the lidar signal and threshold at −0.5, 8.2 and 20.2 km reveal the change of onboard averaging resolution. higher altitudes (Hunt et al., 2009). As an example, signals acquired between 8.2 and 20.2 km are averaged horizontally over three consecutive lidar pulses and vertically for four full-resolution (15 m) range bins. Consequently, the down- linked proﬁle data from within this region have been aver- aged over 12 full-resolution onboard range bins. We com- pute a range-dependent threshold speciﬁcally tailored for the CALIOP proﬁles using higher altitudes (Hunt et al., 2009). As an example, signals acquired between 8.2 and 20.2 km are averaged horizontally over three consecutive lidar pulses and vertically for four full-resolution (15 m) range bins. Consequently, the down- linked proﬁle data from within this region have been aver- aged over 12 full-resolution onboard range bins. We com- pute a range-dependent threshold speciﬁcally tailored for the CALIOP proﬁles using β′ T(r) = \ β′m(r) + k fcorr(r) √N(r) r 1β′ b(r)2 + \ 1β′m(r) 2, (3) (3) (3) where 1β′ b(r) is the standard deviation due to background noise (range independent1), \ 1β′m(r) is the expected standard deviation due to the shot noise (range dependent) in the ex- pected clear sky, and N(r) is the number of bins averaged on- board. The fcorr(r) term is a correction function which takes into account the partial vertical correlation in samples due to the limited electronic bandwidth and the shifting and rebin- ning that can occur in the altitude registration phase of the level 1A processing (details in Appendix A). The number of shot noise standard deviations considered in the thresh- old is quantiﬁed by the factor k, which can be tuned accord- ing to the degree of sensitivity needed to avoid false detec- tions. \ β′m(r) is derived from modeled proﬁles of molecular and ozone number densities. 1β′ b(r) is derived from the on- board computation of the rms of the background signal in the high-altitude background region (HABR) between 65 and 80 km for each shot (Hostetler et al., 2006). T. Vaillant de Guélis et al.: 2D-McDA for CALIOP T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 1596 Figure 3. Pixels of Fig. 1 where the lidar signal is above the range- dependent threshold computed from Eq. (3) with k = 2 are shown in orange. Brown pixels show surface detection. ﬁrst an independent detection of the Earth’s surface. Doing the surface detection in a ﬁrst and separate step allows a bet- ter retrieval of the surface echo and prevents complications in the cloud and aerosol layer detection process. Also, knowing where the surface is detected allows subsequent separation of semi-transparent features from opaque features, which is essential for accurately estimating range-resolved proﬁles of extinction coefﬁcients (Young et al., 2018). Operationally, atmospheric features are deﬁned as being opaque when no surface return or other atmospheric feature can be detected below them. From this deﬁnition, it follows that the signals received from beneath opaque features have been fully at- tenuated within these features. The Earth surface detection algorithm used here is closely akin to the one described in Vaughan et al. (2021) and is applied to the 532 nm paral- lel and 1064 nm channels (details in Appendix B). The sig- nals from the top of the detected surface echo and below this point are removed from the data. To minimize computation times, the surface detection algorithm is not applied to the 532 nm perpendicular channel signal. The backscatter from ocean surfaces (covering ∼70 % of the planet) does not de- polarize and, excluding snow and ice, the depolarization of most land surfaces is relatively low (Lu et al., 2017); hence, the preponderance of the surface backscatter is in the paral- lel channel. The altitude retrieved in the parallel channel is used to remove the signal at and below the estimated surface altitude in the perpendicular channel. Note that there is some small chance that a surface echo can appear in the perpen- dicular channel but not be visible or detected in the parallel channel. The detection of the surface corresponding to Fig. 1 is shown in brown in Fig. 3. Figure 3. Pixels of Fig. 1 where the lidar signal is above the range- dependent threshold computed from Eq. (3) with k = 2 are shown in orange. Brown pixels show surface detection. Jumps at −0.5, 8.2 and 20.2 km reveal the change of onboard averaging resolution. 3.2 Cloud and aerosol detection (6) The detection of cloud and aerosol layers in a single channel curtain of lidar measurements takes place in four main steps: Equation (6) is applied to the three lidar channels (532 nm parallel, 532 nm perpendicular and 1064 nm) during the 2D- McDA process. 1. detecting strong features, i.e., identifying contiguous re- gions of enhanced attenuated scattering ratios that rise above the feature detection threshold (which is repeat- edly decreased from a very large value of k down to k = 1) without applying any signal averaging (i.e., d = 1–4 in Table 1, Sect. 3.2.1); T. Vaillant de Guélis et al.: 2D-McDA for CALIOP However, a few points of the lidar sig- nal still exceed the threshold. Some continuity tests are then needed to determine whether these high signals are due to noise or instead part of an extended feature. Unlike the cur- rent CALIPSO detection algorithm, this continuity test will be applied in two dimensions. Figure 3 shows all pixels of Fig. 1 where the lidar signal is above the range-dependent threshold computed from Eq. (3) with k = 2. p q Like the current CALIOP layer detection algorithm, the 2D-McDA is applied to proﬁles of attenuated scattering ra- tios, deﬁned as R′(r) = β′(r) β′m(r). (5) (5) The attenuated scattering ratio threshold is then obtained from R′ T(r) = β′ T(r) \ β′m(r) = 1 + k fcorr(r) √N(r) v u u t 1 \ β′m(r) 2 1β′ b(r)2 + NSF2 β′ 1 \ β′m(r) . (6) 3 2-D and multi-channel feature detection algorithm The 2D-McDA is applied to the scattering ratio signals at 532 nm parallel, 532 nm perpendicular and 1064 nm. First, the detection of the surface altitude is performed and the sig- nal from this altitude and below is removed from the data. Second, the detection of cloud and aerosol layers is done in each channel based on iterated detection thresholds and 2- D spatial continuity tests. Finally, the masks from the three channels are merged in a composite feature mask. 2. ﬂagging regions below opaque features as “fully attenu- ated” (FA) and regions below transparent features where the signal is strongly attenuated with the low conﬁdence ﬂag “almost fully attenuated” (AFA) (Sect. 3.2.2); 3. averaging of those signals not already ﬂagged using a horizontal sliding window (Sect. 3.2.3); and 4. detecting faint features; features are once again identi- ﬁed as contiguous regions of enhanced signal (i.e., av- eraged attenuated scattering ratios) that rise above the recomputed feature detection threshold (Sect. 3.2.1). 2 Threshold-based feature detection Indeed, although background noise, mainly due to solar radiation, is quite low during night- time, the lidar signal shows large variations around the ex- pected clear-sky return (black). The range-dependent thresh- old correctly keeps most the signal below the detection level. \ 1β′m(r) = NSFβ′ q \ β′m(r). (4) (4) 1The background noise is range- independent in the digitizer- reading domain P . However, it then depends on r when transformed to the β′ domain. https://doi.org/10.5194/amt-14-1593-2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 3.1 Surface detection Before detecting cloud and aerosol layers using the detection threshold as described in the previous section, we perform Atmos. Meas. Tech., 14, 1593–1613, 2021 https://doi.org/10.5194/amt-14-1593-2021 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 1597 Figure 4. Flowchart of the two-dimensional and multi-channel feature detection algorithm (2D-McDA). d is the detection step, k is the number of total noise standard deviations used in the detection threshold Eq. (3), s is the size of the window used for the spatial coherence test, n is the minimum number of pixels in each pattern, and a is the size of the averaging window. See the algorithm description in Sect. 3 and the coefﬁcient values in Table 1. Figure 4. Flowchart of the two-dimensional and multi-channel feature detection algorithm (2D-McDA). d is the detection step, k is the number of total noise standard deviations used in the detection threshold Eq. (3), s is the size of the window used for the spatial coherence test, n is the minimum number of pixels in each pattern, and a is the size of the averaging window. See the algorithm description in Sect. 3 and the coefﬁcient values in Table 1. standard deviation with k = 2 (Fig. 5a; orange pixels). Then, the second substep is to apply a spatial coherence test win- dow (rows a in Table 1) on these detected pixels (Fig. 5a, b). Here, an 11 × 11 pixels window is applied to each pixel of the image, with the window being centered successively on all pixels. If the number of originally detected pixels in the window is greater than half of the total number of pixel in the window (≥61 for a 11 × 11 pixels window), then the center pixel is considered to be detected. If not, the center pixel is considered to be undetected. In this smoothing step, the de- termination of detection status does not rely on a single pixel exceeding its threshold but instead on the fraction of neigh- boring pixels that exceed their thresholds. Consequently, a pixel classiﬁed as detected may not itself exceed the de- tection threshold. Similarly, a pixel that exceeds the thresh- old may not ultimately be classiﬁed as detected. The pixel count within the window is limited to those detected at the current detection level d and at the previous detection level d −1. 3.1 Surface detection This allows detection continuity of similar backscatter intensities and avoids connecting noise encountered during fainter detections to a strong feature detected earlier. Other ﬂagged pixels (i.e., “surface”, detection ≤d −2, “likely arti- fact”, “fully attenuated”, “almost fully attenuated” and “low conﬁdence small strips” (to be described in detail later in this section and Sect. 3.2.2) and pixels outside the window when the top or bottom edges of the image are not consid- ered in the window and the total number of candidate pix- els in the window is decreased accordingly. The shapes of detected features are smoothed by the spatial coherence test window, while the noise (isolated orange pixel) is removed. However, some small clusters of pixels sometimes persist. Those small clusters cannot be conﬁdently declared as fea- tures at this stage. They can be due to noise or they can be part of a larger, fainter feature. Then, we decide not to con- sider these small patterns as detected features and retain these regions for inclusion in the signal averaging used in succes- sive iterations of the algorithm. To be declared as detected features, smoothed detected patterns need to consist of more Figure 4 shows the ﬂowchart of the whole detection algo- rithm. The parameter values used at the different detection levels are given in Table 1. The following subsections give the details of the main steps presented above. Atmos. Meas. Tech., 14, 1593–1613, 2021 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP Table 1. Coefﬁcient k in threshold detection, spatial coherence test window size s, minimum number of pixels in pattern n and averaging window size a used at each detection level d. Window sizes are given in vertical × horizontal pixel counts, with a single pixel resolution of 30 m × 0.33 km. Table 1. Coefﬁcient k in threshold detection, spatial coherence test window size s, minimum number of pixels in pattern n and averaging window size a used at each detection level d. Window sizes are given in vertical × horizontal pixel counts, with a single pixel resolution of 30 m × 0.33 km. d 1 2 3 4 5 532 nm parallel k 100 20 2 1 1 s – – 11 × 11 3 × 21 9 × 51 (330 m × 3.67 km) (90 m × 7 km) (270 m × 17 km) n 1 1 60 200 10 000 a – – – – 1 × 15 (30 m × 5 km) 532 nm perpendicular k 500 100 2 1 1 s – – 11 × 11 3 × 21 9 × 51 (330 m × 3.67 km) (90 m × 7 km) (270 m × 17 km) n 1 1 60 200 1000 a – – – – 1 × 15 (30 m × 5 km) 1064 nm k – 20 2 1 1 s – – 11 × 11 3 × 21 9 × 51 (330 m × 3.67 km) (90 m × 7 km) (270 m × 17 km) n – 1 60 200 10 000 a – – – – 1 × 15 (30 m × 5 km) initial scan is to identify the tops of very strongly scattering liquid clouds and ice clouds containing high fractions of hor- izontally oriented ice (HOI) crystals. The non-ideal transient response by PMTs following these very strong signals often generates a spurious, exponentially decaying signal enhance- ment in the underlying range bins (McGill et al., 2007; Hunt et al., 2009; Lu et al., 2020). The presence of these “noise tails” in the 532 nm signals can introduce large biases into the determination of the apparent bases of opaque water clouds. To exclude this artifact in the detection process, the 600 m below the base of the detected very strong signal are ﬂagged as “likely artifact” and removed from the signal. T. Vaillant de Guélis et al.: 2D-McDA for CALIOP Since the APD used in 1064 nm channel does not produce these noise tails, we rely on the 1064 nm channel for the detection of the apparent base of these strongly scattering layers. than n connected pixels (see Table 1); otherwise, they are re- moved from this detection level. This detection procedure is applied several times (the suc- cessive detection level d of Table 1) with different thresholds, different spatial coherence test windows s and different lim- its on the number of connected pixels required n (Table 1) in order to detect all layers from the most evident, very strong patterns to the very faint ones and from geometrically small patterns to very extended ones. Note that the horizontal spa- tial coherence test window (3 × 21) enables the detection of faint but horizontally extended cirrus such as the layer shown between 50 and 100 km in Fig. 1. The detection of this sub- visible cirrus is presented in Fig. 6. Figure 6a–b show the implementation of the 3 × 21 spatial coherence test window. We see that the cirrus pattern is smoothed and now clearly ap- pears in Fig. 6b due to the fact that most of the noise around has been removed. However, many small clusters of noise pixels persist. By applying the minimum numeric threshold of connected pixel n on the detected pattern, we are able to remove small clusters due to noise while keeping the real cir- rus (Fig. 6c). After detection of the strongest features, i.e., without sig- nal averaging (d = 1−4 in Table 1), we ﬂag all pixels below a detected strong feature where the surface has not been de- tected as “fully attenuated” (FA). In this portion of the pro- ﬁle, the signal is too weak to be further exploited. Second, the contiguous pixels located in the vertical extent between two detected features are ﬂagged as “almost fully attenuated” (AFA) whenever the backscatter intensity falls below an em- pirically determined threshold. For the 532 nm parallel chan- nel, these pixels are ﬂagged as AFA when more than 30 % 3.2.1 Detection The detection phase is performed following three substeps: 1. All pixels within the image that exceed the threshold are ﬁrst ﬂagged as detected (Fig. 3). 2. A spatial coherence test window is applied to the im- age of detected versus undetected pixels. It smooths the shape of detected pattern and removes isolated noisy de- tected pixels by turning some of detected pixels to un- detected or undetected to detected. 3. Smoothed patterns are required to meet a minimum nu- meric threshold of contiguous pixels. Patterns that fail to meet this threshold are removed from consideration for this level of detection. The scattering ratio image used in the layer detection scheme has a spatial resolution of one laser pulse horizon- tally and 30 m vertically, equivalent to the ﬁnest spatial res- olution of the CALIOP data. As described in Hunt et al. (2009), CALIOP data are averaged aboard the satellite with spatial resolutions that vary according to altitude. Scattering ratios in regions where the data resolution is coarser than the image resolution (30 m × 0.33 km horizontally) are du- plicated as necessary to match the image resolution. For ex- ample, between 8.2 and 20.2 km, the spatial resolution of the signal is 1 km horizontally ×60 m vertically. These values are replicated 12 times to populate the corresponding area in the 30 m × 0.33 km scattering ratio image. The ﬁrst sub- step is then to ﬂag all pixels of this image which exceed the detection threshold given by Eq. (6) with the value of k de- ﬁned in Table 1. For example, for the 532 nm parallel chan- nel, at the detection level d = 3, the detected pixels are those where the signal is greater than 1+k times the expected noise Atmos. Meas. Tech., 14, 1593–1613, 2021 https://doi.org/10.5194/amt-14-1593-2021 1598 https://doi.org/10.5194/amt-14-1593-2021 3.2.2 Special ﬂags For the 532 nm channels, a ﬁrst detection of a very strong signal is performed (see d = 1 in Table 1). The aim of this Atmos. Meas. Tech., 14, 1593–1613, 2021 https://doi.org/10.5194/amt-14-1593-2021 1599 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP Figure 5. Illustration of the spatial coherence test window substep of 2D-McDA on the 532 nm parallel channel for the case study shown in Fig. 1. (a) Pixels which are over the detection threshold given by Eq. (6) with the value of k = 2 (orange). (b) Result after applying a 11×11 spatial coherence test window on the detected pixels. Note that the insert at the top is just an illustration and does not show the real content of the image portion. Figure 5. Illustration of the spatial coherence test window substep of 2D-McDA on the 532 nm parallel channel for the case study sh Figure 5. Illustration of the spatial coherence test window substep of 2D-McDA on the 532 nm parallel channel for the case study shown in Fig. 1. (a) Pixels which are over the detection threshold given by Eq. (6) with the value of k = 2 (orange). (b) Result after applying a 11×11 spatial coherence test window on the detected pixels. Note that the insert at the top is just an illustration and does not show the real content of the image portion. After removing all data identiﬁed with these low conﬁ- dence ﬂags from the attenuated scattering ratios, the signal is averaged in order to try to detect fainter features. of the population has backscatter intensities that are less than 10 % of the corresponding detection thresholds. These pix- els are ﬂagged as AFA in the 532 nm perpendicular channel whenever the signals in more than 90 % of the population fall below (100 % of) the corresponding threshold values. To be ﬂagged as AFA in the 1064 nm channel, more than 85 % of the population must have a signal less than (100 %) of the corresponding threshold. In all cases, the AFA thresholds were determined experimentally and are tunable. Finally, the horizontal distance between successive (A)FA columns can be very small and the likelihood of conﬁdently detecting fea- tures in these narrow gaps is very low. 3.2.2 Special ﬂags For this reason, the data in all horizontal extents smaller than 5 km (15 proﬁles) that lie between (A)FA columns are ﬂagged as “low conﬁ- dence small strips”. Atmos. Meas. Tech., 14, 1593–1613, 2021 3.2.3 Signal averaging We then average the remaining signal (here the attenuated scattering ratios) using a Gaussian sliding window that ex- tends over 5 km (15 proﬁles) horizontally and a single range bin vertically (a in Table 1). Using a sliding window, instead of the ﬁxed window used in the CALIOP feature detection al- gorithm, provides much improved resolution of the horizon- tal edges’ position of faint features (0.33 km instead of 5, 20, or 80 km) and makes it possible to detect non-uniform hor- izontal edges. A Gaussian weight with a standard deviation of 1.67 km is applied, thus giving a stronger weight to pixels closer to center of the window than at the edges. We chose a horizontal window here because the spatial extent of very https://doi.org/10.5194/amt-14-1593-2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 1600 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP lustration of the horizontal spatial coherence test window substep and the pattern size threshold rejection substep of 2D nm parallel channel for the case study shown in Fig. 1. (a) Pixels which are over the detection threshold given by Eq. ( k = 1 (orange). (b) Result after applying a 3×21 spatial coherence test window on the detected pixels. (c) Result after r composed by less than n = 200 pixels (note that the insert at the top is just an illustration and does show the real conten on). Before these substeps, surface is detected ﬁrst (brown); then, a very strong signal (k = 100) occurring on highly re ds is detected (black) and the 600 m below is ﬂagged as “likely artifact” (gray), as it is the region where we see artifacts du se of photomultiplier tubes (PMTs) in the 532 nm channels. Two detections were also made: one with k = 20 and anoth × 11 spatial coherence test window and n = 60. Figure 6. Illustration of the horizontal spatial coherence test window substep and the pattern size threshold rejection substep of 2D-McDA on the 532 nm parallel channel for the case study shown in Fig. 1. (a) Pixels which are over the detection threshold given by Eq. (6) with the value of k = 1 (orange). (b) Result after applying a 3×21 spatial coherence test window on the detected pixels. 3.2.3 Signal averaging (c) Result after rejecting all patterns composed by less than n = 200 pixels (note that the insert at the top is just an illustration and does show the real content of the image portion). Before these substeps, surface is detected ﬁrst (brown); then, a very strong signal (k = 100) occurring on highly reﬂecting liquid clouds is detected (black) and the 600 m below is ﬂagged as “likely artifact” (gray), as it is the region where we see artifacts due to the time response of photomultiplier tubes (PMTs) in the 532 nm channels. Two detections were also made: one with k = 20 and another with k = 2, a 11 × 11 spatial coherence test window and n = 60. Pixels ﬂagged as surfaces or features are not considered in the averaging window. However, if the center pixel of the av- eraging window (i.e., the pixel to which the averaging is ap- plied) is a low conﬁdence pixel (i.e., “likely artifact”, (A)FA or “low conﬁdence small strips”), then the averaging window faint layers is mainly in the horizontal direction. Typically, thin cirrus have geometrical thicknesses of a few hundreds of meters but spread horizontally over several hundreds of kilo- meters. The use of a horizontal averaging window thus al- lows the detection of thin layers close to each other vertically. Atmos. Meas. Tech., 14, 1593–1613, 2021 https://doi.org/10.5194/amt-14-1593-2021 1601 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP Figure 7. Final feature mask of the 532 nm parallel channel. Figure 7. Final feature mask of the 532 nm parallel channel. sponse artifacts (note that the light blue color indicates “1064 only” in Fig. 9a). is applied, and, if the average signal value exceeds the detec- tion threshold, this center pixel in the feature detection mask is “unﬂagged” until the end of the detection-level processing, after which its low conﬁdence ﬂag is restored. This allows us to maintain connections between features separated by a few low conﬁdence pixels. Once the averaging is performed, the detection substeps (Sect. 3.2.1) are then applied to the averaged signal. Note too that horizontally adjacent features separated only by a low conﬁdence vertical band (i.e., pix- els classiﬁed as FA, AFA and/or small strips) are considered a single, merged feature when counting the number of con- nected pixels. 3.2.3 Signal averaging Some examples of this horizontal merging are seen in the smaller fragments of the aerosol layer found at about 4 km and an along-track distance of 500 km to 750 km in Fig. 7. 4.1 Variety of cloud type and shape Figure 10 presents the attenuated backscattered lidar signal in the three channels for another case study showing a vari- ety of cloud types and shapes which occurred above Ethiopia on 31 August 2018 during nighttime. We can see that the artifacts below liquid water clouds (close to the surface and up to 8 km) appear in the 532 nm parallel (Fig. 10a) and the 532 nm perpendicular (Fig. 10b) channels but not at 1064 nm (Fig. 10c). We note that thin cirrus clouds, like the one at 17 km in altitude between 1550 and 1850 km, are clearly brought out in the 532 nm perpendicular channel (Fig. 10b). If we look now at the composite feature detections derived from these three signals (Fig. 11a), we note again how well the apparent bases of liquid clouds are retrieved by using the 1064 nm channel. We note also that the successful identiﬁca- tion of thin cirrus can largely be attributed to our use of the 532 nm perpendicular channel. Figure 11b shows the same mask as Fig. 11a but with the same colors that are used for the VFM images (Fig. 11c). This change of colors is intended to facilitate one-to-one comparisons between the two detection schemes. However, note that the yellow and white colors do not discriminate aerosol from cloud, as in the VFM, but in- stead simply differentiate weak from strong features based on whether the feature detection required data averaging (yel- low) or not (white). Finally, Fig. 11d shows the difference between the new composite feature detection mask and the VFM. We see that the contour of features retrieved by the 2D- McDA represents a distinct improvement over the squared boundaries reported by the VFM. We note too that the new Figure 7 shows the ﬁnal mask for the 532 nm parallel chan- nel after the detection of the faint features. 4 Performance assessments and comparisons to version 4 In this section, we present two case studies to show the im- provements made by this new feature detection approach. T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 1602 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP igure 8. Curtain of attenuated scattering ratios measured by CALIOP during nighttime observations on 31 August, 21:46:37 UTC (start oint), daytime observations at (a) 532 nm parallel (same as Fig. 1), (b) 532 nm perpendicular and (c) 1064 nm. e 8. Curtain of attenuated scattering ratios measured by CALIOP during nighttime observations on 31 August, 21:46:37 UTC ( daytime observations at (a) 532 nm parallel (same as Fig 1) (b) 532 nm perpendicular and (c) 1064 nm Figure 8. Curtain of attenuated scattering ratios measured by CALIOP during nighttime observations on 31 August, 21:46:37 UTC (start point), daytime observations at (a) 532 nm parallel (same as Fig. 1), (b) 532 nm perpendicular and (c) 1064 nm. algorithm detects thin clouds that are obviously missed by the VFM and that it eliminates signiﬁcant detection artifacts reported by the VFM between 700 and 900 km. cates that the signals are fully attenuated after detecting (at 532 nm) the apparent base of the smoke layer. However, at 1064 nm, the dense smoke layer is semi-transparent because the 1064 nm signals are attenuated signiﬁcantly less than at 532 nm. Then, the surface is readily detected at 1064 nm (Fig. 12c). This scene clearly illustrates the advantage gained by using a multi-channel feature detection algorithm, since the full vertical extent of the smoke plume can only be re- trieved by inspecting the 1064 nm measurements (light blue color in Fig. 13a). 3.3 Three-channel composite detection The detection algorithm is applied individually to the lidar signal from each of the three channels (Fig. 8), and all pix- els identiﬁed as features in any of the three channels are re- tained in the composite mask (Fig. 9a). Comparing this new feature mask (Fig. 9a, b) to the current version of the VFM (Fig. 9c), we ﬁrst note the improvement in the detected con- tour of the large cirrus. We also note that the 2D-McDA read- ily detects faint cirrus (e.g., as seen between 0 and 75 km) that is missed by the current VFM. The vertical spreading of the clouds seen in the VFM at around 7.5 km in altitude and between 500 and 900 km horizontally is due to the afore- mentioned PMT artifact afﬂicting the 532 nm signals beneath strongly scattering layers. This is not seen in the 2D-McDA feature mask because pixels below the cloud top are ﬂagged as “likely artifact” in the 532 nm channels and so we make no attempt to retrieve the cloud apparent base of such opaque clouds at this wavelength. Instead, in these cases, we retrieve the true penetration depth estimates using the 1064 nm sig- nals (Fig. 8c), which are not affected by detector transient re- https://doi.org/10.5194/amt-14-1593-2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 1604 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP gure 10. Curtain of attenuated scattering ratios measured by CALIOP during nighttime observations on 31 August 2018, 23:25:54 art point), over Ethiopia at (a) 532 nm parallel, (b) 532 nm perpendicular and (c) 1064 nm. 0. Curtain of attenuated scattering ratios measured by CALIOP during nighttime observations on 31 August 2018, 23:25:54 UTC nt), over Ethiopia at (a) 532 nm parallel, (b) 532 nm perpendicular and (c) 1064 nm. Figure 10. Curtain of attenuated scattering ratios measured by CALIOP during nighttime observations on 31 August 2018, 23:25:54 UTC (start point), over Ethiopia at (a) 532 nm parallel, (b) 532 nm perpendicular and (c) 1064 nm. 4.2 Dense smoke Figure 12 presents a dense smoke event in Siberia on 26 July 2006 during daytime. Figure 12 presents a dense smoke event in Siberia on 26 July 2006 during daytime. The smoke layer is opaque at 532 nm, and thus we do not see any surface echo for this channel (Fig. 12a). Note that the smoke is non-depolarizing so there is no perpendicular sig- nal (Fig. 12b). Because the standard CALIOP layer detection only examines the 532 nm channel, the VFM (Fig. 13c) indi- https://doi.org/10.5194/amt-14-1593-2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 1603 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 9. (a) Composite feature detection mask derived from signals shown in Fig. 8. (b) Same as panel (a) but using the same sed for the VFM. “Strong” (white) features are those detected without averaging in at least one channel; others are ﬂagged ). (c) VFM of version 4 of the CALIOP data product. (d) Difference between the new mask and the VFM. Figure 9. (a) Composite feature detection mask derived from signals shown in Fig. 8. (b) Same as panel (a) but using the same colors as those used for the VFM. “Strong” (white) features are those detected without averaging in at least one channel; others are ﬂagged as “weak” (yellow). (c) VFM of version 4 of the CALIOP data product. (d) Difference between the new mask and the VFM. https://doi.org/10.5194/amt-14-1593-2021 https://doi.org/10.5194/amt-14-1593-2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 5 Conclusions tails that are often obscured by CALIPSO’s standard multi- resolution averaging scheme. Relative to the CALIPSO ver- sion 4.2 vertical feature mask (VFM) data product, the 2D- McDA shows the following improvements. This paper describes the architecture and theoretical un- derpinnings of a new two-dimensional, multi-channel fea- ture detection algorithm (2D-McDA) used to identify layer boundaries in the backscatter signals acquired by the elas- tic backscatter lidar aboard the Cloud-Aerosol Lidar and Infrared Pathﬁnder Satellite Observations (CALIPSO) plat- form. The cloud and aerosol layer detection boundaries re- ported in the standard CALIPSO data products are detected by scanning sequences of 532 nm attenuated scattering ratio proﬁles constructed at increasingly coarser horizontal aver- aging resolutions. In contrast, the 2D-McDA is more akin to an image processing algorithm that examines full-resolution lidar scenes and hence can identify many of the ﬁne de- – Because it is applied to single proﬁles averaged over several different horizontal resolutions, the standard CALIOP feature detection produces blocky, rectangu- lar layers. The complex shapes of aerosol and cloud features are better preserved by the 2D-McDA win- dowing and data aggregation operations, which provide the ﬂexibility required to distinguish ﬁne spatial details. It is hoped that this improved feature detection will lead to improvements in classifying features accord- ing to type (e.g., clouds vs. aerosols) and in their op- tical property retrievals. Ideally, separate identiﬁcation – Because it is applied to single proﬁles averaged over several different horizontal resolutions, the standard CALIOP feature detection produces blocky, rectangu- lar layers. The complex shapes of aerosol and cloud features are better preserved by the 2D-McDA win- dowing and data aggregation operations, which provide the ﬂexibility required to distinguish ﬁne spatial details. It is hoped that this improved feature detection will lead to improvements in classifying features accord- ing to type (e.g., clouds vs. aerosols) and in their op- tical property retrievals. Ideally, separate identiﬁcation Atmos. Meas. Tech., 14, 1593–1613, 2021 https://doi.org/10.5194/amt-14-1593-2021 1605 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 11. (a) Composite feature detection mask derived from signals shown in Fig. 10. (b) Same as panel (a) but using the sam sed for the VFM. “Strong” (white) features are those detected without averaging in at least one channel; others are ﬂagged ) (c) VFM of version 4 of the CALIOP data product (d) Difference between the new mask and the VFM Figure 11. T. Vaillant de Guélis et al.: 2D-McDA for CALIOP T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 1606 Figure 12. Curtain of attenuated scattering ratios measured by CALIOP during a dense smoke event which occurred in Siberia on 26 Ju 2006, 06:00:25 UTC (start point), daytime observations at (a) 532 nm parallel, (b) 532 nm perpendicular and (c) 1064 nm. ure 12. Curtain of attenuated scattering ratios measured by CALIOP during a dense smoke event which occurred in Siberia on 26 July 6, 06:00:25 UTC (start point), daytime observations at (a) 532 nm parallel, (b) 532 nm perpendicular and (c) 1064 nm. Figure 12. Curtain of attenuated scattering ratios measured by CALIOP during a dense smoke event which occurred in Siberia on 26 July 2006, 06:00:25 UTC (start point), daytime observations at (a) 532 nm parallel, (b) 532 nm perpendicular and (c) 1064 nm. thermal budget and drive dynamics of the tropopause region (e.g., Hartmann et al., 2001; McFarquhar et al., 2000). of strongly scattering and weakly scattering features by the 2D-McDA will also offer improved discrimination between juxtaposed cloud and aerosol layers or identiﬁ- able regions of ice and liquid water within a cloud. The improvement of the cloud shape detection is by itself important, for example, for studies interested in anvil clouds (e.g., Bony et al., 2016; Hartmann, 2016). – The apparent base altitudes of highly reﬂective clouds, i.e., the levels of complete attenuation of the lidar signal, which are routinely biased low (by several hundred me- ters) due to the non-ideal transient response of 532 nm photomultiplier tubes, are now more correctly retrieved by incorporating measurements made by the 1064 nm channel. The apparent cloud base altitude, which re- sults from both attenuation of the direct beam and mul- tiple scattering effects, has been directly linked with the amount of longwave radiation escaping the Earth at the top of the atmosphere (Vaillant de Guélis et al., 2017a, b), making its accurate estimation very impor- – The apparent base altitudes of highly reﬂective clouds, i.e., the levels of complete attenuation of the lidar signal, which are routinely biased low (by several hundred me- ters) due to the non-ideal transient response of 532 nm photomultiplier tubes, are now more correctly retrieved by incorporating measurements made by the 1064 nm channel. 5 Conclusions (a) Composite feature detection mask derived from signals shown in Fig. 10. (b) Same as panel (a) but using the same colors as those used for the VFM. “Strong” (white) features are those detected without averaging in at least one channel; others are ﬂagged as “weak” (yellow). (c) VFM of version 4 of the CALIOP data product. (d) Difference between the new mask and the VFM. Atmos. Meas. Tech., 14, 1593–1613, 2021 https://doi.org/10.5194/amt-14-1593-2021 https://doi.org/10.5194/amt-14-1593-2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP The apparent cloud base altitude, which re- sults from both attenuation of the direct beam and mul- tiple scattering effects, has been directly linked with the amount of longwave radiation escaping the Earth at the top of the atmosphere (Vaillant de Guélis et al., 2017a, b), making its accurate estimation very impor- – The detection of subvisible cirrus is signiﬁcantly en- hanced by both the 2-D detection scheme and the use of the 532 nm perpendicular channel, which is especially sensitive to the presence of depolarizing ice crystals. Those clouds play an important role in the climate sys- tem as they regulate the vertical transport of water va- por near the upper troposphere–lower stratosphere (e.g., Jensen et al., 1996; Luo et al., 2003), inﬂuence the local Atmos. Meas. Tech., 14, 1593–1613, 2021 https://doi.org/10.5194/amt-14-1593-2021 1607 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 13. (a) Composite feature detection mask derived from signals shown in Fig. 12. (b) Same as panel (a) but using the same sed for the VFM. “Strong” (white) features are those detected without averaging in at least one channel; others are ﬂagged a ). (c) VFM of version 4 of the CALIOP data product. (d) Difference between the new mask and the VFM. Figure 13. (a) Composite feature detection mask derived from signals shown in Fig. 12. (b) Same as panel (a) but using the same colors as those used for the VFM. “Strong” (white) features are those detected without averaging in at least one channel; others are ﬂagged as “weak” (yellow). (c) VFM of version 4 of the CALIOP data product. (d) Difference between the new mask and the VFM. https://doi.org/10.5194/amt-14-1593-2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 1608 tant for cloud feedback studies (Vaillant de Guélis et al., 2018). – The 2D-McDA can retrieve the full vertical extent of dense smoke layers by examining the 1064 nm channel. Within smoke, the 1064 nm signals are attenuated sig- niﬁcantly less than at 532 nm and hence can more often penetrate the full vertical extent of these layers. Those biomass burning aerosols play a signiﬁcant role in the Earth’s radiative balance by their scattering and absorp- tion of incoming solar radiation (e.g., Penner et al., 1992; Christopher et al., 1996) and the interaction they have with clouds (e.g., Kaufman and Fraser, 1997). https://doi.org/10.5194/amt-14-1593-2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 A1 Correction due to electronic bandwidth If Nbin 15 m bins are vertically averaged together to form a larger bin B, then the variance of the mean is given by Var(B) = Var 1 Nbin Nbin X i=1 bi = 1 N2 bin Var Nbin X i=1 bi = 1 N2 bin Nbin X i,j=1 Cov(bi,bj) = 1 N2 bin Nbin X i=1 Var(bi) + 2 X 1≤i<j≤Nbin Cov(bi,bj) , (A1) B′ k = Nbin −Nshift15 Nbin Bk + Nshift15 Nbin Bk+1, (A4) where Bk = b1 + b2 + ... + bNbin Nbin = 1 Nbin Nbin X i=1 bi, (A5) and Bk+1 = bNbin+1 + bNbin+2 + ... + bNbin+Nbin Nbin = 1 Nbin Nbin X j=1 bNbin+j. (A6) (A4) (A1) where Cov(·,·) is the covariance. When averaging Nbin con- secutive 15 m bins, we can consider they have approximately the same range and then that their variance is constant: Var(bi) = Var(b). If the bi bins were uncorrelated, then we would have Cov(bi,bj) = 0, ∀(i ̸= j), and then Var(B) = Var(b) Nbin . However, since each bin is partially correlated with its vertical neighbors, we have Cov(bi,bi+m) = constant∀i for each lag of m range bins. Then, Eq. (A1) can be rewritten following where Cov(·,·) is the covariance. When averaging Nbin con- secutive 15 m bins, we can consider they have approximately the same range and then that their variance is constant: Var(bi) = Var(b). If the bi bins were uncorrelated, then we would have Cov(bi,bj) = 0, ∀(i ̸= j), and then Var(B) = Var(b) Nbin . However, since each bin is partially correlated with its vertical neighbors, we have Cov(bi,bi+m) = constant∀i for each lag of m range bins. Then, Eq. (A1) can be rewritten following (A5) and Bk+1 = bNbin+1 + bNbin+2 + ... + bNbin+Nbin Nbin = 1 Nbin Nbin X j=1 bNbin+j. = 1 Nbin Nbin X j=1 bNbin+j. (A6) (A6) Var(B) = Var(b) Nbin 1 + 2 Nbin X 1≤i<j≤Nbin Cov(bi,bj) Var(b) ! = Var(b) Nbin 1 + 2 Nbin−1 X m=1 Nbin −m Nbin R(m) ! , (A2) (A2) where R(m) = Cov(bi,bi+m) Var(b) is the autocorrelation coefﬁcient for a lag of m range bins. It follows that the correction func- tion to apply on the total noise standard deviation in Eq. A1 Correction due to electronic bandwidth An additional correlation arises from the data redistribution in the altitude registration of level 0 data during the level 1A processing. Indeed, the altitudes of the sample bins of a raw data proﬁle are recalculated with more accurate information about the satellite altitude and laser viewing angle in the data processing on ground. A shift for a few range bins (no more than three in most of the cases) can be needed for the full- resolution (30 m) samples. The number of 30 m bins shifted Nshift30 (which we express below in terms of an equivalent number of 15 m bins shifted; Nshift15) only add correlation to regions in the proﬁle data where the vertical range resolution is coarser than 30 m, i.e., where the vertical range resolutions are 60, 180 and 300 m (Winker et al., 2006). Indeed, in those regions, a vertical shift by Nshift30 30 m bins led to the ne- cessity of rebinning two neighboring bins larger than 30 m which introduce additional correlation to those bins. When there is a shift of Nshift15 15 m bins (an even number since shifts are performed at 30 m resolution), each new shifted bin B′ k, with vertical resolution coarser than 30 m, is computed from the weighted average of the two original bins (with Nbin size resolution) it steps across (Bk and Bk+1; Fig. A1) follow- ing A correction should be applied to Eq. (3) due to the fact that the nominal sample range interval (15 m) of the lidar is smaller than its range resolution (≈40 m) determined by the electronic bandwidth (2 MHz; Hunt et al., 2009). Con- sequently, a 15 m sample bin is partially correlated with the two bins above and the two bins below. As a result, vertical averaging of several 15 m bins Nbin does not reduce the noise standard deviation as much as it would if the samples were independent. A function fa(Nbin) > 1 is then applied to cor- rect from this partial correlation. This function is evaluated as follows. A 15 m bin bi has a variance Var(bi). T. Vaillant de Guélis et al.: 2D-McDA for CALIOP The full detection of those layers will lead to more accurate aerosol optical depth retrievals which will improve es- timates of the radiation budget. Proﬁling the full depth of the smoke layer will also help to understand whether the layer is in contact with underlying clouds and able to affect cloud microphysics. – The 2D-McDA can retrieve the full vertical extent of dense smoke layers by examining the 1064 nm channel. Within smoke, the 1064 nm signals are attenuated sig- niﬁcantly less than at 532 nm and hence can more often penetrate the full vertical extent of these layers. Those biomass burning aerosols play a signiﬁcant role in the Earth’s radiative balance by their scattering and absorp- tion of incoming solar radiation (e.g., Penner et al., 1992; Christopher et al., 1996) and the interaction they have with clouds (e.g., Kaufman and Fraser, 1997). The full detection of those layers will lead to more accurate aerosol optical depth retrievals which will improve es- timates of the radiation budget. Proﬁling the full depth of the smoke layer will also help to understand whether the layer is in contact with underlying clouds and able to affect cloud microphysics. While the current implementation of the algorithm is compu- tationally intensive, numerous optimizations are underway, and it is now feasible to apply the 2D-McDA operationally using CALIPSO’s available computer resources. However, while fundamentally important, feature detection is only the ﬁrst step in extracting a comprehensive suite of geophysical parameters from raw lidar measurements. Taking full advan- tage of the improved spatial analyses delivered by the 2D- McDA thus requires the development of a companion set of 2-D scene processes to replace the 1-D proﬁle-based pro- cesses that are currently used in the CALIPSO retrieval ar- chitecture to perform the essential tasks of discriminating be- tween clouds and aerosols, identifying cloud thermodynamic phase and classifying aerosols by type. Atmos. Meas. Tech., 14, 1593–1613, 2021 1609 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 1610 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP Figure A1. Scheme of redistribution in altitude registration. The variance of a B′ k can be written as Var(B′ k) = Var Nbin −Nshift15 Nbin Bk + Nshift15 Nbin Bk+1 = Nbin −Nshift15 Nbin 2 Var(Bk) + Nshift15 Nbin 2 Var(Bk+1) + 2Nbin −Nshift15 Nbin Nshift15 Nbin Cov(Bk,Bk+1) = "Nbin −Nshift15 Nbin 2 + Nshift15 Nbin 2# Var(b) Nbin fa(Nbin)2 + 2Nbin −Nshift15 Nbin Nshift15 Nbin Var(b) N2 bin Nbin X i,j=1 Cov(bi,bNbin+j) Var(b) = Var(b) Nbin "Nbin −Nshift15 Nbin 2 + Nshift15 Nbin 2# fa(Nbin)2 + 2Nbin −Nshift15 Nbin Nshift15 Nbin 1 Nbin Nbin X i,j=1 R(Nbin + j −i) = Var(b) Nbin "Nbin −Nshift15 Nbin 2 + Nshift15 Nbin 2# fa(Nbin)2 + 2Nbin −Nshift15 Nbin Nshift15 Nbin Nbin X m=1 m Nbin R(m) + Nbin−1 X m=1 Nbin −m Nbin R(Nbin + m) . (A7) Var(B′ k) Var(B′ k) Var(Bk+1) + 2Nbin −Nshift15 Nbin Nshift15 Nbin Cov(Bk,Bk+1) Nbin Figure A1. Scheme of redistribution in altitude registration. d zsurf given by a digital elevation model (DEM). The width of this region will vary according to surface type. Since we are highly conﬁdent of the surface altitude over the ocean (where d zsurf = 0), we will only search in a very narrow range of al- titudes for proﬁles measured over the ocean. On the other hand, we are somewhat less conﬁdent of the DEM surface altitudes reported over land and even much less conﬁdent over Greenland and Antarctica, so our search regions over land will be larger. The surface echo can be very weak due to attenuation by aerosol and cloud layers above. Then, we try to detect even the weakest surface echo as long as it is sub- stantially above background noise. This procedure is applied at single-shot resolution only. For each shot, the method is made up of the following steps: (A7) It follows that the correction function to apply on the stan- dard deviation to take into account both the vertical partial correlation due to the electronic bandwidth and the redistri- bution in altitude registration is 1. Compute d rsurf, the estimated range of the surface, from d zsurf and the satellite altitude zsat. fcorr(Nbin,Nshift15) fcorr(Nbin,Nshift15) 2. Compute 1β′ b(d rsurf), the standard deviation due to back- ground noise in the β′ domain at the range d rsurf. T. Vaillant de Guélis et al.: 2D-McDA for CALIOP fcorr(Nbin,Nshift15) = "Nbin −Nshift15 Nbin 2 + Nshift15 Nbin 2# fa(Nbin)2 + 2Nbin −Nshift15 Nbin Nshift15 Nbin Nbin X m=1 m Nbin R(m) + Nbin−1 X m=1 Nbin −m Nbin R(Nbin + m) 1/2 . (A8) 3. Compute id surf, the bin index of the estimated surface al- titude, i.e., when z(id surf) is closest to d zsurf. 4. Deﬁne id surf ± 1i, the range of the surface search region according to the International Geosphere-Biosphere Programme (IGBP) classiﬁcation of the surface type at the lidar footprint: (A8) a. If surface type is Water and d zsurf = 0, then 1i = 2(≡60m). A1 Correction due to electronic bandwidth (3) to take into account the vertical partial correlation due to the electronic bandwidth is where R(m) = Cov(bi,bi+m) Var(b) is the autocorrelation coefﬁcient for a lag of m range bins. It follows that the correction func- tion to apply on the total noise standard deviation in Eq. (3) to take into account the vertical partial correlation due to the electronic bandwidth is fa(Nbin) = 1 + 2 Nbin−1 X m=1 Nbin −m Nbin R(m) !1/2 . (A3) (A3) https://doi.org/10.5194/amt-14-1593-2021 https://doi.org/10.5194/amt-14-1593-2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 Appendix B: Surface detection The aim of this procedure is to detect a surface echo in the near-neighborhood region of the estimated surface altitude b. Otherwise, if surface type is Permanent-Snow, then 1i = 17(≡510m). Atmos. Meas. Tech., 14, 1593–1613, 2021 https://doi.org/10.5194/amt-14-1593-2021 https://doi.org/10.5194/amt-14-1593-2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 1611 Atmos. Meas. Tech., 14, 1593–1613, 2021 https://doi.org/10.5194/amt-14-1593-2021 Competing interests. The authors declare that they have no conﬂict of interest. Competing interests. The authors declare that they have no conﬂict of interest. Herzfeld, U. C., McDonald, B. W., Wallin, B. F., Neumann, T. A., Markus, T., Brenner, A., and Field, C.: Algorithm for Detection of Ground and Canopy Cover in Micropulse Photon-Counting Lidar Altimeter Data in Preparation for the ICESat-2 Mission, IEEE T. Geosci. Remote, 52, 2109–2125, https://doi.org/10.1109/TGRS.2013.2258350, 2014. Acknowledgements. The authors are grateful to Brian Magill for helping improving the runtime of the algorithm and to Kenneth Beaumont and Brian Getzewich for running the algorithm on the cluster of the Atmospheric Science Data Center. Thibault Vaillant de Guélis’ research was supported by an appointment to the NASA Postdoctoral Program at the NASA Langley Research Center, ad- ministered by Universities Space Research Association under con- tract with NASA. Hostetler, C. A., Liu, Z., Reagan, J., Vaughan, M., Winker, D., Os- born, M., Hunt, W. H., Powell, K. A., and Trepte, C.: Lidar Level I ATBD. Calibration and Level 1 Data Products, Algorithm The- oretical Basis Document, NASA Langley Research Document PC-SCI-201, 66 pp., available at: https://www-calipso.larc.nasa. gov/resources/pdfs/PC-SCI-201v1.0.pdf (last access: 5 February 2021), 2006. Review statement. This paper was edited by Ulla Wandinger and reviewed by two anonymous referees. Huang, J., Fu, Q., Su, J., Tang, Q., Minnis, P., Hu, Y., Yi, Y., and Zhao, Q.: Taklimakan dust aerosol radiative heating derived from CALIPSO observations using the Fu-Liou radiation model with CERES constraints, Atmos. Chem. Phys., 9, 4011–4021, https://doi.org/10.5194/acp-9-4011-2009, 2009. T. Vaillant de Guélis et al.: 2D-McDA for CALIOP T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 1612 Data availability. The CALIPSO level 1 lidar proﬁle prod- uct used throughout this study is publicly distributed by the NASA Langley Research Center Atmospheric Science Data Center: https://www-calipso.larc.nasa.gov/products/ CALIPSO_DPC_Rev4x92.pdf (last access: 5 February 2021, https://doi.org/10.5067/CALIOP/CALIPSO/LID_L1-Standard-V4- 10, Vaughan et al., 2020). tion of Hydrometeor Returns in Micropulse Lidar Data, J. Atmos. Ocean. Tech., 15, 1035–1042, https://doi.org/10.1175/1520- 0426(1998)015<1035:AAAFDO>2.0.CO;2, 1998. Davis, K. J., Gamage, N., Hagelberg, C. R., Kiemle, C., Lenschow, D. H., and Sullivan, P. P.: An Ob- jective Method for Deriving Atmospheric Structure from Airborne Lidar Observations, J. Atmos. Ocean. Tech., 17, 1455–1468, https://doi.org/10.1175/1520- 0426(2000)017<1455:AOMFDA>2.0.CO;2, 2000. Hagihara, Y., Okamoto, H., and Yoshida, R.: Development of a combined CloudSat-CALIPSO cloud mask to show global cloud distribution, J. Geophys. Res.-Atmos., 115, D00H33, https://doi.org/10.1029/2009JD012344, 2010. Author contributions. TVdG led the design of 2D-McDA and wrote the manuscript. MAV and DMW provided the knowledge of the weaknesses of the previous detection algorithm and brought the main ideas to address them. ZL developed the correction function and contributed to discussion and feedback essential to the study. Hartmann, D. L.: Tropical anvil clouds and climate sen- sitivity, P. Natl. Acad. Sci. USA, 113, 8897–8899, https://doi.org/10.1073/pnas.1610455113, 2016. Hartmann, D. L., Holton, J. R., and Fu, Q.: The heat balance of the tropical tropopause, cirrus, and strato- spheric dehydration, Geophys. Res. Lett., 28, 1969–1972, https://doi.org/10.1029/2000GL012833, 2001. T. Vaillant de Guélis et al.: 2D-McDA for CALIOP c. Otherwise, 1i = 5 (≡150 m). c. Otherwise, 1i = 5 (≡150 m). c. Otherwise, 1i = 5 (≡150 m). c. Otherwise, 1i = 5 (≡150 m). 5. Compute the derivatives of the lidar signal for bin index range 289–578 (8.2 to −0.5 km): dβ′ dz i = β′ i −β′ i−1 zi −zi−1 . (B1) (B1) 6. Determine zmin and zmax, the altitudes of the minimum and maximum values of the derivatives in the surface search region, and imin and imax, their respective bin in- dices. 7. Determine β′ max, the maximum signal magnitude lying between zmin and zmax. 8. Sequentially test the three following rules to determine if we have identiﬁed a legitimate surface return: 8. Sequentially test the three following rules to determine if we have identiﬁed a legitimate surface return: a. zmin > zmax; a. zmin > zmax; a. zmin > zmax; b. imin−imax ≤N with N = 2 for the 532 nm channels and N = 4 for the 1064 nm channel; b. imin−imax ≤N with N = 2 for the 532 nm channels and N = 4 for the 1064 nm channel; c. β′ max > 31β′ b(d rsurf). c. β′ max > 31β′ b(d rsurf). 9. If all rules are passed, set surface bin index isurf follow- ing these conditions: a. if dβ′ dz imin−1 > 0 or β′ imin−1 ≤0, then isurf = imin; b. otherwise (i.e., dβ′ dz imin−1 ≤0 and β′ imin−1 > 0), isurf = imin−1 for the 532 nm channels and isurf = imin−2 for the 1064 nm channel. 10. If surface detection arose in a proﬁle (proﬁle horizontal index p) but not in the previous (p−1) and the following (p+1) proﬁles, then the surface detection is canceled if isurf ̸∈id surf ±1. This last step reduces the surface search region for isolated surface detection to prevent false de- tection in a very attenuated region. Atmos. Meas. Tech., 14, 1593–1613, 2021 https://doi.org/10.5194/amt-14-1593-2021 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP 1613 noise in backscatter lidar observations, Appl. Optics, 45, 4437– 4447, https://doi.org/10.1364/AO.45.004437, 2006. noise in backscatter lidar observations, Appl. Optics, 45, 4437– 4447, https://doi.org/10.1364/AO.45.004437, 2006. FeatureMask, KNMI, Algorithm Theoretical Basis Document, 37 pp., available at: https://knmi.nl/research/1/publications/ algorithm-theoretical-basis-document-atbd-for-a-featuremask-atlas (last access: 5 February 2021), 2011. FeatureMask, KNMI, Algorithm Theoretical Basis Document, 37 pp., available at: https://knmi.nl/research/1/publications/ algorithm-theoretical-basis-document-atbd-for-a-featuremask-atlas (last access: 5 February 2021), 2011. Lu, X., Hu, Y., Liu, Z., Rodier, S., Vaughan, M., Lucker, P., Trepte, C., and Pelon, J.: Observations of Arctic snow and sea ice cover from CALIOP lidar measurements, Remote Sens. Environ., 194, 248–263, https://doi.org/10.1016/j.rse.2017.03.046, 2017. Vaughan, M., Pitts, M., Trepte, C., Winker, D., Detweiler, P., Garnier, A., Getzewich, B., Hunt, W., Lambeth, J., Lee, K.-P., Lucker, P., Murray, T., Rodier, S., Trémas, T., Bazureau, A., and Pelon, J.: CALIPSO Lidar Level 1B proﬁle data, V4-10, Cloud-Aerosol LIDAR Infrared Pathﬁnder Satellite Observations (CALIPSO) data manage- ment system data products catalog, NASA Langley Re- search Center, Document No. PC-SCI-503, Release 4.92, 256 pp., https://doi.org/10.5067/CALIOP/CALIPSO/LID_L1- Standard-V4-10, 2020 (data available at: https://www-calipso. larc.nasa.gov/products/CALIPSO_DPC_Rev4x92.pdf, last ac- cess: 5 February 2021). Lu, X., Hu, Y., Vaughan, M., Rodier, S., Trepte, C., Lucker, P., and Omar, A.: New attenuated backscatter proﬁle by removing the CALIOP receiver’s transient response, J. Quant. Spectrosc. Ra., 255, 107244, https://doi.org/10.1016/j.jqsrt.2020.107244, 2020. Luo, B. P., Peter, T., Fueglistaler, S., Wernli, H., Wirth, M., Kiemle, C., Flentje, H., Yushkov, V. A., Khattatov, V., and Rudakov, V.: Dehydration potential of ultrathin clouds at the tropical tropopause, Geophys. Res. Lett., 30, 1557, https://doi.org/10.1029/2002GL016737, 2003. McFarquhar, G. M., Heymsﬁeld, A. J., Spinhirne, J., and Hart, B.: Thin and subvisual tropopause tropi- cal cirrus: Observations and radiative impacts, J. At- mos. Sci., 57, 1841–1853, https://doi.org/10.1175/1520- 0469(2000)057<1841:TASTTC>2.0.CO;2, 2000. Vaughan, M. A., Winker, D. M., and Powell, K. A.: Lidar Level II ATBD. Part 2: Feature Detection and Layer Proper- ties Algorithms, Algorithm Theoretical Basis Document, NASA Langley Research Center Document PC-SCI-202.02, 87 pp., available at: https://www-calipso.larc.nasa.gov/resources/pdfs/ PC-SCI-202_Part2_rev1x01.pdf (last access: 5 February 2021), 2005. McGill, M. J., Vaughan, M. A., Trepte, C. R., Hart, W. D., Hlavka, D. L., Winker, D. M., and Kuehn, R.: Airborne validation of spa- tial properties measured by the CALIPSO lidar, J. Geophys. Res., 112, D20201, https://doi.org/10.1029/2007JD008768, 2007. Vaughan, M. A., Powell, K. A., Winker, D. M., Hostetler, C. A., Kuehn, R. E., Hunt, W. H., Getzewich, B. J., Young, S. A., Liu, Z., and McGill, M. T. Vaillant de Guélis et al.: 2D-McDA for CALIOP J.: Fully Automated De- tection of Cloud and Aerosol Layers in the CALIPSO Li- dar Measurements, J. Atmos. Ocean. Tech., 26, 2034–2050, https://doi.org/10.1175/2009JTECHA1228.1, 2009. Pal, S. R., Steinbrecht, W., and Carswell, A. I.: Auto- mated method for lidar determination of cloud-base height and vertical extent, Appl. Optics, 31, 1488–1494, https://doi.org/10.1364/AO.31.001488, 1992. Penner, J. E., Dickinson, R. E., and O’Neill, C. A.: Effects of aerosol from biomass burning on the global radiation budget, Science, 256, 1432–1434, https://doi.org/10.1126/science.256.5062.1432, 1992. Vaughan, M. A., Lee, K.-P., Garnier, A., Vaillant de Guélis, T., Getzewich, B. J., and Pelon, J.: Surface detection algorithm for space-based lidars, Atmos. Meas. Tech., in preparation, 2021. Shonk, J. K. P. and Hogan, R. J.: Effect of improving representation of horizontal and vertical cloud structure on the Earth’s global radiation budget. Part II: The global effects, Q. J. Roy. Meteor. Soc., 136, 1205–1215, https://doi.org/10.1002/qj.646, 2010. Wang, Z. and Sassen, K.: Cloud Type and Macrophysi- cal Property Retrieval Using Multiple Remote Sensors, J. Appl. Meteorol., 40, 1665–1682, https://doi.org/10.1175/1520- 0450(2001)040<1665:CTAMPR>2.0.CO;2, 2001. Winker, D. M. and Vaughan, M. A.: Vertical distribution of clouds over Hampton, Virginia observed by lidar under the ECLIPS and FIRE ETO programs, Atmos. Res., 34, 117–133, https://doi.org/10.1016/0169-8095(94)90084-1, 1994. Tsushima, Y., Emori, S., Ogura, T., Kimoto, M., Webb, M. J., Williams, K. D., Ringer, M. A., Soden, B. J., Li, B., and An- dronova, N.: Importance of the mixed-phase cloud distribution in the control climate for assessing the response of clouds to car- bon dioxide increase: a multi-model study, Clim. Dynam., 27, 113–126, 2006. Winker, D. M., Hostetler, C. A., Vaughan, M. A., and Omar, A. H.: CALIOP Instrument, and Algorithms Overview, Algorithm Theoretical Basis Document, NASA Langley Re- search Center Document PC-SCI-202.01, 29 pp., avalaible at: https://www-calipso.larc.nasa.gov/resources/pdfs/PC-SCI-202. Part1_v2-Overview.pdf (last access: 5 February 2021) 2006. Vaillant de Guélis, T., Chepfer, H., Noel, V., Guzman, R., Dubuis- son, P., Winker, D. M., and Kato, S.: The link between outgoing longwave radiation and the altitude at which a spaceborne lidar beam is fully attenuated, Atmos. Meas. Tech., 10, 4659–4685, https://doi.org/10.5194/amt-10-4659-2017, 2017a. Winker, D. M., Pelon, J., Coakley, J. A., Ackerman, S. A., Charl- son, R. J., Colarco, P. R., Flamant, P., Fu, Q., Hoff, R. M., Kittaka, C., Kubar, T. L., Le Treut, H., McCormick, M. P., Mégie, G., Poole, L., Powell, K., Trepte, C., Vaughan, M. A., and Wielicki, B. References Bevington, P. R. and Robinson, D. K.: Data Reduction and Error Analysis for the Physical Sciences, McGraw-Hill, New York, NY, USA, 3rd edn., 2003. Hunt, W. H., Winker, D. M., Vaughan, M. A., Powell, K. A., Lucker, P. L., and Weimer, C.: CALIPSO Lidar Description and Per- formance Assessment, J. Atmos. Ocean. Tech., 26, 1214–1228, https://doi.org/10.1175/2009JTECHA1223.1, 2009. Bony, S., Stevens, B., Coppin, D., Becker, T., Reed, K. A., Voigt, A., and Medeiros, B.: Thermodynamic control of anvil cloud amount, P. Natl. Acad. Sci. USA., 113, 8927–8932, https://doi.org/10.1073/pnas.1601472113, 2016. Jensen, E. J., Toon, O. B., Pﬁster, L., and Selkirk, H. B.: Dehy- dration of the upper troposphere and lower stratosphere by sub- visible cirrus clouds near the tropical tropopause, Geophys. Res. Lett., 23, 825–828, https://doi.org/10.1029/96GL00722, 1996. Brooks, I. M.: Finding Boundary Layer Top: Ap- plication of a Wavelet Covariance Transform to Lidar Backscatter Proﬁles, J. Atmos. Ocean. Tech., 20, 1092–1105, https://doi.org/10.1175/1520- 0426(2003)020<1092:FBLTAO>2.0.CO;2, 2003. Kaufman, Y. J. and Fraser, R. S.: The effect of smoke parti- cles on clouds and climate forcing, Science, 277, 1636–1639, https://doi.org/10.1126/science.277.5332.1636, 1997. Lewis, J. R., Campbell, J. R., Welton, E. J., Stewart, S. A., and Haftings, P. C.: Overview of MPLNET Version 3 Cloud Detection, J. Atmos. Ocean. Tech., 33, 2113–2134, https://doi.org/10.1175/JTECH-D-15-0190.1, 2016. Campbell, J. R., Sassen, K., and Welton, E. J.: Elevated Cloud and Aerosol Layer Retrievals from Micropulse Li- dar Signal Proﬁles, J. Atmos. Ocean. Tech., 25, 685–700, https://doi.org/10.1175/2007JTECHA1034.1, 2008. Liu, Z. and Sugimoto, N.: Simulation study for cloud detection with space lidars by use of analog detec- tion photomultiplier tubes, Appl. Opt., 41, 1750–1759, https://doi.org/10.1364/AO.41.001750, 2002. p g Christopher, S. A., Kliche, D. V., Chou, J., and Welch, R. M.: First estimates of the radiative forcing of aerosols generated from biomass burning using satellite data, J. Geophys. Res.-Atmos., 101, 21265–21273, https://doi.org/10.1029/96JD02161, 1996. Liu, Z., Hunt, W., Vaughan, M., Hostetler, C., McGill, M., Powell, K., Winker, D., and Hu, Y.: Estimating random errors due to shot Clothiaux, E. E., Mace, G. G., Ackerman, T. P., Kane, T. J., Spin- hirne, J. D., and Scott, V. S.: An Automated Algorithm for Detec- https://doi.org/10.5194/amt-14-1593-2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 Atmos. Meas. Tech., 14, 1593–1613, 2021 T. Vaillant de Guélis et al.: 2D-McDA for CALIOP A.: The CALIPSO Mission: A Global 3D View of Aerosols and Clouds, B. Am. Meteorol. Soc., 91, 1211–1229, https://doi.org/10.1175/2010BAMS3009.1, 2010. Vaillant de Guélis, T., Chepfer, H., Noel, V., Guzman, R., Winker, D. M., and Plougonven, R.: Using Space Lidar Ob- servations to Decompose Longwave Cloud Radiative Effect Variations Over the Last Decade: Space lidar decomposes LWCRE variations, Geophys. Res. Lett., 44, 11994–12003, https://doi.org/10.1002/2017GL074628, 2017b. Vaillant de Guélis, T., Chepfer, H., Guzman, R., Bonazzola, M., Winker, D. M., and Noel, V.: Space lidar observations constrain longwave cloud feedback, Sci. Rep.-UK, 8, 16570, https://doi.org/10.1038/s41598-018-34943-1, 2018. Young, S. A., Vaughan, M. A., Garnier, A., Tackett, J. L., Lambeth, J. D., and Powell, K. A.: Extinction and optical depth retrievals for CALIPSO’s Version 4 data release, Atmos. Meas. Tech., 11, 5701–5727, https://doi.org/10.5194/amt-11-5701-2018, 2018. van Zadelhoff, G.-J., Donovan, D. P., and Berthier, S.: ATLID Algorithms and Level 2 System Aspects: ATBD for A- Atmos. Meas. Tech., 14, 1593–1613, 2021 https://doi.org/10.5194/amt-14-1593-2021
https://openalex.org/W3176311105	https://hal.science/hal-03690148/document	English	null	Youth climate activism in the United States	E-REA	2,021	cc-by	15,006	Youth climate activism in the United States Mélanie Meunier Mélanie Meunier To cite this version: Mélanie Meunier. Youth climate activism in the United States. E-rea - Revue électronique d’études sur le monde anglophone, 2021, 1. ”Letters to Swift” / 2. “ État d’urgence environnemental : comprendre, agir, représenter ”, 18.2, 10.4000/erea.12490. hal-03690148 HAL Id: hal-03690148 https://hal.science/hal-03690148v1 Submitted on 7 Jun 2022 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. E-rea Revue électronique d’études sur le monde anglophone 18.2 \| 2021 1. "Letters to Swift" / 2. « État d’urgence environnemental : comprendre, agir, représenter » Publisher Brought to you by Université de Strasbourg Youth climate activism in the United States Melanie MEUNIER Electronic reference Electronic reference Melanie MEUNIER, “Youth climate activism in the United States”, E-rea [Online], 18.2 \| 2021, Online since 15 June 2021, connection on 21 November 2021. URL: http://journals.openedition.org/erea/ 12490 ; DOI: https://doi.org/10.4000/erea.12490 This text was automatically generated on 21 November 2021. E-rea est mis à disposition selon les termes de la licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International. Youth climate activism in the United States 1 Youth climate activism in the United States Youth climate activism in the United States Tilly and Tarrow explain in Contentious Politics that an upward scale shift, i.e. the coordination of collective action at a higher level (from local to regional, national, international), is one of the most important processes in the success of a movement (Tilly & Tarrow 95). The scale shift can be explained by mechanisms such as “diffusion” and “emulation” when activists learn of episodes of contention and derive inspiration from them, or “brokerage” whereby contention leaders bring formerly isolated groups into contact. Social media tools, a mainstay of most youths’ daily lives, enable individuals to meet and mobilize online. The pervasive use of social media is creating a cultural shift to the political arena (Gamber-Thompson 236). In addition, participating in horizontal, peer-led groups creates a strong sense of collective identity and of belonging to a community (Jenkins et al 3). These processes have spawned and developed numerous youth-led initiatives in the United States in parallel to movements in other countries (Thunberg’s Fridays for Future, Extinction Rebellion), which have mutually inspired one another. 4 In this context of contentious politics, the stage is set for an upward scale shift, as concerned youths, who have successfully channeled their fear and frustration into positive action, spread their enthusiasm to their peers. Tilly and Tarrow explain in Contentious Politics that an upward scale shift, i.e. the coordination of collective action at a higher level (from local to regional, national, international), is one of the most important processes in the success of a movement (Tilly & Tarrow 95). The scale shift can be explained by mechanisms such as “diffusion” and “emulation” when activists learn of episodes of contention and derive inspiration from them, or “brokerage” whereby contention leaders bring formerly isolated groups into contact. Social media tools, a mainstay of most youths’ daily lives, enable individuals to meet and mobilize online. The pervasive use of social media is creating a cultural shift to the political arena (Gamber-Thompson 236). In addition, participating in horizontal, peer-led groups creates a strong sense of collective identity and of belonging to a community (Jenkins et al 3). These processes have spawned and developed numerous youth-led initiatives in the United States in parallel to movements in other countries (Thunberg’s Fridays for Future, Extinction Rebellion), which have mutually inspired one another. 5 In order to better understand youth activism in the United States, a framework categorizing various means of action would be helpful. Youth climate activism in the United States Melanie MEUNIER “The recent attention on climate policy is not because members of the media or political establishment woke up one day. It is a direct result of the active energy and the demand from thousands of young people on the front lines of the crisis.” Varshini Prakash, Co-founder, Sunrise Movement 1 Climate change is often referred to as the defining crisis of the twenty-first century. The international and American scientific communities have published reports that confirm beyond doubt that climate change exists, is accelerating, and will have devastating effects if current rates of greenhouse gas emissions continue. North America being a continent of extreme weather events (Huret & Tourbe), the diverse regions of the United States experience hurricanes, wildfires, droughts, flooding, and peaks of severe heat and cold. Due to climate change, the intensity and frequency of these extreme weather events have increased in recent years (National Centers for Environmental Information). Mounting numbers of Americans have been directly affected (homes and livelihoods lost, communities devastated, etc.), which has raised the salience of the issue. In a 2019 US Conference of Mayors’ survey of 182 cities, 96% of respondents stated that they had experienced the impact of climate change (Durr), and a 2020 Yale Climate Change Communication report found that 51% of respondents think that global warming poses a risk to their community over the next ten years (Leiserowitz et al 17). 2 Young Americans take the issue much more seriously than the population as a whole: in a recent nationwide online survey of 1,000 youths aged 18 to 29 years old, 80% of the respondents felt that “global warming is a major threat to human life on earth as we know it” (Barnet et alt). Climate change also has a greater impact on young people for many reasons, including physical and mental health. As Judy Wu et al stated in The E-rea, 18.2 \| 2021 Youth climate activism in the United States 2 Lancet, “They are at a crucial point in their physical and psychological development, when enhanced vulnerability to the effects of stress and everyday anxiety elevate their risk of developing depression, anxiety, and substance use disorders” (Wu et al e435). Climate anxiety, or eco-anxiety, defined as “a chronic fear of environmental doom” (Clayton et al 27, Huizen) can result from the trauma of experiencing natural disasters firsthand or learning about them through written and video accounts. Youth climate activism in the United States Vasser Seydel, a young director of the Oxygen Project, remarked that the alternative to eco-anxiety and depression is action and empowerment: “If you can get yourself to start taking action, you can overcome the paralysis and the overwhelming nature of the climate emergency” (Climate Academy). 3 Compounding these youths’ anxiety is the observation that the United States government has often been reluctant to address the climate crisis: many anti- environmental policies under the G.W. Bush and Trump administrations, and in Republican-dominated Congresses have been enacted since 1994. Even as the crisis became more apparent, the Trump administration adopted an anti-environmental, climate-denying, anti-scientific stance. Thus, it rolled back the Obama administration’s climate mitigation efforts such as the Clean Power Plan, the refusal of the Keystone XL Pipeline Project and automotive emissions standards, and announced withdrawal from the Paris Agreement (Pitt et al). The urgency of the situation combined with federal inaction has led to the mobilization of scores of young people, many of whom are too young to vote. As Henry Jenkins et al stated in By Any Media Necessary: the New Youth Activism, there is a widespread perception that: On the one hand, the institutions historically associated with American democracy are dysfunctional, public trust in core institutions is eroding, civic organizations no longer bring us together, elected representatives are more beholden to big contributors than to voters... and very little is likely to emerge at the level of institutional politics that is going to shift conditions very much. On the other hand, we have seen an expansion of the communicative and organizational resources available to everyday people (and grassroots organizations). (Jenkins et al 3) On the one hand, the institutions historically associated with American democracy are dysfunctional, public trust in core institutions is eroding, civic organizations no longer bring us together, elected representatives are more beholden to big contributors than to voters... and very little is likely to emerge at the level of institutional politics that is going to shift conditions very much. On the other hand, we have seen an expansion of the communicative and organizational resources available to everyday people (and grassroots organizations). (Jenkins et al 3) 4 In this context of contentious politics, the stage is set for an upward scale shift, as concerned youths, who have successfully channeled their fear and frustration into positive action, spread their enthusiasm to their peers. Youth climate activism in the United States In a 2018 article titled “Exploring youth activism on climate change,” published in Ecology and Society, Karen E-rea, 18.2 \| 2021 Youth climate activism in the United States 3 O’Brien et al present a typology to account for the various ways in which youths are expressing dissent. Based largely on Western modes of activism, their study distinguishes three types of dissent: disruptive, dutiful, and dangerous. “Disruptive” dissent, according to O’Brien et al, is oppositional, where activists organize protests to contest prevailing policies or norms. “Dutiful” dissent is reformist and works within existing power and economic structures. It challenges the status quo but does not challenge existing decision-making processes. Membership in political and environmental organizations is the most common example of dutiful dissent in climate activism. While disruptive dissent focuses on the underlying causes of climate change, raises awareness and encourages participation of newcomers to fight against the system, dutiful dissent offers direct access to power holders and builds legitimacy within the system. The third type of dissent, “dangerous,” is propositional: activists propose new norms and values which communities embrace. The dangerous aspect is not because dissenters’ actions are violent, but rather, dangerous in the eyes of existing power holders, because the activists’ ideas subvert existing systems and pose a threat to their authority or practice (O’Brien et al). Reformist environmentalism, i.e. dutiful dissent, was the mainstay of the environmental movement from the 1970s to 2000, according to sociologist Robert J. Brulle. The “Big Ten” environmental organizations like the Sierra Club, the Wilderness Society, the National Wildlife Federation, and Friends of the Earth, opened offices in Washington D.C. and became part of the lobbying establishment (Andrews 238). Since the 1990s, however, environmental organizations have expended most of their energy in defending gains against neoconservatives’ attacks rather than achieving new advances (Brulle). Critics have found them to be too technocratic, too top down, and too entrenched in outmoded discourses. Indeed, the failures of the major environmental associations have been much documented (Brulle, Shellenberger & Nordhaus). Today’s youths have largely preferred to create their own organizations, “set the agenda and frame the message,” rather than to work through the institutionalized environmental associations (Jenkins et al 289). Youth climate activism in the United States 9 Her words speak for many young activists who have become involved in one of the many groups that have sprung up in the last few years, such as the Sunrise Movement (2017), the Youth Climate Strike (2018), the Down Ballot Disruption Project (2020), Zero Hour (2017), American Conservation Coalition (2017), Earth Uprising (2019), and many more. Not only are these youths fearful for their future because of climate change and frustrated with the inertia of the political class, they, for the most part, have solely experienced a society divided by inequality, endless wars, a financial collapse, and a student debt crisis. Faced with a combination of so many disruptive factors and inspired by politician Bernie Sanders and activists like Naomi Klein who criticize unbridled capitalism that put “profits over people” (Evelyn), many youths have concluded that old models are not working. Youths are forging new visions of their future. Spurred by a feeling of urgency and necessity, they are willing to fight for a “livable planet” (sunrisemovement.org). Observing the 2019 Youth Climate Marches in March and September, Fisher remarked, “Something different is happening here. We have a new wave of contention in society that’s being led by women…and the youth climate movement is leading this generational shift” (Kaplan). 10 The following pages will examine the diverse strategies the young generations are utilizing to fight for climate action. Analysis will primarily be based on the mainstream press coverage to assess the impact of youths’ actions. For the purposes of this paper, youths will be loosely defined as people under thirty years old. The majority of their tactics could be considered “disruptive”: mobilizing an “army of climate activists,” engaging in classic street rallies and sit-ins. Other strategies could be categorized as “dutiful” dissent because they work within existing power structures, like the electoral process and the courts. Third, a discussion of the “dangerous” aspects of youth activism that threaten power holders will conclude this analysis. 10 The following pages will examine the diverse strategies the young generations are utilizing to fight for climate action. Analysis will primarily be based on the mainstream press coverage to assess the impact of youths’ actions. For the purposes of this paper, youths will be loosely defined as people under thirty years old. The majority of their tactics could be considered “disruptive”: mobilizing an “army of climate activists,” engaging in classic street rallies and sit-ins. Youth climate activism in the United States Given that many founders of these movements are teenagers – most of whom do not have the right to vote – the multitudes of loosely organized groups of youths that have sprung up in recent years belong principally to the classic model of street marches, sit- ins and protests. Disruptive dissent corresponds to the anger and frustration with lawmakers’ lack of action, and is the easiest way for relatively unorganized, little- financed groups of youths to express themselves. As sociologist Dana Fisher remarked to a PBS journalist during the September 2019 Youth Climate March, “In some ways, the school strike for climate change is what the sit-in was for the civil rights movement. It is a tactic that is doable for people in a very local way, where they can get involved in a movement” (Fisher). Given that many founders of these movements are teenagers – most of whom do not have the right to vote – the multitudes of loosely organized groups of youths that have sprung up in recent years belong principally to the classic model of street marches, sit- ins and protests. Disruptive dissent corresponds to the anger and frustration with lawmakers’ lack of action, and is the easiest way for relatively unorganized, little- financed groups of youths to express themselves. As sociologist Dana Fisher remarked to a PBS journalist during the September 2019 Youth Climate March, “In some ways, the school strike for climate change is what the sit-in was for the civil rights movement. It is a tactic that is doable for people in a very local way, where they can get involved in a movement” (Fisher). 8 While protests may raise public awareness about climate change, the political elites and establishment generally ignore or discount such efforts (O’Brien et al). Jamie Margolin, who was 15 years old in 2017 when she founded the activist group Zero Hour, admitted that the threat of climate change had been a constant fixture in her life, and that the lack of government action amidst all the disasters that were happening during the summer of 2017 (Hurricane Maria in Puerto Rico, wildfire smoke from Canada blanketing her city of Seattle, Trump announcing the withdrawal from the Paris Agreement) triggered her activism. Youth climate activism in the United States Interviewed on CBS News, she expressed her E-rea, 18.2 \| 2021 Youth climate activism in the United States frustration with the failure of older people to take youths seriously, to view the climate crisis like an emergency and act in consequence: frustration with the failure of older people to take youths seriously, to view the climate crisis like an emergency and act in consequence: Enough is enough. They’ll just pat us on the head and say...maybe one day you should run for office and you can take action on climate change. By the time I grow up, it’ll be too late! Right now, the people who are holding power – whether we like it or not – they’re the ones in this tiny, tiny window. The door is closing and we have this tiny window to take action before it’s too late. (CBS News 19 Sept. 2019) 9 Her words speak for many young activists who have become involved in one of the many groups that have sprung up in the last few years, such as the Sunrise Movement (2017), the Youth Climate Strike (2018), the Down Ballot Disruption Project (2020), Zero Hour (2017), American Conservation Coalition (2017), Earth Uprising (2019), and many more. Not only are these youths fearful for their future because of climate change and frustrated with the inertia of the political class, they, for the most part, have solely experienced a society divided by inequality, endless wars, a financial collapse, and a student debt crisis. Faced with a combination of so many disruptive factors and inspired by politician Bernie Sanders and activists like Naomi Klein who criticize unbridled capitalism that put “profits over people” (Evelyn), many youths have concluded that old models are not working. Youths are forging new visions of their future. Spurred by a feeling of urgency and necessity, they are willing to fight for a “livable planet” (sunrisemovement.org). Observing the 2019 Youth Climate Marches in March and September, Fisher remarked, “Something different is happening here. We have a new wave of contention in society that’s being led by women…and the youth climate movement is leading this generational shift” (Kaplan). Youth climate activism in the United States Other strategies could be categorized as “dutiful” dissent because they work within existing power structures, like the electoral process and the courts. Third, a discussion of the “dangerous” aspects of youth activism that threaten power holders will conclude this analysis. 1.1 Fossil Fuel Divestment 12 The Fossil Fuel Divestment campaign began in 2011 on several university campuses and has maintained a significant presence on campuses ever since. Students demonstrate to put pressure on their university boards to divest from any holdings in coal and other fossil fuels. Amplified by climate advocacy group 350.org’s efforts internationally, the movement grew rapidly, encompassing 181 universities and other institutions by September 2014, adding hundreds more each year. Today, after ten years, the divestment movement has spread from American college campuses to many other countries and many other institutions. As of April 2021, the latest divestment figures comprise ten categories of institutions, out of which faith-based, philanthropic and educational ones constitute 64% of 1,319 institutions and a total of $14.56 trillion, according to the website gofossilfree.org. 13 University campuses are prime sites where the “diffusion” mechanism of contentious politics works particularly well to spread ideas from one campus to another. They have long been sites of contention, from the 1960s antiwar and other protests to the shanties against the South African apartheid movement in the 1980s (Tilly & Tarrow 38-41). The fossil fuel divestment movement carried the idea forward to 2011 as a means to raise awareness about climate change, with the potential of impacting the fossil fuel industry. 14 To mention a few examples, at Harvard University, student protesters have staged protests over several years, so far without success. The youth-led movement has inspired adult support, thereby expanding its influence. After it became public in 2015 that Harvard had considerably increased its investments in fossil fuel companies, support for divestment flourished amongst alumni, as well as faculty members whose numbers jumped from 100 to 250 professors. In 2019, led by Divest Harvard and Fossil Free Yale, a crowd of protesters took over the football field at the halftime of a Harvard-Yale game to raise awareness about the issue (O’Daly). Harvard Forward, a campaign created by concerned alumni, supported three pro-divestment candidates for the Harvard Board of Overseers in 2020. The website of the Harvard Forward platform states: In the 1980s, alumni successfully ran petition candidates for the Harvard Board of Overseers to push Harvard to divest from apartheid-era South Africa and become more socially responsible. Over three decades later, alumni still want the same things from Harvard: leadership & accountability in facing the biggest challenges of our time. 1. “Disruptive” Dissent: working from outside the system 11 These youth movements rely on their energy, enthusiasm and the sheer number of peers they are able to mobilize and who take part in street marches and peaceful protests, targeting particular officials or institutions. In this section, we will present some of the movements and profile a few of their leaders, both from the progressive and the conservative sides of the political spectrum: specifically, on the left, the Divestment movement, and U.S. Youth Climate Strikes, and on the right, the American Conservation Coalition, and Young Evangelicals for Climate Change. E-rea, 18.2 \| 2021 Youth climate activism in the United States 5 1.2 U.S. Youth Climate Strike 18 Secondly, inspired by Swedish youth activist Greta Thunberg’s Fridays-for-Future student strikes initiated in 2018, three American teenagers who had been striking on their own in their respective cities joined forces to organize the first countrywide youth climate strike in March 2019. The executive director of the U.S. Youth Climate Strike is a sixteen-year-old (in 2018) named Isra Hirsi from Minneapolis, Minnesota. She became an activist due to the drinking water crisis in Flint, Michigan, which primarily affected poor communities of color. Her identity as a Black Muslim woman drove her to seek environmental justice. Twelve-year-old (in 2018) co-founder Haven Coleman from Denver, Colorado was inspired by a schoolteacher who taught about deforestation which, she realized, was putting her favorite animal, the sloth, in danger. She started striking alone but sought contact with other strikers, which deepened her commitment to climate action. The third co-founder is Alexandria Villaseñor, a thirteen-year-old (in 2018) who had been striking alone in front of the United Nations since December 2018. A native of California, her activism was triggered by the simultaneous events in November 2018 of the Camp Fire, a wildfire which completely razed the town of Paradise, California, and of Greta Thunberg’s speech at COP 24, the United Nations’ annual meeting on climate change, which took place in Poland. In order to motivate more young people, she founded a new group, Earth Uprising, which focuses on educating youths in schools, peer-to-peer (The Climate Group). 9 All three noted that staying connected through social media with likeminded activists was crucial in sustaining their motivation. The communication and joining of these three youths to found the U.S. Youth Climate Strike is a classic example of coordinated action according to the framework proposed by Tilly and Tarrow (Tilly & Tarrow 31). As momentum built, eight U.S.-based youth-led climate groups formed an alliance, the U.S. Youth Climate Strike Coalition. The groups worked together to plan the American marches organizing more than 1,000 events in all fifty states, Puerto Rico, and Washington, D.C. The third global Youth Climate Strike, with multiple events planned on the week of 20-27 September 2019, was timed to coincide with the U.N. Youth Climate Summit and the U.N. Climate Action Summit in New York City. The latter was to be attended by most major world leaders and Greta Thunberg, the “poster girl” of the youth climate movement, who would address them (Sky News). 1.1 Fossil Fuel Divestment (harvardforward.org/) 15 All three candidates were elected on a platform of climate action and social justice, with fossil fuel divestment at the center of their campaign, which indicates broad support for divestment. On the West coast, Stanford University divested its $18.7- billion endowment in 2014, which was a major victory for the youth movement. More recently, the University of California announced in May 2020 that it had fully divested from fossil fuels (Watanabe). The University of California is the largest American university with 285,000 students and a portfolio of $2.6 billion. 16 As important as these divestments may be for the movement, observers assert that a tarnished image of fossil fuel companies could have even more damaging consequences to them (Atif et al). The exposure of ExxonMobil’s knowledge of the impact of fossil fuel emissions on human health and the company’s efforts to manufacture doubt about E-rea, 18.2 \| 2021 Youth climate activism in the United States 6 global warming in direct contradiction to the results of its studies, is a case in point (Banerjee et al). As Bill McKibben stated in 2015, after years of domination of the energy sector, the fossil fuel industry is starting to receive some serious setbacks: Keystone XL pipeline, multiple lawsuits, and with the Biden administration, the probable loss of fossil fuel subsidies. The current administration’s $2.3 trillion infrastructure plan includes a proposal to turn those subsidies into incentives for renewable energy (McKibben 6 Nov 2015, Gardner). 17 The movement’s ability to sustain itself despite the high turnover (as graduating students leave) has made the divestment movement a model for youths aspiring to start their own movements. 1.3 American Conservative Coalition 20 On the conservative side of the political spectrum, a Pew Research survey conducted in the spring of 2020 showed that 79% of millennial and younger Republicans believed the U.S. should invest in clean energy sources. It also showed that nearly twice as many millennial and younger Republicans as older Republicans believe human activity contributes a great deal to climate change (Fry). 21 Republican-leaning conservatives like Benji Backer, (aged 19 in 2017), from Appleton, Wisconsin founded the American Conservation Coalition in 2017. Backer declared that the health of the environment affects everyone, regardless of where they live, their background or political affiliation (60 Minutes). Along with Greta Thunberg and Jamie Margolin and other youths, Benji Backer testified on Capitol Hill in September 2019, pressing elected officials to take serious action on climate change. The coalition’s website states that it is “dedicated to mobilizing young people around environmental action through common-sense, market-based, and limited-government ideals.” It seeks to overcome partisan inaction by “promoting a mix of free-market, pro-business, and limited-government environmentalism in legislatures, college campuses, the political arena, and beyond.” As of May 2021, the coalition counts 58 branches throughout the United States (American Conservation Coalition). 22 The American Conservative Coalition produced an alternative to the Green New Deal in 2020, called the American Climate Contract. Its FAQ page states that the conservative plan focuses on action now, using “carrots, not sticks” and that the Green New Deal is “a wish list of progressive policies.” The website announces “an unprecedented coalition of experts and activists, liberals and conservatives, environmental organizations and energy groups,” auguring a nonpartisan approach (American Conservation Coalition, acc.eco). 1.2 U.S. Youth Climate Strike Turnout across the United States and the world for the global youth climate strike on 15 March 2019 was much higher than at previous marches and skyrocketed on 20 September 2019. It is estimated that nearly two million attended in March and between 6 and 7.6 million on E-rea, 18.2 \| 2021 Youth climate activism in the United States 7 20 September (Climate Academy). Impressed by the numbers of young people at these marches, Fisher conducted a survey and found that half of the people who attended the 20 September protest were of college age or younger, and fully 25% was under eighteen years of age (Kaplan). 1.4 Young Evangelicals for Climate Action However, initiatives of the executive branch not enshrined in law by Congress can easily be overturned by the president’s successor, which is exactly what President Trump did upon arriving in office (Pitt et al). Furthermore, the Trump administration proved to be particularly inattentive to voices of opposition. Clearly, taking to the streets, even en masse, is insufficient to effect lasting change. It will take more to move an entire country like the United States away from fossil fuels, especially in the current contentious political and social climate. Some young activists have concluded that, rather than simply expressing opposition to the current state of affairs, working within the system could produce more results. 1.4 Young Evangelicals for Climate Action 23 The American Conservation Coalition is silent on religious matters, but studies find a positive correlation between conservatism and religiosity. Indeed, Pew Research surveys consistently find that evangelicals are some of the staunchest skeptics of climate change science. Young evangelicals, however, support action to mitigate climate change (Meyaard-Schaap). According to a November 2019 poll on Americans’ views about climate and energy, 52% of young Republicans responded that the government is doing too little to address climate change (Funk & Hefferon). Reverend Kyle Meyaard-Schaap, a spokesperson for Young Evangelicals for Climate Change who crisscrosses the country talking to members, believes that a “tectonic shift” is taking place in American churches. He no longer hears debates about the science of climate change amongst the young generations. Rather, like the vast majority of climate activists, they are mobilizing on college campuses, marching in the streets, meeting E-rea, 18.2 \| 2021 Youth climate activism in the United States 8 elected officials and "speaking up for the people's right to clean air and water and a stable climate" (Meyaard-Schaap). 24 These movements are examples of an upward scale shift as youths turn their energy into action, emulating earlier actors and inspiring later ones. They demonstrate the depth and breadth of youth climate activism which, judging by the numbers cited above of participation at the youth climate strikes in March and September 2019 (Climate Interactive), have increased significantly in the last few years. Mass mobilization and widespread protests are key in raising awareness and influencing public opinion and have been quite successful in the divestment movement. Yet, pragmatists question the effectiveness of mass protests: all this action and passion is exciting and seems promising to many, but have these actions resulted in real change? In a 2017 New Yorker article, Nathan Heller wonders, “Is there any point to protesting?” (Heller). In his review of recent protests – against the war in Iraq in 2003, Occupy Wall Street in 2011, Black Lives Matter in 2014, Women’s March in 2017 – he finds that none of them had concrete success. 25 So what have the climate protests accomplished? Has the United States taken steps to reduce its emissions? The Obama administration took significant steps near the end of its term: the Clean Power Plan, signing the Paris Agreement, refusal of the Keystone XL Pipeline and of the Dakota Access pipeline (PBS 2017). 2. “Dutiful” Dissent: Working from within, through the electoral system 26 Voting is a primary tool in a democracy: if enough elected officials in Congress vote for a climate action plan, a comprehensive program will finally be passed after years of gridlock. Those who are too young to vote can still influence the results of votes by putting pressure on elected officials in other ways than classic rallies. In this section, we will examine two other types of initiatives that youth groups are implementing to force a change in climate policy. One focuses on the electoral process while the other targets the courts. E-rea, 18.2 \| 2021 2.1 The Sunrise Movement Social media presence – informing, educating, motivating followers – is an integral part of the Sunrise Movement’s communication strategy, both to enhance cohesion amongst those who are members and to attract new activists. Innovative young volunteers have devised a new style of digital campaigning, creating “stan” accounts to publicize campaign messages through memes when door-to-door canvassing became tricky, if not impossible, during the Coronavirus pandemic (Higgins and Chávez). 29 In addition to demonstrating, organizing sit-ins, handing out flyers door-to-door and the like, Sunrise developed many tools to galvanize and train an “army” of youth activists. Their website offers training programs so that newcomers can quickly learn the basics of becoming an activist. Getting few mentions in the press, these training retreats of a day up to a week are fundamental in helping young, motivated people gain confidence and skills in order to become key actors in their respective local areas. Now under the heading of “Sunrise School,” the website proposes both online support and in-person training programs for newcomers and trainers. The level of support offered to volunteers appears to be well above that on other websites and contributes to the movement’s success. According to a Sunrise volunteer, Ben Bristol, teenagers strongly desire to be part of the solution to the climate crisis: Kids are already suspicious of the establishment. They have strong opinions and rebellious streaks. When you’re a high schooler, environmentalism too often just means recycling programs and classroom lessons on sustainability. The chance to think about that problem – and themselves – in political terms is new for them. Even for students who study politics or history or government, activism isn’t really a dimension of their curriculum. (Turrentine) 30 Thus, many youth movements, especially Sunrise, have provided means for those who desire to take action. Public opinion is a powerful tool as well: the practice of “naming and shaming” is an effective way to force climate change onto the agenda (Lakhani). During the 2020 presidential primaries, Sunrise called upon candidates to sign the “No fossil fuel money” pledge. Twenty-four Democratic presidential candidates (all but one) signed the pledge in 2020, and as of 9 May 2021, the website lists over 3,293 elected officials at all levels of government who have signed the pledge. Sunrise has also adopted and extensively used a technique that early environmentalists pioneered in 1970: identifying elected officials with poor environmental voting records. 2.1 The Sunrise Movement 27 One group in particular amongst the youth activists realized that peaceful protests were insufficient and, from the outset, has had a strategy to both mobilize large numbers of energetic activists and elect committed individuals to Congress and state legislatures. Co-founder of the Sunrise Movement, Varshini Prakash started out in the divestment movement at her university, so was familiar with demonstrations, sit-ins, E-rea, 18.2 \| 2021 Youth climate activism in the United States 9 rallies. She and a group of friends spent a full year studying previous successes of protest movements, like the civil rights movement of the 1960s, and previous failures, like the Occupy Wall Street movement and Bernie Sanders’ 2016 presidential campaign, in order to develop a comprehensive strategy before launching the Sunrise Movement in 2017. Not only would these youths protest against inaction, they would also propose a solution for the future. What distinguishes this group from the majority of youth climate action groups then, is its multi-pronged strategy and wide-ranging program to adapt the American economy to the twenty-first century. By framing the fight for the planet as “a battle between underdogs and moneyed special interests,” Sunrise “taps into reservoirs of youthful passion and energy, turning students into student activists” (Turrentine). 28 Sunrise is structured to encourage the creation of “hubs,” which are active at the local level throughout the country. According to their website, there are more than 400 hubs and although no official figure is available, it is believed that a million volunteers have joined the movement (Higgins & Chávez). Social media presence – informing, educating, motivating followers – is an integral part of the Sunrise Movement’s communication strategy, both to enhance cohesion amongst those who are members and to attract new activists. Innovative young volunteers have devised a new style of digital campaigning, creating “stan” accounts to publicize campaign messages through memes when door-to-door canvassing became tricky, if not impossible, during the Coronavirus pandemic (Higgins and Chávez). 28 Sunrise is structured to encourage the creation of “hubs,” which are active at the local level throughout the country. According to their website, there are more than 400 hubs and although no official figure is available, it is believed that a million volunteers have joined the movement (Higgins & Chávez). 2.1 The Sunrise Movement In 1972, environmentalists contributed to the defeat of seven out of twelve anti-environmental E-rea, 18.2 \| 2021 10 Youth climate activism in the United States 10 incumbents, the “dirty dozen”. Today, with the help of the League of Conservation Voters, which tracks the voting records of all congressional members, Sunrise targets those elected officials who have poor voting records on environmental issues, publicizes the records in their home districts, and if the officials are up for re-election, strongly supports challengers whose platforms coincide with Sunrise’s objectives. To summarize, picket signs read: “Dear Democrats: Step Up or Step Aside” (Sunrise website). 31 Once convinced of a candidate’s platform and commitment to the group’s core objectives, Sunrise deploys an “army” of activists to call potential voters and canvas neighborhoods door-to-door. In 2018 for example, they were instrumental in Alexandria Ocasio-Cortez’s upset win of the U.S. House of Representatives seat for the Queens and Bronx areas of New York City. Joseph Crowley, the incumbent Democrat, had held the seat for twenty years and did not take his challenger, Ocasio-Cortez, a 28- year-old female bartender at the time, seriously. AOC, as she is known, won the primary by 15 points and went on to win the election. Since then, she has become an influential progressive leader in the House of Representatives. Ten of the twenty candidates Sunrise endorsed for Congress in 2018 were elected and six more won seats in the state legislatures of Pennsylvania, New York and Florida (Influence Watch). 32 In 2020, Sunrise endorsed seventeen candidates: Bernie Sanders for president and the others for congressional seats. Once convinced of candidates’ merits, the volunteers put great energy into promoting them. Notable successes include the elections of Senator Ed Markey and Representative Cori Bush. Markey, a longtime incumbent senator of Massachusetts, was challenged by Joe Kennedy, 39, in the Democratic primary of the 2020 election (Higgins & Chávez). Aged 74, Markey is a veteran politician and co- sponsor of the Green New Deal in the Senate, but the Kennedy name and the allure of a young, new voice had Markey trailing by a significant margin a year before the primary. 2.1 The Sunrise Movement Sunrise volunteers engaged in intensive voter turnout efforts (mostly digital and telephone contacts due to Covid restrictions) to bolster Markey’s image: “In a clever twist, Sunrise was able to cast the veteran lawmaker not as a stale incumbent, but rather as a longtime ally in developing and fighting for the Green New Deal” (Higgins and Chávez). Markey eventually won the primary handily, despite influential House Speaker Nancy Pelosi’s endorsement of Joe Kennedy. Commentators remarked that the turnabout would have been impossible without Sunrise’s help (Tierstein, Ward & Wilson, White, Holden, Manchester). 33 Cori Bush ran for a seat in the U.S. House of Representatives in Missouri in 2020. Sunrise threw its support behind Democratic challenger Bush, who was running for the second time against Lacy Clay, the ten-term incumbent, for a seat that had been held by a member of the Clay family for 52 years. Sunrise volunteers called up 150,000 voters in the final weeks of the primary campaign, contributing significantly to Bush’s upset win. While in 2018 Bush lost to Clay by twenty points, in 2020 Bush won by three points, which the Sunrise volunteers considered proof of the efficacy of their endeavors. 33 Cori Bush ran for a seat in the U.S. House of Representatives in Missouri in 2020. Sunrise threw its support behind Democratic challenger Bush, who was running for the second time against Lacy Clay, the ten-term incumbent, for a seat that had been held by a member of the Clay family for 52 years. Sunrise volunteers called up 150,000 voters in the final weeks of the primary campaign, contributing significantly to Bush’s upset win. While in 2018 Bush lost to Clay by twenty points, in 2020 Bush won by three points, which the Sunrise volunteers considered proof of the efficacy of their endeavors. 34 Overall, the majority of Sunrise’s endorsed candidates won their primaries in 2020, consolidating the Sunrise Movement’s position as a force to be reckoned with in the electoral field. “Dutiful” may not be the best word to characterize Sunrise’s tactics in the electoral process. At least from the standpoint of targeted incumbents, Sunrise constitutes a danger to the status quo. 2.1 The Sunrise Movement In hopes of furthering their influence, Sunrise activists posted on their website panels of people that they would like to see in the 2021 34 Overall, the majority of Sunrise’s endorsed candidates won their primaries in 2020, consolidating the Sunrise Movement’s position as a force to be reckoned with in the electoral field. “Dutiful” may not be the best word to characterize Sunrise’s tactics in the electoral process. At least from the standpoint of targeted incumbents, Sunrise constitutes a danger to the status quo. In hopes of furthering their influence, Sunrise activists posted on their website panels of people that they would like to see in the 2021 E-rea, 18.2 \| 2021 11 11 Youth climate activism in the United States Biden cabinet. Sunrise’s support for the Green New Deal, another important aspect of the group’s activism, will be discussed in the “dangerous” dissent section. 2.2 Working through the courts 35 Another strategy employed by youth climate activists – and perhaps the most effective one – is litigation, a widespread tool of contestation in the United States. The knowledge that human activity (primarily the burning of fossil fuels) is impacting the climate was known by the late 1950s and significant enough for policymakers to take note by the 1960s. In 1965 President Johnson said in a speech to the nation: "this generation has altered the composition of the atmosphere on a global scale through ... a steady increase in carbon dioxide from the burning of fossil fuels" (Lavelle). 36 A watershed moment occurred when in 1988, James Hansen, a physicist at NASA testified to the U.S. Senate committee: “The greenhouse effect has been detected and is changing our climate now” (Yale Climate Connections). Thus, the federal government, if not the general public, has been aware of the phenomenon for decades. Hansen’s 1988 testimony was instrumental in convincing the George H.W. Bush administration to sign the first treaty on climate change in 1992 – before it became a partisan issue. This treaty, called the United Nations Framework Convention on Climate Change (UNFCCC) remains the basis for a coordinated international effort to reduce greenhouse gas emissions and reduce the threat. 36 A watershed moment occurred when in 1988, James Hansen, a physicist at NASA testified to the U.S. Senate committee: “The greenhouse effect has been detected and is changing our climate now” (Yale Climate Connections). Thus, the federal government, if not the general public, has been aware of the phenomenon for decades. Hansen’s 1988 testimony was instrumental in convincing the George H.W. Bush administration to sign the first treaty on climate change in 1992 – before it became a partisan issue. This treaty, called the United Nations Framework Convention on Climate Change (UNFCCC) remains the basis for a coordinated international effort to reduce greenhouse gas emissions and reduce the threat. 37 On the domestic scene, the country has a solid basis from which to combat rising emissions through wide reaching environmental protection legislation that was passed during the late 1960s and early 1970s. A crucial new provision that was written into the 1970s’ environmental legislation states that citizens “have a right to know” and can have access to documents pertaining to a particular issue. 2.2 Working through the courts In addition, developers using federal funds are required to produce an Environmental Impact Statement before proceeding with their project (Andrews 241-2). Environmental associations understood the potential of this principle and have used it extensively to block projects through litigation. Thus, the courts have played a major part in environmental protection since the 1970s. 38 Currently, several cities, counties and a state are suing major fossil fuel companies for selling harmful products and misleading the public about their dangers (Banerjee et al). InsideClimate News has conducted in-depth research which has proven that Exxon Mobil knew of the negative impact of fossil fuels on human health and the environment decades ago (Banerjee et al). Rather than inform the public and find ways to make their products less polluting, they launched misinformation campaigns and cast doubt on the science. As Rhode Island Attorney General Peter F. Kilmartin said in 2018, announcing his state’s suit against companies like Exxon Mobil, Chevron, BP, and Royal Dutch Shell, Instead of working to reduce that harm, these companies chose to conceal the dangers, undermine public support for greenhouse gas regulation, and engage in massive campaigns to promote the ever-increasing use of their products and ever- increasing revenues in their pockets. (Kusnetz 2 July 2018) his state’s suit against companies like Exxon Mobil, Chevron, BP, and Royal Dutch Sh Instead of working to reduce that harm, these companies chose to conceal the dangers, undermine public support for greenhouse gas regulation, and engage in massive campaigns to promote the ever-increasing use of their products and ever- increasing revenues in their pockets. (Kusnetz 2 July 2018) 39 The fossil fuel industry’s misinformation campaign has been quite successful: judging by the surveys of public opinion in various countries, the American public shows significantly greater doubt about the human responsibility for climate change. According to Pew Research, in 2019, only 59% of Americans compared to 83% of the E-rea, 18.2 \| 2021 12 Youth climate activism in the United States 12 French or 90% of Greeks believe that human activity is the driver of global warming (Fagan & Huang). And this figure is considerably higher than a few years ago (Saad). The prevalence of climate skepticism and denialism in the United States points strongly to the success of the misinformation campaign and has been a major factor in the federal government’s lack of climate action. 2.2 Working through the courts With gridlock in Congress and President Obama’s “all-of-the-above” energy strategy followed by the Trump administration’s support for the coal industry and systematic repeal of as many environmental protections as possible, (Fox News, Pitt et al) the federal government has been unable or unwilling to take serious action on climate change for decades. 40 In an attempt to force the federal government to act on climate change, twenty-one young plaintiffs, accompanied by Dr. James Hansen and lawyer Julia Olson, filed a lawsuit against the United States government in 2015. The lawsuit, Juliana vs. the United States, based on the Public Trust Doctrine1, charges that the government is guilty of allowing and encouraging greenhouse gas emissions over the past fifty years despite knowing about global climate change, thus violating the plaintiffs’ due process rights under the Fifth Amendment of the Constitution. The litigants were chosen from diverse locations in the United States (Oregon, Florida, Colorado, etc.), all of which had been damaged due to climate change (60 Minutes). The plaintiffs contend that federal support for fossil fuel extraction is causing climate change, endangering their future and violating their constitutional rights to life, liberty and property. 40 In an attempt to force the federal government to act on climate change, twenty-one young plaintiffs, accompanied by Dr. James Hansen and lawyer Julia Olson, filed a lawsuit against the United States government in 2015. The lawsuit, Juliana vs. the United States, based on the Public Trust Doctrine1, charges that the government is guilty of allowing and encouraging greenhouse gas emissions over the past fifty years despite knowing about global climate change, thus violating the plaintiffs’ due process rights under the Fifth Amendment of the Constitution. The litigants were chosen from diverse locations in the United States (Oregon, Florida, Colorado, etc.), all of which had been damaged due to climate change (60 Minutes). The plaintiffs contend that federal support for fossil fuel extraction is causing climate change, endangering their future and violating their constitutional rights to life, liberty and property. 41 In an interview on 60 Minutes, Julia Olson stated, The lawsuit claims the executive and legislative branches of government have proven incapable of dealing with climate change. It argues that the government has failed in its obligation to protect the nation's air, water, forests and coastlines. 2.2 Working through the courts And it petitions the federal courts to intervene and force the government to come up with a plan that would wean the country off fossil fuels by the middle of this century.” (60 Minutes) 42 After multiple stays, the case was heard in court in June 2019 and a decision was published in January 2020 (Schwartz). Two of the three judges stated that they recognized the gravity of the evidence provided by the plaintiffs, the government’s role in causing injury from climate change, and that the government was violating the plaintiff’s constitutional rights. However, the two judges concluded that the plaintiffs did not have standing and that the executive and legislative branches – not the courts – should address the requests, while the third judge dissented, strongly affirming the youths’ constitutional rights. The decision was a disappointment for the plaintiffs, who responded by filing a petition to request a review of the decision by the full panel of eleven circuit court judges. Groups including twenty-four members of the U.S. Congress, experts in constitutional law, climate change and public health filed ten amicus curiae briefs in support of their petition (Schwartz). 42 After multiple stays, the case was heard in court in June 2019 and a decision was published in January 2020 (Schwartz). Two of the three judges stated that they recognized the gravity of the evidence provided by the plaintiffs, the government’s role in causing injury from climate change, and that the government was violating the plaintiff’s constitutional rights. However, the two judges concluded that the plaintiffs did not have standing and that the executive and legislative branches – not the courts – should address the requests, while the third judge dissented, strongly affirming the youths’ constitutional rights. The decision was a disappointment for the plaintiffs, who responded by filing a petition to request a review of the decision by the full panel of eleven circuit court judges. Groups including twenty-four members of the U.S. Congress, experts in constitutional law, climate change and public health filed ten amicus curiae briefs in support of their petition (Schwartz). 43 Our Children’s Trust is the non-profit organization that underwrites and provides the legal representation for Juliana and many similar actions. 2.2 Working through the courts Analysts are even more skeptical since President Trump appointed three conservative justices to the Supreme Court, bringing the count to six conservative and three progressive justices. But as mentioned above, the number of lawsuits against fossil fuel corporations and governments for failing to protect public health and common resources has multiplied, and a victory would set a precedent for future litigation. 45 In sum, working through the electoral system and through the courts are two avenues that youth climate activists are exploring. Battles in these institutions require determination and patience but offer more likelihood of long-term success. Indeed, electing enough climate action proponents to Congress to enact laws may take years. A case like Juliana vs. the United States has already been in the works for six years and may not result in a victory. Analysts are even more skeptical since President Trump appointed three conservative justices to the Supreme Court, bringing the count to six conservative and three progressive justices. But as mentioned above, the number of lawsuits against fossil fuel corporations and governments for failing to protect public health and common resources has multiplied, and a victory would set a precedent for future litigation. 3. “Dangerous” Dissent 46 The majority of the youth actions examined so far could be considered “disruptive:” mobilizing an “army of climate activists,” engaging in classic street rallies and sit-ins. Some could be categorized as “dutiful” dissent because they work within existing power structures, like the electoral process or the courts. Unconventional techniques, however, are employed to produce successes: “naming and shaming” for example. And in the courts, youths are using an innovative approach by suing their federal, state and local governments for violating their right to a livable planet. Reparations are unlikely, but the courts could judge that the state and federal governments must devise and implement climate action plans, which is what the Juliana plaintiffs are demanding (60 Minutes). In fact, states and cities have been far more active than the federal government in this domain. As of 2020, thirty-two states have climate action plans with targets to reduce greenhouse gas emissions by implementing various policies that favor clean energy sources, clean transportation and buildings, as well as strategies to cope with climate disasters and support economic and social goals (Ricketts et al). And the Biden administration appears to be committed to action on climate change, which may result in more accountability for corporations and utilities (Gardner). 2.2 Working through the courts Its website declares that it is “the world’s only non-profit public interest law firm dedicated exclusively to securing the legal rights of youth to a healthy atmosphere and safe climate, based on the best available science.” The website offers information on seven pending cases in the states of Alaska, Colorado, Maine, Massachusetts, New Mexico, North Carolina and Oregon as well as details concerning legal actions in all fifty states. 43 Our Children’s Trust is the non-profit organization that underwrites and provides the legal representation for Juliana and many similar actions. Its website declares that it is “the world’s only non-profit public interest law firm dedicated exclusively to securing the legal rights of youth to a healthy atmosphere and safe climate, based on the best available science.” The website offers information on seven pending cases in the states of Alaska, Colorado, Maine, Massachusetts, New Mexico, North Carolina and Oregon as well as details concerning legal actions in all fifty states. E-rea, 18.2 \| 2021 13 Youth climate activism in the United States 13 44 Juliana vs. the United States has received significant coverage in the media, both mainstream (New York Times, Washington Post, CBS, ABC, NBC) and specialized (InsideClimate News, Grist) and shone a moral spotlight on the country’s government and on powerful corporations. If Juliana were to prevail, it could have a far-reaching impact on climate action. Indeed, if the tobacco industry’s experience is any indication, the case could lead to multibillion-dollar payouts, and force the federal government to make greenhouse gas emission reductions a central priority (Oreskes and Conway). Although few legal observers expect the case to prevail, some believe that it has opened the door to more favorable rulings because the court clearly acknowledged that there has been injury to the youths and that the federal government has promoted fossil fuel use despite the harm fossil fuels are known to cause (Banerjee et al). 45 In sum, working through the electoral system and through the courts are two avenues that youth climate activists are exploring. Battles in these institutions require determination and patience but offer more likelihood of long-term success. Indeed, electing enough climate action proponents to Congress to enact laws may take years. A case like Juliana vs. the United States has already been in the works for six years and may not result in a victory. 3.1 A threat to the fossil fuel industry 47 One indication that the fossil fuel industry feels threatened is its draconian use of force against groups of oil pipeline protesters, comprising significant numbers of youths (Cagle). The Keystone XL and Dakota Access pipeline projects to transport Canadian tar E-rea, 18.2 \| 2021 Youth climate activism in the United States 14 sands to Texas refineries are a case in point. Many powerful interests are at stake: the federal government, to ensure energy supplies in the interest of the nation; the states that the pipeline passes through, for the jobs and wealth that the project brings; the corporations, to maximize profits from such projects. The perceived threat of the protests, regarded as “dangerous” dissent, could then become the object of repression, even criminalization. According to an article published by InsideClimate News and the Washington Post, since January 2017 elected officials in at least thirty-one states have proposed measures to limit the right to protest (Sadasivam). 48 For example, the Standing Rock Indian Reservation groups protested against the construction of an oil pipeline that would despoil sacred burial sites and endanger their drinking water (Kusnetz 22 August 2018). Despite legal action that was pending, the Dakota Access Pipeline Company, Energy Transfer Partners, proceeded with construction and brought in a private security firm which deployed dogs and pepper spray against protesters. According to Democracy Now’s journalist Amy Goodman, who filmed the conflict, the project had already faced months of resistance from the Standing Rock Sioux tribe as well as members of nearly 100 more tribes from across the U.S. and Canada (Democracy Now). Numerous protesters and Goodman were targeted for arrest under new laws in North Dakota that criminalize protesting within a few feet of a contested site. Since these confrontations in 2016, activists have increasingly been treated as terrorists and arrests have become frequent (Cagle). "They're basically trying to silence and minimize the impact of environmental organizations," said Pamela Spees, a senior staff attorney at the Center for Constitutional Rights, which represents activists in Louisiana (Kusnetz 22 August 2018). 3.2 A threat to the capitalist establishment 49 Furthermore, progressive climate activists integrate social justice and equity into their calls for climate action, deeming that one is unattainable without the other. Climate activists have values that come into conflict with society’s mainstream tenets – the set of political values and ideals that constitute the bedrock of American identity. Sociologist Dennis C. Pirages defined the dominant social paradigm as comprising four elements: a commitment to individualism, property rights and laissez-faire government; faith in material abundance, progress and economic growth; faith in science and technology; and a view of nature as something to be subdued (Pirages 37). The vast majority of environmentalists and climate activists contest at least two of these beliefs: that of the eternal growth model and that of humans being superior to nature (Dunlap et al). As these ideas run counter to the dominant social paradigm, they put into question the extractive model on which the American economy is based. Surveys show that young generations believe less in humans’ ability to dominate nature or in the limitless growth model. 50 This melding of the environmental with the social brings to mind the words of an early environmentalist, Barry Commoner. In The Closing Circle, he spoke of the environmental crisis in terms that are entirely applicable to the current situation: Human beings have broken out of the circle of life, driven not by biological need, but by the social organization which they have devised to “conquer” nature: means of gaining wealth that are governed by requirements conflicting with those which govern nature. The end result is the environmental crisis, a crisis of survival… the world is being carried to the brink of ecological disaster… by the phalanx of 50 This melding of the environmental with the social brings to mind the words of an early environmentalist, Barry Commoner. In The Closing Circle, he spoke of the environmental crisis in terms that are entirely applicable to the current situation: Human beings have broken out of the circle of life, driven not by biological need, but by the social organization which they have devised to “conquer” nature: means of gaining wealth that are governed by requirements conflicting with those which 50 This melding of the environmental with the social brings to mind the words of an early environmentalist, Barry Commoner. 3.2 A threat to the capitalist establishment According to Naomi Klein, Bernie Sanders, and many others, the neoconservative revolution flourished, sweeping away as many environmental and trade regulations as possible, leaving in its wake a hyper-consumerist society and a badly degraded planet. As Klein remarks, We have not done the things that are necessary to lower emissions because those things fundamentally conflict with deregulated capitalism, the reigning ideology for the entire period that we have been struggling to find a way out of this crisis. We are stuck because the actions that would give us the best chance of averting catastrophe, and that would benefit the vast majority, are extremely threatening to an elite minority that has a stranglehold over our economy, our political process, and most of our media outlets. (Klein 2019 249) 3.2 A threat to the capitalist establishment In The Closing Circle, he spoke of the environmental crisis in terms that are entirely applicable to the current situation: Human beings have broken out of the circle of life, driven not by biological need, but by the social organization which they have devised to “conquer” nature: means of gaining wealth that are governed by requirements conflicting with those which govern nature. The end result is the environmental crisis, a crisis of survival… the world is being carried to the brink of ecological disaster… by the phalanx of Human beings have broken out of the circle of life, driven not by biological need, but by the social organization which they have devised to “conquer” nature: means of gaining wealth that are governed by requirements conflicting with those which govern nature. The end result is the environmental crisis, a crisis of survival… the world is being carried to the brink of ecological disaster… by the phalanx of E-rea, 18.2 \| 2021 15 15 Youth climate activism in the United States powerful economic, political, and social forces that constitute the march of history. …sweeping social change can be designed only in the workshop of rational, informed, collective social action. That we must act is now clear. (Commoner 222-23) powerful economic, political, and social forces that constitute the march of history. …sweeping social change can be designed only in the workshop of rational, informed, collective social action. That we must act is now clear. (Commoner 222-23) powerful economic, political, and social forces that constitute the march of history. …sweeping social change can be designed only in the workshop of rational, informed, collective social action. That we must act is now clear. (Commoner 222-23) 51 It is worth noting that, fifty years later, the conversation seems to be stuck in a time warp: this statement could as easily have been written today, with “climate” in place of “environmental.” What happened between 1971 and 2020? According to Naomi Klein, Bernie Sanders, and many others, the neoconservative revolution flourished, sweeping away as many environmental and trade regulations as possible, leaving in its wake a hyper-consumerist society and a badly degraded planet. As Klein remarks, 51 It is worth noting that, fifty years later, the conversation seems to be stuck in a time warp: this statement could as easily have been written today, with “climate” in place of “environmental.” What happened between 1971 and 2020? 3.3 A threat to the political establishment: the Green New Deal Thus, the progressive wing of the Democratic Party faces opposition in both houses of Congress – and not only with Republicans. 54 This longstanding congressional reticence to tackle the climate emergency has undermined trust in the political system and galvanized youth activism. Youths like Solei Yoder Salim, aged 15, said that, leading up to the November election, he was spending five hours a day on phonebanking: “The Green New Deal is worth fighting for, it’s important for the world and our country. I like coming together with my community and other high schools, it’s important but also fun” (Lakhani). Even if Biden was not their ideal candidate, 60% of 18-29-year-olds responded that they would vote for him vs. 27% for Trump in the Fall 2020 Harvard Youth Poll (Ranalli). Many youths saw the choice between Biden and Trump as an existential emergency, the former representing the potential for a livable planet whereas the latter represented “the apocalypse” (Lakhani). While youth climate activists were relieved to see Biden elected and feel that their work had contributed to this victory, they remain skeptical: will the current administration take the drastic measures necessary to avert the worst effects of climate change? (Ebbs et al). 3.3 A threat to the political establishment: the Green New Deal 52 An alternative model that puts equity and respect for the natural environment over growth and profits is a threat to those with an interest in perpetuating the current model. The sheer number of people turning out to demonstrate their support for climate action and the Green New Deal gives cause for worry to all the stakeholders in the status quo of a fossil-fuel, hyper-consumptive growth model of society. This is perhaps why the proposition of the Green New Deal by the progressive wing of the Democratic Party has been soundly rejected, not only by Republicans, but also by the Democratic Party establishment (Roberts 2018). The Sunrise Movement does not only contest the status quo; these youths propose an alternative vision of society. Their vision is detailed in the clean energy and job creation plan called the Green New Deal, of which the Sunrise political director Evan Weber was one of the main architects (Matthews et al). Inspired by Franklin Delano Roosevelt’s New Deal, the Green New Deal is a program to achieve net-zero greenhouse gas emissions, secure clean air and water, invest in the country’s infrastructure and industry, and create millions of good, high- wage jobs in the process. According to Green New Dealers’ reckoning, this overhauling of the nation’s economic and energy systems would also promote justice and equity (Roberts 2019). 53 To these youth activists, most members of Congress seem either out of touch with reality or fail to appreciate the urgency of the climate crisis. When youths approached Senator Dianne Feinstein, a well-known Democratic senator from California, to obtain her support for the Green New Deal, her dismissal made them realize that they could not count on elected officials in Washington, D.C. to address the climate emergency (Guardian 23 Feb 2019). The Republican and Democratic opposition charge that it would cost trillions of dollars to accomplish. Seasoned Democratic politicians like Senators Feinstein and Joe Manchin, a ranking member of the Senate Energy Committee, and Representatives Speaker of the House Nancy Pelosi and the four co-chairs of the House’s Climate Change Task Force all oppose the Green New Deal (Roberts 2019). The House’s Climate Change Task Force published its alternative plan in June 2020, which is more incremental and less disruptive of the status quo than the Green New Deal (House E-rea, 18.2 \| 2021 16 Youth climate activism in the United States 16 Select Committee). 4. Biden’s response to climate activism and young voters 55 As noted earlier, since Joe Biden’s election, the federal administration has signaled its commitment to climate action. Appointments of highly qualified climate actors to key positions like the heads of the Department of the Interior, Energy, the Environmental Protection Agency, the Council on Environmental Quality, as well as the creation of two new positions, Special Presidential Envoy for Climate and National Climate Advisor, demonstrate a willingness to tackle the climate emergency. Second, Biden announced that the United States would rejoin the Paris Agreement and commit to emissions reductions goals and take on a leading role in the international fight against climate change (Davenport & Friedman). Third, in March 2021, the President unveiled a $2.3- trillion green infrastructure plan which focuses on rebuilding the country’s infrastructure and shifts to greener energy and transportation, two prime polluting sectors. While the bill may undergo necessary compromises in order to pass the Senate, the Biden administration’s plan is far more comprehensive than many climate activists imagined possible (Gardner). Fourth, President Biden hosted a Global Climate Change Summit on 22 April 2021 and announced that the United States would reduce its greenhouse gas emissions by 50%-52% below its 2005 emissions levels by 2030 (Mahtani et al). 56 From the perspective of climate campaigners and in particular of youth climate activists, this flurry of executive branch climate action is a long-awaited turn of events. The calls for action coming from youths on both sides of the political spectrum are being answered, and the nonpartisan nature of these demands signals the potential for American politics to move toward a less polarized approach. The youths’ calls in unison for action against climate change confirm the “generationally shared values of truth, empathy, and patriotism,” according to Nikayla Jefferson, of the Sunrise Movement (Kahn). 56 From the perspective of climate campaigners and in particular of youth climate activists, this flurry of executive branch climate action is a long-awaited turn of events. The calls for action coming from youths on both sides of the political spectrum are being answered, and the nonpartisan nature of these demands signals the potential for American politics to move toward a less polarized approach. The youths’ calls in unison for action against climate change confirm the “generationally shared values of truth, empathy, and patriotism,” according to Nikayla Jefferson, of the Sunrise Movement (Kahn). BIBLIOGRAPHY BIBLIOGRAPHY 4. Biden’s response to climate activism and young voters 57 From the perspective of elected officials, the young generations represent a significant number of potential voters. Demographic forecasts predict that as early as 2028, Gen Z 57 From the perspective of elected officials, the young generations represent a significant number of potential voters. Demographic forecasts predict that as early as 2028, Gen Z E-rea, 18.2 \| 2021 17 Youth climate activism in the United States 17 (born since 1997) and millennials (born 1981-1996) will dominate the electorate: 49% vs. 46% comprised of baby boomers (born 1946-1964) and Gen X (born 1965-1980) (Brownstein). According to the U.S. Census Bureau, voter turnout for the 2020 elections was higher across all age groups, but the increase was greatest amongst young voters, aged 18 to 34, rising from 49% to 57% of eligible citizens (Fabina). Higher turnout indicates greater engagement and may prove to be a formative “stage in the process of acquiring the skills necessary for political participation” (Jenkins et al 7). As Ruy Texeira writes in the Atlantic, the Democratic Party has a golden opportunity to bring the young voters into its fold, on condition that it opens up to their ideas and shares leadership (Griffin et al). In an interview in August 2020, young conservative Benji Backer warned that unless the Republican Party started dealing with the climate crisis, it would lose scores of young members for whom the issue is primordial and nonpartisan. 58 In sum, large numbers of American youths have joined the ranks of climate activists in the last few years, particularly since the election of President Donald Trump. Their actions have taken various forms, from street marches and school strikes to phone banking and canvassing communities to get out the vote, to suing their governments, both state and federal, for violating the public trust doctrine, to supporting the Green New Deal. Characterized as “disruptive, dutiful or dangerous” dissent, their actions helped elect progressive candidates in 2018 and 2020. These youth movements span the political spectrum from progressive to conservative and, as a generation, are putting substantial pressure on older elected officials to pass legislation to limit emissions and ensure an equitable society on a livable planet. 4. Biden’s response to climate activism and young voters The victory of Joe Biden and Kamala Harris in the 2020 presidential election – to which the uncharacteristically high number of young voters contributed – has vastly increased the chances of the United States meeting its emissions reductions goals and playing a positive role in the fight against climate change. Youth climate activism Climate Academy. “How youth activism is moving the needle on climate change.” Guardian, 23 Sept 2020, https://www.theguardian.com/climate-academy/2020/sep/23/youth-activism- climate-crisis-action. Accessed 8 May 2021. Ebbs, Stephanie, Ginger Zee, Lindsey Griswold and Jon Schlosberg. “Why youth climate change activists are pushing Biden to do more.” ABC News, 21 Apr 2021; https://abcnews.go.com/US/ youth-climate-change-activists-pushing-biden/story?id=77167042. Accessed 22 May 2021. “Explained: The rise of Greta Thunberg.” Sky News, 20 Sept 2019. https://www.youtube.com/ watch?v=itAu7Uw0sn8. Accessed 25 Sept. 2019. Fisher, Dana R. “The broader importance of #FridaysForFuture.” Nature, 6 May 2019, https:// www.nature.com/articles/s41558-019-0484-y. Accessed 30 Oct 2020. Fisher, Dana R. “The broader importance of #FridaysForFuture.” Nature, 6 May 2019, https:// www.nature.com/articles/s41558-019-0484-y. Accessed 30 Oct 2020. E-rea, 18.2 \| 2021 18 Youth climate activism in the United States “Influence Watch: Sunrise Movement.” https://www.influencewatch.org/non-profit/sunrise- movement/. Accessed 1 Nov 2020. “Influence Watch: Sunrise Movement.” https://www.influencewatch.org/non-profit/sunrise- movement/. Accessed 1 Nov 2020. Kaplan, Sarah. “Teen girls are leading the climate strikes and helping change the face of environmentalism.” Washington Post, 24 Sept 2019, https://www.washingtonpost.com/science/ 2019/09/24/teen-girls-are-leading-climate-strikes-helping-change-face-environmentalism/. Accessed 31 May 2021. Klein, Naomi. On Fire, The Burning Case for a Green New Deal. Allen Lane, 2019. Lakhani, Nina. “’We don’t have any choice’: The young climate activists naming and shaming politicians’.” Guardian, 16 Oct 2020, https://www.theguardian.com/environment/2020/oct/16/ sunrise-movement-wide-awakes-us. Accessed 30 Oct 2020. Matthews, Mark, Nick Bowlin and Benjamin Hulak. “Inside the Sunrise Movement.”, E&E News, 3 Dec 2018, https://www.eenews.net/stories/1060108439. Accessed 2 Nov 2020. Meyaard-Schaap, Kyle. “Young evangelicals used to be skeptical of climate change. Not anymore. And we’re voting.” Grist, 23 Sept 2020, https://grist.org/fix/young-evangelicals-climate-change- voting/. Accessed 30 Oct 2020. O'Brien, Karen, Elin Selboe and Bronwyn M. Hayward. “Exploring Youth Activism on climate change: dutiful, disruptive, and dangerous dissent.” Ecology and Society 23(3):42, 2018, https:// doi.org/10.5751/ES-10287-230342. Accessed 2 Nov 2020. The Climate Group. “Youth activism in climate in USA.” 23 Sep 2020, https://www.youtube.com/ watch?v=5NJ-p0vMEEU. Accessed 31 Oct 2020. The Climate Group. “Youth activism in climate in USA.” 23 Sep 2020, https://www.youtube.com watch?v=5NJ-p0vMEEU. Accessed 31 Oct 2020. Turrentine, Jeff. “Will the Kids Save us from Climate Change?” NRDC, 23 March 2019, https:// www.nrdc.org/onearth/will-kids-save-us-climate-change. Accessed 2 Apr 2019. “Young climate activists testify on Capitol Hill ahead of U.N. Climate Action Summit.” CBS News, 19 Sept 2019, https://www.youtube.com/watch?v=ecuiQXSv5pc. Accessed 31 May 2021. Lawsuits Lawsuits 60 Minutes. “Juliana vs. the United States: the climate change lawsuit that could stop the US government from supporting fossil fuels.” CBS News, 3 March 2019, https://www.cbsnews.com/ news/juliana-versus-united-states-the-climate-change-lawsuit-that-could-stop-the-u-s- government-from-supporting-fossil-fuels-60-minutes/. Accessed 29 Sept 2019. Kusnetz, Nicholas. “Faced with the costs of climate change, R.I. sues fossil-fuel companies.” InsideClimate News, 2 July 2018, https://www.bostonglobe.com/metro/2018/07/02/faced-with- costs-climate-change-sues-fossil-fuel-companies/mGflVecLlTflLHCmBHYzAM/story.html. Accessed 31 May 2021. Schwartz, John. “Court Quashes Youth Climate Change Case Against Government,” New York Times, 17 Jan 2020 https://www.nytimes.com/2020/01/17/climate/juliana-climate-case.html. Accessed 4 Nov 2020. “The ‘climate kids’ suing the government.” 60 Minutes, 3 March 2019, https:// www.youtube.com/watch?v=1Qyo6T5xd4A. Accessed 31 May 2021. “The ‘climate kids’ suing the government.” 60 Minutes, 3 March 2019, https:// www.youtube.com/watch?v=1Qyo6T5xd4A. Accessed 31 May 2021. US Energy and Climate Policies Andrews, Richard L. Managing the Environment, Managing Ourselves, Yale University Press, 2006. Andrews, Richard L. Managing the Environment, Managing Ourselves, Yale University Press, 2006. E-rea, 18.2 \| 2021 19 Youth climate activism in the United States Banerjee, Neela, Lisa Song and David Hasemyer. “Exxon: The Road Not Taken.” InsideClimate News, 16 Sept 2015, https://insideclimatenews.org/content/Exxon-The-Road-Not-Taken. Accessed 10 Nov 2020. Davenport, Coral and Lisa Friedman. “Biden Cancels XL Pipeline and Rejoins Paris Climate Agreement.” New York Times, 21 Jan 2021, https://www.nytimes.com/2021/01/20/climate/biden- paris-climate-agreement.html. Accessed 22 May 2021. Davenport, Coral and Lisa Friedman. “Biden Cancels XL Pipeline and Rejoins Paris Climate Agreement.” New York Times, 21 Jan 2021, https://www.nytimes.com/2021/01/20/climate/biden- paris-climate-agreement.html. Accessed 22 May 2021. Gardner, Timothy. “Biden tax plan replaces U.S. fossil fuel subsidies with clean energy incentives.” Reuters, 7 Apr 2021 https://www.msn.com/en-us/news/us/biden-tax-plan-replaces- u-s-fossil-fuel-subsidies-with-clean-energy-incentives/ar-BB1foARo. Accessed 8 May 2021. Gardner, Timothy. “Biden tax plan replaces U.S. fossil fuel subsidies with clean energy incentives.” Reuters, 7 Apr 2021 https://www.msn.com/en-us/news/us/biden-tax-plan-replaces- u-s-fossil-fuel-subsidies-with-clean-energy-incentives/ar-BB1foARo. Accessed 8 May 2021. House Select Committee on the Climate Crisis. www.climatecrisis.house.gov/report. Accessed 5 Nov 2020. Mahtani, Melissa, Meg Wagner and Melissa Macaya. “Biden hosts global climate change summit on Earth Day.” CNN, 23 Apr 2021, https://edition.cnn.com/politics/live-news/biden-climate- change-summit-04-22-22/index.html. Accessed 22 May 2021. McKibben, Bill. “What Exxon Knew About Climate Change.” New Yorker, 18 Sept 2015, https:// www.newyorker.com/news/daily-comment/what-exxon-knew-about-climate-change. Accessed 31 May 2021. McKibben, Bill. “What Exxon Knew About Climate Change.” New Yorker, 18 Sept 2015, https:// www.newyorker.com/news/daily-comment/what-exxon-knew-about-climate-change. Accessed 31 May 2021. McKibben, Bill. “Exxon, Keystone, and the Turn Against Fossil Fuels,” New Yorker, 6 Nov 2015, https://www.newyorker.com/news/news-desk/exxon-keystone-and-the-turning-tide-on-fossil- fuels. Accessed 31 May 2021. Pitt, Hannah, Kate Larsen, and Maggie Young. “The Undoing of US Climate Policy: The Emissions Impact of Trump-Era Rollbacks.” Rhodium Group, 17 Sep 2020, https://rhg.com/research/the- rollback-of-us-climate-policy/. Lawsuits Accessed 15 Oct 2020. Pitt, Hannah, Kate Larsen, and Maggie Young. “The Undoing of US Climate Policy: The Emissions Impact of Trump-Era Rollbacks.” Rhodium Group, 17 Sep 2020, https://rhg.com/research/the- rollback-of-us-climate-policy/. Accessed 15 Oct 2020. Roberts, David. “Utilities have a problem: the public wants 100% renewable energy, and quick.” Vox, 11 Oct 2018, https://www.vox.com/energy-and-environment/2018/9/14/17853884/utilities- renewable-energy-100-percent-public- Roberts, David. “Utilities have a problem: the public wants 100% renewable energy, and quick.” Vox, 11 Oct 2018, https://www.vox.com/energy-and-environment/2018/9/14/17853884/utilities- renewable-energy-100-percent-public- opinion#:~:text=100%25%20renewable%20energy%3A%20the%20public%20wants%20it%2C%20and, %40drvox%20david%40vox.com%20Updated%20Oct%2011%2C%202018%2C%209%3A19am%20EDT. Accessed 31 May 2021. opinion#:~:text=100%25%20renewable%20energy%3A%20the%20public%20wants%20it%2C%20and, %40drvox%20david%40vox.com%20Updated%20Oct%2011%2C%202018%2C%209%3A19am%20EDT. Accessed 31 May 2021. “Trump: We've ended the war on coal.” Fox News, 23 Feb 2018, https://www.youtube.com/watch? v=kmvW25iKd9M. Accessed 31 May 2021. “Trump: We've ended the war on coal.” Fox News, 23 Feb 2018, https://www.youtube.com/watch? v=kmvW25iKd9M. Accessed 31 May 2021. “Trump: We've ended the war on coal.” Fox News, 23 Feb 2018, https://www.youtube.com/watch? v=kmvW25iKd9M. Accessed 31 May 2021. Yale Climate Connections. “James Hansen’s 1988 Testimony after 30 years. How did he do?” 20 June 2018, https://www.youtube.com/watch?v=UVz67cwmxTM. Accessed 31 May 2021. US Politics Yale Climate Connections. “James Hansen’s 1988 Testimony after 30 years. How did he do?” 20 June 2018, https://www.youtube.com/watch?v=UVz67cwmxTM. Accessed 31 May 2021. US Politics US Politics “Dianne Feinstein rebuffs young climate activists' calls for Green New Deal.” Guardian News, 23 Feb 2019, https://www.youtube.com/watch?v=jEPo34LCss8. Accessed 10 Nov 2020. “Dianne Feinstein rebuffs young climate activists' calls for Green New Deal.” Guardian News, 23 Feb 2019, https://www.youtube.com/watch?v=jEPo34LCss8. Accessed 10 Nov 2020. Evelyn, Kenya. “Warren and Sanders: Amazon must end culture that puts profit over people.” Guardian, 11 Feb 2020 https://www.theguardian.com/technology/2020/feb/11/warren-sanders- amazon-end-culture-profit-over-people-worker-safety-bezos. Accessed 1 Nov 2020. Evelyn, Kenya. “Warren and Sanders: Amazon must end culture that puts profit over people.” Guardian, 11 Feb 2020 https://www.theguardian.com/technology/2020/feb/11/warren-sanders- amazon-end-culture-profit-over-people-worker-safety-bezos. Accessed 1 Nov 2020. Evelyn, Kenya. “Warren and Sanders: Amazon must end culture that puts profit over people.” Guardian, 11 Feb 2020 https://www.theguardian.com/technology/2020/feb/11/warren-sanders- amazon-end-culture-profit-over-people-worker-safety-bezos. Accessed 1 Nov 2020. Fabina, Jacob. “Despite Pandemic Challenges, 2020 Election Had Largest Increase in Voting Between Presidential Elections on Record.” US Census Bureau, 29 Apr 2021, https:// www.census.gov/library/stories/2021/04/record-high-turnout-in-2020-general-election.html. Accessed 9 May 2021. Fabina, Jacob. “Despite Pandemic Challenges, 2020 Election Had Largest Increase in Voting Between Presidential Elections on Record.” US Census Bureau, 29 Apr 2021, https:// www.census.gov/library/stories/2021/04/record-high-turnout-in-2020-general-election.html. Accessed 9 May 2021. E-rea, 18.2 \| 2021 20 20 Youth climate activism in the United States Higgins, Eoin and Aída Chávez. “The Sunrise Movement, a Growing Electoral Force, Faces ‘Painful Moment’.” The Intercept, 9 Sept 2020 https://theintercept.com/2020/09/08/sunrise-movement- ed-markey-election/. Accessed 10 Nov 2020. Higgins, Eoin and Aída Chávez. “The Sunrise Movement, a Growing Electoral Force, Faces ‘Painful Moment’.” The Intercept, 9 Sept 2020 https://theintercept.com/2020/09/08/sunrise-movement- ed-markey-election/. Accessed 10 Nov 2020. Holden, Emily. “'Bold action is a winning message': climate advocates hail Ed Markey win.” Guardian, 3 Sept 2021, https://www.theguardian.com/us-news/2020/sep/03/ed-markey-primary- victory-climate-crisis. Accessed 23 May 2021. Holden, Emily. “'Bold action is a winning message': climate advocates hail Ed Markey win.” Guardian, 3 Sept 2021, https://www.theguardian.com/us-news/2020/sep/03/ed-markey-primary- victory-climate-crisis. Accessed 23 May 2021. Lavelle, Marianne. “A Fiftieth Anniversary Few Remember: LBJ’s Warning on Carbon Dioxide.” Jackson Free Press, 2 Feb 2015, https://www.jacksonfreepress.com/news/2015/feb/02/50th- anniversary-few-remember-lbjs-warning-carbon-/. Accessed 2 Nov 2020. Manchester, Julia. “How Markey took down a Kennedy.” The Hill, 1 Sept 2021, https:// thehill.com/homenews/campaign/514702-how-markey-took-down-a-kennedy. Accessed 23 May 2021. Manchester, Julia. “How Markey took down a Kennedy.” The Hill, 1 Sept 2021, https:// thehill.com/homenews/campaign/514702-how-markey-took-down-a-kennedy. Accessed 23 May 2021. Ricketts, Sam, Rita Cliffton, Lola Oduyeru and Bill Holland. “States Are Laying a Road Map for Climate Leadership.” Center for American Progress, 30 Apr 2020, https:// Ricketts, Sam, Rita Cliffton, Lola Oduyeru and Bill Holland. US Politics “States Are Laying a Road Map for Climate Leadership.” Center for American Progress, 30 Apr 2020, https:// www.americanprogress.org/issues/green/reports/2020/04/30/484163/states-laying-road-map- climate-leadership/. Accessed 20 March 2021. www.americanprogress.org/issues/green/reports/2020/04/30/484163/states-laying-road-map- climate-leadership/. Accessed 20 March 2021. Roberts, David. “The Green New Deal, Explained,” Vox, 7 Jan 2019, https://www.vox.com/energy- and-environment/2018/12/21/18144138/green-new-deal-alexandria-ocasio-cortez. Accessed 31 May 2021. Tierstein, Zoya. “This 74-year-old senator has a secret weapon: Climate activists.” Grist, 2 Sept 2020, https://grist.org/politics/ed-markey-wins-massachusetts-primary-over-joseph-kennedy/. Accessed 23 May 2021. Ward, Jon and Christopher Wilson. “Markey defeats Kennedy in Massachusetts Senate race.” MSN, 2 Sept 2020, https://www.msn.com/en-us/news/politics/markey-defeats-kennedy-in- massachusetts-senate-race/ar-BB18BZXw. Accessed 23 May 2021. White, Chris. “Ocasio-Cortez, Green New Deal Activists Score Another Big Win After Sen Markey Beats A Kennedy.” Daily Caller, 2 Sept 2020, https://dailycaller.com/2020/09/02/ed-markey-joe- kennedy-primary-green-new-deal-ocasio-cortez/. Accessed 31 May 2021. Brownstein, Robert. “The GOP’s Demographic Doom.” The Atlantic, 23 Oct 2020, https:// www.theatlantic.com/politics/archive/2020/10/millennials-and-gen-z-will-soon-dominate-us- elections/616818/. Accessed 2 Nov 2020. Clayton, Susan, Christie Manning, Kirra Krygsman and Meighen Speiser. Mental Health and Our Changing Climate: Impacts, Implications, and Guidance. American Psychological Association, Climate for Health and ecoAmerica, March 2017, https://www.apa.org/news/press/releases/2017/03/ mental-health-climate.pdf. Accessed 31 May 2021. Dunlap, Riley E., Aaron M. McCright and Jerrod H. Yarosh. “The Political Divide on Climate Change: Partisan Polarization Widens in the U.S.” Environment 58, no.5, 2016: 4–22. Fagan, Moira and Christine Huang. “A look at how people around the world view climate change.” Pew Research, 18 April 2019, https://www.pewresearch.org/fact-tank/2019/04/18/a- look-at-how-people-around-the-world-view-climate-change. Accessed 5 Nov 2020. Fagan, Moira and Christine Huang. “A look at how people around the world view climate change.” Pew Research, 18 April 2019, https://www.pewresearch.org/fact-tank/2019/04/18/a- look-at-how-people-around-the-world-view-climate-change. Accessed 5 Nov 2020. Funk, Cary and Meg Hefferon. “U.S. Public Views on Climate and Energy.” Pew Research, 25 November 2019, https://www.pewresearch.org/science/2019/11/25/u-s-public-views-on- climate-and-energy/. Accessed 31 May 2021. Funk, Cary and Meg Hefferon. “U.S. Public Views on Climate and Energy.” Pew Research, 25 November 2019, https://www.pewresearch.org/science/2019/11/25/u-s-public-views-on- climate-and-energy/. Accessed 31 May 2021. E-rea, 18.2 \| 2021 21 Youth climate activism in the United States Fry, Richard. “Millennials Overtake Baby Boomers as America’s Largest Generation.” Pew Research, 28 Apr 2020, https://www.pewresearch.org/fact-tank/2020/04/28/millennials- overtake-baby-boomers-as-americas-largest-generation. Accessed 2 Nov 2020. Fry, Richard. “Millennials Overtake Baby Boomers as America’s Largest Generation.” Pew Research, 28 Apr 2020, https://www.pewresearch.org/fact-tank/2020/04/28/millennials- overtake-baby-boomers-as-americas-largest-generation. Accessed 2 Nov 2020. Griffin, Rob, William H. Frey and Ruy Teixeira. “America’s Electoral Future: The Coming Generational Transformation.” Center for American Progress, 19 Oct 2020, https:// www.americanprogress.org/issues/politics-and-elections/reports/2020/10/19/491870/americas- electoral-future-3. Accessed 2 Nov 2020. Huizen, Jennifer. “What to know about eco-anxiety.” Medical News Today, 19 Dec 2019, https:// www.medicalnewstoday.com/articles/327354. Accessed 8 May 2021. Ranalli, Ralph. US Politics “Young voters are highly motivated, flocking to Biden, and worried about the economy.” Harvard Kennedy School, 23 Sept 2020 https://www.hks.harvard.edu/faculty- research/policy-topics/politics/young-voters-are-highly-motivated-flocking-biden-and. Accessed 2 Nov 2020. Saad, Lydia. “Global Warming Concern at Three-Decade High in US.” Gallup, 14 March 2017, www.gallup.com/poll/206030/global-warming-concern-three-decade-high.aspx. Accessed 31 May 2021. Durr, Sarah. “New Poll: Climate Change, Gun Violence the Defining Issues for Young Voters.” 24 Jan 2020, https://www.usmayors.org/2020/01/24/new-poll-climate-change-gun-violence-the- defining-issues-for-young-voters/. Accessed 10 Nov 2020. Wu, Judy, Gaelen Snell and Hasina Samji. “Climate anxiety in young people: a call to action,” The Lancet, Vol 4, Issue 10, 1 Oct 2020 https://www.thelancet.com/journals/lanplh/article/ PIIS2542-5196(20)30223-0/fulltext. Accessed 8 May 2021. Banerjee, Neela, Lisa Song and David Hasemyer. Exxon: The Road Not Taken, InsideClimate News, 2015. Climate Interactive, Climate Scoreboard. https://www.climateinteractive.org/ci-topics/climate- energy/scoreboard/frequently-asked-questions/. Accessed 17 Nov 2020. National Centers for Environmental Information (NCEI) formerly National Climatic Data Center (NCDC) of the National Oceanic and Atmospheric Administration (NOAA). Extreme Events https:// www.ncdc.noaa.gov/climate-information/extreme-events. Accessed 17 Nov 2020. Atif Ansar, Ben Caldecott and James Tilbury. “Stranded assets and the fossil fuel divestment campaign: what does divestment mean for the valuation of fossil fuel assets?” University of Oxford, 11 March 2015, https://www.smithschool.ox.ac.uk/publications/reports/SAP- divestment-report-final.pdf. Accessed 5 May 2021. O’Daly, Britton. “Climate Change Protesters Disrupt Yale-Harvard Football Game.” The New York Times, 23 Nov 2019, https://www.nytimes.com/2019/11/23/us/harvard-yale-game-protest.html. Accessed 5 Nov 2020. Watanabe, Teresa. “UC becomes nation’s largest university to divest fully from fossil fuels.” Los Angeles Times, 19 May 2020, https://www.latimes.com/california/story/2020-05-19/uc-fossil-fuel- divest-climate-change. Accessed 31 May 2021. NOTES 1. The principle that certain natural and cultural resources are preserved for public use, and that the government owns and must protect and maintain these resources for the public's use (Legal Information Institute). https://www.law.cornell.edu/wex/public_trust_doctrine. Accessed 31 May 2021. U.S. Youth Climate Strike. strikewithus.org Young Evangelicals for Climate Action. yecaction.org Zero Hour. thisiszerohour.org U.S. Youth Climate Strike. strikewithus.org Young Evangelicals for Climate Action. yecaction.org Zero Hour. thisiszerohour.org U.S. Youth Climate Strike. strikewithus.org Pipeline debate E-rea, 18.2 \| 2021 22 Youth climate activism in the United States “Dakota Access Pipeline Company Attacks Native American Protesters with Dogs & Pepper Spray.” Democracy Now, Sep 4 2016, https://www.youtube.com/watch?v=kuZcx2zEo4k, 4. Accessed 14 Nov 2020. Cagle, Susie. “Protesters as terrorists': growing number of states turn anti-pipeline activism into a crime.” Guardian, 8 July 2019, https://www.theguardian.com/environment/2019/jul/08/wave- of-new-laws-aim-to-stifle-anti-pipeline-protests-activists-say. Accessed 31 May 2021. Kusnetz, Nicolas. “How Energy Companies and Allies Are Turning the Law Against Protesters.” InsideClimate News, 22 Aug 2018, https://insideclimatenews.org/news/22082018/pipeline-protest- laws-felony-free-speech-arrests-first-amendment-oklahoma-iowa-louisiana. Accessed 14 Nov 2020. Sadasivam, Naveena. “After Standing Rock, protesting pipelines can get you a decade in prison and $100K in fines.” Grist, 14 May 2019, https://grist.org/article/after-standing-rock-protesting- pipelines-can-get-you-a-decade-in-prison-and-100k-in-fines/. Accessed 31 May 2021. Other References Sadasivam, Naveena. “After Standing Rock, protesting pipelines can get you a decade in prison and $100K in fines.” Grist, 14 May 2019, https://grist.org/article/after-standing-rock-protesting- pipelines-can-get-you-a-decade-in-prison-and-100k-in-fines/. Accessed 31 May 2021. Other References Brulle, Robert J. “The U.S. Environmental Movement”, Twenty Lessons in Environmental Sociology, New York, NY, Oxford University Press, 2013. Commoner, Barry. The Closing Circle, New York, Mineola, Dover Publications, 1971. Earth Day Network. www.earthday.net/education. Earth Day Network. www.earthday.net/education. Gamber-Thompson, Liana. “Bypassing the Ballot Box,” By Any Media Necessary: The New Youth Activism, New York University Press, 2016, 219-252. Heller, Nathan. “Is There any Point to Protesting?” New Yorker, 14 Aug 2017, https:// www.newyorker.com/magazine/2017/08/21/is-there-any-point-to-protesting. Accessed 31 May 2021. Huret, Romain and Cédric Tourbe. “Climat, une guerre américaine : quand la ‘wilderness’ tourne au drame.” Documentary film, 2020. Jenkins, Henry, Sangita Shresthova, Liana Gamber-Thompson, Neta Kligler-Vilenchik and Arely Zimmerman. By Any Media Necessary: The New Youth Activism. New York University Press, 2016. Zimmerman. By Any Media Necessary: The New Youth Activism. New York University Press, 2016. Oreskes, Naomi and Erik M. Conway. Merchants of Doubt. Bloomsbury Press, 2010. Pirages, Dennis C. Sustainable Society: Implications for Limited Growth. Praeger Publishers, 1977. Shellenberger, Michael and Ted Nordhaus. The Death of Environmentalism. 2004, https:// static1.squarespace.com/static/56a45d683b0be33df885def6/t/5c75569d104c7bbf4b4ef48e/ 1551194797246/Death_of_Environmentalism.pdf. Accessed 31 May 2021. Tilly, Charles and Sidney Tarrow. Contentious Politics, Oxford University Press, New York, 2007. Youth climate associations American Conservation Coalition. acc.eco, climatesolution.eco Divest from Fossil Fuels. gofossilfree.org Earth Uprising. earthuprising.org No Fossil Fuel Money Pledge. nofossilfuelmoney.org One Mind Youth Movement. omym.org Sunrise Movement. sunrisemovement.org, sunrisemovement.org/green-new-deal E-rea 18 2 \| 2021 Oreskes, Naomi and Erik M. Conway. Merchants of Doubt. Bloomsbury Press, 2010. Pirages, Dennis C. Sustainable Society: Implications for Limited Growth. Praeger Publishers, 1977. Pipeline debate E-rea, 18.2 \| 2021 Youth climate activism in the United States 23 INDEX Keywords: youth, climate, dissent, disruption, activism, fossil fuels, environment, Green New Deal, Sunrise Movement, protest, divestment, justice Mots-clés: Mots-clés: jeunesse, climat, activisme, énergies fossiles, environnement, Green New Deal, Sunrise Movement, justice, contestation, désinvestissement Keywords: youth, climate, dissent, disruption, activism, fossil fuels, environment, Green New Deal, Sunrise Movement, protest, divestment, justice Mots-clés: Mots-clés: jeunesse, climat, activisme, énergies fossiles, environnement, Green New Deal, Sunrise Movement, justice, contestation, désinvestissement Keywords: youth, climate, dissent, disruption, activism, fossil fuels, environment, Green New Deal, Sunrise Movement, protest, divestment, justice Mots-clés: Mots-clés: jeunesse, climat, activisme, énergies fossiles, environnement, Green New Deal, Sunrise Movement, justice, contestation, désinvestissement Keywords: youth, climate, dissent, disruption, activism, fossil fuels, environment, Green New Deal, Sunrise Movement, protest, divestment, justice ABSTRACTS E-rea, 18.2 \| 2021 24 Youth climate activism in the United States ABSTRACTS Youths are on the front line when considering their future on a planet that is becoming less and less livable. As the frequency and intensity of extreme weather events devastate regions and impact millions, the unapologetic anti-environmentalism of the Trump administration may have been the ultimate spark that ignited mass action from the young generations. Realizing, too, that many Congress members lack the political will to effect change, an increasing number of passionate youth leaders have initiated movements to raise awareness about the urgency of acting to mitigate the worst effects of climate change. Ranging from classic street protests to electoral canvassing to litigation, their actions may be characterized in terms of “disruptive, dutiful or dangerous dissent” (O’Brien et al). They have already played a key role in electing climate-conscious, progressive candidates, put pressure on governments through lawsuits and proposed an overhaul of the American economy on greener, more equitable terms by supporting the Green New Deal. In 2019 millennials (37% of US voters) surpassed baby boomers as the largest voting demographic group and could have a significant impact on climate policy. Très préoccupés par le changement climatique, les jeunes sont en première ligne sur une planète qui devient de moins en moins vivable. Alors que la fréquence et l’intensité des catastrophes naturelles ont un impact dévastateur sur des régions et des millions d’habitants, l’anti- environnementalisme avéré de l’administration Trump a sans doute été l’étincelle ultime déclenchant une mobilisation de masse des jeunes Américains. Ayant fait le constat que le Congrès manque de volonté politique pour initier le changement, un nombre croissant de leaders, jeunes et passionnés, ont créé des mouvements pour informer le public de l’urgence de la situation. Ils agissent dans l’optique de réduire les conséquences les plus graves du réchauffement de la terre. Leurs actions contestataires peuvent se ranger dans trois catégories : « disruptive, dutiful, dangerous » (O’Brien et al). Ces jeunes ont déjà joué un rôle clef dans les victoires électorales de candidats progressistes et pro-climat ; ils ont poursuivi des gouvernements en justice ; et ils ont proposé une refonte de l’économie américaine en termes verts et équitables, le « Green New Deal. » En 2019, la génération née entre 1981 et 1996 a dépassé celle des baby-boomers pour devenir le groupe démographique dominant, ce qui pourrait avoir un impact important sur la politique climatique. MELANIE MEUNIER Melanie Meunier is an American language and civilization teacher and researcher at Sciences Po Strasbourg where she teaches courses on environmental politics, climate change and the evolution of the media. She received degrees from Dartmouth College (USA), Sciences Po Paris and the University of Strasbourg, and wrote a doctoral dissertation on American environmental politics. E-rea, 18.2 \| 2021
https://openalex.org/W2574923855	https://www.banglajol.info/index.php/IMCJMS/article/download/31102/20950	English	null	Prophylactic intracameral vancomycin: efficacy in preventing endophthalmitis after cataract surgery	IMC journal of medical science	2,017	cc-by	2,369	Address for Correspondence: Dr. Manash Kumar Goswami, Associate Professor, Department of Ophthalmology, BIRDEM General Hospital, 122 Kazi Nazrul Islam Avenue, Shahbag, Dhaka. Email: manashkg@yahoo.com Original Article Original Article Introduction Intraocular infection following cataract surgery and lens implantation is a rare but dreadful complication that can have devastating consequences for sight. The Endophthalmitis Vitrectomy Study (EVS) estimated that up to half of patients developing endophthalmitis following cataract surgery had visual acuity worse than 20/40 and 15% had worse than 5/200 [1]. The incidence of endophthalmitis following cataract surgery varies between 0.04 and 0.27% [2-4]. Even though several risk factors have been identified, this complication is often unexpected, with variable response to standard treatment. The number of cataract surgeries is expected to rise with an increasing aging population in Bangladesh; the burden of endophthalmitis is likely to rise. A robust preventative strategy is needed to contain Abstract Background and objective: Post Operative endophthalmitis is rare but devastating complication in ocular surgery. The present study determined the efficacy of intracameral vancomycin after phaco-emusification cataract surgery to prevent endophthalmitis. Method: A total of 768 cases who had undergone phaco-emusification cataract surgery were included in the study. Every alternate patient received 0.5 ml injection of vancomycin (1mg in 0.1 ml) in the anterior chamber after completion of phaco-emulcification and formation of anterior chamber. All the patients were examined for symptoms and signs of bacterial endophthalmitis at 24 hrs, 7 days, 15 days and subsequently at 1, 3 and 6 months following surgery. Results: No endophthalmitis case was recorded at any time period during 6 month follow up in either group. However, significantly higher number of cases in vancomycin group had cells in anterior chamber and disturbances in visual acuity at day 15 following surgery. Conclusion: Vancomycin did not have any prophylactic role in preventing endophthalmitis. Proper aseptic measures are important to prevent any infection in ocular surgery. IMC J Med Sci 2016; 10(1): 24-28 this serious complication. The European Society of Cataract and Refractive Surgeons (ESCRS) in a multinational, partially-masked placebo-controlled trial have provided strong evidence for using intracameral antibiotics in preventing postoperative endophthalmitis following cataract surgery [5]. In this study, we determined the prophylactic effect of vancomycin injection in the anterior chamber at the end of phacoemulsification surgery in preventing post operative bacterial endophthalmitis. Manash Kumar Goswami, Md. Ferdous Hossain and Md. Asaduzzaman Department of Ophthalmology, Bangladesh Institute of Research and Rehabilitation in Diabetes, Endocrine and Metabolic Disorder (BIRDEM), Dhaka Prophylactic intracameral vancomycin: efficacy in preventing endophthalmitis after cataract surgery Manash Kumar Goswami, Md. Ferdous Hossain and Md. Asaduzzaman Original Article Original Article Methods A total of 768 cases undergoing phacoemulsification cataract surgery from January 2014 to December 2015 in the Department of Ophthalmology at BIRDEM general hospital were enrolled in the study. Every alternate eye (patient) received IMC J Med Sci 2016; 10(1): 24-28 Goswami MK et al. 25 day 15 and at 6 month were 44 and 23 cases in control group; 10 and 15 cases in vancomycin group respectively. No patients in any group developed features of bacterial endophthalmitis during this period. At 24 hrs, significantly higher proportion of cases had cells in anterior chamber in vancomycin group compared to control (22.3% vs 9.3%; p<0.05). Visual acuity was significantly (p<0.05) poor in vancomycin group (20.6%) compared to control (8.3%) at 15 day follow up. However, the visual acuity was similar in both groups at 6 month. intracameral injection of vancomycin (vancomycin group) while others did not receive any intracameral antibiotic and served as a control. The cases in control arm were treated according to the exiting protocol followed in the hospital for cataract surgery as described below. In the anesthetic room, 10 minutes prior to cataract surgery, each patient had periocular skin cleaning with 5% povidone iodine. Two drops of 5% povidone iodine were instilled into the conjunctival sac at this stage. The surgeon, after scrubbing and draping, used 5% povidone iodine to cleanse the eyelids and periocular skin. The eyelids including the eye lashes were draped. All cataract surgeries were performed in a dedicated ophthalmic theatre assisted by resident doctor and ophthalmic specialist nurses. Phacoemulsification was performed using infiniti machine (Alcon). Every alternate patient got 0.5 ml injection of vancomycin (1mg in 0.1 ml) in the anterior chamber of the eye after completion of phaco-emulcification and formation of anterior chamber [5]. Postoperatively, patients were prescribed prednisolone acetate 1% eye drop two- hourly for two weeks which was then tapered off over a further four weeks. In addition, all cases received topical moxifloxacin eye drop two hourly for 1 week and 6 hourly for 2 weeks. All the patients were examined for symptoms and signs of bacterial endophthalmitis [4] at 24 hrs, 7days, and 15 days and subsequently at 1, 3 and 6 months following surgery. Methods Table-1: Results of follow up of study population at 24 hrs, 15 day and 6 month following phacoemulsification cataract surgery Parameters Without vancomycin (Control) N (%) With vancomycin N (%) At 24 hrs Number of cases (n) Visual acuity 6/6-6/9 6/12-6/24 6/36-above Conjunctival congestion A/C cells ++ Endophthalmitis At 15 day Number of cases Drop out Visual acuity 6/6-6/9 6/12-6/24 6/36-above Endophthalmitis At 6 month Number of case Drop out Visual acuity 6/6-6/9 6/12-6/24 6/36-above Endophthalmitis 384 201 (52.3) 105 (27.3) 78 (20.3) 67 (17.4) 36 (9.3) 0 340 44 308 (90.6) 25 (7.4) 7 (2.1) 0 361 23 318 (88.1) 37 (10.2) 6 (1.7) 0 384 187 (48.6) 151 (39.3) 46 (11.9) 78 (20.3) 86 (22.3) 0 374 10 297 (79.4) 47 (12.6) 30 (8.0) 0 369 15 315 (85.4) 49 (13.3) 5 (1.4) 0 Note: A/C: anterior chamber; p<0.05 when compared between vancomycin and control group for visual acuity at day 15. Table-1: Results of follow up of study population at 24 hrs, 15 day and 6 month following phacoemulsification cataract surgery The study was approved by the IRB of BADAS and informed written consent was obtained from each patient. Discussion operation theater. In the present series there was no single case of endophthalmitis in either group. However, cells in the anterior chamber immediately following surgery was significantly higher in vancomycin group compared to control who did not receive vancomycin. Similarly, the visual acuity was significantly poor in vancomycin group at 15 day follow up. But the visual acuity in vancomycin group gradually improved over time and was same as the control at 6 month. The reduced visual acuity observed among the patients in both group could be due to the presence of other co-morbid conditions like diabetes, aging, hypertension, etc. Though minimal, patient should be monitored if vancomycin is used as prophylactic antibiotic in cataract surgery. There was no adverse effect like cystoid macular edema in intracameral vancomycin group. Significant increase in incidence of cystoid macular edema has been reported with the use of intracameral vancomycin during extracapsular cataract extraction [20]. It is interesting to note that no endophthalmitis has occurred in our cases who did not receive vancomycin as prophylaxis indicating that proper aseptic technique was adequate alone to prevent endophthalmitis and other infections following eye surgery. A recent survey on antibiotic prophylaxis practice after cataract surgery in the UK has reported that 37% of the respondents did not use intracameral antibiotic prophylaxis [21]. Anterior segment intraocular surgery frequently introduces bacteria into the anterior chamber [6]. In 65% to 100% of healthy subjects, conjunctiva and adnexa were found to be colonized by microorganisms with the potential to cause endophthalmitis [7, 8]. In 82% of cases of endophthalmitis, the organism isolated was the same species as that found in the conjunctiva and lids of the same patient [9]. In the 420 cases of endophthalmitis in the EVS series, 70% of the isolates were coagulase negative staphylococci. Prophylactic measures to reduce postoperative endophthalmitis after cataract surgery are wide and varied. Several perioperative and postoperative methods have been recognized. Preoperative iodine antisepsis combined with preoperative and postoperative topical antibiotic therapy is considered the standard of care for this purpose [10]. In the last decade intracameral administration of antibiotics has evolved as a prophylactic to prevent postoperative endophthalmitis after cataract surgery, claiming superior effectiveness compared with standard care. Intracameral antibiotics can achieve high concentrations in the anterior chamber but have the potential to cause corneal endothelial toxicity [11]. Discussion However, cefuroxime and vancomycin have been reported safe on human corneal endothelial cells in clinically used concentration, but at higher concentrations it could cause irreversible cell death. The present study demonstrated that proper pre and per operative aseptic measures together with the post operative use of standard topical antibiotics is adequate to prevent the bacterial endophthalmitis in cataract surgery. Routine use of prophylactic intracameral vancomycin did not have any additive value for the prevention of post operative bacterial endophthalmitis. The choice of intracameral antibiotic prophylaxis appears to be empirical, based on the fact that the majority of isolates are Gram-positive organisms. The first report of intracameral antibiotic prophylaxis with gentamicin was by Peyman et al [12]. Later, Gimbel et al [13] reported no case of endophthalmitis in a series of 12,000 patients after using an infusion fluid of gentamicin together with vancomycin during cataract surgery. The choice of prophylactic antibiotics included cefuroxime, cefazolin and vancomycin based on the pattern of organisms, antibiotic resistance profile, cost and possible adverse effects [14-19]. Results A total of 768 cases were enrolled in the study of which 384 were in vancomycin group and 384 were in control group. The age range of the study population was 45-80 years. Male female distribution in vancomycin group was 283 and 101 respectively while in control group it was 237 and 147. All had type 2 diabetes mellitus of various durations. During the follow up, visual acuity, conjunctival congestion, cells in the anterior chamber and features of endophthalmitis were recorded. Table 1 shows the results of follow up at 24 hrs, 15 day and 6 month. Lost to follow up at 26 IMC J Med Sci 2016; 10(1): 24-28 Goswami MK et al. References 1. Endophthalmitis Vitrectomy Study Group. Results of the Endophthalmitis Vitrectomy Study: A randomized trial of immediate vitrectomy and of intravenous antibiotics for the treatment of postoperative bacterial endophthalmitis. Arch Ophthalmol 1995; 113: 1479–1496. [PubMed] In our study, we serially enrolled 384 patients undergoing phacoemulsification cataract surgery at BIRDEM general hospital over two year’s period. Meticulous aseptic technique was followed in the IMC J Med Sci 2016; 10(1): 24-28 Goswami MK et al. 2. Taban M, Beherens A, Newcomb RL, et al. Acute endophthalmitis following cataract surgery: A systematic review of the literature. Arch Ophthalmol 2005; 123: 613– 620. [PubMed] 10. Ciulla TA, Starr MB, Masket S. Bacterial endophthalmitis prophylaxis for cataract surgery: An evidence-based update. Ophthalmology 2002: 10913-10924. [PubMed] 11. Yoeruek E, Spitzer MS, Saygili O, et al. Comparison of in vitro safety profiles of vancomycin and cefuroxime on human corneal endothelial cells for intracameral use. J Cataract Refract Surg 2008; 34: 2139- 2145. [PubMed] 3. Hatch WV, Cernat G, Wong D, Devenyi R, Bell CM. Risk factors for acute endophthalmitis after cataract surgery: A population-based study. Ophthalmology 2009; 116: 425-430. [PubMed] 12. Peyman GA, Sathar MI, May DR. Intraocular gentamicin as intraoperative prophylaxis in South India eye camps. Br J Ophthalmol 1977; 61: 260-262. [PMC free article] [PubMed] 4. Miller JJ, Scott IU, Flynn HW, Jr, Smiddy WE, Newton J, Miller D. Acute onset endophthalmitis after cataract surgery (2000– 2004): Incidence, clinical settings, and visual acuity outcomes after treatment. Am J Ophthalmol 2005; 139: 983-987. [PubMed] 13. Gimbel HV, Sun R, DeBroff BM. Prophylactic intracameral antibiotics during cataract surgery: The incidence of endophthalmitis and corneal endothelial cell loss. Eur J Implant Refract Surg 1994; 6: 280-285. 5. Barry P, Seal DV, Gettinby G, Lees F, Peterson M, Revie CW. ESCRS study of prophylaxis of postoperative endophthalmitis after cataract surgery: Preliminary report of principal results from a European multicenter study; the ESCRS Endophthalmitis Study Group. J Cataract Refract Surg 2006; 32: 407- 410. [PubMed] 14. Garat M, Moser CL, Alonso-Tarres C, Martin-Baranera M, Alberdi A. Intracameral cefazolin to prevent endophthalmitis in cataract surgery: 3-year retrospective study. J Cataract Refract Surg 2005; 31: 2230-2234. [PubMed] 6. Dickey JB, Thompson KD, Jay WN. Anterior chamber aspirate cultures after uncomplicated cataract surgery. Am J Ophthalmol 1991; 112: 278-282. [PubMed] 15. Montan PG, Wejde G, Setterquist H, Rylander M, Zetterstom C. Prophylactic intracameral cefuroxime: Evaluation of safety and kinetics in cataract surgery. References J Cataract Refract Surg 2002; 28: 982-987. [PubMed] 7. Speaker MG, Milch FA, Shah MK, Eisner W, Kreiswirth BN. Role of external bacterial flora in the pathogenesis of acute postoperative endophthalmitis. Ophthalmology 1991; 98: 639-649. [PubMed] 16. Hospital Infection Control Practices Advisory Committee (HICAC) Recommendations for preventing the spread of vancomycin resistance. Infect Control Hosp Epidemiol 1995; 16: 105–113. [PubMed] 8. Boes DA, Lindquist TD, Firtsche TR, Kalima RE. Effects of povidone iodine chemical preparation and saline irrigation on the perlimbal flora. Ophthalmology 1992; 99: 1569-1573. [PubMed] 17. Garat M, Moser CL, Alonso-Tarres C, Martin-Baranera M, Alonso-Tarres C, Alvarez-Rubio L. Prophylactic intracameral cefazolin after cataract surgery: Endophthalmitis risk reduction and safety results in a 6-year study. J Cataract Refract Surg 2009; 35: 637-642. [PubMed] 9. Bannerman TL, Rhoden DL, McAllister SK, Miller JM, Wilson LA. The source of coagulase-negative staphylococci in the Endophthalmitis Vitrectomy Study: A comparison of eyelid and intraocular isolates using pulsed-field gel electrophoresis: The Endophthalmitis Vitrectomy Study Group. Arch Ophthalmol 1997; 115: 357- 361. [PubMed] 18. Montan PG, Wejde G, Koranyi G, Rylander M. Prophylactic intracameral cefuroxime: efficacy in preventing endophthalmitis after cataract surgery. J Cataract Refract Surg 2002; 28: 977-981. [PubMed] 28 IMC J Med Sci 2016; 10(1): 24-28 28 IMC J Med Sci 2016; 10(1): 24-28 Goswami MK et al. Cystoid macular edema after cataract surgery with intraocular vancomycin. Ophthalmology 1999; 106(9): 1660-1664. [PubMed] 19. Seal DV, Barry P, Gettinby G, et al. ESCRS study of prophylaxis of postoperative endophthalmitis after cataract surgery: Case for a European multicenter study; the ESCRS Endophthalmitis Study Group. J Cataract Refract Surg 2006; 32: 396-406. [PubMed] 21. Gore DM, Angunawela RI, Little BC. United Kingdom survey of antibiotic prophylaxis practice after publication of the ESCRS Endophthalmitis Study. J Cataract Refract Surg 2009; 35: 770-773. [PubMed] 20. Axer-Siegel R, Stiebel-Kalish H, Rosenblatt I, Strassmann E, Yassur Y, Weinberger D.
https://openalex.org/W2784514039	https://figshare.com/articles/journal_contribution/Deciphering_Functions_of_Intracellular_Formaldehyde_-_Linking_Cancer_and_Aldehyde_Metabolism/10211633/1/files/18411584.pdf	English	null	Deciphering Functions of Intracellular Formaldehyde: Linking Cancer and Aldehyde Metabolism	Biochemistry	2,018	cc-by	1,691	Deciphering Functions of Intracellular Formaldehyde - Linking Cancer and Aldehyde Metabolism Richard J. Hopkinsona and Christopher J. Schofieldb* aLeicester Institute of Chemical and Structural Biology, University of Leicester, Henry Wellcome Building, Lancaster Road, Leicester, LE1 7RH, United Kingdom. aLeicester Institute of Chemical and Structural Biology, University of Leicester, Henry Wellcome Building, Lancaster Road, Leicester, LE1 7RH, United Kingdom. bDepartment of Chemistry, University of Oxford, Chemistry Research Laboratory, 12 Mansfield Road, Oxford, OX1 3TA, United Kingdom. E-mail: christopher.schofield@chem.ox.ac.uk Formaldehyde, the simplest aldehyde, is an environmental pollutant and human toxin. Acute exposure to exogenous formaldehyde can cause irritation, nausea, renal failure, cardiac arrhythmia, and coma. Chronic formaldehyde exposure correlates with increased cancer incidence, in particular of nasopharyngeal cancer and leukaemia. In addition to exogenous sources, formaldehyde is produced endogenously in cells; eukaryotic pathways producing formaldehyde include xenobiotic metabolism and enzyme-catalysed N-methyl demethylation of the N-methylated histone and DNA components of chromatin, as well as of RNA. Thus, endogenously produced formaldehyde may have biological roles; there have been very few studies to understand such functions at the biochemical level and even fewer connecting the biochemistry of formaldehyde with physiology. The carbonyl group of formaldehyde is electrophilic, reacting efficiently with nucleophiles, including water; in aqueous solution, formaldehyde is predominantly in its hydrated form, (H2C(OH)2). It has long been known that reactions of formaldehyde with biologically relevant nucleophiles have potential for complexity, such as the formation of oligomeric, cyclised, and disproportionated products, including via Mannich and Cannizzaro type reactions.1, 2 Many, but not all, of formaldehyde’s reactions, including hydration, are readily reversible with the degree of reversibility depending on the nucleophile, conditions, and nature of the reaction product. Formaldehyde’s toxicity is proposed to result from its reaction with nucleophiles in cells, i.e. with nucleic acid bases and nucleophilic amino acid side-chains (such as the cysteinyl thiol and lysyl Nε-amino groups). Reactions with small molecules are important in formaldehyde metabolism and may be relevant to its pathophysiological effects. Formaldehyde detoxification involves its reaction with nucleophilic thiols: in plants, animals, and some bacteria, formaldehyde reacts with glutathione to give S-hydroxymethylglutathione, which is the substrate for an NAD-dependent dehydrogenase (ADH5 in humans). The product, S- formylglutathione, is then subjected to enzymatic hydrolysis, giving formate, which is less reactive than formaldehyde, and recycling glutathione (Figure 1). A similar pathway in mycobacteria uses mycothiol instead of glutathione. Figure 1. Recent work has shown how formaldehyde links the glutathione and tetrahydrofolate (THF) ‘metabolic systems’ (only selected elements of these are shown). Deciphering Functions of Intracellular Formaldehyde - Linking Cancer and Aldehyde Metabolism HCHO is toxic above threshold levels and is detoxified via its reaction with the cysteinyl thiol of glutathione. HCHO can also act as a C1 precursor for THF-mediated metabolism. THF and some of its derivatives are susceptible to oxidation, releasing formaldehyde (dashed arrows). SHMT, serine hydroxymethyltransferase; ADH5, alcohol dehydrogenase 5. Figure 1. Recent work has shown how formaldehyde links the glutathione and tetrahydrofolate (THF) ‘metabolic systems’ (only selected elements of these are shown). HCHO is toxic above threshold levels and is detoxified via its reaction with the cysteinyl thiol of glutathione. HCHO can also act as a C1 precursor for THF-mediated metabolism. THF and some of its derivatives are susceptible to oxidation, releasing formaldehyde (dashed arrows). SHMT, serine hydroxymethyltransferase; ADH5, alcohol dehydrogenase 5. Figure 1. Recent work has shown how formaldehyde links the glutathione and tetrahydrofolate (THF) ‘metabolic systems’ (only selected elements of these are shown). HCHO is toxic above threshold levels and is detoxified via its reaction with the cysteinyl thiol of glutathione. HCHO can also act as a C1 precursor for THF-mediated metabolism. THF and some of its derivatives are susceptible to oxidation, releasing formaldehyde (dashed arrows). SHMT, serine hydroxymethyltransferase; ADH5, alcohol dehydrogenase 5. Insights into the mechanisms by which elevated levels of formaldehyde and acetaldehyde cause toxicity have emerged from ongoing studies from K.J. Patel and co-workers concerning the origins of Fanconi Anaemia (FA). FA is a rare genetic disorder leading to an increased cancer risk, which is characterised by mutations causing deficiencies in a specific DNA Insights into the mechanisms by which elevated levels of formaldehyde and acetaldehyde cause toxicity have emerged from ongoing studies from K.J. Patel and co-workers concerning the origins of Fanconi Anaemia (FA). FA is a rare genetic disorder leading to an increased cancer risk, which is characterised by mutations causing deficiencies in a specific DNA damage repair pathway. In the breakthrough work, robust evidence linking the FA-associated DNA repair pathway and impaired aldehyde metabolism has been accrued. The research demonstrates that a functional glutathione-dependent formaldehyde metabolism pathway is required for survival of FA models (Rosado et al).3 The observations suggest that elevated levels of endogenous formaldehyde lead to DNA damage, which is not tolerated in FA repair- deficient cells. The initial work has been supported by subsequent studies with ADH5-deficient FA mice (Pontel et al),4 which manifest toxic phenotypes including kidney failure, bone marrow dysfunction, and liver cancer. Deciphering Functions of Intracellular Formaldehyde - Linking Cancer and Aldehyde Metabolism The same symptoms are observed in human DNA cross-link repair deficient syndromes; DNA cross-linking is induced by drugs such as cis-platin and electrophiles, notably including formaldehyde. The identification of links between FA and aldehydes stimulated efforts to identify physiologically relevant sources of cellular formaldehyde in mammalian cells (Burgos- Barragan et al).5 Given the relatively high levels of formaldehyde in blood (20-100µm), the Patel group hypothesised that the ubiquitous folate-utilising C1 metabolic cycle, which is crucial in nucleotide/nucleic acid and amino acid/protein biosynthesis (Figure 1), is a source of endogenous formaldehyde. 5,10-Methylenetetrahydrofolate, which is formed by reaction of tetrahydrofolate (THF) with formaldehyde, fragments to give formaldehyde and tetrahydrofolate (THF); it was initially proposed that addition of THF to cells would sequester formaldehyde and so reduce its concentration. However, it was observed that addition of THF was cytotoxic, especially to cells deficient in ADH5 and in FA models. Sensitivity to added THF was tested in cells deficient in both ADH5 and C1 metabolism genes, including serine hydroxymethyltransferases 1/2 (which enable formation of 5,10- methylenetetrahydrofolate), mitochondrial folate transporter (which regulates mitochondrial THF levels), and folylpolyglutamate synthase (which is crucially involved in folate homeostasis). Diminished C1 metabolism is toxic; however, supplementation with glycine, hypoxanthine, or thymidine enables cell growth. All the tested cells types were similarly THF sensitive, suggesting that THF toxicity is independent of ‘direct’ C1 metabolism. It was also noted that deletion of ADH5 exacerbated the growth defect in the C1 metabolism-deficient cells without supplementation. Studies on the oxidative decomposition of THF (and derivatives) reveal it reacts to release formaldehyde, the production of which is promoted or inhibited by hydrogen peroxide or ascorbate, respectively. Assessment of the toxicity of THF derivatives, folate, dihydrofolate, 5-methyltetrahydrofolate and 5-formyltetrahydrofolate, using C1 metabolism/ADH5-deficient cell lines implies that only the derivatives susceptible to oxidative formaldehyde release are cytotoxic. Whilst the folate derivative associated toxicity appears to be principally due to formaldehyde release, it was observed that addition of either formaldehyde or THF rescued growth in C1 metabolism-deficient cells (when not appropriately supplemented). The THF rescue effect was dependent on the presence of active ADH5, implying a role for the glutathione-dependent metabolism. Mass spectrometric analysis of cells treated with 13C-formaldehyde revealed that a significant proportion of cellular formate is derived from ADH5-dependent oxidation of formaldehyde (blood formate levels are decreased in ADH5-deficient mice). Interestingly, incorporation of 13C into deoxyadenosine, adenosine triphosphate, and thymidine was observed. Deciphering Functions of Intracellular Formaldehyde - Linking Cancer and Aldehyde Metabolism Incorporation of the 13C label into the nucleotides was most prevalent in C1 metabolism-deficient cells and inhibited by loss of ADH5; though some incorporation was observed in ADH5-deficient cells. These observations define ADH5-dependent links between formaldehyde metabolism and the C1 metabolic cycle – as well as being toxic, formaldehyde can act as a C1 precursor, including to mediate nucleotide synthesis (via formate). Upregulating formaldehyde levels (e.g. by drug or diet therapy) in combination with ADH5 inhibition may be a way to selectively target cancer cells with defective DNA repair pathways, such as in BRCA mutant cancers. The Patel group’s work has thus identified physiologically relevant links between formaldehyde metabolism and DNA damage repair / genotoxicity. Formaldehyde can act as both a genotoxin and as a benign metabolically relevant C1 metabolite. The balance between these roles is presumably altered by context-dependent factors, including concentration, localisation, pH, and, as hinted at in the work, redox state. The findings provide a mechanistic rationale for formaldehyde genotoxicity, and reveal the potential of altered levels of formaldehyde, and other aldehydes including acetaldehyde, to regulate multiple cell functions in health and disease. Folate is widely used as dietary supplement because its deficiency causes birth and neural defects. This, coupled with the widespread societal use of the alcohol precursors of aldehydes, and of formaldehyde/formaldehyde derivatives, means the results may have implications far beyond academic work on genetic diseases. Funding We thank the Biotechnology and Biological Sciences Research Council (BB/E527620/1), Cancer Research UK (C8717/A18245), and the Wellcome Trust (091857/7/10/7) for funding our work with formaldehyde. RJH acknowledges a William R. Miller Junior Research Fellowship, St. Edmund Hall, Oxford, UK. References References [1] Mannich, C., and Krösche, W. (1912) Ueber ein Kondensationsprodukt aus Formaldehyd, Ammoniak und Antipyrin, Archiv der Pharmazie 250, 647-667. [2] Cannizzaro, S. (1853) Ueber den der Benzoësäure entsprechenden Alkohol, Justus Liebigs Annalen der Chemie 88, 129-130. [3] Rosado, I. V., Langevin, F., Crossan, G. P., Takata, M., and Patel, K. J. (2011) Formaldehyde catabolism is essential in cells deficient for the Fanconi anemia DNA- repair pathway, Nat Struct Mol Biol 18, 1432-1434. [4] Pontel, Lucas B., Rosado, Ivan V., Burgos-Barragan, G., Garaycoechea, Juan I., Yu, R., Arends, Mark J., Chandrasekaran, G., Broecker, V., Wei, W., Liu, L., Swenberg, James A., Crossan, Gerry P., and Patel, Ketan J. (2015) Endogenous Formaldehyde Is a Hematopoietic Stem Cell Genotoxin and Metabolic Carcinogen, Molecular Cell 60, 177-188. [5] [5] Burgos-Barragan, G., Wit, N., Meiser, J., Dingler, F. A., Pietzke, M., Mulderrig, L., Pontel, L. B., Rosado, I. V., Brewer, T. F., Cordell, R. L., Monks, P. S., Chang, C. J., Vazquez, A., and Patel, K. J. (2017) Mammals divert endogenous genotoxic formaldehyde into one-carbon metabolism, Nature 548, 549. TOC Graphic TOC Graphic TOC Graphic
https://openalex.org/W4319920571	https://seer.ufu.br/index.php/sociedadenatureza/article/download/67033/35570	Portuguese	null	A susceptibilidade geoambiental da bacia hidrográfica do Ribeirão Lajeado - Tocantins	Sociedade & natureza	2,023	cc-by	7,605	Resumo O estudo geoambiental através da integração dos elementos naturais com o uso e a cobertura da terra é uma proposta metodológica para o planejamento estratégico. Nesse sentido, o objetivo deste artigo é realizar uma análise da susceptibilidade geoambiental, a partir do uso de análise multicritério, na bacia hidrográfica do Ri- beirão Lajeado. Para isso, realizou-se a organização de materiais cartográficos: rede de drenagem, a declividade, a litologia, os solos, o uso e a cobertura da terra e as vias de acesso. Foram estabelecidos pesos normalizados a partir da técnica do pro- cesso analítico hierárquico (AHP). Esse cruzamento gerou o mapa de suscepti- bilidade geoambiental, definida em quatro unidades, a saber: I) baixa (área de topo plano com cobertura vegetal e nenhuma proximidade com as vias acessos); II) mé- dia (locais com algumas ocupações humanas ou com relevo suave ondulado); III) alta (áreas onde apresentam ampla ocorrência de estradas e rede de drenagem abundante, susceptíveis à intervenção antrópica, relacionadas principalmente as atividades agropecuária e áreas urbanas) e IV) muito alta (locais com estradas e porções de uso agropecuário com declividades maiores que 15%, presença de cica- trizes de fogo e áreas urbanas. Assim, a técnica AHP apresenta-se como uma im- portante ferramenta para o geoambiental, principalmente na definição de pesos para os fatores analisados. Keywords Land Use and Cover Relief Integrated Analysis AHP Method DOI: 10.14393/SN-v35-2023-67033 Recebido: 21 Setembro, 2022 \|Aceito: 10 Janeiro, 2023 \| Publicado: 09 Fevereiro, 2023 Artigos The geoenvironmental susceptibility of the Ribeirão Lajeado watershed – Tocantins Lucas da Silva Ribeiro1 Luis Eduardo de Souza Robaina2 Sandro Sidnei Vargas de Cristo3 Palavras-chave: Uso e Cobertura da Terra Relevo Análise Integrada Método AHP 1 1 Universidade Federal de Santa Maria – UFSM, Santa Maria - RS, Brasil. lucassilvaribeiro12@gmail.com 2 Universidade Federal de Santa Maria – UFSM, Santa Maria - RS, Brasil. lesrobaina@yahoo.com.br 3 Universidade Federal do Tocantins – UFT, Porto Nacional - RS, Brasil. sidneicristo@mail.uft.edu.br Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 Palavras-chave: Uso e Cobertura da Terra Relevo Análise Integrada Método AHP A susceptibilidade geoambiental da bacia hidrográfica do Ribeirão Lajeado - Tocantins The geoenvironmental susceptibility of the Ribeirão Lajeado watershed – Tocantins 1 Universidade Federal de Santa Maria – UFSM, Santa Maria - RS, Brasil. lucassilvaribeiro12@gmail.com Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 Keywords Land Use and Cover Relief Integrated Analysis AHP Method , , @g 2 Universidade Federal de Santa Maria – UFSM, Santa Maria - RS, Brasil. lesrobaina@yahoo.com.br 3 Universidade Federal do Tocantins – UFT, Porto Nacional - RS, Brasil. sidneicristo@mail.uft.edu.br INTRODUÇÃO e planejamento de uma área. Para Abreu et al. (2020), o geoambiental é uma proposta metodológica fundamental para o planejamento estratégico do território, em seus diferentes usos e níveis de exploração. A partir da integração das atividades humanas com os elementos naturais é possível desenvolver um zoneamento de susceptibilidades ambientais que possibilita planejar estratégias que auxiliam na conservação da natureza, no planejamento e na gestão estratégica para ocupação e uso humano. Esse estudo realiza uma análise geoambiental na bacia hidrográfica do Ribeirão Lajeado, através do zoneamento, aplicando o Processo Analítico Hierárquico (AHP) com a modelagem espacial em ambiente SIG (Sistemas de informações geográficas) (SAATY, 1991). O processo AHP focaliza o estudo de sistemas por meio de uma sequência de comparações, aos pares, dos condicionantes (elementos) que têm influência no sistema considerado. Marques e Zuquette (2004), Marchiori-Faria e Augusto Filho (2010), Paula e Cerri (2012), França et al. (2019) discorrem sobre o emprego do AHP na seleção e priorização de áreas com variados objetivos. Análise integrada dos dados obtidos e conhecimento do potencial dos recursos naturais incluem estudos de todos os componentes do estrato geográfico, considerando o princípio de que a natureza possui funcionalidade intrínseca entre seus componentes (ROSS, 1994). Dessa forma, destaca-se o estudo geoambiental, que pode ser entendido como todo processo envolvido na obtenção, análise, representação, comunicação e aplicação de dados e informações do meio físico, considerando as potencialidades e as fragilidades do terreno, bem como os perigos, os riscos, os impactos e os conflitos decorrentes da interação entre as ações humanas e o meio fisiográfico (VEDOVELLO, 2004). Segundo Medeiros e Cestaro (2020) o geoambiental identifica unidades de forma integrada, considerando a compartimentação de um sistema a partir da interação dinâmica de elementos físicos e antrópicos. A bacia hidrográfica do Ribeirão Lajeado (Figura 1) abrange uma área com cerca de 616,29 km², e está localizada na porção central do estado do Tocantins, nos municípios de Aparecida do Rio Negro, Lajeado, Tocantínia e Palmas. Destaca-se, a presença de duas Unidades de Conservação da Natureza (UCs), uma de Proteção Integral, o Parque Estadual do Lajeado (PEL) e a outra de Uso Sustentável, APA Serra do Lajeado, demonstrando sua importância por ser o principal sistema hidrográfico das UCs. Carvalho et al. (2021) apontam que o geoambiental tem ganhado importância nos estudos científicos, pois auxilia na compreensão das condições ambientais relacionadas às ações humanas. Abstract The geoenvironmental study through the integration of natural elements with the use and land cover is a methodological proposal for strategic planning. In this sense, the objective of this article is to perform an analysis of geoenvironmental susceptibility from the use of multicriteria in the watershed of Ribeirão Lajeado. For this, the organization of cartographic materials was carried out, such as the drainage network, slope, lithology, soils, land use and cover and access routes, where normalized weights were established from the technique of the Hierarchical Analytical Process (AHP). This crossing generated the geoenvironmental susceptibility map, defined in four units, namely: I) low (flat top area with vegetation cover and no proximity to the access roads); II) medium (places with some human occupations or with wavy smooth relief); III) high (areas with wide occurrence of roads and abundant drainage network, susceptible to anthropic intervention, mainly related to agricultural activities and urban areas) and IV) very high (places with roads and portions of agricultural use with slopes greater than 15%, presence of fire scars and urban areas. Thus, the AHP technique presents itself as an important tool for geoenvironmental, especially in defining weights for the factors analyzed. 1 1 Universidade Federal de Santa Maria – UFSM, Santa Maria - RS, Brasil. lucassilvaribeiro12@gmail.com 2 Universidade Federal de Santa Maria – UFSM, Santa Maria - RS, Brasil. lesrobaina@yahoo.com.br 3 Universidade Federal do Tocantins – UFT, Porto Nacional - RS, Brasil. sidneicristo@mail.uft.edu.br Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 1 RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 INTRODUÇÃO Destacando o estudo integrado como principal característica do geoambiental, Robaina e Trentin (2021) relatam que a análise dos elementos do substrato rochoso, relevo e solos, no que envolve o uso e a cobertura da terra, é fundamental para a compreensão da dinâmica geomorfológica e no auxílio ao manejo As UCs são espaços territoriais que apresentam recursos naturais com características relevantes. Tais áreas são utilizadas como estratégia para a proteção da natureza. É neste contexto que o uso de geotecnologias, como os Sistemas de Informações Geográficas e o Sensoriamento Remoto, otimiza a diagnose e o monitoramento ambiental (DIAS; MARTINS; BARROS, 2020). 2 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental Figura 1 - Localização da área de estudo. Fonte: Os autores (2020). Figura 1 - Localização da área de estudo. Fonte: Os autores (2020). Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 MATERIAIS E MÉTODOS As informações litológicas foram obtidas no site do Serviço Geológico do Brasil (CPRM) na escala de 1: 250.000 em formato shapefile que correspondem às cartas geológicas de Porto Nacional (SC.22-Z-B) e Miracema do norte (SC- 22-X -D). Na bacia hidrográfica, foram identificados a Cobertura Detrito Laterítica, a Formação Pimenteiras, a Formação Jaicós e a Granitóides. Para o estudo da susceptibilidade geoambiental, realizou-se a integração de informações sobre a rede de drenagem, declividades, litologias, solos e uso e cobertura da terra. Esses dados foram organizados no software Qgis na versão 3.14.16, onde foi possível aplicar o cruzamento automático dessas informações a partir de pesos normalizados para cada mapa, definido pelo método do AHP. Os dados sobre os solos foram obtidos através do trabalho Sousa et al. (2012), representando a SEPLAN do Tocantins. Ao analisar as informações de solos e comparando com a declividade na bacia hidrográfica, notou-se a necessidade de realizar alguns ajustes conforme a relação solo-declividade. Segundo Nowatzki e Santos (2014), os diferentes tipos de solos podem ser compartimentados de acordo com as características do relevo. Nesse sentido, os dados de solos foram ajustados com base na declividade e com apoio de trabalhos de campo, definindo-se as seguintes classes: Latossolo, Plintossolo, Associação Neossolo-Cambissolo e Associação Argissolo-Cambissolo. A imagem de radar da missão Shuttle Radar Topographic Mission (SRTM), com resolução espacial de 30 metros, foi obtida na plataforma topodata, junto ao Instituto Nacional de Pesquisas Espaciais (INPE). Esse dado foi utilizado como base para a elaboração do Modelo Digital de Elevação (MDE) que possibilitou gerar informações sobre a rede de drenagem e a delimitação da bacia e relevo (declividade). As declividades das vertentes foram definidas com base no trabalho de Ponçano et al. (1981), que estabeleceram as classes <5%, 5 – 15% e >15%. Os autores Santos et al. (2020) também utilizaram esses mesmos intervalos na compartimentação geomorfométrica da bacia hidrográfica do Rio Jaguari – Oeste do RS. Por meio do site da United States Geological Survey (USGS), realizou-se o download da imagem do Landsat 8, do mês de julho de 2021, com resolução espacial de 30 metros. Na sequência, utilizou-se as bandas 6, 5 e 4 na 3 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental partir desses procedimentos, mapeou-se a espacialização das estradas não pavimentadas. Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 MATERIAIS E MÉTODOS respectiva ordem para gerar a composição RGB, e assim, gerar a classificação supervisionada, a qual oferece informações do uso e cobertura da terra sobre a bacia hidrográfica do Ribeirão Lajeado nas seguintes classes: Formação Florestal, Formação Savânica-Campestre, Agropecuária, Cicatrizes de Fogo, Corpos d’Água e Áreas Urbanas. As informações lineares como estradas e rede de drenagem foram transformadas em raster com pixel de 30 metros, utilizando-se um algoritmo que calcula a distância euclidiana, ou seja, em linha reta no plano horizontal. A distância euclidiana (Figura 2) determina os graus de influência, ou seja, quando a distância de um determinado ponto (L) for próximo da linha do vetor de interesse, maior será a sua influência e, quando o ponto de interesse estiver distante, menor será a influência. Para ter acesso às informações de estradas na bacia hidrográfica do Ribeirão Lajeado, foi necessário realizar uma vetorização sobre as imagens de satélites do Google Earth, disponível no plugin Quickmapservices no software Qgis. A Figura 2 - Exemplo da distância euclidiana Fonte: Sousa et al. (2015). Figura 2 - Exemplo da distância euclidiana Fonte: Sousa et al. (2015). Figura 2 - Exemplo da distância euclidiana. (autovetor) que indicam, ao tomador de decisão, quais elementos devem ter maior importância (hierarquia), para o cruzamento automático. (autovetor) que indicam, ao tomador de decisão, quais elementos devem ter maior importância (hierarquia), para o cruzamento automático. Para atribuir a priorização pela técnica AHP é necessário calcular a matriz ponderada, ou seja, cada elemento da coluna é dividido pela soma dos valores da própria coluna. Os valores obtidos a partir desse procedimento fazem com que a soma de cada coluna seja igual a um (1). Em seguida, utiliza-se a função matriz.multi da planilha eletrônica para obtenção do produto, que vai ser dividido pelo autovetor e gerar o lambda. Desse modo, é gerada a Razão de Consistência (RC), obtida pela divisão do IC pelo Índice de randômico (IR) do AHP, definido em função do número (n) de elementos comparados, segundo Saaty (1994) (Quadro 2). Quadro 2 - Valores recalculados de IR. Fonte: adaptado de Saaty (1994). N 3 4 5 6 7 8 9 10 11 12 13 14 15 IR 0.52 0.89 1.11 1.25 1.35 1.4 1.45 1.49 1.51 1.54 1.56 1.57 1.58 Quadro 2 - Valores recalculados de IR. Fonte: adaptado de Saaty (1994). N 3 4 5 6 7 8 9 10 11 12 13 14 15 IR 0.52 0.89 1.11 1.25 1.35 1.4 1.45 1.49 1.51 1.54 1.56 1.57 1.58 Quadro 2 - Valores recalculados de IR. Fonte: adaptado de Saaty (1994). N 3 4 5 6 7 8 9 10 11 12 13 14 15 IR 0.52 0.89 1.11 1.25 1.35 1.4 1.45 1.49 1.51 1.54 1.56 1.57 1.58 De modo geral, a figura 3 exemplifica a rotina do método AHP para definir os pesos normalizados para o uso e cobertura da terra, a declividade, os solos e as litologias na bacia hidrográfica do Ribeirão Lajeado. Os julgamentos dos diferentes tipos de solos basearam-se nas informações que estão no trabalho de Silva e Oliveira (2015), observando o grau de erodibilidade. Deste modo, considerou- se o Latossolo com peso 0,04; o Plintossolo peso 0,13; Associação Argissolo-Cambissolo peso 0,32 e Neossolo-Cambissolo com peso 0,51. As classes de declividade foram julgadas com base em diferentes autores que discutem a formação de voçorocas (ALMEIDA FILHO, 2004; OLIVEIRA, 1994; 2007; MARCHIORO et al., 2016). Portanto a classe menor que 5% tiveram o peso 0,07; a classe de 5 a 15% o peso 0,20 e a classe maior que 15% peso 0,73. Figura 2 - Exemplo da distância euclidiana. Figura 2 - Exemplo da distância euclidiana. Fonte: Sousa et al. (2015). A partir da organização de todos os dados cartográficos, aplicou-se o método do AHP que representa o julgamento ou a comparação pareada entre dois elementos que estão no mesmo par (que formam o par) (SAATY, 1991). A escala numérica de comparação varia de 1 a 9, significando a importância de um critério sobre outro (Quadro 1). Esses julgamentos são representados em uma matriz quadrada, com n (n - 1) /2, organizado em uma matriz n x n, sendo n o número de linhas e colunas, onde os analistas participantes julgam se o elemento (A) domina o elemento (B). Em estudos ambientais, observa-se o trabalho de França et al. (2019), que aplicaram esse método para analisar a fragilidade ambiental do município de Capelinha em Minas Gerais. Quadro 1 - Escala de comparação de critérios. Intensidade de importância em escala absoluta Definição Explicação 1 Igual importância Os dois critérios contribuem de forma idêntica para o objetivo 3 Importância moderada de um sobre o outro A análise e a experiência mostram que um critério é um pouco mais importante que o outro 5 Importância essencial ou forte A análise e a experiencia mostram que um critério é claramente mais importante que o outro 7 Importância muito forte A análise e a experiencia mostram que um dos critérios é predominante para o objetivo 9 Extremamente mais importante Sem qualquer dúvida um dos critérios é absolutamente predominante para o objetivo 2, 4, 6, 8 Valores intermediários entre os dois adjacentes julgamentos Quando o compromisso é necessário Fonte: adaptado de Saaty (1991). O método do AHP se inicia com a elaboração de uma matriz ponderada, a partir da qual se obtém o cálculo dos pesos normalizados (vetor de prioridades) para cada critério, calculando-se a média dos valores de cada linha da matriz ponderada. São esses pesos normalizados Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental Na coluna do lambda realiza-se a média. geométrica, que é subtraída pela quantidade de elementos analisados. Por fim, dividindo a mesma quantidade de elementos subtraindo por 1, será gerado o Índice de Consistência (IC). O IC mede a coerência dos julgamentos – quanto mais próximo o índice estiver de zero, maior será a consistência global da matriz de comparação. Figura 2 - Exemplo da distância euclidiana. As diferentes susceptibilidades das litologias estão relacionadas aos processos de desagregação e erosão. Assim, foram consideradas as rochas Cristalinas com peso 0,05; Formação Pimenteiras peso 0,14; Cobertura Detrito Laterítica com peso 0,30 e a Formação Jaicós peso 0,51. Figura 3 – Procedimentos da técnica do AHP. Figura 3 – Procedimentos da técnica do AHP. Fonte: Os autores (2022). Fonte: Os autores (2022). 5 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 R é o raster que está sendo ser ponderado. R é o raster que está sendo ser ponderado. Quadro 3 – Valores dos julgamentos e pesos dos fatores. Uso da terra estradas declividade rede de drenagem solos litologia pesos uso cobertura da terra 1 3 3 5 7 9 0,36 estradas 0,33 1 3 3 5 9 0,26 declividade 0,33 0,33 1 3 3 7 0,18 rede de drenagem 0,20 0,33 0,33 1 3 5 0,12 solos 0,14 0,20 0,33 0,33 1 3 0,06 litologia 0,11 0,11 0,14 0,20 0,33 1 0,02 razão de consistência: 0,06 Fonte: Os autores (2022). Fonte: Os autores (2022). estão representados pelos Córregos Mutum e Algodão, Ribeirões do Lajeado e Agem como de 5° ordem e, ainda, os Córregos Cedro, Cachorro, Serrinha, Brejo da Passagem e Ribeirão Agem como de 4° ordem. estão representados pelos Córregos Mutum e Algodão, Ribeirões do Lajeado e Agem como de 5° ordem e, ainda, os Córregos Cedro, Cachorro, Serrinha, Brejo da Passagem e Ribeirão Agem como de 4° ordem. A última etapa para análise da susceptibilidade geoambiental da bacia hidrográfica consistiu na realização de três trabalhos de campo em dezembro de 2021 para a validação das informações levantadas e cruzadas de forma automática em SIG. RESULTADOS E DISCUSSÃO O mapeamento da declividade das encostas (Figura 4B) apresentam que locais com <5% ocupam cerca de 33,23%, com predomínio na planície de inundação do Ribeirão Lajeado e nos topos que marcam as nascentes das drenagens principais da bacia. Declividades entre 5 - 15% abrangem cerca de 37,51%, distribuídas principalmente no médio e baixo curso. As áreas com declividades >15%, abrangem 29,26%, espacializando na borda dos relevos tabulares onde se encontram as encostas escarpadas. Localizam-se principalmente no médio e baixo curso do Ribeirão Lajeado com maior concentração nas porções leste e sudeste da área de pesquisa Analisando as características das variáveis (Figura 4), pode-se observar informações sobre os aspectos naturais e das atividades humanas na bacia hidrográfica. RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental No uso e na cobertura da terra, consideraram-se as classes com os seguintes pesos: Formação Florestal peso 0,03; Formação Savânica-Campestre peso 0,08; Agropecuária peso 0,21; Cicatrizes de Fogo com peso 0,28 e Áreas Urbanas peso 0,40. Os fatores utilizados para a análise geoambiental são: uso e cobertura da terra, estradas, declividade, rede de drenagem, solos e litologias. Por meio desses fatores, os pesos normalizados (Quadro 3) foram obtidos. Assim, os produtos ponderados são multiplicados pelos seus respectivos pesos, alcançados a partir do método AHP, conforme apresentado na Equação 2. Para esse cruzamento, utilizou-se a calculadora raster do software Qgis. Após a obtenção dos pesos de cada classe, realizou-se a agregação dos fatores condicionantes através do método da Combinação Linear Ponderada (WLC). Para tanto, os arquivos raster de uso e cobertura da Terra, estradas, declividade, rede de drenagem, solos e litologias foram padronizados dentro do intervalo de 0 a 1, utilizando-se a equação 1 na calculadora raster. S = (Fus * 0,36 + Fe * 0,26 + Fd * 0,18 + Frd * 0,12 + Fs * 0,06 + Fl * 0,02) (Equação 2) Onde: S é a susceptibilidade, Fus é o uso e cobertura da terra, Fe é a estrada, Fd é a declividade, Frd é a rede de drenagem, Fs é o solo, Fl é a litologia. (R – menor valor de R) / (maior valor de R – menor valor de R) (Equação 1) Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 Rede de Drenagem A bacia hidrográfica do Ribeirão Lajeado apresenta hierarquia fluvial de 6° ordem, apresentando o canal principal com extensão de 27km. A susceptibilidade é indicada pela distância euclidiana da rede de drenagem (Figura 4A). Os principais canais de drenagem 6 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental base em Ribeiro e Alves (2017) tem a cobertura Detrito Laterítica, rochas da Formação Pimenteiras, rochas da Formação Jaicós e rochas Granitoides (Figura 4C). Uso e Cobertura Vegetal O uso e a cobertura da terra na bacia hidrográfica do Ribeirão Lajeado podem ser observados nas seguintes classes: Formação Florestal, Formação Savânica-Campestre, Agropecuária, Cicatrizes de Fogo, Corpos d’Água e Áreas Urbanas (Figura 4E). Arenitos conglomeráticos e conglomerado da Formação Jaicós são dominantes e constituem grande parte das escarpas verticalizadas da Serra do Lajeado. Na área de estudo representa 13,19%, estando espacializada em uma faixa estreita no alto curso e, a partir do médio curso, alarga-se representando maior concentração em termos de área. A Formação Florestal, correspondendo a 21,48%, é dividida em matas ciliares, matas de galerias, mata seca e cerradão (RIBEIRO; WALTER, 2008). As rochas cristalinas na bacia hidrográfica do Ribeirão Lajeado são representadas pelos granitos e gabronoritos. O granito é um extenso corpo que aflora por mais de 60 km e desaparece sob as camadas sub-horizontalizadas das rochas sedimentares da Bacia do Parnaíba (CHAVES, 2003). As rochas cristalinas na bacia hidrográfica correspondem a 19,65%, ocorrendo a partir do alto e médio curso junto ao Ribeirão Lajeado, alargando-se no baixo curso. A Formação Savânica é dividida em Cerrado stricto sensu, parque cerrado, palmeiral e vereda. Já Formação Campestre é dividida em campo sujo, campo rupestre, campo limpo (RIBEIRO; WALTER, 2008). Esse tipo de cobertura vegetal (Savânica-Campestre) ocupa cerca de 57,17 % da área de estudo. O uso Agropecuário representa as atividades agrícolas (plantio de soja, milho e sorgo) e pecuária (criação de gado e cavalo) em cerca de 19,94%. As cicatrizes de fogo, na bacia hidrográfica do Ribeirão Lajeado, são locais que sofreram ação do fogo (por queimadas ou incêndios florestais) em um determinado período do ano. Assim, foi identificada uma abrangência de 1,13% da área. Á RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental As coberturas Detrito Lateríticas tratam-se da lateritização, predominantemente, ferruginosa que quando exposta ou exumadas são litificadas, formando ironstones. No topo da serra do Lajeado, elas se encontram expostas em vasta área plana em cotas ao redor de 600 m (RIBEIRO; ALVES, 2017). Na bacia hidrográfica do Ribeirão Lajeado, essa cobertura ocupa uma área com cerca de 3,01%, estando espacializada na porção leste, do médio e baixo curso. permeabilidade que restringe a circulação. Além disso, o clima com sazonalidade gera variações do lençol freático e a formação de plintitas no horizonte subsuperficial. Os solos mal desenvolvidos são os Neossolos e Cambissolos que se caracterizam pela baixa intensidade de atuação dos processos pedogenéticos. Na bacia hidrográfica corresponde a 22,30%, localizado no alto curso, acompanhando o canal do Ribeirão Lajeado, no médio e baixo curso abrange outros afluentes e apresentam a maior concentração em termo de área. Os folhelhos vermelhos e cinza escuros, com intercalações de arenitos e siltitos brancos a cinza claro da Formação Pimenteira, ocorrem especialmente em direção ao topo da sequência. Representam uma deposição em plataforma rasa dominada por tempestitos (VAZ et al., 2007). A Formação Pimenteira corresponde a 64,15%, sendo a principal litologia em termo de abrangência de área. Localiza-se principalmente no alto curso, enquanto no médio e baixo curso apresenta-se em menores áreas. A Associação Argissolo-Cambissolo, na área de estudo, ocupa cerca de 14,66%. Em suma, ocorre no médio e baixo curso, acompanhando o canal do Ribeirão Lajeado e outros afluentes no médio e baixo curso. Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 Litologias A bacia hidrográfica do Ribeirão Lajeado está inserida na bacia sedimentar do Parnaíba. Com ( g ) Figura 4 - Fatores analisados para a susceptibilidade geoambiental. Fonte: Os autores (2022). Figura 4 - Fatores analisados para a susceptibilidade geoambiental. Fonte: Os autores (2022). 7 7 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 RIBEIRO; ROBAINA; CRISTO Solos Os solos na bacia hidrográfica do Ribeirão Lajeado podem ser classificados como, Latossolos, Plintossolos, Associação Neossolo- Cambissolo e Associação Argissolo-Cambissolo, apresentado na figura 4D. Os Corpos d’Água, por sua vez, são áreas que representam os barramentos e tanques escavados relacionados às atividades de piscicultura, que também contribuem para a supressão da cobertura vegetal natural local (FURTADO; CRISTO, 2018). Os Latossolos na bacia hidrográfica, representa cerca de 18,67%, espacializando-se no alto curso, porções leste e oeste. No médio e baixo curso, ocupam as áreas da porção oeste dos afluentes do Ribeirão Lajeado e uma parte nos divisores na porção norte. Na área de estudo, encontra-se a cidade de Lajeado, próxima à foz da bacia hidrográfica, que corresponde a 0,22%. Local que passou por várias transformações econômicas, principalmente com a construção da usina Os Plintossolos na área de estudo, abrange cerca de 44,37%, espacializando principalmente no alto e médio; no baixo curso (porção leste) é associado a um relevo suave ondulado sobre um substrato de folhelhos com baixa 8 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental Susceptibilidade geoambiental da bacia hi- drográfica do Ribeirão Lajeado Susceptibilidade geoambiental da bacia hi- drográfica do Ribeirão Lajeado hidrelétrica de Luís Eduardo Magalhães, inaugurada no ano de 2002 (SEPLAN, 2017). A susceptibilidade geoambiental da bacia hidrográfica do Ribeirão Lajeado (Figura 5 e Tabela 1) foi definida a partir dos cruzamentos dos fatores de uso e da cobertura da terra (peso 0,36), estradas (peso 0,26), declividade (peso 0,18), rede de drenagem (peso 0,12), solos (peso 0,06) e litologias (peso 0,02), multiplicados pelos seus respectivos Observa -se que o uso e as vias de acessos tiveram mais pesos em relação aos outros elementos, pois as ocupações humanas nem sempre são precedidas de estudos que considerem as restrições dos recursos naturais, especialmente relacionadas à fragilidade das litologias e dos solos, quando submetidos a determinados usos (ROBAINA; TRENTIN, 2019). Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 Estradas As estradas não pavimentadas estão distribuídas principalmente no baixo e no médio curso, onde apresentam uma extensão de 732,39 km (Figura 4F). A menor concentração de estradas na porção oeste do alto curso ocorre devido ao fato de ser uma área destinada ao PEL. A maior concentração de estradas na porção leste do alto curso pode ser justificada pela presença de chácaras e pelo acesso as áreas de plantios de atividades agrícolas. O resultado desses cruzamentos apresentou a susceptibilidade geoambiental em quatro unidades: I baixa, II média, III alta e IV muito alta. Tabela 1 - Dados de susceptibilidade geoambiental da bacia hidrográfica. Susceptibilidade Geoambiental Área (km²) Área (%) Baixa 4,86 0,80 Média 157,78 25,60 Alta 431,81 70,06 Muito alta 21,84 3,54 Total 616,29 100 Fonte: Os autores (2022). Tabela 1 - Dados de susceptibilidade geoambiental da bacia hidrográfica. 9 II Média susceptibilidade Nessas áreas, os tipos de solos são Latossolo e Plintossolo, e a litologia predominante é da Formação Pimenteiras. A unidade de média susceptibilidade (Figura 7) tem uma área de 157,78 km² que corresponde a 25,60%. A unidade de baixa susceptibilidade ocorre no alto curso (em alguns locais dentro do PEL). No médio curso, próximos aos divisores de águas. No baixo curso, surgem alguns locais na porção leste da bacia hidrográfica. Figura 7 – Média susceptibilidade Fonte: Os autores (2021). Figura 7 – Média susceptibilidade Fonte: Os autores (2021). De maneira geral, são locais onde tem a presença de cobertura vegetal, principalmente do tipo Formação Savânica-Campestre, onde as características de solos, litologias e declividades são importantes para este tipo de unidade e sem nenhuma proximidade com estradas. No alto curso, as maiores concentrações dessa unidade estão na porção leste, área que pertence ao PEL. No médio curso, ocupam principalmente os divisores de água de alguns afluentes do Lajeado. No baixo curso, as maiores ocupações estão na sub-bacia do Córrego Mutum e nos divisores desse córrego com os afluentes do Ribeirão Lajeado. As características mais marcantes dessa unidade são as áreas de declividade menor que 15%, especificamente nas áreas de topos dos morros, onde ocorrem predominantemente os solos do tipo Plintossolo, a presença de cobertura vegetal natural e a distância das estradas. A baixa concentração de vias de acesso é uma característica importante dessa unidade que permite a conservação das condições naturais. Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental Figura 5 – Susceptibilidade geoambiental da bacia hidrográfica Figura 5 – Susceptibilidade geoambiental da bacia hidrográfica Fonte: Os autores (2022). Baixa susceptibilidade Baixa susceptibilidade (Figura 6) ocupa cerca 4,86 km², ou seja, cerca de 0,80%. Nessa nidade as principais características são o predomínio de vertentes com baix declividades, menor ou igual a 5%, com presença de cobertura vegetal natural e s nenhuma proximidade com as estradas. Fonte: Os autores (2022). predomínio de vertentes com baixas declividades, menor ou igual a 5%, com a presença de cobertura vegetal natural e sem nenhuma proximidade com as estradas. predomínio de vertentes com baixas declividades, menor ou igual a 5%, com a presença de cobertura vegetal natural e sem nenhuma proximidade com as estradas. I Baixa susceptibilidade A Baixa susceptibilidade (Figura 6) ocupa cerca de 4,86 km², ou seja, cerca de 0,80%. Nessa unidade, as principais características são o 10 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental Figura 6 - Baixa susceptibilidade Fonte: Os autores (2021). Figura 6 - Baixa susceptibilidade Fonte: Os autores (2021). Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 III Alta susceptibilidade A unidade de alta susceptibilidade (Figura 7) tem uma área de 431,81 km², isto é, 70,06%. Ocupa a maior classe dentro da bacia hidrográfica do Ribeirão Lajeado, ocorrendo em 11 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 A susceptibilidade geoambiental RIBEIRO; ROBAINA; CRISTO susceptível a intervenção antrópica pelos acessos (Figura 8). áreas onde o acesso é facilitado pela ampla ocorrência de estradas. A rede de drenagem é gura 8 – Alta susceptibilidade ocasionado pela confluência de estradas e rede de drenagem. Fonte: Os autores (2021). Fonte: Os autores (2021). As concentrações dessa unidade no alto curso estão principalmente na porção leste, onde ocorre significativa ocorrência de estradas, devido ao acesso para chácaras (Figura 9). As concentrações dessa unidade no alto curso estão principalmente na porção leste, onde ocorre significativa ocorrência de estradas, devido ao acesso para chácaras (Figura 9). Figura 9 – Alta susceptibilidade estradas e atividades agrícolas. Fonte: Os autores (2021). Fonte: Os autores (2021). IV Muito Alta susceptibilidade principais características são determinadas, pelo tipo de ocupação da terra e as vias acessos em locais de declividade com maior que 15% (Figura 10). principais características são determinadas, pelo tipo de ocupação da terra e as vias acessos em locais de declividade com maior que 15% (Figura 10). A Muito alta susceptibilidade tem uma área de 21,84 km² que corresponde a 3,54%, as 12 12 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental Figura 10 – Alta susceptibilidade em estradas e declividades maior que 15%. lta susceptibilidade em estradas e declividades ma Fonte: Os autores (2021). Fonte: Os autores (2021). Alguns locais (Figura 11) que já está desencadeando um significativo processo erosivo, em que a jusante está ocasionando entulhamento do canal de rede de drenagem. Outros aspectos associados a esta unidade são as áreas com a presença de cicatrizes de fogo e área urbana. De maneira geral as susceptibilidades ambientais de um determinado ambiente podem acontecer em decorrência de causas naturais ou antrópicas. Assim, alguns locais já têm áreas Susceptíveis devido às causas naturais (tipos de solos, declividade, etc.). Com a participação da atividade humana, tornam-se ainda mais frágeis e desencadeiam os processos erosivos. Alta susceptibilidade em estradas com significativos avanços dos processos erosivos. Fonte: Os autores (2021). Fonte: Os autores (2021). Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 CONSIDERAÇÕES FINAIS ABREU, W. L. et al. Zoneamento Geoambiental a partir das Unidades de Conservação: subsídios para a Gestão Integrada da Zona Costeira Paraense - Brasil. Revista Brasileira de Geografia Física, v. 13, n. 06, p. 3042-3059, dez. 2020. As observações realizadas na bacia hidrográfica do Ribeirão Lajeado mostraram a importância dos estudos integrados com cruzamentos automatizados para identificação da susceptibilidade geoambiental. Destaca-se que as características do uso e cobertura da terra, comparadas de maneira integrada com os elementos naturais, possibilitaram mapear as unidades de Baixa, Média, Alta e Muito Alta susceptibilidade. https://doi.org/10.26848/rbgf.v13.6.p3042-3059 ALMEIDA FILHO, G. S.; SANTORO, J.; GOMES, L. A. Estudo da dinâmica evolutiva da boçoroca São Dimas no município de São Pedro, SP. In: Anais do Simpósio Brasileiro de Desastres Naturais, 1., Florianópolis, p. 73-86, 2004. Assim, um dos aspectos que foram importantes na definição das áreas mais susceptíveis se deu através dessa integração, principalmente dos acessos e dos usos sobre a declividade. ALVES, M. R.; SILVA, J. C. Caracterização e avaliação da susceptibilidade a erosão de uma propriedade rural no município de Álvares Machado-SP. Colloquium Exactarum, v. 9, n. 3, jul./set., p. 57-64, 2017. https://doi.org/10.5747/ce.2017.v09.n3.e204 Nesse sentido, as áreas com declividades maiores que 15%, contendo atividades humanas (Agropecuária e Estradas), foram mapeadas como unidades de Muito Alta susceptibilidade. Outros fatores importantes para essa classificação foram as Áreas Urbanas e as Cicatrizes de Fogo. Em suma, podem ser desenvolvidas ações que minimizem esses problemas, principalmente frente às queimadas, pois é uma área de UC que pode ter um maior controle na questão do fogo. CARVALHO, A. P. P.; PEJON, O. J.; COLLARES, E. G. Geoenvironmental mapping and integrated analysis of the units within a protected area: municipality of Delfnópolis and the Serra da Canastra National Park, Brazil. Environmental Earth Sciences, online, v. 80, p. 1-18, fev. 2021. https://doi.org/10.1007/s12665-021-09469-x CHAVES, C. L. Caracterização geológica, petrográfica, geoquímica e geocronológica do magmatismo granítico da região de Porto Nacional – TO. 2003. 150 f. Tese (Doutorado em Geologia e Geoquímica) – Universidade Federal do Para, Belém, PA, 2003. No cruzamento automatizado, a técnica do AHP é considerada uma ferramenta fundamental na análise geoambiental, tendo em vista que é possível definir os graus de importância de cada elemento, a partir de um breve conhecimento sobre a área de estudo, considerando os aspectos específicos de cada um. CUNHA, N. R. S. et. al. FINANCIAMENTO O presente trabalho foi realizado com apoio da Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES), código de financiamento 001, e do Conselho Nacional de Desenvolvimento Científico e Tecnológico. III Alta susceptibilidade Para Alves e Silva (2017), um dos fatores que podem causar um aumento desta susceptibilidade, de causa antrópica, é a utilização das propriedades rurais para a agricultura e a pecuária. contribuindo para a perda de qualidade ambiental por meio, diminuição das áreas centrais e aumento do isolamento entre os fragmentos remanescentes. Pina et al. (2021), relatam a retirada da cobertura vegetal, onde os proprietários normalmente fazem a opção por um retorno de curto espaço de tempo, retirando a madeira e, posteriormente, implantando pastagens exóticas para a criação de gado. Essas atividades no Cerrado Brasileiro, segundo Cunha et al. (2008), são marcadas pelo uso intensivo de agrotóxicos, fertilizantes e corretivos; irrigação sem controle; pisoteio excessivo de animais; monocultura e cultura em grande escala; uso inadequado de fatores de produção. Outros aspectos podem ser associados as vias de acessos, pois são locais que influenciam no aumento das susceptibilidades, devido serem locais onde o fluxo superficial é aumentado, contribuindo para a geração de processos erosivos. Pires e Carmo Junior (2018), apontam Mascarenhas e Farias (2018) destacam a retirada da cobertura vegetal no Cerrado, que está ligada à fragmentação do bioma, 13 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental que alguns danos ambientais ocasionados pelas estradas são, muitas vezes, irreversíveis, devido à supressão de vegetação, à exposição do solo, à interferência nos habitats de animais nativos, às modificações no próprio solo, aos desvios das drenagens naturais e aos barramentos de rios. e para a gestão territorial, podendo vir a ser um importante recurso na identificação de unidades susceptíveis ao processo erosivo e no auxílio de criação de políticas ambientais que visam à conservação da natureza, seja em bacias hidrográficas, UCs, municípios ou outra área de interesse. Em relação às unidades de conservação da natureza, destaca-se que na área do PEL a principal classe foi a de média susceptibilidade, que pode ser justificada pela presença da cobertura vegetal natural e poucas estradas. As demais partes da bacia hidrográfica que pertence à APA Serra do Lajeado encontram-se como principal classe alta susceptibilidade, relacionadas às fragilidades naturais e com a participação das atividades humanas, principalmente, em função das variáveis e dos pesos adotados, pois neste tipo de unidade é permitido o “uso de forma sustentável”. Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 CONSIDERAÇÕES FINAIS A Intensidade da Exploração Agropecuária como Indicador da Degradação Ambiental na Região dos Cerrados, Brasil. RER, Piracicaba, SP, v. 46, Além disso, o AHP é concebido como um método que pode contribuir para o planejamento 14 14 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental n. 02, p. 291-323, abr./jun. 2008. https://doi.org/10.1590/S0103- 20032008000200002 n. 02, p. 291-323, abr./jun. 2008. https://doi.org/10.1590/S0103- 20032008000200002 Natureza, [S. l.], v. 32, p. 404-415, 2020. https://doi.org/10.14393/SN-v32-2020-46691 NOWATZKI, A.; SANTOS, L. J. C. Mapeamento digital de Solos por pedometria com base em atributos topográficos da bacia hidrográfica do Rio Pequeno-Paraná. Revista Ra’e Ga – Curitiba, v.31, p.185-211, dez./2014. https://doi.org/10.5380/raega.v32i0.33769 DIAS, N. O.; MARTINS, F. C. M.; BARROS; K. O. Geotecnologia aplicada à diagnose ambiental: Reserva Biológica de Pinheiro Grosso, Barbacena -MG. Sociedade e Natureza. Uberlândia, MG, v.32, p.126-140, 2020. https://doi.org/10.14393/SN-v32-2020- 45716 OLIVEIRA, M. A. T. et al. Morfometria de encostas e desenvolvimento de boçorocas no médio vale do rio Paraíba do Sul. Geociências, São Paulo, v. 13, n. 1, p. 9-23, 1994. FRANÇA, L. C. J. et al. Zoneamento da fragilidade ambiental de ecossistemas naturais e antropizados por meio de avaliação multicritério. Nativa, Sinop, v. 7, n. 5, p. 589- 599, set./out. 2019. h //d i /10 31413/ i 7i5 7300 OLIVEIRA, M. A. T. Processos erosivos e preservação de áreas de risco de erosão por voçorocas. In: GUERRA, A. J. T. et al (orgs), Erosão e conservação dos solos: conceitos, temas e aplicações, 2. ed. Rio de Janeiro: Bertrand Brasil, 2007. p. 57-99. FURTADO, S. E.; CRISTO, S. S. V. Análise das transformações ambientais no entorno do Parque Estadual do Lajeado, Palmas – Tocantins. Geografia, Ensino e Pesquisa, v. 22, e. 13, p. 01-11, 2018. https://doi.org/10.5902/2236499429699. PAULA, B. L.; CERRI, L.E.S. Aplicação do Processo Analítico Hierárquico (AHP) para priorização de obras de intervenção em áreas e setores de risco geológico nos municípios de Itapecerica Da Serra E Suzano (SP), São Paulo, UNESP, Geociências, v. 31, n. 2, p. 247-257, 2012. p g INPE Instituto Nacional de Pesquisas Espaciais. Portal do Topodata. Disponível em: http://www.webmapit.com.br/inpe/topodata/. Acesso em: 10 ago. 2020. PINA, J. C.; OLIVEIRA, A. K. M.; BOCCHESE, R. A. Composição florística e potencial de uso das espécies em uma área do bioma Cerrado em Bandeirantes – MS. Research, Society and Development, v. 10, n. 1, 2021. https://doi.org/10.33448/rsd-v10i1.11425 g MARCHIORI-FARIA, D. G.; AUGUSTO FILHO, O. CONSIDERAÇÕES FINAIS Mapeamento de perigo associado a escorregamentos em encostas urbanas utilizando o Processo de Análise Hierárquica (AHP). In: Simpósio Brasileiro de Cartografia Geotécnica e Geoambiental, Maringá, v. 7, 2010. PIRES, R. R.; CARMO JUNIOR, G. N. R. Processos erosivos em rodovias: uma revisão sistemática sobre os métodos de previsão e monitoramento. Engineering and Science, v. 4, ed. 7, 2018. https://doi.org/10.18607/ES201876853 MARCHIORO E.; ANDRADE E. E.; OLIVEIRA J. C. Evolução Espaço-Temporal de Voçorocas no Espírito Santo: Estudo de caso nos Municípios de Afonso Cláudio e Alegre. Rev. Bras. Geomorfol. [online], São Paulo, v.17, n.1, jan-mar, p. 191-204, 2016. https://doi.org/10.20502/rbg.v17i1.712 PONÇANO, W.L. et al. Mapa Geomorfológico do Estado de São Paulo. São Paulo: Instituto de Pesquisas Tecnológicas. 1981, 94p. p g g MARQUES, G. N.; ZUQUETTE, L. V. Aplicação da técnica AHP para seleção de áreas para aterros sanitários – Região de Araraquara (SP), Brasil. In: PEJON, O. J.; ZUQUETTE, L. V. (eds.), Cartografia Geotécnica e Geoambiental. São Carlos. Suprema Gráfica Editora, p. 263-272, 2004. MASCARENHAS, I. G. B.; FARIAS, K. M. S. Dinâmica da paisagem e relações com o uso do Solo e fragmentação da cobertura vegetal no município de Flores de Goiás (GO) entre 1985 e 2017. Élisée, Revista de Geografia da UEG, Porangatu, v. 7, n. 2, p.115-135, jul./dez, 2018. MARQUES, G. N.; ZUQUETTE, L. V. Aplicação da técnica AHP para seleção de áreas para aterros sanitários – Região de Araraquara (SP), Brasil. In: PEJON, O. J.; ZUQUETTE, L. V. (eds.), Cartografia Geotécnica e Geoambiental. São Carlos. Suprema Gráfica Editora, p. 263-272, 2004. RIBEIRO, J. F.; WALTER, B. M. T. As principais fitofisionomias do bioma Cerrado. In: SANO, S. M.; ALMEIDA, S. P.; RIBEIRO, J. F. (org.). Cerrado: ecologia e flora. Ecologia e flora. Brasília: Embrapa, 2008. p. 151-222. RIBEIRO, P. S. E.; ALVES, C. L. Geologia e Recursos Minerais da Região de Palmas: Folha Miracema do Norte (SC.22-X-D), Porto nacional (SC.22-Z-B) e Santa teresinha (SC.22-Z-a). escala 1:250.000. Goiânia: CPRM, 2017. MASCARENHAS, I. G. B.; FARIAS, K. M. S. Dinâmica da paisagem e relações com o uso do Solo e fragmentação da cobertura vegetal no município de Flores de Goiás (GO) entre 1985 e 2017. Élisée, Revista de Geografia da UEG, Porangatu, v. 7, n. 2, p.115-135, jul./dez, 2018. ROBAINA, L. E. S.; TRENTIN, R. Estudos e zoneamento geoambiental do município de São Francisco de Assis – Oeste do Rio Grande do Sul. Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 CONSIDERAÇÕES FINAIS Revista de Geografia e Ordenamento do Território (GOT), n.16, março, p. 323-344, 2019. MEDEIROS, J. F.; CESTARO, L. A. Using statistical techniques to conduct the geo- environmental compartmentalization of Serra de Martins-RN, Brazil. Sociedade e http://dx.doi.org/10.17127/got/2019.16.014 e 15 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 RIBEIRO; ROBAINA; CRISTO A susceptibilidade geoambiental ROBAINA, L. E. S.; TRENTIN, R. Compartimentação Geoambiental no município de São Vicente do Sul, RS, Brasil. Terr@Plural, Ponta Grossa, v. 15, p. 1-15, 2021. https://doi.org/10.5212/TerraPlural.v.15.21136 45.008 SOUSA, P. A. B.; BORGES, R. S. T.; DIAS, R. R. Atlas do Tocantins: subsídios ao planejamento da Gestão Territorial. Palmas: SEPLAN, 2012. 80p. SOUSA, T. M. I.; SARAIVA, A. G. S.; PAZ, A. R. Distâncias relativas à rede de drenagem: euclidiana x caminho de fluxo. Anais XVII Simpósio Brasileiro de Sensoriamento Remoto - SBSR, João Pessoa-PB, Brasil, 2015. ROSS, J. L. S. Análise empírica da fragilidade dos ambientes naturais e antropizados. Revista Departamento de Geografia, São Paulo: USP, p. 63-74, 1994. https://doi.org/10.7154/RDG.1994.0008.0006 USGS - United States Geological Survey. (2021). Disponível em: https://earthexplorer.usgs.gov/ Acesso em: 11 agosto de 2021. Acesso em: 11 agosto de 2021. SAATY, T. L. Método de análise hierárquica. São Paulo: McGraw-Hill Publisher, 1991. 367 p. VAZ, P. T., RESENDE, N. G. A. M., WANDERLEY FILHO, J. R., TRAVASSOS, W. A., Bacia do Parnaíba. Boletim de Geociências da Petrobras, Rio de Janeiro 15, 253-263, 2007. SAATY, T.L. How to Make a Decision: The Analytic Hierarchy Process. The Institute for Operations Research and the Management Sciences. Interfaces, v. 24, n. 6, p. 19-43, 1994. https://doi.org/10.1287/inte.24.6.19 VEDOVELLO, R. Aplicações da cartográfica geoambiental. In: 5° Simpósio Brasileiro de Cartografia Geotécnica. (org.) PEJON, O. J., ZUQUETTE, L. V. São Carlos: Anais, 2004. SANTOS, V. S.; ROBAINA, L. E. S.; TRENTIN, R. Compartimentação geomorfométrica da bacia hidrográfica do Rio Jaguari - Oeste do RS. Geosul, v. 35 n. 76, p.87-106, 2020. https://doi.org/10.5007/2177- 5230.2020v35n76p87 CONTRIBUIÇÃO DOS AUTORES SEPLAN - Secretaria de Planejamento do estado do Tocantins. Perfil socioeconômico dos municípios: Lajeado. 2017. Disponível em: https://central3.to.gov.br/arquivo/348406/. Acesso em: 9 set. 2021. SEPLAN - Secretaria de Planejamento do estado do Tocantins. Perfil socioeconômico dos municípios: Lajeado. 2017. Disponível em: https://central3.to.gov.br/arquivo/348406/. Acesso em: 9 set. 2021. CPRM - Serviço Geológico do Brasil. Disponível em: Lucas da Silva Ribeiro concebeu o estudo, organizou os dados para o cruzamento automático, fez o trabalho de campo, analisou as informações e redigiu o texto. Luís Eduardo de Souza Robaina, orientou Lucas da Silva Ribeiro indicando a literatura, participou da organização dos dados para o cruzamento automático e revisou os resultados obtidos. Sandro Sidnei Vargas de Cristo coorientou Lucas da Silva Ribeiro, indicando literatura e participou do trabalho de campo e revisão das informações. CPRM - Serviço Geológico do Brasil. Disponível em: CPRM - Serviço Geológico do Brasil. Disponível em: https://geosgb.cprm.gov.br/geosgb/downloads. html. Acesso em: 15 jun. 2021. https://geosgb.cprm.gov.br/geosgb/downloads. html. Acesso em: 15 jun. 2021. SILVA, G. G.; OLIVEIRA, L. N. Análise da suscetibilidades e potencial à erosão laminar no município de São Miguel do Araguaia – GO. Anais XVII Simpósio Brasileiro de Sensoriamento Remoto - SBSR, João Pessoa-PB, Brasil, 2015. E u Este é um artigo de acesso aberto distribuído nos termos da Licença de Atribuição Creative Commons, que permite o uso irrestrito, distribuição e reprodução em qualquer meio, desde que o trabalho original seja devidamente citado. 16 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513 Soc. Nat. \| Uberlândia, MG \| v.35 \| e67033 \| 2023 \| ISSN 1982-4513
https://openalex.org/W2972164631	https://trepo.tuni.fi/bitstream/10024/117424/2/Engineered_carbonic_anhydrase_2019.pdf	English	null	Engineered Carbonic Anhydrase VI-Mimic Enzyme Switched the Structure and Affinities of Inhibitors	Scientific reports	2,019	cc-by	16,764	OPEN Received: 17 February 2019 Accepted: 15 August 2019 Published: xx xx xxxx Justina Kazokaitė1,6, Visvaldas Kairys2, Joana Smirnovienė1, Alexey Smirnov1, Elena Manakova3, Martti Tolvanen4, Seppo Parkkila5 & Daumantas Matulis 1 Secretory human carbonic anhydrase VI (CA VI) has emerged as a potential drug target due to its role in pathological states, such as excess acidity-caused dental caries and injuries of gastric epithelium. Currently, there are no available CA VI-selective inhibitors or crystallographic structures of inhibitors bound to CA VI. The present study focuses on the site-directed CA II mutant mimicking the active site of CA VI for inhibitor screening. The interactions between CA VI-mimic and a series of benzenesulfonamides were evaluated by fluorescent thermal shift assay, stopped-flow CO2 hydration assay, isothermal titration calorimetry, and X-ray crystallography. Kinetic parameters showed that A65T, N67Q, F130Y, V134Q, L203T mutations did not influence catalytic properties of CA II, but inhibitor affinities resembled CA VI, exhibiting up to 0.16 nM intrinsic affinity for CA VI-mimic. Structurally, binding site of CA VI-mimic was found to be similar to CA VI. The ligand interactions with mutated side chains observed in three crystallographic structures allowed to rationalize observed variation of binding modes and experimental binding affinities to CA VI. This integrative set of kinetic, thermodynamic, and structural data revealed CA VI-mimic as a useful model to design CA VI-specific inhibitors which could be beneficial for novel therapeutic applications. Human carbonic anhydrases (CAs) are widespread enzymes known for over 80 years1. CAs regulate both intra- cellular and extracellular pH homeostasis through the catalysis of reversible carbon dioxide hydration to bicar- bonate and proton. To date, there are twelve catalytically active human CAs, which display diverse sub-cellular localization, tissue-specific expression, and kinetic properties2,3. Among a broad spectrum of CA-linked research areas, clinical investigation is a major focus due to the implication of abnormal CA levels or their activities in diseases, such as glaucoma4, epilepsy5, obesity6, and cancer7. Therefore, many efforts have been dedicated over years to design CA isoform-selective compounds exhibiting sufficient affinity properties8. These derivatives would be prospective for the translation into the clinic because of therapeutic efficacy without inducing undesired side effects caused by inhibited vital off-target CAs. However, it is a challenging task because of the high structural homology among human CAs9. gy g CA VI is the only secreted human CA isoform found in saliva10, serum11, milk12, respiratory airways13, and ali- mentary canal14. www.nature.com/scientificreports www.nature.com/scientificreports Received: 17 February 2019 Accepted: 15 August 2019 Published: xx xx xxxx www.nature.com/scientificreports/ www.nature.com/scientificreports/ limitations, such as the largely unknown CO2 concentration and unfeasibility to measure inhibition constant below several nM26. Therefore, biophysical techniques, such as the fluorescent thermal shift assay (FTSA) and isothermal titration calorimetry (ITC), are promising alternatives to screen CA-targeting derivatives. FTSA is a high-throughput method exhibiting minimized biomolecule consumption and low limitations for binding affin- ity, thereby both strong (picomolar) and weak (millimolar) compounds can be identified during the same experi- ment26–29. ITC allows the direct determination of stoichiometry and thermodynamic parameters, such as affinity, enthalpy, entropy, and heat-capacity, during a single or several titration experiments but it demands relatively large quantities of proteins and has limitations for assessing the binding affinity26,30,31. g q p g gfi y Importantly, two types of variables can be distinguished when binding reactions are carried out by FTSA or ITC: the observed parameters obtained from experimental setup and the intrinsic values calculated according to the corresponding observed data. Most studies on the development of CA inhibitors usually provide only observed binding parameters, which are dependent on experimental conditions and might be misleading. Both the CA and inhibitor exist in different protonation states in the solution compared with ones in the complex. Therefore, protonation-deprotonation reactions are required to initiate the binding of inhibitor to CA. Only intrinsic values subtract energetic contribution of binding-linked protonation events and thus are relevant for the rational drug design32–35. g g Due to the recent advances in the structural and in silico biology, production of target recombinant proteins, including CAs, in large quantities is of high demand for in vitro inhibitor screening of drug-candidates during preclinical research. The literature lists a number of host cells for expression of recombinant proteins. Among microorganisms, the enterobacterium Escherichia coli (E. coli) is selected frequently owing to numerous advan- tages, such as rapid growth, easy genetic manipulation, and relative cost effectiveness36,37. However, the stability of heterologous protein in E. coli can be influenced by the several factors, including mRNA instability, codon bias, protein aggregation, toxicity, and lack of post-translational modification38,39. Results E t Enzymatic activity of CA VI-mimic correlates with CA II, but not CA VI. Studies on inhibitor selec- tivity towards diverse human CA isoforms are important to develop efficient compounds for the treatment of diseases caused by abnormal levels or activities of a particular CA isoform. Therefore, it is essential to evaluate inhibitor affinity to all human CAs, including CA VI. Since our previous study40 indicated a low yield of recom- binant CA VI from E. coli, we generated CA II mutant as a CA VI model protein (CA VI-mimic) for inhibitor screening. Inhibitor affinities towards CA II and CA VI-mimic were expected to differ in the way imitating inhibi- tor binding to CA VI, but not CA II (Fig. 1A). Thus, negligible differences between inhibitor affinities towards CA VI and CA VI-mimic were presumed. According to computational modeling, five point mutations A65T, N67Q, F130Y, V134Q, L203T were chosen (Figs 1B–D and S1) and introduced into the active site of CA II.h g The catalytic activity of CA VI and CA VI-mimic to catalyze CO2 hydration reaction was measured by SFA (Figs 2A and S3). Analysis of kinetic data showed that site-directed mutagenesis did not significantly affect either the catalytic activity or pKa of zinc-bound water molecule of CA II. Catalytic constants (kcat) of CA II and CA VI-mimic did not differ (kcat values were 6.0 × 105 s−1), whereas kcat for CA VI was lower than CA VI-mimic by 3-fold (1.9 × 105 s−1). In the pH range 5.9–7.0 Michaelis constants (KM) as well as kcat values of the carbon dioxide hydration reaction were comparable: 7.3 ± 2.9 mM for CA II, 8.8 ± 1.8 mM for CA VI-mimic and 9.9 ± 3.2 mM for CA VI. However, in the pH range 7.1–8.4 KM values of CA VI-mimic (6.8 ± 2.0 mM) were closer to CA II (4.7 ± 1.0 mM) than to CA VI (11.3 ± 0.7 mM). Interestingly, maximum catalytic activity of CA VI was observed at pH 7.0–8.0 and it decreased at pH above 8.0. The determined pKa value of zinc-bound water molecule of CA VI was 6.6 ± 0.2. The observed inhibition constants by SFA correlated with dissociation constants determined by FTSA. Typical SFA curves of CA II, CA VI-mimic and CA VI inhibition by compound 39 are shown in Fig. 2B. www.nature.com/scientificreports/ Therefore, different, more efficient strategies to obtain functionally active recombinant proteins in high yield are required for screens of chemical compounds with the aim to identify hits in the initial stages of drug discovery.h p y g g y The goal of the present study was to design a CA II-based CA VI model protein, named as CA VI-mimic, for the search of CA-isoform selective inhibitors. As CA VI-mimic, mutant of CA II containing five point mutations, such as A65T, N67Q, F130Y, V134Q, L203T, was generated via site-directed mutagenesis. CA II was selected as a core for CA VI-mimic because purification yield of CA II from E. coli is ~10-fold higher than CA VI, CA II has highest catalytic efficiency among CAs, and CA II is confirmed as a stable CA protein for X-ray crystallography. Here enzymatic activity and inhibition of CA II, CAVI-mimic, and CA VI was determined by SFA. Biophysical studies on inhibitor binding to CA II, CA VI-mimic, and CA VI were carried out by ITC and FTSA. X-ray crys- tallography and computational modeling were used to compare positions of several inhibitors in the active sites of CA II, CA VI-mimic, and CA VI. Observed and intrinsic thermodynamics were in line with structural results which confirmed the relevance of CA VI-mimic as a CA VI model protein. The most tested benzenesulfonamides bound to CA VI-mimic in a manner corresponding to their interactions with CA VI but not CA II, thereby emphasizing suitability of the investigated CA II mutant mimicking CA VI for inhibitor screening. Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 OPEN Several studies have indicated the immunological CA VI function15,16 and have presented asso- ciations of CA VI with bitter taste perception17,18 or protection of excess acidity-caused complications, including dental caries19,20 and injuries of esophageal or gastric epithelium21. The link of CA VI with certain cancers, such as that of salivary glands, has been speculated by gene comparison study22, which have shown close relation of CA VI with CA IX, a marker of tumors23. Thus, there is a demand for effective and selective CA VI inhibitors, which would be relevant to determine the exact physiological role of CA VI.i p y g For more than five decades, the most widely applied method in the search of CA isoform selective inhibitors has been the stopped-flow assay of the catalytic activity of CO2 hydration (SFA)24,25. However, SFA has several 1Department of Biothermodynamics and Drug Design, Institute of Biotechnology, Vilnius University, Saulėtekio 7, Vilnius, LT-10257, Lithuania. 2Department of Bioinformatics, Institute of Biotechnology, Vilnius University, Saulėtekio 7, Vilnius, LT-10257, Lithuania. 3Department of Protein-DNA Interactions, Institute of Biotechnology, Vilnius University, Saulėtekio 7, Vilnius, LT-10257, Lithuania. 4Department of Information Technology, University of Turku, FI-20520, Turku, Finland. 5Tampere University, Faculty of Medicine and Health Technology; Fimlab Ltd, Tampere University Hospital, Arvo Ylpön katu 34, FI-33520, Tampere, Finland. 6Present address: Division of Biochemistry, the Netherlands Cancer Institute, Amsterdam, the Netherlands. Correspondence and requests for materials should be addressed to J.K. (email: kazokaite@ibt.lt) Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 Results E t Influence of buffer and pH for the observed binding affinity of ethoxzolamide to CA VI-mimic. Biophysical methods, such as ITC and FTSA, enable measurements of observed thermodynamics and thereafter calculations of intrinsic affinities. The observed binding profiles are altered by linked reactions and therefore, only intrinsic binding parameters can be correlated with compound structures, thereby revealing structural reasons for protein-ligand binding affinity.if fi Binding energetics are significantly affected by several protonation-deprotonation events which are necessary for the binding of sulfonamide derivative to CA. Only deprotonated sulfonamides can interact with the zinc cat- ion in the active site of pronated CA, containing zinc-coordinated water molecule (protonated hydroxy group). In this study, observed and intrinsic affinities of inhibitor binding to CA VI-mimic were determined and compared Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 www.nature.com/scientificreports/ Figure 1. (A) The mimic of CA VI was prepared from CA II by site-directed mutagenesis of amino acids that differ between two CA isoforms. The CA VI-mimic protein served as a model of compound binding to CA VI (B) Active site of CA II (PDB ID: 3KS3). Dark red molecular surfaces mark the positions of point mutations introduced in the active site of CA II to resemble CA VI by making a multiple-residue mutant of CA II (CA V mimic). (C) Active site of CA VI (PDB ID: 3FE4). The light blue areas are buried molecular surfaces between interacting molecules in the homodimeric complex. Dark red molecular surfaces mark the equivalent position between multiple-residue mutant of CA II (CA VI-mimic) and CA VI. The labels belong to CA VI (CA II numbering). (D) Superposed structures of the binding pockets of CA II (rose; PDB ID: 3M96), CA VI (blue; ) d ( ) Th d d f l d d k d Figure 1. (A) The mimic of CA VI was prepared from CA II by site-directed mutagenesis of amino acids that differ between two CA isoforms. The CA VI-mimic protein served as a model of compound binding to CA VI. (B) Active site of CA II (PDB ID: 3KS3). Dark red molecular surfaces mark the positions of point mutations introduced in the active site of CA II to resemble CA VI by making a multiple-residue mutant of CA II (CA VI- mimic). (C) Active site of CA VI (PDB ID: 3FE4). The light blue areas are buried molecular surfaces between interacting molecules in the homodimeric complex. Results E t Dark red molecular surfaces mark the equivalent positions between multiple-residue mutant of CA II (CA VI-mimic) and CA VI. The labels belong to CA VI (CA II numbering). (D) Superposed structures of the binding pockets of CA II (rose; PDB ID: 3M96), CA VI (blue; PDB ID: 3FE4), and CA VI-mimic (green; PDB ID: 6QL2). The mutated residues of CA II are colored dark red. The zinc ion in the active site of each CA isoform is shown as a grayish sphere in panels (B–D). Figure 1. (A) The mimic of CA VI was prepared from CA II by site-directed mutagenesis of amino acids that differ between two CA isoforms. The CA VI-mimic protein served as a model of compound binding to CA VI. (B) Active site of CA II (PDB ID: 3KS3). Dark red molecular surfaces mark the positions of point mutations introduced in the active site of CA II to resemble CA VI by making a multiple-residue mutant of CA II (CA VI- mimic). (C) Active site of CA VI (PDB ID: 3FE4). The light blue areas are buried molecular surfaces between interacting molecules in the homodimeric complex. Dark red molecular surfaces mark the equivalent positions between multiple-residue mutant of CA II (CA VI-mimic) and CA VI. The labels belong to CA VI (CA II numbering). (D) Superposed structures of the binding pockets of CA II (rose; PDB ID: 3M96), CA VI (blue; PDB ID: 3FE4), and CA VI-mimic (green; PDB ID: 6QL2). The mutated residues of CA II are colored dark red. The zinc ion in the active site of each CA isoform is shown as a grayish sphere in panels (B–D). to their affinities towards CA II and CA VI. The obtained experimental data by FTSA on interactions between ethoxzolamide (EZA) and CA VI-mimic in buffers with different pH showed that pH remarkably influenced the observed binding Gibbs energy (ΔbGobs, Fig. 3A). The dependence of ΔbGobs on pH has also been observed previously40,41 when EZA binding to CA II or CA VI was measured. The strongest interaction was determined near neutral pH and became weaker both in acidic and alkaline pH. Sulfonamide group usually has pKa in the range between 7 and 10, whereas CA isoforms have pKa around 7. Therefore, diminished EZA affinity in acidic solution was because the fraction of binding-ready deprotonated form of EZA decreased by 10-fold with every pH unit. Results E t Similarly, EZA affinity decreased in alkaline solution because the fraction of binding-ready CA with the to their affinities towards CA II and CA VI. The obtained experimental data by FTSA on interactions between ethoxzolamide (EZA) and CA VI-mimic in buffers with different pH showed that pH remarkably influenced the observed binding Gibbs energy (ΔbGobs, Fig. 3A). The dependence of ΔbGobs on pH has also been observed previously40,41 when EZA binding to CA II or CA VI was measured. The strongest interaction was determined near neutral pH and became weaker both in acidic and alkaline pH. Sulfonamide group usually has pKa in the range between 7 and 10, whereas CA isoforms have pKa around 7. Therefore, diminished EZA affinity in acidic solution was because the fraction of binding-ready deprotonated form of EZA decreased by 10-fold with every pH unit. Similarly, EZA affinity decreased in alkaline solution because the fraction of binding-ready CA with the Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 www.nature.com/scientificreports/ Figure 2. Catalytic activity, inhibition and binding profiles of CA II (red squares), CA VI-mimic (wine triangles) and CA VI (royal circles). (A) The plot of kcat dependence on pH by stopped-flow CO2 hydration assay (SFA). Solid lines were fit using single protonation model. (B) Inhibition of CAs by compound 39 using SFA. Data points were fit to the Morrison eq. (solid lines)74,75. The insets show raw activity curves of CA catalyzed reaction without added inhibitor (red, wine, royal lines), CA inhibited reaction with 313 nM added compound 39 (magenta, dark yellow, purple lines) and spontaneous CO2 hydration reaction (pink, orange, cyan lines) in the absence of CA. (C) Dosing curves of compound 39 binding to CAs by fluorescent thermal shift assay (FTSA). Data points show the ΔTm as a function of the total concentration of compound 39 added and the lines are simulated using fitting parameters when temperature is 37 °C, CA concentration is 10 µM, enthalpy of unfolding is 690 kJ/mol for CA II and CA VI-mimic, and 480 kJ/mol for CA VI, enthalpy of binding is −42 kJ/mol, heat capacity of binding is −0.8 J/ (molK) and the reference melting temperature is 56.8 °C for CA II, 63.1 °C for CA VI-mimic, and 47.6 °C for CA VI. The ΔTm shift is equal for CA VI-mimic and CA VI, but Kd’s differ due to different enthalpies of unfolding. Results E t The insets show CA unfolding curves at 0 and 200 µM inhibitor 39 concentrations. (D) Isothermal titration calorimetry (ITC) curves of EZA binding to CA VI-mimic in phosphate (Pi, pH 6.1 (▲) and 8.4 (●)) and TRIS buffer (pH 6.9 (■)) at 25 °C. Lines were fitted using single binding site model. The insets show raw data ITC curves at 10 µM CA VI-mimic concentration. Different observed enthalpies of binding illustrate the presence of binding-linked protonation reactions that must be accounted for the determination of intrinsic binding parameters. Figure 2. Catalytic activity, inhibition and binding profiles of CA II (red squares), CA VI-mimic (wine triangles) and CA VI (royal circles). (A) The plot of kcat dependence on pH by stopped-flow CO2 hydration assay (SFA). Solid lines were fit using single protonation model. (B) Inhibition of CAs by compound 39 using SFA. Data points were fit to the Morrison eq. (solid lines)74,75. The insets show raw activity curves of CA catalyzed reaction without added inhibitor (red, wine, royal lines), CA inhibited reaction with 313 nM added compound 39 (magenta, dark yellow, purple lines) and spontaneous CO2 hydration reaction (pink, orange, cyan lines) in the absence of CA. (C) Dosing curves of compound 39 binding to CAs by fluorescent thermal shift assay (FTSA). Data points show the ΔTm as a function of the total concentration of compound 39 added and the lines are simulated using fitting parameters when temperature is 37 °C, CA concentration is 10 µM, enthalpy of unfolding is 690 kJ/mol for CA II and CA VI-mimic, and 480 kJ/mol for CA VI, enthalpy of binding is −42 kJ/mol, heat capacity of binding is −0.8 J/ (molK) and the reference melting temperature is 56.8 °C for CA II, 63.1 °C for CA VI-mimic, and 47.6 °C for CA VI. The ΔTm shift is equal for CA VI-mimic and CA VI, but Kd’s differ due to different enthalpies of unfolding. The insets show CA unfolding curves at 0 and 200 µM inhibitor 39 concentrations. (D) Isothermal titration calorimetry (ITC) curves of EZA binding to CA VI-mimic in phosphate (Pi, pH 6.1 (▲) and 8.4 (●)) and TRIS buffer (pH 6.9 (■)) at 25 °C. Lines were fitted using single binding site model. The insets show raw data ITC curves at 10 µM CA VI-mimic concentration. Results E t Different observed enthalpies of binding illustrate the presence of binding-linked protonation reactions that must be accounted for the determination of intrinsic binding parameters. zinc-bound protonated hydroxide (water molecule) decreased. According to U-shaped curve as the global fit of experimental data extrapolated to 25 °C, intrinsic binding Gibbs energy (ΔbGintr) change upon EZA interaction with CA VI-mimic was determined to be −52.4 kJ/mol which was 7.8 kJ/mol greater than the highest experimen- tally observed value (−44.6 kJ/mol at pH 7.1). Difference of ΔbGintr (ΔΔbGintr) between EZA interaction with CA VI-mimic and CA VI were smaller (ΔΔbGintr = −1.5 kJ/mol) compared to that between CA VI-mimic and CA II (ΔΔbGintr = 6.1 kJ/mol). zinc-bound protonated hydroxide (water molecule) decreased. According to U-shaped curve as the global fit of experimental data extrapolated to 25 °C, intrinsic binding Gibbs energy (ΔbGintr) change upon EZA interaction with CA VI-mimic was determined to be −52.4 kJ/mol which was 7.8 kJ/mol greater than the highest experimen- tally observed value (−44.6 kJ/mol at pH 7.1). Difference of ΔbGintr (ΔΔbGintr) between EZA interaction with CA VI-mimic and CA VI were smaller (ΔΔbGintr = −1.5 kJ/mol) compared to that between CA VI-mimic and CA II (ΔΔbGintr = 6.1 kJ/mol). ( b intr J ) Observed standard enthalpy changes (ΔbHobs) upon EZA binding to CA VI-mimic formed an X-shaped curve which depended on pH and buffer (Fig. 3B). The same tendency has been found previously40,41 when ΔbHobs of EZA binding to CA II or CA VI was analyzed. Results were obtained by ITC titration at 25 °C in two buffers exhibiting different protonation enthalpies: sodium phosphate (Pi) and TRIS. Upon EZA-CA VI-mimic titration, more than 20 kJ/mol difference in ΔbHobs was observed in same buffer at different pHs (in TRIS buffer: −81.6 kJ/ mol at pH 6.5, −57.2 kJ/mol at pH 8.5; in Pi buffer: −38.9 kJ/mol at pH 5.2, −84.9 kJ/mol at pH 8.4). To dissect protonation influence, intrinsic enthalpy (ΔbHintr) of EZA interaction with CA VI-mimic was globally fitted to be −62.0 kJ/mol. Difference of ΔbHintr (ΔΔbHintr) between EZA interaction with CA VI-mimic and CA VI were smaller (4.0 kJ/mol) compared to that between CA VI-mimic and CA II (11.0 kJ/mol). Thus, ΔbHintr were in line with ΔbGintr, confirming that CA II mutant was mimicking CA VI for EZA binding. Results E t Protein pKa ΔpG, kJ/mol ΔpH, kJ/mol TΔpS, kJ/mol CA IIa 7.1 −40.5 −26.0 14.5 CA VI-mimic 7.2 −41.1 −38.0 3.1 CA VIb 6.2 −35.4 −32.0 3.4 Table 1. Thermodynamic parameters of protonation of zinc-bound hydroxide anion of studied CA isoforms as determined by FTSA and ITC at 25 °C. The uncertainty of the pKa values determined by FTSA and ITC is approximately 0.2 pH units, while for the change in Gibbs energies and enthalpies it is approximately 2 kJ/mol. aData taken from41; bData taken from40. Protein pKa ΔpG, kJ/mol ΔpH, kJ/mol TΔpS, kJ/mol CA IIa 7.1 −40.5 −26.0 14.5 CA VI-mimic 7.2 −41.1 −38.0 3.1 CA VIb 6.2 −35.4 −32.0 3.4 Table 1. Thermodynamic parameters of protonation of zinc-bound hydroxide anion of studied CA isoform as determined by FTSA and ITC at 25 °C. The uncertainty of the pKa values determined by FTSA and ITC is approximately 0.2 pH units, while for the change in Gibbs energies and enthalpies it is approximately 2 kJ/mo aData taken from41; bData taken from40. Table 1. Thermodynamic parameters of protonation of zinc-bound hydroxide anion of studied CA isoforms as determined by FTSA and ITC at 25 °C. The uncertainty of the pKa values determined by FTSA and ITC is approximately 0.2 pH units, while for the change in Gibbs energies and enthalpies it is approximately 2 kJ/mol. aData taken from41; bData taken from40. Table 1. Thermodynamic parameters of protonation of zinc-bound hydroxide anion of studied CA isoforms as determined by FTSA and ITC at 25 °C. The uncertainty of the pKa values determined by FTSA and ITC is approximately 0.2 pH units, while for the change in Gibbs energies and enthalpies it is approximately 2 kJ/mol. aData taken from41; bData taken from40. Table 1. Thermodynamic parameters of protonation of zinc-bound hydroxide anion of studied CA isoforms as determined by FTSA and ITC at 25 °C. The uncertainty of the pKa values determined by FTSA and ITC is approximately 0.2 pH units, while for the change in Gibbs energies and enthalpies it is approximately 2 kJ/mol. aData taken from41; bData taken from40. (ΔpH) of the zinc-bound water molecule (Table 1). The pKa of CA VI-mimic was determined to be 7.2 at 25 °C as the average of two pKa values evaluated independently by two techniques: 7.1 by FTSA and 7.3 by ITC. Results E t b intri g g g Furthermore, analysis of U- and X-shaped curves obtained by FTSA and ITC, respectively, led to the charac- terization of two important parameters of CA VI-mimic: ionization constant (pKa) and enthalpy of protonation Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 www.nature.com/scientificreports/ Figure 3. (A) Comparison of observed Gibbs energy changes (ΔbGobs) upon EZA binding to CA II (red), CA VI-mimic (dark red), and CA VI (blue) as a function of pH (25 °C). Experiments were performed by FTSA in universal buffer (50 mM sodium phosphate, 50 mM sodium acetate, and 25 mM sodium borate). The pKa for CA VI-mimic was determined to be 7.1. (B) The observed enthalpy changes (ΔbHobs) upon EZA binding to CA II (red), CA VI-mimic (dark red), and CA VI (blue) as a function of pH in two different buffers (sodium phosphate (Pi) and TRIS), which have different protonation enthalpies. Experiments were performed by isothermal titration calorimetry (ITC) at 25 °C. The dashed line shows the intrinsic binding enthalpy (ΔbHintr), which is independent of pH. The pKa for CA VI-mimic was determined to be 7.3. Thermodynamic binding parameters of EZA binding to CA II and CA VI have been previously published40,41. Red arrows indicate difference in ΔbGintr or ΔbHintr of EZA binding to CA VI-mimic compared to CA II or CA VI. Figure 3. (A) Comparison of observed Gibbs energy changes (ΔbGobs) upon EZA binding to CA II (red), CA VI-mimic (dark red), and CA VI (blue) as a function of pH (25 °C). Experiments were performed by FTSA in universal buffer (50 mM sodium phosphate, 50 mM sodium acetate, and 25 mM sodium borate). The pKa for CA VI-mimic was determined to be 7.1. (B) The observed enthalpy changes (ΔbHobs) upon EZA binding to CA II (red), CA VI-mimic (dark red), and CA VI (blue) as a function of pH in two different buffers (sodium phosphate (Pi) and TRIS), which have different protonation enthalpies. Experiments were performed by isothermal titration calorimetry (ITC) at 25 °C. The dashed line shows the intrinsic binding enthalpy (ΔbHintr), which is independent of pH. The pKa for CA VI-mimic was determined to be 7.3. Thermodynamic binding parameters of EZA binding to CA II and CA VI have been previously published40,41. Red arrows indicate difference in ΔbGintr or ΔbHintr of EZA binding to CA VI-mimic compared to CA II or CA VI. Results E t The pKas of CA II and CA VI-mimic matched each other within the error margin of 0.2 pH unit42, whereas pKas of CA VI and CA VI-mimic significantly differed by 1.0 pH (25 °C). Thus, target five point mutations of CA II, which were introduced to design CA VI-mimic, did not affect amino acids surrounding zinc in active sites of CA II at the level causing significant difference of pKas between CA II and CA VI-mimic. Moreover, ΔpH for CA VI-mimic was assessed to be −38.0 kJ/mol at 25 °C. The difference of ΔpH between CA VI and CA VI-mimic (6.0 kJ/mol) was 2-fold lower than difference of ΔpH between CA II and CA VI-mimic (12.0 kJ/mol). Therefore, experiments with one inhibitor EZA resulted in both pKa and ΔpH for CA VI-mimic, which are essential parameters to determine intrinsic energetics of any other inhibitor binding to CA VI-mimic. Hydrophobic substituents and fluorine substituents significantly affected intrinsic inhibitor binding affinity for CA VI-mimic. Here 43 benzenesulfonamide derivatives binding to CA VI-mimic was measured by FTSA and inhibition constants of several selected compounds were confirmed by SFA. Trifluoromethanesulfonamide (TFS), EZA, and methazolamide (MZM) were used as controls. Structures of tested compounds are shown in Fig. 4, while dissociation constants (Kd) are listed in Tables 2 and S1 (examples Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 www.nature.com/scientificreports/ www.nature.com/scientificreports/ www.nature.com/scientificreports cientificreports/ of raw and integrated data of inhibitor binding to CA VI-mimic by FTSA and ITC at different pHs are indicated Figure 4. Chemical structures of 1-46 compounds designed as CA inhibitors. Compounds 1-3 are standard inhibitors of CAs that we used here as control compounds (TFS, EZA, and MZM). Figure 4. Chemical structures of 1-46 compounds designed as CA inhibitors. Compounds 1-3 are standard inhibitors of CAs that we used here as control compounds (TFS, EZA, and MZM). Figure 4. Chemical structures of 1-46 compounds designed as CA inhibitors. Compounds 1-3 are standard inhibitors of CAs that we used here as control compounds (TFS, EZA, and MZM). of raw and integrated data of inhibitor binding to CA VI-mimic by FTSA and ITC at different pHs are indicated in Fig. 2C,D, respectively). of raw and integrated data of inhibitor binding to CA VI-mimic by FTSA and ITC at different pHs are indicated in Fig. 2C,D, respectively). d b d h d h b h b d h According to observed thermodynamics, EZA was shown to be the strongest binder to CA VI-mimic with observed Kd (Kd_obs) of 17 nM. From a series of fluorinated benzenesulfonamides, compounds 22 and 26 bear- ing substituents at para position were characterized to be the most potent CA VI-mimic inhibitors bound with observed Kd (Kd_obs) in the range of 50–67 nM. The comparison between binding affinities of corresponding fluor- inated and nonfluorinated compounds (6 vs 30, 15 vs 16, and 22 vs 23) showed that fluorination significantly increased observed binding affinity and diminished pKa of inhibitor sulfonamide amino group. For instance, Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 www.nature.com/scientificreports/ Inhibitor Lab. name pKa_SA Kd_obs (nM) Kd_intr (nM) CA II CA VI-mimic CA VI CA II CA VI-mimic CA VI 1. TFS 6.02 20 33 14 8.0 15 1.2 2. EZA 7.82 1.3 (<5.0) 17 (<54) 33 (<54) 0.073 1.1 0.40 3. MZM 6.86 100 330 830 27 97 44 4. VD10-9 8.12 46 130 430 1.4 4.7 2.7 5. VD10-12 8.61 18 100 1100 0.19 1.2 2.4 6. VD10-13 8.14 11 130 200 0.32 4.2 1.2 7. VD10-14 8.84 91 330 1000 0.57 2.4 1.3 8. VD10-16 8.47 9.1 200 1200 0.13 3.3 3.7 9. VD10-18 7.80 3.4 140 200 0.20 9.8 2.5 10. VD10-35 7.28 17 130 67 2.5 22 2.1 11. VD10-39b 7.85 83 140 130 4.6 8.8 1.5 12. Table 2. The Kd_obs and Kd_intr values (nM) for interactions between inhibitor and three CA proteins: CA II, CA VI-mimic, and CA VI. The observed inhibitor affinities for CA VI-mimic were obtained experimentally by FTSA (pH 7.0, 37 °C), whereas the intrinsic parameters were calculated from the corresponding observed data using pKa of 7.0 for CA VI-mimic at 37 °C as explained in the methods part. The standard error of Kd measurements is ±2-fold. The pKa values of applied sulfonamide amino group (pKa_SA) and inhibitor affinities towards CA II and CA VI have been already reported58,86. Dissociation constants Kds of selected compounds were confirmed by SFA (pH 7.5). Experiments were performed at 23 °C and observed Kds were extrapolated to 37 °C using van’t Hoff equation when enthalpy of binding is −42 kJ/mol. The values at 37 °C are given in parentheses. The determined Kds at 23 °C are given in Table S2. www.nature.com/scientificreports/ VD10-45 7.69 5.8 140 140 0.43 12 2.2 13. VD10-49 7.83 0.65 110 200 0.037 7.2 2.3 14. VD10-50 8.02 9.6 (<43) 330 (140) 1000 (630) 0.37 15 7.9 15. VD11-9 8.05 1.7 100 400 0.061 4.1 3.0 16. VD11-36 10.1 12 500 1100 0.0039 0.20 0.077 17. VD11-67 8.67 5900 100 000 100 000 55 1000 190 18. VD11-51 7.07 3.3 130 160 0.68 31 6.6 19. VD11-56 7.97 20 330 500 0.86 16 4.4 20. VD11-71 8.67 500 2500 3300 4.6 26 6.3 21. VD12-04 8.67 1800 3300 20 000 17 35 38 22. VD12-05 8.15 2.2 50 140 0.065 1.7 0.86 23. VD12-10 10.2 25 (<36) 500 (200) 830 (3900) 0.0070 0.16 0.048 24. VD12-17 8.67 1300 5000 11 000 12 52 21 25. VD11-4-2 8.01 56 100 67 2.2 4.5 0.54 26. VD11-10 7.22 1.2 67 140 0.21 13 4.8 27. VD11-16 7.87 35 1000 2000 1.9 59 22 28. VD11-17 7.87 50 140 330 2.6 8.5 3.6 29. VD11-28 7.87 6.7 200 110 0.35 12 1.2 30. VD11-31 9.96 140 1000 5000 0.070 0.55 0.50 31. VD11-39 7.87 33 2000 1000 1.8 120 11 32. VD11-61 8.53 3.3 140 330 0.042 2.1 0.87 33. E46 8.90 50 50 200 0.28 0.31 0.23 34. E11-6 8.70 5.6 330 1100 0.048 3.3 2.0 35. E11-11 9.40 1000 2000 13 000 1.8 4.0 4.5 36. E11-12 8.90 8.5 500 5000 0.047 3.1 5.7 37. E11-14 8.90 7.1 330 3300 0.039 2.1 3.8 38. E11-15 9.40 560 1300 5000 0.98 2.5 1.8 39. E11-18 8.90 28 (<54) 500 (300) 2500 (1600) 0.15 3.1 2.8 40. E11-37 9.60 3600 3300 6700 4.0 4.2 1.5 41. E11-36 8.30 8.3 500 3300 0.16 12 14 42. E11-52 8.70 2.9 330 1400 0.025 3.3 2.5 43. E11-53 9.60 1700 3300 10 000 1.9 4.2 2.3 44. E11-57 9.60 140 500 3300 0.16 0.63 0.75 45. E11-60 8.70 250 500 3300 2.2 4.9 5.9 46. E11-64 9.40 100 1100 17 000 0.16 2.2 5.5 T bl Th d l ( ) f b h b d h Table 2. The Kd_obs and Kd_intr values (nM) for interactions between inhibitor and three CA proteins: CA II, CA VI-mimic, and CA VI. www.nature.com/scientificreports/ For example, upon chlorination Kd_obs for 40 and 41 interaction with CA VI-mimic increased 430-fold (from 3600 nM to 8.3 nM, respectively), while pKa values were lowered by 1.30 unit (from 9.60 to 8.30, respectively). y ( p y) To investigate structure-activity relationships, intrinsic Kd (Kd_intr) values for interactions between CA VI-mimic and investigated series of compounds were calculated. The largest differences between Kd_obs and Kd_intr values were determined for nonfluorinated benzenesulfonamides (16, 23, and 30), where the binding to CA VI-mimic differed 2500, 3200, and 1800-fold, respectively. Only five compounds (10, 18, 26, TFS, and MZM) exhibited lower than 10-fold difference between the Kd_obs and Kd_intr. According to intrinsic thermodynamics, the strongest binders were inhibitors 16, 23, 30, and 33 with Kd_intr in the range of 0.16–0.55 nM. Therefore, the strongest intrinsic interaction between inhibitor and CA VI-mimic was observed when inhibitor did not possess any fluorines in benzenesulfonamide scaffold and contained a hydrophobic substituent at para position, such as SCH2CH2CH3 (Kd_intr for inhibitor 23 was 0.16 nM) and SCH2CH2Ph (Kd_intr for inhibitor 16 was 0.20 nM). Exceptionally, inhibitor 33 was the strongest binder to CA VI-mimic with chlorine at ortho position and large hydrophilic group at meta position (Kd_intr was 0.31 nM). Replacement of the methyl group (inhibitor 23) by hydrophilic hydroxyl group (inhibitor 30) weakened intrinsic binding affinity more than 3-fold (from to 0.16 nM to 0.55 nM). Moving on to the structural analysis of fluorinated benzenesulfonamides, two inhibitors were deter- mined to be the strongest binders to CA VI-mimic: compound 5 bearing 4-Morpholinyl group at para position (Kd_intr was 1.2 nM) and 22 with SCH2CH2CH3 group at para position (Kd_intr was 1.7 nM). In line with results obtained from nonfluorinated compounds, hydrophobic contacts between inhibitors and CA VI were identified to be significant because the exchange of methyl group (inhibitor 22) by hydroxyl group (inhibitor 6) or car- boxyl group (inhibitor 19) weakened intrinsic interaction by 2 and 9-fold, respectively. Apparently, the number of methyl groups of substituents at para position had significant effect on intrinsic binding affinity. The inhibitor 32 with SCH2Ph bound to CA VI-mimic 2-fold stronger than inhibitor 15 with SCH2CH2Ph and 4-fold stronger than inhibitor 9 with SPh. www.nature.com/scientificreports/ The observed inhibitor affinities for CA VI-mimic were obtained experimentally by FTSA (pH 7.0, 37 °C), whereas the intrinsic parameters were calculated from the corresponding observed data using pKa of 7.0 for CA VI-mimic at 37 °C as explained in the methods part. The standard error of Kd measurements is ±2-fold. The pKa values of applied sulfonamide amino group (pKa_SA) and inhibitor affinities towards CA II and CA VI have been already reported58,86. Dissociation constants Kds of selected compounds were confirmed by SFA (pH 7.5). Experiments were performed at 23 °C and observed Kds were extrapolated to 37 °C using van’t Hoff equation when enthalpy of binding is −42 kJ/mol. The values at 37 °C are given in parentheses. The determined Kds at 23 °C are given in Table S2. Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 www.nature.com/scientificreports/ Figure 5. Comparison of logKd values representing observed (A) and intrinsic (B) inhibitor binding affinities towards CA VI-mimic and CA II (red squares) or CA VI-mimic and CA VI (blue squares). Straight line represents a model of equal affinity of inhibitor binding to pairwise proteins. Red and blue dashed lines show linear regression models for inhibitor binding to CA II and CA VI, respectively. R2 values and linear equations are indicated. Experiments were performed by FTSA (pH 7.0, 37 °C). Figure 5. Comparison of logKd values representing observed (A) and intrinsic (B) inhibitor binding affinities towards CA VI-mimic and CA II (red squares) or CA VI-mimic and CA VI (blue squares). Straight line represents a model of equal affinity of inhibitor binding to pairwise proteins. Red and blue dashed lines show linear regression models for inhibitor binding to CA II and CA VI, respectively. R2 values and linear equations are indicated. Experiments were performed by FTSA (pH 7.0, 37 °C). Kd_obs for 30 and 6 binding to CA VI-mimic increased 8-fold upon fluorination (from 1000 nM to 130 nM), whereas Kd_obs for 23 and 22 increased affinity 10-fold (from 500 nM to 50 nM). Fluorines reduced pKa of sulfon- amide group significantly: from 9.96 to 8.14 for inhibitors 30 and 6, respectively, and from 10.2 to 8.15 for com- pounds 23 and 22, respectively (Table 2). Correspondingly, chlorine in most compounds also increased observed affinity and reduced pKa of inhibitor sulfonamide amino group. www.nature.com/scientificreports/ For the intrinsic data, only 6 compounds were identified as CA VI-mimic binders with the affinity more alike CA II compared to CA VI. Moreover, mean absolute errors (MAEs) as the averages for each AE were also evaluated. MAEs of AEobs,CA II and AEintr,CA II were equal to 1.1, while MAEs of AEobs,CA VI and AEintr,CA VI were significantly smaller, 0.47 and 0.41, respectively. Therefore, CA VI-mimic designed via site-directed mutagenesis from CA II was characterized to be a proper model of CA VI for observed and intrinsic inhibitor binding reactions. Differences in inhibitor binding affinities are due different binding modes as determined by crystallographic analysis of CA II and CA VI-mimic. Despite numerous attempts, the crystal structures of recombinant CA VI complexes with sulfonamide-based inhibitors were not obtained by soaking. Even though CA VI crystals survived soaking procedure, crystals did not contain the clear electron densities of inhibitors. The co-crystallization of CA VI protein with several inhibitors failed, as we did not obtain any crystals suitable for X-ray diffraction experiment. Most likely, CA VI complexes with sulfonamide-based inhibitors cannot be crys- tallized using crystallization conditions that are effective for the unbound CA VI protein. g yf p To structurally investigate the binding of benzenesulfonamides with CA VI, we have engineered CA VI-mimic and applied in crystallographic studies. We have solved crystal structures of CA VI-mimic complexes with three inhibitors (Fig. S2): EZA (PDB ID: 6QL2), inhibitor 14 (PDB ID: 6QL1), and 25 (PDB ID: 6QL3). These com- plexes were compared with the corresponding complexes composed of CA II and same ligands (EZA (PDB IDs: 3CAJ (X-ray), 6BCC (neutron diffraction)), inhibitor 14 (PDB ID: 4HT0), and 25 (4PYY)). The space groups and unit cell parameters of CA VI-mimic crystals were similar to those of CA II (Table 3). There was one unique protein-ligand complex in the asymmetric unit. CA VI-mimic binding pocket was found to be similar to CA VI according to crystallographic studies (Fig. 1D) followed by thermodynamic analysis. For this reason, the insights into the compound binding mode to CA VI-mimic are likely to be valid for analyzing the ligand binding data to CA VI.h The comparison of binding mode of inhibitor 14 in the active sites of CA VI-mimic and CA II is shown in Fig. 6A. www.nature.com/scientificreports/ Inhibitor 14 in the active site of CA VI-mimic had two alternative binding modes characterized by differ- ent positions of the fluorinated ring: the ring was either lodged between Leu198 and Thr200 side chains (colored cyan in Fig. 6A), or located in the hydrophilic part of active site (colored blue). On the other hand, in the active site of CA II we had only one position of fluorinated ring – between Leu198 and Thr200. It looks like the replace- ment of Phe130 in CA II with tyrosine in CA VI-mimic enabled additional position of fluorinated ring of ligand due to a steric collision between the fluorine atom of fluorinated ring and the oxygen atom of Tyr130 side chain (the close contact found in the structure was 2.5 Å). The alternative position of the fluorinated ring of compound 14 in the active site of CA VI-mimic probably was available only due to spatial fluctuations of Tyr130 side chain. Also, due to a significantly larger size and the hydrophilicity of the side chain of Gln134 in CA VI-mimic com- pared to Val134 in CA II, the hydrophobic dimethylpyrimidine tail of inhibitor 14 was repelled in CA VI-mimic (see para-group of the cyan ligand, Fig. 6A). Therefore, the change of size and the hydrophobicity/hydrophilicity of the residues 130 and 134 upon mutation could be rationalized as the main causes for the relatively significant difference in the binding affinities: inhibitor 14 bound to CA II 40-fold better than to CA VI-mimic (Kd_intr values were 0.37 nM and 15 nM for CA II and CA VI-mimic, respectively; Table 2).h p y ) The model compound EZA was bound similarly in active sites of CA II and CA VI-mimic (Fig. 6B). Some discrepancy was present only in the positions of highly flexible ethoxy moiety. The aliphatic-aromatic interactions between the methyl group of Leu198 and the first ring of EZA was present in both cases. The larger side chain of Tyr130 slightly changed the position of EZA aromatic ring in CA VI-mimic as compared with CA II due to steric conflicts. It is important also to note the role of the residue 134 interacting with the hydrophobic tail of EZA, similarly to the observation for inhibitor 14 above. The hydrophobic Val134 sidechain in CA II was mutated into hydrophilic Gln134 in CA VI, leading to the worsening of the interaction. www.nature.com/scientificreports/ Most often, introducing diverse substituents at meta position did not change intrinsic binding affinity significantly (6 vs 25, 26 vs 28, and 26 vs 29), except for 26 vs 31 bearing large hydrophobic group which weakened interaction 9-fold. The compound 17 bearing two large and highly hydrophobic substituents at ortho and para positions was the weakest binder not only according to the observed parameters (Kd_obs was 100 µM), but also intrinsic data (Kd_intr of 1000 nM). Thermodynamically CA VI-mimic binds benzenesulfonamides similarly to CA VI but differing from CA II. To evaluate if CA VI-mimic based on CA II is a suitable CA VI model protein for inhibitor screening, observed and intrinsic affinities represented by logarithmic Kd values of inhibitor binding to CA II, CA VI-mimic, and CA VI were compared by applying linear regression. A higher linear correlation was deter- mined between observed affinities of inhibitor binding to CA VI and CA VI-mimic (R2 = 0.79) compared to the observed affinities of inhibitor interaction with CA II and CA VI (R2 = 0.61; Fig. 5A). Analysis of the calculated intrinsic parameters were in line with experimentally measured observed data, emphasizing a stronger correla- tion of the intrinsic thermodynamics of inhibitor binding to CA VI and CA VI-mimic (R2 = 0.74) compared to that of CA II and CA VI-mimic (R2 = 0.56; Fig. 5B). Furthermore, regression line slopes indicating the compar- ison of inhibitor binding to CA VI and CA VI-mimic (0.74 for observed affinity, 0.95 for intrinsic affinity) were larger than the corresponding slopes for CA II and CA VI (0.54 for observed affinity, 0.58 for intrinsic affinity), Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 www.nature.com/scientificreports/ hereby indicating a lower difference between the inhibitor binding towards CA VI-mimic and CA VI compared o CA II. thereby indicating a lower difference between the inhibitor binding towards CA VI-mimic and CA VI compared to CA II.hl The influence of investigated CA II mutations on inhibitor binding thermodynamics was further analyzed by calculating the absolute error (AE) values from logarithmic observed or intrinsic Kds of inhibitor binding to CA VI-mimic, CA II, and CA VI. According to the observed thermodynamics, binding affinities of only 12 inhibitors out of 46 tested compounds towards CA VI-mimic was more similar to CA II than CA VI. www.nature.com/scientificreports/ Thus, the mutations of residues 130 and 134 were the likely reasons for 15-fold stronger binding of EZA to CA II, as determined by intrinsic thermody- namics (Kd intr values were 0.073 nM and 1.1 nM for CA II and CA VI-mimic, respectively; Table 2).h _ p y The intrinsic binding parameters of inhibitor 25 towards CA II and CA VI-mimic were comparable (2.2 nM vs 4.5 nM, respectively; Table 2). In contrast, the binding modes of the compound found in crystal structures were different in these active sites. In CA II, inhibitor 25 had two alternative conformations: (1) the fluorinated ring located between Leu198 and Thr200, whereas the cyclooctyl ring replaced the side chain of Phe130 (Fig. 6C, pink ligand); (2) the fluorinated ring positioned in the hydrophobic part of active site, while the cyclooctyl ring – in the hydrophilic part (Fig. 6D, pink ligand). In CA VI-mimic, compound 25 had one well-defined conformation (Fig. 6C, D, blue) in which the cyclooctyl ring replaced the Tyr130 side chain, whereas the fluorinated ring occu- pied the hydrophilic part of active site. We can explain the presence of the single conformation of inhibitor 25 bound to CA VI-mimic. It seems that the mutation in position 67 (asparagine to glutamine) allows the position of fluorinated ring in hydrophilic part of active site when the para-substituent of ligand does not have sterical col- lision with side chain of residue 67 (compare side chain conformations of asparagine (CA II) and glutamine (CA VI-mimic), Fig. 6C, D). The same position of glutamine is found in the complexes of CA VI-mimic with inhibitor 14 and EZA ligands which means that compound 25 does not influence the position of side chain of residue 67 in CA VI-mimic. The replacement of asparagine (CA II) to glutamine (CA VI-mimic) creates the additional free space in the active site and allows for another binding mode. www.nature.com/scientificreports/ Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 9 www.nature.com/scientificreports/ Isoform-ligand CA VI-mimic – inhibitor 14 CA VI-mimic - EZA CA VI-mimic – inhibitor 25 PDB ID 6QL1 6QL2 6QL3 Data-collection statistics Space group P1211 P1211 P1211 Unit-cell parameters (Å) a = 42.3, b = 41.4, c = 71.2, β = 104.3° a = 42.1, b = 41.3, c = 71.4, β = 104.2° a = 42.2, b = 41.4, c = 71.9, β = 104.2° Resolution range (Å) 1.42–69.0 1.30–40.9 1.35–69.7 Wavelength (Å) 0.976300 0.975522 0.975522 Radiation source EMBL, P14 EMBL, P14 EMBL, P14 Unique reflections number 42097 56403 52382 Rmerge, overall (outer shell) 0.042(0.241) 0.067 (0.334) 0.088 (0.338) I/σ overall (outer shell) 22.7(7.2) 13.1 (5.0) 10.8 (4.1) Multiplicity overall (outer shell) 7.0 (6.6) 6.9 (6.9) 6.8 (6.7) Completeness (%) overall (outer shell) 92.8 (74.1) 96.4 (94.5) 98.8 (99.0) Wilson B-factor 13.2 13.1 9.3 Refinement statistics: Rwork 0.157 0.119 0.116 Rfree 0.185 0.157 0.156 RMSD bond lengths, (Å) 0.011 0.013 0.033 RMSD bond angles (°) 2.000 1.991 2.195 Average B factors (Å2): all 16.7 20.4 15.0 main-chain 13.5 16.3 10.3 side-chain 16.2 21.7 14.8 inhibitors 26.9 13.3 17.3 waters 27.6 33.1 31.3 zinc 7.7 9.1 4.7 other molecules 40.5 37.0 36.9 Number of atoms: all 2562 2380 2482 protein 2181 2111 2127 inhibitor 69 16 28 water 287 226 275 zinc 1 1 1 other molecules 24 26 51 Ramachandran statistics (%): most favored regions 96 97 97 additionally allowed regions 4 3 3 outliers 0 0 0 Table 3. Data collection and refinement statistics of human CA VI-mimic and its complexes with inhibitors inhibitor 14, 25, and EZA. All datasets were collected at 100 K, test set size was 10%. Table 3. Data collection and refinement statistics of human CA VI-mimic and its complexes with inhibitors inhibitor 14, 25, and EZA. All datasets were collected at 100 K, test set size was 10%. Discussion Hydrophobic part of active site is shown as white surface, whereas hydrophilic part is shown as gray surface. (A) Compound 14 bound to active sites of CA VI-mimic (two alternative conformations of 14 are shown in cyan and blue, PDB ID: 6QL1) and CA II. The second ring of the “cyan” conformation of 14 in CA VI-mimic is not resolved in crystal structure, and not shown. (B) EZA bound to active site of CA VI-mimic (PDB ID: 6QL2) and CA II (PDB ID 3CAJ). Interaction of L198 with the first ring of compound is marked by thick dashed yellow line. Yellow line designates hydrophobic pocket for binding of para-substituent. The typical interactions for sulfonamide moiety are indicated, whereas they are omitted for clarity in other panels. (C, D) Compound 25 bound to active site of CA VI-mimic (PDB ID: 6QL3) and CA II (two alternative conformations of the ligand are indicated, PDB ID 4PYY). Blue line designates a hydrophilic part of the active site of CA VI-mimic. Figure 6. Differences in the binding structural modes of three compounds in CA II and CA VI-mimic as determined by X-ray crystallography. The zinc ion in active sites of CAs is shown as a light blue sphere. CA II side chain residues and ligands bound to CA II are colored pink and are shown transparent. CA VI-mimic side chains as well as its ligands are colored blue and also shown transparent. Mutated CA II side chains are labeled red for CA II, blue for CA VI-mimic. Hydrogen bonds are shown as dashed lines and distances are marked. Hydrophobic part of active site is shown as white surface, whereas hydrophilic part is shown as gray surface. (A) Compound 14 bound to active sites of CA VI-mimic (two alternative conformations of 14 are shown in cyan and blue, PDB ID: 6QL1) and CA II. The second ring of the “cyan” conformation of 14 in CA VI-mimic is not resolved in crystal structure, and not shown. (B) EZA bound to active site of CA VI-mimic (PDB ID: 6QL2) and CA II (PDB ID 3CAJ). Interaction of L198 with the first ring of compound is marked by thick dashed yellow line. Yellow line designates hydrophobic pocket for binding of para-substituent. The typical interactions for sulfonamide moiety are indicated, whereas they are omitted for clarity in other panels. Discussion Nowadays enzymes encompass over one-third of drug targets investigated by large pharmaceutical companies43, thereby emphasizing the relevance of target-based drug approach. This strategy aims to identify the compounds which would exhibit the most therapeutically beneficial effect via modulating catalytic activity or expression levels of disease-associated enzymes. The present study is focused on CA VI isoform as a drug target due to the link of CA VI with several pathologies19,20,22. Even though several studies on the design of compounds targeting CA VI have been reported40,44,45, CA VI-selective inhibitor has not been discovered so far. Therefore, there is an interest in inhibitors with high affinity and selectivity against CA VI which would be crucial to reveal biological function of CA VI. During preclinical development, numerous high-throughput screening assays are employed to design and optimize hits toward a target protein. Therefore, in vitro techniques require high quantities of recombinant pro- teins for the proper evaluation of compound quality, efficacy, and safety before testing in humans. Despite recent advances in molecular sciences, difficulties in the production of recombinant proteins in large scale are observed. Our previous study40 indicated low yield of CA VI from E. coli. Here we have presented a strategy to apply CA VI-mimic as CA VI model protein for the investigation on enzymatic inhibition and inhibitor binding thermody- namics. The CA VI-mimic was designed via site-directed mutagenesis from CA II by introducing five point muta- tions, such as A65T, N67Q, F130Y, V134Q, and L203T. Such approach to obtain CA protein based on the CA II mutant for the search of CA-selective inhibitors has been successfully applied previously. It was due to troubles to obtain sufficient amounts of recombinant CA isoforms and enabled by high structural homology between human Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 10 www.nature.com/scientificreports/ Figure 6. Differences in the binding structural modes of three compounds in CA II and CA VI-mimic as determined by X-ray crystallography. The zinc ion in active sites of CAs is shown as a light blue sphere. CA II side chain residues and ligands bound to CA II are colored pink and are shown transparent. CA VI-mimic side chains as well as its ligands are colored blue and also shown transparent. Mutated CA II side chains are labeled red for CA II, blue for CA VI-mimic. Hydrogen bonds are shown as dashed lines and distances are marked. Discussion (C, D) Compound 25 bound to active site of CA VI-mimic (PDB ID: 6QL3) and CA II (two alternative conformations of the ligand are indicated, PDB ID 4PYY). Blue line designates a hydrophilic part of the active site of CA VI-mimic. CAs. McKenna’s group reported a number of structural studies on CA IX-mimic based on CA II mutant with 2 mutations (S65A, Q67N)46–48, 7 mutations (A65S, N67Q, E69T, I91L, F131V, K170E, L204A)49–51 or 8 mutations (A9K, S65A, Q67N, T69E, L91I, V131F, E170K, A204L)52. Our group also published the exploration on inhibition parameters and binding thermodynamics via the application of CA IX-mimic as CA II mutant with 6 mutations (S65A, Q67N, L91I, V130F, L134V, A203L) and CA XII-mimic as CA II mutant with 6 mutations (S65A, K67N, T91I, A130F, S134V, N203L)53. The significant findings of the listed studies promoted the present investigation of CA II mutant mimicking CA VI. Over 10-fold higher purification yield of CA VI-mimic compared to CA VI from E. coli allowed kinetic, thermodynamic, and structural analyses of 43 benzenesulfonamides binding to CA VI-mimic. Even though the most tested inhibitors exhibited moderate affinities towards CA VI-mimic, this study provided insight into the structure-based design of inhibitors with better affinity and selectivity towards CA VI. The observed kinetic parameters of CA VI and CA VI-mimic are consistent with previous works. The deter- mined catalytic constant of CA VI compares reasonable well to published kcat value (3.4 × 105 s−1)44. The differ- ence of kcat values most likely arise from the uncertain CO2 concentration in the previous work. The CA VI-mimic had the same catalytic activity as CA II (kcat – 6.0 × 105 s−1) and confirmed previously published results that A65, N67, F130, V134, L203 amino acids in the active site of CA II are not important for catalytic activity54–56. Three CAs. McKenna’s group reported a number of structural studies on CA IX-mimic based on CA II mutant with 2 mutations (S65A, Q67N)46–48, 7 mutations (A65S, N67Q, E69T, I91L, F131V, K170E, L204A)49–51 or 8 mutations (A9K, S65A, Q67N, T69E, L91I, V131F, E170K, A204L)52. Our group also published the exploration on inhibition parameters and binding thermodynamics via the application of CA IX-mimic as CA II mutant with 6 mutations (S65A, Q67N, L91I, V130F, L134V, A203L) and CA XII-mimic as CA II mutant with 6 mutations (S65A, K67N, T91I, A130F, S134V, N203L)53. www.nature.com/scientificreports/ times higher catalytic activity of CA VI-mimic than CA VI is an advantage in measuring nanomolar inhibition constants by SFA, because similar to all enzymatic methods it is limited by both CA activity and concentration26. y y y y The relevance of protonation-deprotonation reactions occurring additionally upon inhibitor binding to CA has been reviewed by several groups57,58. Such protonation events have been recently confirmed by neutron crys- tallography59,60. To generate compounds with great affinities by rational design, it is essential to understand the structural reasons for the changes in binding affinities of the investigated compounds towards the target. Only intrinsic parameters subtract the contribution of protonation reactions occurring in conjunction with the bind- ing reaction between the CA and inhibitor. In the present study, nonfluorinated benzenesulfonamides exhibited stronger intrinsic and weaker observed binding affinity than corresponding fluorinated compounds. This result is in line with the previous investigation34, emphasizing the impact of fluorine electronegativity on the lowering of the pKa of inhibitor sulfonamide group but not the direct recognition of CA VI-mimic surface. The diminished pKa of fluorinated inhibitors led to the elevated observed affinity due to the increased fraction of inhibitor in the deprotonated form that bound to CA VI-mimic with the protonated zinc-bound hydroxide ion in the active site. Furthermore, substituents at ortho and para, but not meta positions were identified to be significant for the molecular recognition between the compound and CA VI. However, ortho and meta-substituted benzenesulfona- mides have been recently shown to act as tight CA IX binders32. Therefore, such findings confirmed that intrinsic, but not observed parameters should be applied to analyze the dependence of binding efficiency on compound chemical structures, thereby allowing important structure-thermodynamics correlations to design CA-isoform selective inhibitors. Inhibitor binding affinity can be significantly affected by structural properties of CA VI which have been widely investigated. The crystal structure of recombinant CA VI catalytic domain, lacking signal sequence and C-terminal region, has revealed its dimeric arrangement with the active sites of monomers facing each other and directed towards the center of the dimer61. Interestingly, the recent study62 has indicated that pentraxin domain (PTX) is present in non-mammalian CA VI, whereas PTX-coding exon is not found in mammalian CA6 gene most likely due to rearrangements occurring upon the duplication of the adjacent glucose transporter genes. www.nature.com/scientificreports/ Instead of the PTX domain, mammalian CA VI contains a C-terminal region of at least 25 residues which is not detected in other vertebrate CA isoforms. This part of CA VI may be important to form oligomers and bind other proteins affecting CA VI enzymatic activity or causing biological effect as a consequence of CA VI-protein inter- actions. Therefore, in vitro and in vivo studies on targeting CA VI can yield discrepancies in results because of the structural differences between recombinant CA VI applied for inhibitor screening and endogenous CA VI of live model organisms, such as mice or zebrafish.h gi The combination of data obtained from enzymatic inhibition and biophysical binding methods, such as FTSA and ITC, significantly strengthens the conclusions of compound structure-activity relationships. Since techniques are based on different strategies to characterize inhibitor efficacies, the precision and accuracy of the measurements are necessary to reliably select the most potent and strongest inhibitors/binders. The uncertainty and repeatability of FTSA42 and ITC63,64 measurements have been previously discussed. Furthermore, correlation between the affinities determined by FTSA and SFA or FTSA and ITC have been recently reported26,58 and con- firmed in the present study. Therefore, both enzymatic inhibition and biophysical binding techniques are neces- sary for precise identification of inhibitors with great affinity and selectivity towards the particular CA isoform, thereby leading to the success in clinical development. y g p In conclusion, this study on site-directed mutagenesis of residues in the active site of CA II to resemble CA VI gave clues to the basis for isoform specificity of benzenesulfonamides towards CA VI over CA II. The characteri- zation of numerous properties, such as kinetics of binding, inhibition profiles and the mechanism of action, pro- vided the deeper insight into the efficacy of CA VI-targeting inhibitors. This in vitro step is crucial because only sufficiently characterized compound can result in the success on translating experimental data to a clinical disease setting. Moreover, the present kinetic, thermodynamic, and structural information is important for experiments in silico, such as machine learning, when current binding information will be determined, filtered, and extracted. Discussion The significant findings of the listed studies promoted the present investigation of CA II mutant mimicking CA VI. Over 10-fold higher purification yield of CA VI-mimic compared to CA VI from E. coli allowed kinetic, thermodynamic, and structural analyses of 43 benzenesulfonamides binding to CA VI-mimic. Even though the most tested inhibitors exhibited moderate affinities towards CA VI-mimic, this study provided insight into the structure-based design of inhibitors with better affinity and selectivity towards CA VI. Th b d k f d h k Th d Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 www.nature.com/scientificreports/ www.nature.com/scientificreports/ each sense and antisense primer, and 1.5 U DNA polymerase. Thermal cycling conditions: initial denaturation – 95 °C for 10 min, then 18 cycles: 95 °C for 3 min, annealing – 66 °C (A65T and F130Y), 63 °C (N67Q), 71 °C (V134Q), or 72 °C (L203T) for 2 min, extension – 72 °C for 8 min, final extension – one time 72 °C for 10 min. After temperature cycling, PCR product was treated with Dpn I restriction endonuclease in order to digest the parental DNA template and to select new synthesized mutated DNA73. The mutations were confirmed by DNA sequencing. each sense and antisense primer, and 1.5 U DNA polymerase. Thermal cycling conditions: initial denaturation – 95 °C for 10 min, then 18 cycles: 95 °C for 3 min, annealing – 66 °C (A65T and F130Y), 63 °C (N67Q), 71 °C (V134Q), or 72 °C (L203T) for 2 min, extension – 72 °C for 8 min, final extension – one time 72 °C for 10 min. After temperature cycling, PCR product was treated with Dpn I restriction endonuclease in order to digest the parental DNA template and to select new synthesized mutated DNA73. The mutations were confirmed by DNA sequencing. q g Expression of CA VI-mimic protein was carried out in E. coli BL21(DE3) strain. Transformed cells colony was transferred to LB medium, containing 100 µg/ml ampicillin, grown at 37 °C and 220 rpm for 16 h. Then the satu- rated culture was diluted (1:50) in fresh LB medium, containing 100 µg/ml ampicillin and 60 µM ZnSO4 and grown to OD600 ≈ 0.8. The expression of CA VI-mimic protein was induced with 0.2 mM isopropyl-β-D-thiogalactoside (IPTG) and 0.4 mM ZnSO4. The cells were grown over night at 19 °C, 220 rpm and harvested by centrifugation at 4000 g for 20 min at 4 °C. g The biomass was suspended in the lysis buffer (20 mM HEPES, 0.15 M NaCl, and 1 mM PMSF, pH 7.4), incu- bated at 4 °C for 60 min and then disrupted by sonication. Debris of cells and insoluble proteins precipitated after centrifugation at 30 000 g for 25 min. The soluble CA VI-mimic protein was purified using a metal chelate and CA-affinity chromatography. For the metal chelate chromatography the column was equilibrated with 20 mM HEPES, 0.15 M NaCl (pH 7.4). www.nature.com/scientificreports/ Inhibitor concentration was 0–20 µM in <0.2% DMSO. Raw curves were fitted using a single exponential model and the inhibition constants were determined using Morrison equation74,75: . = − ⋅ + + − + + − CA act (%) 1 100% CA I IC CA I IC CA I CA ([ ] [ ] ([ ] [ ] ) 4[ ][ ]) 2[ ] 50 50 2 where [CA] is the total concentration of the active CA molecules, [I] is the total added inhibitor concentration, and IC50 is the concentration of inhibitor that achieves 50% inhibition of enzymatic activity. A dose-response curve was fitted using fixed CA concentration and assuming that it is equal to the active enzyme concentration. where [CA] is the total concentration of the active CA molecules, [I] is the total added inhibitor concentration, and IC50 is the concentration of inhibitor that achieves 50% inhibition of enzymatic activity. A dose-response curve was fitted using fixed CA concentration and assuming that it is equal to the active enzyme concentration. Inhibitor binding by FTSA. FTSA measurements were performed using a Corbett Rotor-Gene 6000 (Qiagen Rotor-Gene Q) instrument using the blue channel (excitation 365 ± 20 nm, detection 460 ± 15 nm). Samples contained 10 µL of 10 µM CA VI-mimic protein, 10 µL of 0–200 µM inhibitor in 50 mM phosphate buffer at pH 7.0 containing 100 mM NaCl, 50 µM solvatochromic dye 8-anilino-1-naphthalene sulfonate (ANS) and a final DMSO concentration of 2%. The applied heating rate was 1 °C/min. The pH dependence of the observed binding constant was measured in universal buffer containing 50 mM sodium phosphate, 50 mM sodium acetate, 25 mM sodium borate at pH 5.0–10.0. Data were fitted and analyzed as previously described72,76. Experiments were repeated at least twice. Inhibitor binding by ITC. ITC measurements were performed using a VP-ITC instrument (Microcal Inc., Northampton, USA) with 1.4 mL of 4–6 µM CA VI-mimic protein solution in the cell and 300 µL of 40–60 µM ligand solution in the syringe. A typical experiment consisted of 25–30 injections (10 µL each) added at 200–240 s intervals. In order to determine the pH dependence of the observed binding enthalpy, experiments were per- formed at 25 °C in 50 mM phosphate or 50 mM TRIS buffer containing 100 mM NaCl at pH 5.0–10.0 with a final DMSO concentration of 1%, equal in the syringe and the cell. www.nature.com/scientificreports/ For elution of CA VI-mimic protein, solution composed of 20 mM HEPES, 0.15 M NaCl, and 0.2 M imidazole (pH 7.4) was used. Eluted protein was purified using a CA-affinity column containing p-aminomethylbenzene sulfonamide-agarose (Sigma-Life Science Aldrich). Sorbent was equilibrated with 20 mM HEPES, 0.15 M NaCl (pH 7.4). For the protein elution, solution composed of 0.1 M sodium acetate and 0.5 M sodium perchlorate (pH 5.6) was used. Eluted CA VI-mimic protein was dialyzed into storage buffer containing 20 mM HEPES, 0.05 M NaCl, pH 7.4, and stored at −80 °C.h p The purity of CA VI-mimic protein was analyzed by SDS-PAGE. Protein concentrations were determined by UV-vis spectrophotometry using extinction coefficient ɛ280 = 51910 M−1 cm−1 and confirmed by standard Bradford method. Molecular mass of CA VI-mimic protein was confirmed by Mass spectrometer: observed – 29192.4 Da, theoretically predicted – 29323.0 Da. The difference is due to Met residue removed during production. Enzymatic activity and inhibition by SFA. Enzymatic activity and inhibition experiments were per- formed using an Applied Photophysics SX.18MV-R stopped-flow spectrophotometer at 23 °C. Saturated CO2 solution was prepared by bubbling the CO2 gas in Milli-Q water at 23 °C for 1 h. The concentration of CO2 was determined using a model described previously26. Catalytic constants kcat and Michaelis constants KM of CA VI and CA VI-mimic were determined in a pH range from 6.0 to 8.4 using 25 mM buffer and 30–300 µM indicator systems with similar pKa values: MES (pKa 6.1) and Bromocresol Purple (pKa 6.4, λ - 590 nm, pH 6.0–6.4), MOPS (pKa 7.2) and Bromothymol Blue (pKa 7.1, λ - 615 nm, pH 6.8–7.1), HEPES (pKa 7.5) and Phenol Red (pKa 7.5, λ - 557 nm, pH 7.2–7.8), TRIS (pKa 8.06) and m-Cresol Purple (pKa 8.3, λ - 575 nm, pH 8.0–8.4). CA VI and CA VI -mimic concentration was 50–100 nM. The ionic strength of solution was maintained at 0.2 M by the addition of sodium sulfate. Maximal velocities vmax were obtained using Lineweaver-Burk coordinates, and kcat, pKa values were determined using single ionization model: = + − − k k 1 10 cat cat max pH pKa Enzyme inhibition experiments were performed using 25 mM HEPES buffer containing 0.2 M sodium sulfate and 50 µM Phenol Red indicator, pH 7.5. Enzyme concentration was 10–30 nM for CA II, and 50 nM for CA VI and 50–107 nM CA VI-mimic. Methods S h i Synthesis of CA inhibitors. The synthesis of CA inhibitors has been previously described65–68. EZA, MZM nd TFS were purchased from Sigma-Aldrich (St. Louis, MO, USA) and were used without further purification. Synthesis of CA inhibitors. The synthesis of CA inhibitors has been previously described65–68. EZA, M and TFS were purchased from Sigma-Aldrich (St. Louis, MO, USA) and were used without further purificati Production of CA VI-mimic protein. The structural superpositions of proteins for Figs 1D and 6(A–D) were performed using UCSF Chimera v. 1.1269. The residues within 5 Å from the typical CA II inhibitor in PDB entry 3M96 in both isoforms were analyzed, and five residues which were different between the isoforms were selected to create CA VI-mimic. The structure-based alignment was generated using PROMALS3D web server70. The sequence alignment figure was prepared using TeXshade package71.h h gi g g g The expression vector pET15b-CA II72, encoding full length CA II (1-260), was used in site-directed mutagen- esis. The residues located in CA II active site, A65, N67, F130, V134, and L203, were replaced to T, Q, Y, Q, and T, respectively. For each mutagenesis reaction two oligonucleotide primers (sense and antisense) with target muta- tion were used: A65T_s: 5‘CTC AAC AAT GGT CAT ACT TTC AAC GTG GAG3’ and A65T_a: CTC CAC GTT GAA AGT ATG ACC ATT GTT GAG; N67Q_s: CAA TGG TCA TAC TTT CCA GGT GGA GTT TGA TGA C and N67Q_a: GTC ATC AAA CTC CAC CTG GAA AGT ATG ACC ATT G; F130Y_s: CCA AAT ATG GGG ATT ATG GGA AAG CTG TGC AG and F130Y_a: CTG CAC AGC TTT CCC ATA ATC CCC ATA TTT GG; V134Q_s: GAT TAT GGG AAA GCT CAG CAG CAA CCT GAT GG and V134Q_a: CCA TCA GGT TGC TGC TGA GCT TTC CCA TAA TC; L203T_s: GAC CAC CCC TCC TCT TAC GGA ATG TGT GAC CTG and L203T_a: CAG GTC ACA CAT TCC GTA AGA GGA GGG GTG GTC. PCR was carried out with high fidelity Pfu DNA polymerase (Thermo Fisher Scientific), except V134Q created with Phusion DNA polymerase (Thermo Fisher Scientific). Composition of PCR: 1× polymerase buffer, 50 ng template DNA, 0.2 mM dNTPmix, 125 ng Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 www.nature.com/scientificreports/ www.nature.com/scientificreports/ = + − − − f 10 1 10 _ _ RSO NH pH pK pH pK a SA a SA 2 = − + − − f 1 10 1 10 _ _ CAZnH O pH pK pH pK a CA a CA 2 = + − − − f 10 1 10 _ _ RSO NH pH pK pH pK a SA a SA 2 = − + − − f 1 10 1 10 _ _ CAZnH O pH pK pH pK a CA a CA 2 The intrinsic Gibbs energy change (Δ G b intr) is associated with the change in Kd_intr for the binding reaction. Δ = G RT K ln _ b intr d intr Δ = G RT K ln _ b intr d intr The Kd_intr values for the tested benzenesulfonamide binding to CA VI-mimic were calculated using the pKa of 7.0 for CA VI-mimic at 37 °C.h The Kd_intr values for the tested benzenesulfonamide binding to CA VI-mimic were calculated using the pKa f 7.0 for CA VI-mimic at 37 °C.h The Δ H b obs was measured by ITC as the sum of enthalpies caused by inhibitor binding to CA and protonation events, such as protonation enthalpies of buffer (Δ H p buf ), sulfonamide inhibitor (Δ H p SA), and hydroxide bound to zinc in the active site of CA (Δ H p CA): Δ = Δ − Δ − Δ + Δ H H n H n H n H , b intr b obs SA p SA CA p CA buf p buf where = − − n f 1 SA RSO NH 2 is the number of protons released from the inhibitor to buffer, = − n f 1 CA CAZnH O 2 is the number of protons bound to zinc-bound hydroxide of CA, and = + n n n buf CA SA is the sum of uptaken or released protons by buffer. The enthalpy of protonation of TRIS and sodium phosphate buffers at 25 °C is equal to −47.4 kJ/mol and −5.1 kJ/mol, respectively78. is Crystallization. The CA VI-mimic was concentrated by ultrafiltration to 19 mg/mL. Crystallization condi- tion (buffer) was 0.1 M sodium BICINE (pH 9.0), 0.2 M ammonium sulfate and 2 M sodium malonate (pH 7.0). www.nature.com/scientificreports/ Data were integrated, fitted and analyzed as previ- ously described77. Experiments were repeated at least twice. Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 13 www.nature.com/scientificreports/ Calculation of the intrinsic thermodynamics. The enzymatic or biophysical assays allow the determi- nation of observed inhibitor affinity for CA. However, observed parameters depend on buffer or pH. Therefore, observed values are only relevant for the comparison of inhibitor binding affinities towards the target using the same experimental conditions and should not be used in structure-thermodynamics correlations in drug design. Calculation of the intrinsic thermodynamics. The enzymatic or biophysical assays allow the determi- nation of observed inhibitor affinity for CA. However, observed parameters depend on buffer or pH. Therefore, observed values are only relevant for the comparison of inhibitor binding affinities towards the target using the same experimental conditions and should not be used in structure-thermodynamics correlations in drug design. p y g g Several protonation events take place upon the interaction between the inhibitor and CA: protonation of zinc-bound hydroxide in the active site of CA, deprotonation of inhibitor sulfonamide group, bond formation between CA and inhibitor, and compensating protonation-deprotonation reactions of buffer. To develop com- pounds with great affinities in the rational drug design, intrinsic parameters must be determined by subtracting the contribution of protonation reactions occurring in the conjunction with the inhibitor binding to CA58.h p g j g The parameter of Kd_intr is directly related to Kd_obs and fractions of deprotonated sulfonamide-based inhibitor − f( ) RSO NH 2 and CA with protonated zinc-bound hydroxide (water molecule) in the active site f( ) CAZnH O 2 . = × × − K K f f _ _ d intr d obs RSO NH CAZnH O 2 2 The fractions of binding-ready inhibitor and CA depend on the pKa of sulfonamide amino group (pKa_SA) and the pKa of water molecule in the active site of CA (pKa_CA), respectively. www.nature.com/scientificreports/ The ligand solutions for crystal soaking were made by mixing of 50 μL of corresponding reservoir solution and 1 μL of 50 mM ligand solution (in DMSO). Data collection and crystallographic structure determination. Three datasets of X-ray diffrac- tion (CA VI-mimic in complex with inhibitor 14 (VD10-50), 25 (VD11-4-2), and EZA) were collected at the EMBL beamline P14. The datasets were processed by XDS program79. The molecular replacement was made by MOLREP program80 using as initial model 4HT0. The 3D models of compounds were created by AVOGADRO program81. The library files which contain complete chemical and geometric descriptions of compounds were created using LIBCHECK program82,83. The models were prepared using COOT84 and refined using REFMAC85. All represented graphics were made using Pymol programs (PyMOL, version 1.8.4.0). Coordinates and structure factors have been deposited to the RCSB Protein Data Bank (PDB). The PDB access codes are listed in Table 3. 7. Mboge, M. Y., Mahon, B. P., McKenna, R. & Frost, S. C. Carbonic Anhydrases: Role in pH Control and Cancer. Metabolites 8 (2018) 8. Alterio, V., Di Fiore, A., D’Ambrosio, K., Supuran, C. T. & De Simone, G. Multiple Binding Modes of Inhibitors to Carbonic Anhydrases: How to Design Specific Drugs Targeting 15 Different Isoforms? Chem. Rev. 112, 4421–4468 (2012).i Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 y gi g g gf 9. Pinard, M. A. et al. Probing the Surface of Human Carbonic Anhydrase for Clues towards the Design of Isoform Specific Inhibitors, Probing the Surface of Human Carbonic Anhydrase for Clues towards the Design of Isoform Specific Inhibitors. BioMed Research International, BioMed Research International 2015, 2015, e453543 (2015). 5. Thiry, A., Dogné, J.-M., Supuran, C. T. & Masereel, B. Carbonic anhydrase inhibitors as anticonvulsant agents. Curr Top Med Chem 7, 855–864 (2007). 6. Scozzafava, A., Supuran, C. T. & Carta, F. Antiobesity carbonic anhydrase inhibitors: a literature and patent review. Expert Opinion on Therapeutic Patents 23, 725–735 (2013). 2. Frost, S. C. Physiological functions of the alpha class of carbonic anhydrases. Subcell. Biochem. 75, 9 30 (2014). 3. Aggarwal, M., Boone, C. D., Kondeti, B. & McKenna, R. Structural annotation of human carbonic anhydrases. J Enzyme Inhib Med Chem 28, 267–277 (2013). 4 Mincione F et al Carbonic anhydrase inhibitors: design of thioureido sulfonamides with potent isozyme II and XII inhibitory 4. Mincione, F. et al. Carbonic anhydrase inhibitors: design of thioureido sulfonamides with potent isozyme II and XII inhibitory properties and intraocular pressure lowering activity in a rabbit model of glaucoma. Bioorg. Med. Chem. Lett. 15, 3821–3827 (2005) 5 Thiry A Dogné J M Supuran C T & Masereel B Carbonic anhydrase inhibitors as anticonvulsant agents Curr Top Med Chem References 1. Meldrum, N. & Roughton, F. Some properties of carbonic anhydrase, the CO2 enzyme present in blood. J Physiol 75, 15–6 (1932). 2. Frost, S. C. Physiological functions of the alpha class of carbonic anhydrases. Subcell. Biochem. 75, 9–30 (2014). 3. Aggarwal, M., Boone, C. D., Kondeti, B. & McKenna, R. Structural annotation of human carbonic anhydrases. J Enzyme Inhib Med Chem 28, 267–277 (2013). 4. Mincione, F. et al. Carbonic anhydrase inhibitors: design of thioureido sulfonamides with potent isozyme II and XII inhibitory properties and intraocular pressure lowering activity in a rabbit model of glaucoma. Bioorg. Med. Chem. Lett. 15, 3821–3827 (2005). 5. Thiry, A., Dogné, J.-M., Supuran, C. T. & Masereel, B. Carbonic anhydrase inhibitors as anticonvulsant agents. Curr Top Med Chem 7, 855–864 (2007). 6. Scozzafava, A., Supuran, C. T. & Carta, F. Antiobesity carbonic anhydrase inhibitors: a literature and patent review. Expert Opinion on Therapeutic Patents 23, 725–735 (2013). 7. Mboge, M. Y., Mahon, B. P., McKenna, R. & Frost, S. C. Carbonic Anhydrases: Role in pH Control and Cancer. Metabolites 8 (2018). 8. Alterio, V., Di Fiore, A., D’Ambrosio, K., Supuran, C. T. & De Simone, G. Multiple Binding Modes of Inhibitors to Carbonic Anhydrases: How to Design Specific Drugs Targeting 15 Different Isoforms? Chem. Rev. 112, 4421–4468 (2012). 9. Pinard, M. A. et al. Probing the Surface of Human Carbonic Anhydrase for Clues towards the Design of Isoform Specific Inhibitors, Probing the Surface of Human Carbonic Anhydrase for Clues towards the Design of Isoform Specific Inhibitors. BioMed Research International, BioMed Research International 2015, 2015, e453543 (2015). 1. Meldrum, N. & Roughton, F. Some properties of carbonic anhydrase, the CO2 enzyme present in blood. J Physiol 75, 15–6 (1932). 2. Frost, S. C. Physiological functions of the alpha class of carbonic anhydrases. Subcell. Biochem. 75, 9–30 (2014). 3. Aggarwal, M., Boone, C. D., Kondeti, B. & McKenna, R. Structural annotation of human carbonic anhydrases. J Enzyme Inhib Med Chem 28, 267–277 (2013). 1. Meldrum, N. & Roughton, F. Some properties of carbonic anhydrase, the CO2 enzyme present in blood. J Physiol 75, 15–6 (1932). 2. Frost, S. C. Physiological functions of the alpha class of carbonic anhydrases. Subcell. Biochem. 75, 9–30 (2014). Chem 28, 267 277 (2013). 4. Mincione, F. et al. Carbonic anhydrase inhibitors: design of thioureido sulfonamides with potent isozyme II and XII inhibitory properties and intraocular pressure lowering activity in a rabbit model of glaucoma. Bioorg. Med. Chem. Lett. 15, 3821–3827 (2005).h References 1. Meldrum, N. & Roughton, F. Some properties of carbonic anhydrase, the CO2 enzyme present in blood. J Physiol 75, 15–6 (1932). 2. Frost, S. C. Physiological functions of the alpha class of carbonic anhydrases. Subcell. Biochem. 75, 9–30 (2014). 3. Aggarwal, M., Boone, C. D., Kondeti, B. & McKenna, R. Structural annotation of human carbonic anhydrases. J Enzyme Inhib Med Chem 28, 267–277 (2013). 4. Mincione, F. et al. Carbonic anhydrase inhibitors: design of thioureido sulfonamides with potent isozyme II and XII inhibitory properties and intraocular pressure lowering activity in a rabbit model of glaucoma. Bioorg. Med. Chem. Lett. 15, 3821–3827 (2005). 5. Thiry, A., Dogné, J.-M., Supuran, C. T. & Masereel, B. Carbonic anhydrase inhibitors as anticonvulsant agents. Curr Top Med Chem 7, 855–864 (2007). 6. Scozzafava, A., Supuran, C. T. & Carta, F. Antiobesity carbonic anhydrase inhibitors: a literature and patent review. Expert Opinion on Therapeutic Patents 23, 725–735 (2013). 7. Mboge, M. Y., Mahon, B. P., McKenna, R. & Frost, S. C. Carbonic Anhydrases: Role in pH Control and Cancer. Metabolites 8 (2018). References Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 www.nature.com/scientificreports/ 10. Murakami, H. & Sly, W. S. Purification and characterization of human salivary carbonic anhydrase. J. Biol. Chem. 262, 1382– (1987). ( ) 11. Kivelä, J. et al. Secretory carbonic anhydrase isoenzyme (CA VI) in human serum. Clin. Chem. 43, 2318–2322 (1997). ( ) 11. Kivelä, J. et al. Secretory carbonic anhydrase isoenzyme (CA VI) in human serum. Clin. Chem. 43, 2318–2322 (1997).hi 2. Karhumaa, P. et al. The identification of secreted carbonic anhydrase VI as a constitutive glycoprotein of human and rat milk. Pro Natl Acad Sci USA 98, 11604–11608 (2001). hi Natl Acad Sci USA 98, 11604–11608 (2001). 3. Leinonen, J. S., Saari, K. A., Seppänen, J. M., Myllylä, H. M. & Rajaniemi, H. J. Immunohistochemical demonstration of carbonic anhydrase isoenzyme VI (CA VI) expression in rat lower airways and lung. J. Histochem. Cytochem. 52, 1107–1112 (2004). y y p y g y 4. Kaseda, M., Ichihara, N., Nishita, T., Amasaki, H. & Asari, M. Immunohistochemistry of the bovine secretory carbonic anhydras isozyme (CA-VI) in bovine alimentary canal and major salivary glands. J. Vet. Med. Sci. 68, 131–135 (2006).ii aseda, M., Ichihara, N., Nishita, T., Amasaki, H. & Asari, M. Immunohistochemistry of the bovine secretory carbonic anhydrase ozyme (CA-VI) in bovine alimentary canal and major salivary glands. J. Vet. Med. Sci. 68, 131–135 (2006).ii y y j y g 5. Pan, P. et al. Gene expression profiling in the submandibular gland, stomach, and duodenum of CAVI-deficient mice. Transgenic Res 20, 675–698 (2011). 16. Patrikainen, M. S., Pan, P., Barker, H. R. & Parkkila, S. Altered gene expression in the lower respiratory tract of Car6 (−/−) mice. Transgenic Res. 25, 649–664 (2016).h g 7. Patrikainen, M., Pan, P., Kulesskaya, N., Voikar, V. & Parkkila, S. The role of carbonic anhydrase VI in bitter taste perception evidence from the Car6−/− mouse model. J Biomed Sci 21, 82 (2014).h 8. Melis, M. et al. The gustin (CA6) gene polymorphism, rs2274333 (A/G), as a mechanistic link between PROP tasting and fungiform taste papilla density and maintenance. PLoS One 8, e74151 (2013). p p y 19. Kivela, J., Parkkila, S., Parkkila, A. K., Leinonen, J. & Rajaniemi, H. Salivary carbonic anhydrase isoenzyme VI. J. Physiol. (Lond.) 520(Pt 2), 315–320 (1999). 0. Kimoto, M., Kishino, M., Yura, Y. & Ogawa, Y. A role of salivary carbonic anhydrase VI in dental plaque. Arch. Oral Biol. References et al. Intrinsic thermodynamics of 4-substituted-2,3,5,6-tetrafluorobenzenesulfonamide binding to carbonic anhydrases by isothermal titration calorimetry. Biophys. Chem. 205, 51–65 (2015). 5. Baranauskienė, L. & Matulis, D. Intrinsic thermodynamics of ethoxzolamide inhibitor binding to human carbonic anhydrase XIII BMC Biophys 5, 12 (2012). p y ( ) 36. Rosano, G. L. & Ceccarelli, E. A. Recombinant protein expression in Escherichia coli: advances and challenges. Front Microbiol 5 (2014). 37. Jia, B. & Jeon, C. O. High-throughput recombinant protein expression in Escherichia coli: current status and future perspec Open Biol 6 (2016). p 38. Gopal, G. J. & Kumar, A. Strategies for the production of recombinant protein in Escherichia coli. Protein J. 32, 419–425 (2013 39. Idicula-Thomas, S. & Balaji, P. V. Understanding the relationship between the primary structure of proteins and its propensity soluble on overexpression in Escherichia coli. Protein Sci. 14, 582–592 (2005). 40. Kazokaitė, J., Milinavičiūtė, G., Smirnovienė, J., Matulienė, J. & Matulis, D. Intrinsic binding of 4-substituted-2,3,5,6- tetrafluorobenezenesulfonamides to native and recombinant human carbonic anhydrase VI. FEBS J. 282, 972–983 (2015). l 41. Morkūnaitė, V. et al. Intrinsic thermodynamics of sulfonamide inhibitor binding to human carbonic anhydrases I and II. J Enzyme Inhib Med Chem 30, 204–211 (2015).h 2. Cimmperman, P. & Matulis, D. Chapter 8:Protein Thermal Denaturation Measurements via a Fluorescent Dye. In Chapter 8:Protein Thermal Denaturation Measurements via a Fluorescent Dye 247–274 (2011). h y 43. Holdgate, G. A., Meek, T. D. & Grimley, R. L. Mechanistic enzymology in drug discovery: a fresh perspective. Nature Reviews Drug Discovery 17, 115–132 (2018).i 44. Nishimori, I. et al. Carbonic anhydrase activators: the first activation study of the human secretory isoform VI with amino acids and amines. Bioorg. Med. Chem. 15, 5351–5357 (2007). g 5. Winum, J.-Y., Montero, J.-L., Vullo, D. & Supuran, C. T. Carbonic anhydrase inhibitors: glycosylsulfanilamides act as subnanomolar inhibitors of the human secreted isoform VI. Chem Biol Drug Des 74, 636–639 (2009). 46. Tars, K. et al. Sulfocoumarins (1,2-Benzoxathiine-2,2-dioxides): A Class of Potent and Isoform-Selective Inhibitors of Tumor- Associated Carbonic Anhydrases. J. Med. Chem. 56, 293–300 (2013). 47. Genis, C. et al. Design of a carbonic anhydrase IX active-site mimic to screen inhibitors for possible anticancer properties. Biochemistry 48, 1322–1331 (2009).i y 8. Sippel, K. H. et al. Characterization of Carbonic Anhydrase Isozyme Specific Inhibition by Sulfamated 2-Ethylestra Compounds Lett. Drug Des. Discovery 8, 1–25 (2011).i 49. Moeker, J. et al. References 51, 117–122 (2006). 1. Parkkila, S. et al. Salivary carbonic anhydrase protects gastroesophageal mucosa from acid injury. Dig. Dis. Sci. 42, 1013–1019 (1997). 22. Hassan, M. I., Shajee, B., Waheed, A., Ahmad, F. & Sly, W. S. Structure, function and applications of carbonic anhydrase isozymes. Bioorg Med Chem 21, 1570–1582 (2013). 23. Pastorek, J. & Pastorekova, S. Hypoxia-induced carbonic anhydrase IX as a target for cancer therapy: From biology to clinical use. Seminars in Cancer Biology 31, 52–64 (2015). gy 4. Gibbons, B. H. & Edsall, J. T. Rate of Hydration of Carbon Dioxide and Dehydration of Carbonic Acid At 25 Degrees. J. Biol. Chem 238, 3502–3507 (1963).hl 5. Khalifah, R. G. The carbon dioxide hydration activity of carbonic anhydrase. I. Stop-flow kinetic studies on the native human isoenzymes B and C. J. Biol. Chem. 246, 2561–2573 (1971). 26. Smirnovienė, J., Smirnovas, V. & Matulis, D. Picomolar inhibitors of carbonic anhydrase: Importance of inhibition and binding assays. Anal. Biochem. 522, 61–72 (2017).t y 7. Kranz, J. K. & Schalk-Hihi, C. Protein thermal shifts to identify low molecular weight fragments. Meth. Enzymol. 493, 277–298 (2011).hfl 8. Niesen, F. H., Berglund, H. & Vedadi, M. The use of differential scanning fluorimetry to detect ligand interactions that promote protein stability. Nat Protoc 2, 2212–2221 (2007).h y 29. Matulis, D., Kranz, J. K., Salemme, F. R. & Todd, M. J. Thermodynamic stability of carbonic anhydrase: measurements of binding affinity and stoichiometry using ThermoFluor. Biochemistry 44, 5258–5266 (2005). fi y y gh y 0. Rogez-Florent, T. et al. Label-free characterization of carbonic anhydrase-novel inhibitor interactions using surface plasmon resonance, isothermal titration calorimetry and fluorescence-based thermal shift assays. J. Mol. Recognit. 27, 46–56 (2014).h ylt y g 31. Garbett, N. C. & Chaires, J. B. Thermodynamic studies for drug design and screening. Expert Opin Drug Discov 7, 299 31. Garbett, N. C. & Chaires, J. B. Thermodynamic studies for drug design and screening. Expert Opin Drug Discov 7, 299–314 (2012). b l Th d d f d b d l d lf d d h y g g g p p g 2. Zubrienė, A. et al. Intrinsic Thermodynamics and Structures of 2,4- and 3,4-Substituted Fluorinated Benzenesulfonamides Binding to Carbonic Anhydrases. ChemMedChem 12, 161–176 (2017). y 33. Linkuvienė, V. et al. Intrinsic thermodynamics of inhibitor binding to human carbonic anhydrase IX. Biochimica et Biophysica Acta (BBA) - General Subjects 1860, 708–718 (2016). j 4. Zubrienė, A. Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 Acknowledgements Authors thank Gleb Bourenkov for the help with data collection at P14 EMBL beamline at PETRA III ring of the DESY synchrotron. The access to the beamline was supported by iNEXT (grant number 653706) funded by the Horizon 2020 program of the European Commission. The study was also supported by the grant S-MIP-17-87 from the Research Council of Lithuania. References Structural Insights into Carbonic Anhydrase IX Isoform Specificity of Carbohydrate-Based Sulfamates. Journal of Medicinal Chemistry 57, 8635–8645 (2014).i 50. Pinard, M. A., Aggarwal, M., Mahon, B. P., Tu, C. & McKenna, R. A sucrose-binding site provides a lead towards an isoform-specific inhibitor of the cancer-associated enzyme carbonic anhydrase IX. Acta Crystallogr F Struct Biol Commun 71, 1352–1358 (2015). 51. Mahon, B. P. et al. Saccharin: A lead compound for structure-based drug design of carbonic anhydrase IX inhibitors. Bioorganic & Medicinal Chemistry 23 849–854 (2015) 50. Pinard, M. A., Aggarwal, M., Mahon, B. P., Tu, C. & McKenna, R. A sucrose-binding site provides a lead towards an isoform-specific inhibitor of the cancer-associated enzyme carbonic anhydrase IX. Acta Crystallogr F Struct Biol Commun 71, 1352–1358 (2015). y y y g 1. Mahon, B. P. et al. Saccharin: A lead compound for structure-based drug design of carbonic anhydrase IX inhibitors. Bioorganic & Medicinal Chemistry 23, 849–854 (2015). 15 Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 www.nature.com/scientificreports/ 52. Pinard, M. A., Boone, C. D., Rife, B. D., Supuran, C. T. & McKenna, R. Structural study of interaction between brinzolamide dorzolamide inhibition of human carbonic anhydrases. Bioorg Med Chem 21, 7210–7215 (2013). 53. Dudutienė, V. et al. Discovery and characterization of novel selective inhibitors of carbonic anhydrase IX. J. Med. Chem. 57, 9435–9446 (2014). 54. Krebs, J. F. & Fierke, C. A. Determinants of catalytic activity and stability of carbonic anhydrase II as revealed by random mutagenesis. J. Biol. Chem. 268, 948–954 (1993). g 5. Behravan, G., Jonasson, P., Jonsson, B. H. & Lindskog, S. Structural and functional differences between carbonic anhydrase isoenzymes I and II as studied by site-directed mutagenesis. Eur. J. Biochem. 198, 589–592 (1991). 56. Sjöblom, B., Polentarutti, M. & Djinovic-Carugo, K. Structural study of X-ray induced activation of carbonic anhydrase. Proc. Natl. Acad. Sci. USA 106, 10609–10613 (2009). 57. Krishnamurthy, V. M. et al. Carbonic anhydrase as a model for biophysical and physical-organic studies of proteins and protein- ligand binding. Chem. Rev. 108, 946–1051 (2008).h g g 8. Linkuvienė, V. et al. Thermodynamic, kinetic, and structural parameterization of human carbonic anhydrase interactions toward enhanced inhibitor design. Quarterly Reviews of Biophysics 51 (2018). g g . Linkuvienė, V. et al. Thermodynamic, kinetic, and structural par . Kovalevsky, A. et al. ‘To Be or Not to Be’ Protonated: Atomic Det 59. Kovalevsky, A. et al. References ‘To Be or Not to Be’ Protonated: Atomic Details of Human Carbonic Anhydrase-Clinical Drug Complexes by Neutron Crystallography and Simulation. Structure 26, 383–390.e3 (2018). 59. Kovalevsky, A. et al. ‘To Be or Not to Be’ Protonated: Atomic Details of Human Carbonic Neutron Crystallography and Simulation. Structure 26, 383–390.e3 (2018). y g p y ( ) 60. Koruza, K. et al. Using neutron crystallography to elucidate the basis of selective inhibition of carbonic anhydrase by saccharin a derivative Journal of Structural Biology https://doi org/10 1016/j jsb 2018 12 009 (2019) y g y 60. Koruza, K. et al. Using neutron crystallography to elucidate the basis of selective inhibition of carbonic anhydrase by saccharin and a derivative. Journal of Structural Biology, https://doi.org/10.1016/j.jsb.2018.12.009 (2019). rnal of Structural Biology, https://doi.org/10.1016/j.jsb.2018.12.009 a derivative. Journal of Structural Biology, https://doi.org/10.1016 1. Pilka, E. S., Kochan, G., Oppermann, U. & Yue, W. W. Crystal structure of the secretory isozyme of mammalian carbonic anhydrase CA VI: implications for biological assembly and inhibitor development Biochem Biophys Res Commun 419 485 489 (2012) 61. Pilka, E. S., Kochan, G., Oppermann, U. & Yue, W. W. Crystal structure of the secretory isozyme of mammalian carbonic an CA VI: implications for biological assembly and inhibitor development. Biochem. Biophys. Res. Commun. 419, 485–489 (2 pp y y y y CA VI: implications for biological assembly and inhibitor development. Biochem. Biophys. Res. Commun. 419, 485–489 (2012). k l d fi d h f l b fi h ( ) b h d 62. Patrikainen, M. S. et al. Identification and characterization of a novel zebrafish (Danio rerio) pentraxin-carbonic anhydrase. PeerJ 5, e4128 (2017). 63. Linkuvienė, V., Krainer, G., Chen, W.-Y. & Matulis, D. Isothermal titration calorimetry for drug design: Precision of the enthalpy and binding constant measurements and comparison of the instruments. Anal. Biochem. 515, 61–64 (2016). 4. Paketurytė, V., Linkuvienė, V., Krainer, G., Chen, W.-Y. & Matulis, D. Repeatability, precision, and accuracy of the enthalpies and Gibbs energies of a protein–ligand binding reaction measured by isothermal titration calorimetry. European Biophysics Journal 64. Paketurytė, V., Linkuvienė, V., Krainer, G., Chen, W.-Y. & Matulis, D. Repeatability, precision, and accuracy of the enthalpie Gibbs energies of a protein–ligand binding reaction measured by isothermal titration calorimetry. European Biophysics Jou https://doi.org/10.1007/s00249-018-1341-z (2018). p g 5. Dudutienė, V. et al. Functionalization of fluorinated benzenesulfonamides and their inhibitory properties toward carbonic anhydrases. ChemMedChem 10, 662–687 (2015). y 66. Zubrienė, A. References PLoS One 7, e36899 (2012). 78. Goldberg, R. N., Kishore, N. & Lennen, R. M. Thermodynamic Quantities for the Ionization Reactions of Buffers. Journal of Physical d Ch i l R f D t 31 (2002) 77. Kazlauskas, E. et al. Thermodynamics of Aryl-Dihydroxyphenyl-Thiadiazole Binding to Human Hsp90. PLoS One 7, e36899 (2012). 78. Goldberg, R. N., Kishore, N. & Lennen, R. M. Thermodynamic Quantities for the Ionization Reactions of Buffers. Journal of Physical and Chemical Reference Data 31 (2002). b h ( ) f 79. Kabsch, W. XDS. Acta Cryst D 66, 125–132 (2010). 79. Kabsch, W. XDS. Acta Cryst D 66, 125–132 (2010). 0. Vagin, A. & Teplyakov, A. MOLREP: an Automated Program for Molecular Replacement. J Appl Cryst, J Appl Crystallogr 30 1022–1025 (1997). 81. Hanwell, M. D. et al. Avogadro: An advanced semantic chemical editor, visualization, and analysis platform. J Cheminform 4, 17 (2012). ( ) 2. Vagin, A. A. et al. REFMAC5 dictionary: organization of prior chemical knowledge and guidelines for its use. Acta Crystallogr D Bio Crystallogr 60, 2184–95 (2004). y g 3. Winn, M. D. et al. Overview of the CCP4 suite and current developments. Acta Crystallogr D Biol Crystallogr 67, 235–242 (2011). 83. Winn, M. D. et al. Overview of the CCP4 suite and current developments. Acta Crystallogr D Biol Crystallogr 67, 235–242 (2011). 84. Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486–501 (2010).i 5. Murshudov, G. N., Vagin, A. A. & Dodson, E. J. Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. D Biol. Crystallogr. 53, 240–255 (1997). y g y g ( ) 86. Čapkauskaitė, E. et al. Combinatorial Design of Isoform-Selective N-Alkylated Benzimidazole-Based Inhibitors of Carbonic Anhydrases. ChemistrySelect 2, 5360–5371 (2017). y g y g 6. Čapkauskaitė, E. et al. Combinatorial Design of Isoform-Selective N-Alkylated Benzimidazole-Based Inhibitors of Carbonic Anhydrases. ChemistrySelect 2, 5360–5371 (2017). References et al. Benzenesulfonamides with benzimidazole moieties as inhibitors of carbonic anhydrases I, II, VII, XII and XIII. J Enzyme Inhib Med Chem 29, 124–131 (2014).l y ( ) 7. Dudutienė, V. et al. 4-Substituted-2,3,5,6-tetrafluorobenzenesulfonamides as inhibitors of carbonic anhydrases I, II, VII, XII, and XIII. Bioorg. Med. Chem. 21, 2093–2106 (2013). g 8. Čapkauskaitė, E. et al. Benzenesulfonamides with pyrimidine moiety as inhibitors of human carbonic anhydrases I, II, VI, VII, XII and XIII. Bioorganic & Medicinal Chemistry 21, 6937–6947 (2013). and XIII. Bioorganic & Medicinal Chemistry 21, 6937–6947 (20 g y ( ) 9. Pettersen, E. F. et al. UCSF Chimera–a visualization system for exploratory research and analysis. J Comput Chem 25, 1605–1612 (2004). 0. Pei, J., Kim, B.-H. & Grishin, N. V. PROMALS3D: a tool for multiple protein sequence and structure alignments. Nucleic Acids Re 36, 2295–2300 (2008). 71. Beitz, E. TeXshade: shading and labeling of multiple sequence alignments using LaTeX2e. Bioinformatics 16, 135–139 (2000). l d l f h l b l d d b l f b l d h 72. Cimmperman, P. et al. A quantitative model of thermal stabilization and destabilization of proteins by ligands. Biophys. 3222–3231 (2008). 73. Carrigan, P. E., Ballar, P. & Tuzmen, S. Site-directed mutagenesis. Methods Mol. Biol. 700, 107–124 (2011). 74. Morrison, J. F. Kinetics of the reversible inhibition of enzyme-catalysed reactions by tight-binding inhibitors. Biochim. Bioph 185, 269–286 (1969).h 75. Williams, J. W. & Morrison, J. F. The kinetics of reversible tight- 75. Williams, J. W. & Morrison, J. F. The kinetics of reversible tight-binding inhibition. Meth. Enzymol. 63, 437–467 (1979). h 76. Baranauskienė, L. et al. Inhibition and binding studies of carbonic anhydrase isozymes I, II thiadiazole-7-sulphonamides. J Enzyme Inhib Med Chem 25, 863–870 (2010).hh 76. Baranauskienė, L. et al. Inhibition and binding studies of carbonic anhydrase isozymes I, II and IX with benzimidazo[1,2-c thiadiazole-7-sulphonamides. J Enzyme Inhib Med Chem 25, 863–870 (2010).hh y 77. Kazlauskas, E. et al. Thermodynamics of Aryl-Dihydroxyphenyl-Thiadiazole Binding to Human Hsp90. PLoS One 7, e36899 (2012).hf skas, E. et al. Thermodynamics of Aryl-Dihydroxyphenyl-Thiadiazo 77. Kazlauskas, E. et al. Thermodynamics of Aryl-Dihydroxypheny 77. Kazlauskas, E. et al. Thermodynamics of Aryl-Dihydroxyphenyl-Thiadiazole Binding to Human Hsp90. PLoS One 7, e36899 (2012). 78 Goldberg R N Kishore N & Lennen R M Thermodynamic Quantities for the Ionization Reactions of Buffers Journal of Physical 77. Kazlauskas, E. et al. Thermodynamics of Aryl-Dihydroxyphenyl-Thiadiazole Binding to Human Hsp90. Author Contributions J.K., J.S., V.K., A.S., M.T., S.P. and D.M. participated in the conception or design of the study; J.K. produced recombinant proteins, carried out biophysical assays and analyzed thermodynamic parameters; V.K. performed computational modelling; J.S. carried out enzymatic activity and inhibition measurements; A.S. and E.M. were responsible for X-ray crystallographic analysis; J.K. wrote the first version of manuscript; J.K., V.K., J.S., A.S., E.M., M.T., S.P. and D.M. contributed to manuscript drafting and approved the final version of manuscript. Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 16 www.nature.com/scientificreports/ Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-019-49094-0. Competing Interests: D.M. declares that he has patents and patent applications pending on CA inhibitors. Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2019 Additional Information upplementary information accompanies this paper at https://doi.org/10.1038/s41598-019-49094-0. Competing Interests: D.M. declares that he has patents and patent applications pending on CA inhibitors. Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and i tit ti l ffili ti Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2019 Scientific Reports \| (2019) 9:12710 \| https://doi.org/10.1038/s41598-019-49094-0 17
https://openalex.org/W2145561697	https://research.rug.nl/files/48741171/De_Mol_2010_Accuracy_of_handheld_blood_glucose_.pdf	English	null	Accuracy of Handheld Blood Glucose Meters at High Altitude	PloS one	2,010	cc-by	6,278	University of Groningen University of Groningen Accuracy of Handheld Blood Glucose Meters at High Altitude de Mol, Pieter; Krabbe, Hans G.; de Vries, Suzanna T.; Fokkert, Marion J.; Dikkeschei, Bert D.; Rienks, Rienk; Bilo, Karin M.; Bilo, Henk J. G. de Mol, Pieter; Krabbe, Hans G.; de Vries, Suzanna T.; Fokkert, Marion J.; Dikkeschei, Bert D.; Rienks, Rienk; Bilo, Karin M.; Bilo, Henk J. G. DOI: 10.1371/journal.pone.0015485 DOI: 10.1371/journal.pone.0015485 IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Document Version Publisher's PDF, also known as Version of record Document Version Publisher's PDF, also known as Version of record Publication date: 2010 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): de Mol, P., Krabbe, H. G., de Vries, S. T., Fokkert, M. J., Dikkeschei, B. D., Rienks, R., Bilo, K. M., & Bilo, H. J. G. (2010). Accuracy of Handheld Blood Glucose Meters at High Altitude. PLoS ONE, 5(11), Article 15485. https://doi.org/10.1371/journal.pone.0015485 Citation for published version (APA): de Mol, P., Krabbe, H. G., de Vries, S. T., Fokkert, M. J., Dikkeschei, B. D., Rienks, R., Bilo, K. M., & Bilo, H. J. G. (2010). Accuracy of Handheld Blood Glucose Meters at High Altitude. PLoS ONE, 5(11), Article 15485. https://doi.org/10.1371/journal.pone.0015485 Abstract This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. funded by the Isala Clinics, Zwolle, The Netherlands. The funders had no role in study design, data collection and analysis, decisions to the manuscript. nding: This study was funded by the Isala Clinics, Zwolle, The Netherlands. The funders had no role in study design, data collection an blish or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. Competing Interests: The authors have declared that no competing interests exist. * E il P d M l@ i l * E-mail: P.deMol@aig.umcn.nl . These authors contributed equally to this work. . These authors contributed equally to this work. Previous studies have examined the accuracy and reliability of various glucose meters at (simulated) altitudes up to 5,000 m. These and other studies with type 1 diabetes mellitus patients indicated a considerable subset of the handheld blood glucose meters (BGMs) to be unreliable at high altitude [1–6]. One could hypothesize that glucose oxidase (GOX) BGMs, which use oxygen as one of the substrates, would underestimate true glucose values since less oxygen is available under hypobaric conditions. This effect has been demonstrated by some [5]. Others have de- monstrated oxygen independent BGMs using glucose dehydroge- nase (GDH), to perform better under hypobaric conditions and GOX based BGMs to underestimate true blood glucose levels [1]. However, these studies were limited by field- or simulated altitude settings only [3–5], test protocols using only GOX based BGMs [3], test protocols up to simulated altitudes ,4500 m [3–5] or with outdated material [3,4]. Pieter de Mol1., Hans G. Krabbe2., Suzanna T. de Vries3, Marion J. Fokkert2, Bert D. Dikkeschei2, Rienk Rienks4, Karin M. Bilo5, Henk J. G. Bilo5,6 1 Department of Internal Medicine, Canisius Wilhelmina Hospital, Nijmegen, The Netherlands, 2 Department of Clinical Chemistry, Isala Clinics, Zwolle, The Netherlands, 3 Department of Cardiology, Isala Clinics, Zwolle, The Netherlands, 4 Centre for Human Aviation, Dutch Airforce, Soesterberg, The Netherlands, 5 Department of Internal Medicine, Isala Clinics, Zwolle, The Netherlands, 6 Department of Internal Medicine, University Medical Centre, Groningen, The Netherlands PLoS ONE \| www.plosone.org Abstract Background: Due to increasing numbers of people with diabetes taking part in extreme sports (e.g., high-altitude trekking), reliable handheld blood glucose meters (BGMs) are necessary. Accurate blood glucose measurement under extreme conditions is paramount for safe recreation at altitude. Prior studies reported bias in blood glucose measurements using different BGMs at high altitude. We hypothesized that glucose-oxidase based BGMs are more influenced by the lower atmospheric oxygen pressure at altitude than glucose dehydrogenase based BGMs. Methodology/Principal Findings: Glucose measurements at simulated altitude of nine BGMs (six glucose dehydrogenase and three glucose oxidase BGMs) were compared to glucose measurement on a similar BGM at sea level and to a laboratory glucose reference method. Venous blood samples of four different glucose levels were used. Moreover, two glucose oxidase and two glucose dehydrogenase based BGMs were evaluated at different altitudes on Mount Kilimanjaro. Accuracy criteria were set at a bias ,15% from reference glucose (when .6.5 mmol/L) and ,1 mmol/L from reference glucose (when ,6.5 mmol/L). No significant difference was observed between measurements at simulated altitude and sea level for either glucose oxidase based BGMs or glucose dehydrogenase based BGMs as a group phenomenon. Two GDH based BGMs did not meet set performance criteria. Most BGMs are generally overestimating true glucose concentration at high altitude. Conclusion: At simulated high altitude all tested BGMs, including glucose oxidase based BGMs, did not show influence of low atmospheric oxygen pressure. All BGMs, except for two GDH based BGMs, performed within predefined criteria. At true high altitude one GDH based BGM had best precision and accuracy. Citation: de Mol P, Krabbe HG, de Vries ST, Fokkert MJ, Dikkeschei BD, et al. (2010) Accuracy of Handheld Blood Glucose Meters at High Altitude. PLoS ONE 5(11): e15485. doi:10.1371/journal.pone.0015485 Citation: de Mol P, Krabbe HG, de Vries ST, Fokkert MJ, Dikkeschei BD, et al. (2010) Accuracy of Handheld Blood Glucose Meters at High Altitude. PLoS ONE 5(11): e15485. doi:10.1371/journal.pone.0015485 bbe HG, de Vries ST, Fokkert MJ, Dikkeschei BD, et al. (2010) Accuracy of Handheld Blood Glucose Meters at High Altitude. PLoS 0.1371/journal.pone.0015485 Editor: Marian Ludgate, Cardiff University, United Kingdom Received August 4, 2010; Accepted October 1, 2010; Published November 12, 2010 Received August 4, 2010; Accepted October 1, 2010; Published November 12, 2010 Copyright: 2010 de Mol et al. Copyright Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date: 24-10-2024 November 2010 \| Volume 5 \| Issue 11 \| e15485 Accuracy of Handheld Blood Glucose Meters at High Altitude Pieter de Mol1., Hans G. Krabbe2., Suzanna T. de Vries3, Marion J. Fokkert2, Bert D. Dikkeschei2, Rienk Rienks4, Karin M. Bilo5, Henk J. G. Bilo5,6 Hypobaric chamber experiment Each sample was tested in duplicate under hypobaric and normobaric conditions simultaneously, and results are presented as the mean glucose levels with absolute (mmol/L) or relative (%) bias from GHex glucose determination. Four of six tested GDH and all three GOX based BGMs performed within predefined performance criteria of 61.0 mmol/ L when reference glucose was ,6.5 mmol/L and 615% when reference glucose was .6.5 mmol/L at all simulated altitudes. There were no mechanical failures due to hypobaric conditions. To prevent glucose variation due to time-dependent glycolysis by cellular uptake in blood samples (5–7% per hour [7]) a blood sample was prepared for GHex reference testing using a perchloric acid preparation method. The latter was done at the time spiked glucose samples were prepared – prior to BGM and GHex testing. These blood samples were used for both GHex and BGM glucose testing under normobaric and hypobaric conditions. One GDH based BGM (Freestyle mini) showed biases (range 15.2 to 18.5%) outside of predefined criteria for 7 out of 20 tested glucose samples at simulated altitudes of 2000 m and 3000 m. Also, this BGM showed biases outside of predefined criteria at 4 out of 20 sea level readings (range 16.0 to 18.5%). Furthermore, another GDH based BGM (Precision X-ceed) only showed biases outside of predefined criteria at sea level for 5, 10 and 15 mmol/L glucose samples (Table 1). In general, under hypobaric conditions (at simulated altitude) both the GOX and GDH based BGMs tended to overestimate glucose levels com- pared to normobaric conditions (bias range; GDH BGMs: 28.2 to 16.9% and GOX BGMs: 0 to 10.8%) (Figure 1, 2 and Table 1, 2). Table 1 and 2 demonstrate the relative bias (%) of the 10 mmol/L test sample, and is representative for the remaining tested glucose samples (5, 15 and 20 mmol/L). Hypobaric chamber experiment We used a hypobaric chamber of the Dutch Airforce Research facility in Soesterberg, The Netherlands. Six GDH and three GOX based BGMs were tested at simulated altitudes: sea level, 2000 m, 3000 m, 4000 m and 5000 m while temperature (20uC) and humidity were kept constant. GDH based BGMs used were; Freestyle Mini and Precision X-ceed (both Abbott Diabetes Care, USA), Hemocue 201+ (Quest Diagnostics U.K.), Accu-Chek Aviva and Accu-Chek Compact Plus (both Roche Diagnostics, Switzerland) and the Contour Link (Bayer, Germany). GOX based BGMs used were; Glucocard Memory (Menarini Diagnos- tics, Italy) Statstrip (Nova Biomedical, USA) and the Klinion (Klinion Diabetes Care, Medeco the Netherlands) All BGMs used were new - i.e. the same BGM was not extensively used for medical practice before the experiments took place - to prevent errors from wear and tear. Venous blood drawn from a healthy individual spiked to a glucose concentration of 5, 10, 15 and 20 mmol/L was used as substrate to test BGMs. True glucose values were determined with the laboratory certified Glucose Hexokinase (GHex) based method (Roche Diagnostics, Manheim, Germany). This GHex method is aligned with the GC-MS glucose reference method. BGMs were simultaneously tested inside and outside the hypobaric chamber, and samples were tested in duplicate. Methods Performance criteria for BGMs were set at a difference of 61 mmol/L of the reference method when ,6.5 mmol/L, or 615% when the reference sample read $6.5 mmol/L (Dutch Organisation of Applied Physics and Science (‘‘TNO’’) guideline, the Netherlands). Furthermore, Clarke’s error grid analysis was used to determine clinically relevant accuracy. [8] The error grid plots the reference glucose and the glucose measured by the BGM on a x- y plot. The grid is divided into 5 zones (A–E) of clinical accuracy and corresponding treatment assumptions to correct the glucose value. Zone A represents the target glucose range and glucose values are clinically accurate; Zone B represents glucose values that deviate .15% from the reference glucose leading to no or benign treatment; Zone C glucose values would lead to overcorrection of acceptable glucose values causing hyper- or hypoglycemia; Zone D represents glucose values that could lead to dangerous misdetection and failure to treat; Zone E glucose values are opposite from the reference glucose leading to very dangerous treatment decisions. Values in zone A and B are considered clinically acceptable. Glucose measured by BGMs is presented in means (mmol/L) 6 SEM unless stated otherwise. Deviation of mean BGM glucose from the reference method is presented in percentage (%) and it’s range. Overestimation is presented as the actual percentage and underestimation is preceded by a minus (-) sign. Two experiments were performed; First, all BGMs were tested at simulated altitudes in a hypobaric chamber. Second, a selection of GOX and GDH based BGMs in the first experiment were used during the ascent of Mount Kilimanjaro (5895 m). The local ethics committee of the Isala Clinics Zwolle, the Netherlands, approved the study protocol and all participants gave written informed consent. Introduction Regular exercise and a healthy life style should be part of daily life of everybody, but definitely of any person with diabetes mellitus. Minimal requirements for exercise have been formulated, with a minimum of 30 minutes of vigorous exercise or brisk walking, five times weekly. However, increasing numbers of people with diabetes do participate in more strenuous forms of physical activity, amongst others high-altitude trekking and mountain climbing. These kinds of activities do pose special challenges to subjects with diabetes mellitus treated with insulin. Glucose levels have to be kept under good control, and it is a challenge to find a balance between energy intake, energy expenditure, blood glucose levels, and insulin requirements. In general, frequent assessment of blood glucose levels will be necessary to allow proper adjustments. Therefore, both accuracy and easy accessibility of frequent blood glucose measurements in subjects with type 1 diabetes mellitus are paramount for a safe activity at high altitude. Since we planned an expedition to climb Mount Kilimanjaro with a team of people with type 1 diabetes mellitus, we were in November 2010 \| Volume 5 \| Issue 11 \| e15485 November 2010 \| Volume 5 \| Issue 11 \| e15485 1 Measuring Blood Glucose at High Altitude need of the most reliable handheld BGMs up to 6000 m altitude. The purpose of this study was to investigate whether modern GDH or GOX based BGMs are reliable under simulated and true high altitude conditions. altitudes. These control solutions are made for the Statstrip BGM and were primarily used to observe possible consistent biases within that same BGM (precision) with repeated measures at different altitudes. For between-BGM performance comparison the best-tested BGM from the hypobaric chamber experiment – the GDH based Accu-Chek Compact Plus- was used as a reference. No air humidity was measured at altitude. We hypothesized, that GDH based BGMs would be more accurate than GOX based BGMs at high altitude, since GDH based BGMs are by definition less oxygen dependent in their reaction than GOX based BGMs. Field experiment Relative bias (%) of GDH based BGMs compared to the Hexokinase laboratory reference method at different simulated altitudes (10 mmol/L glucose sample). Note: normobaric BGMs stay at sea level. (normo = bias under normobaric conditions (sea level); hypo = bias under hypobaric conditions (simulated altitude), * = reference method). Note: normobaric BGMs stay at sea level. (normo = bias under normobaric conditions (sea level); hypo = bias under hypobaric conditions (simulated altitude), * = reference method). doi:10.1371/journal.pone.0015485.t001 specific results of the Accu-chek Compact Plus, there was a trend of greater overestimation of 10 and 15 mmol/L glucose samples at higher simulated altitudes. Furthermore, the smallest of biases was found in the 20 mmol/L sample. However, the bias of this BGM (bias range 21.1 to 6.7% at all glucose concentrations and simulated altitudes) was well within predefined criteria for acceptable bias and within meter variation was minimal. Moreover, the meter was user-friendly, which is important, especially under extreme conditions such as high altitude mountaineering. Indeed, differences observed were seen in all tested glucose concentrations and were therefore not sample dependent. Moreover, no trend in bias was seen for the individual BGMs with increasing simulated altitude. However, one GDH-based BGM performed better under hypobaric than under normobaric conditions (Precision Xceed). The best performing BGM under hypo- and normobaric conditions and at all glucose concentrations tested was the GDH based BGM Accu-chek Compact Plus. When further assessing the Figure 1. Error grid analysis of 5, 10, 15 and 20 mmol/L glucose testing samples of BGMs (y- axis) against reference (GHex) (x-axis) at sea level. Note: Leftward deviation from dashed line means overestimation of true glucose by tested BGM compared to GHex reference glucose and vice versa. No differences were noted outside of predefined criteria due to testing conditions at sealevel (0 m) in and outside the hypobaric chamber. Black diamonds = Contour; Black squares = Accu Chek Aviva; Black circles = Accuchek Compact Plus; Black triangles = Freestyle Mini; Black asterixes = Precision; Black stripes = Hemocue; White squares = Klinion; White circles = Statstrip; White diamonds = Glucocard; Dashed line = GHex reference glucose. doi:10.1371/journal.pone.0015485.g001 Figure 1. Error grid analysis of 5, 10, 15 and 20 mmol/L glucose testing samples of BGMs (y- axis) against reference (GHex) (x-axis) at sea level. Field experiment On Mount Kilimanjaro testing took place at incremental altitudes (range 1300–4600 m), using capillary blood samples from eight subjects with type 1 diabetes and eight healthy control subjects. Care was taken to take a full droplet of blood from a subject’s warm and clean index finger. Four BGMs, two GDH based BGMs (Accu-Chek Compact Plus and Contour) and two GOX based BGMs (Klinion and Glucocard) were used. Samples were tested at temperatures ranging from 610–30uC using one single sample of each subject. GHex reference testing was not available at altitude. Instead, standard reference solutions (glucose range 2.4–4.1; 4.4–6.7 and 14.0–17.9 mmol/L; NOVA biomed- ical, USA) were used to test within-BGM performance at different As a group phenomenon, no significant difference was observed between GOX- and GDH-BGMs under hypobaric conditions (Bias at 2000–5000 m: 5.867.9 (25.0 to 19.1) vs. 7.566.5 (23.5 to 22.1) respectively (mean % 6 SEM (range %) (p = NS)). PLoS ONE \| www.plosone.org November 2010 \| Volume 5 \| Issue 11 \| e15485 PLoS ONE \| www.plosone.org 2 Measuring Blood Glucose at High Altitude Table 1. Relative bias (%) of GDH based BGMs compared to the Hexokinase laboratory reference method at different simulated altitudes (10 mmol/L glucose sample). GDH Hexokinase * Contour Link Accu-chek Aviva Accu-chek Compact Plus Freestyle Mini Precision Xceed Hemocue Altitude (m) Glucose (mmol/L) normo (%) hypo (%) normo (%) hypo (%) no rmo (%) hypo (%) normo (%) hypo (%) normo (%) hypo (%) normo (%) hypo (%) 0 10.05 3.8 4.3 3.8 1.5 2.9 1.5 2.4 15.5 11.1 9.5 22.6 20.5 2000 9.73 24.0 23.5 2.3 3.7 2.8 0.8 10.4 16.9 15.4 6.5 22.9 20.8 3000 9.38 22.5 2.8 0.8 23.0 2.8 2.8 12.8 15.2 15.9 8.1 26.5 23.0 4000 8.93 3.0 4.5 20.3 6.1 3.0 3.5 18.5 8.9 12.9 28.2 20.3 3.5 5000 8.50 6.1 0.0 6.1 5.0 4.5 4.5 17.1 11.5 12.8 0.6 2.9 7.6 Note: normobaric BGMs stay at sea level. (normo = bias under normobaric conditions (sea level); hypo = bias under hypobaric conditions (simulated altitude), * = reference method). doi:10.1371/journal.pone.0015485.t001 Table 1. Relative bias (%) of GDH based BGMs compared to the Hexokinase laboratory reference method at different simulated altitudes (10 mmol/L glucose sample). Table 1. Relative bias (%) of GDH based BGMs compared to the Hexokinase laboratory reference method at different simulated altitudes (10 mmol/L glucose sample). Table 1. Field experiment Note: Leftward deviation from dashed line means overestimation of true glucose by tested BGM compared to GHex reference glucose and vice versa. No differences were noted outside of predefined criteria due to testing conditions at sealevel (0 m) in and outside the hypobaric chamber. Black diamonds = Contour; Black squares = Accu Chek Aviva; Black circles = Accuchek Compact Plus; Black triangles = Freestyle Mini; Black asterixes = Precision; Black stripes = Hemocue; White squares = Klinion; White circles = Statstrip; White diamonds = Glucocard; Dashed line = GHex reference glucose. doi:10.1371/journal.pone.0015485.g001 November 2010 \| Volume 5 \| Issue 11 \| e15485 PLoS ONE \| www.plosone.org 3 Measuring Blood Glucose at High Altitude Figure 2. Error grid analysis of 5, 10, 15 and 20 mmol/L glucose testing samples of BGMs (y- axis) against reference (GHex) (x-axis) at 5000 m simulated altitude. Note: Leftward deviation from dashed line means overestimation of true glucose by tested BGM compared to GHex reference glucose and vice versa. Black diamonds = Contour; Black squares = Accu Chek Aviva; Black circles = Accuchek Compact Plus; Black triangles = Freestyle Mini; Black asterixes = Precision; Black stripes = Hemocue; White squares = Klinion; White circles = Statstrip; White diamonds = Glucocard; Dashed line = GHex reference glucose. doi:10.1371/journal.pone.0015485.g002 Figure 2. Error grid analysis of 5, 10, 15 and 20 mmol/L glucose testing samples of BGMs (y- axis) against reference (GHex) (x-axis) at 5000 m simulated altitude. Note: Leftward deviation from dashed line means overestimation of true glucose by tested BGM compared to GHex reference glucose and vice versa. Black diamonds = Contour; Black squares = Accu Chek Aviva; Black circles = Accuchek Compact Plus; Black triangles = Freestyle Mini; Black asterixes = Precision; Black stripes = Hemocue; White squares = Klinion; White circles = Statstrip; White diamonds = Glucocard; Dashed line = GHex reference glucose. doi:10.1371/journal.pone.0015485.g002 doi:10.1371/journal.pone.0015485.t002 10.1371/journal.pone.0015485.t002 Discussion Overestimation of glucose by BGMs at true altitude could have dangerous consequences when hypoglycemia is falsely not shown by BGM readings and hypoglycemic symptoms might not be fully sensed due to extreme conditions. At the hyperglycemic range, the therapeutic decisions made due to overestimation of true blood glucose have less consequences than in the hypoglycemic range, where not treating hypoglycemia could lead to life threatening conditions. In this perspective, the findings of underestimation by the Glucocard GOX based BGM in the low glycemic range are not considered as dangerous as vice versa. On the contrary, if overestimation of glucose by BGMs at simulated altitude is truly present at true altitude this might lead to dangerous situations. However, at simulated altitude overestimation was limited to 17 (9%) of all measurements, all of which fell in the lower B zone of the error grid leading to no or benign treatment errors. The present study was conducted to test accuracy and reliability of GOX and GDH based BGMs at simulated and true high altitude. We hypothesized that oxygen independent GDH based BGMs would be more reliable and accurate at high altitude. In contrast to previous reports and our hypothesis, no significant differences were observed between GOX- and GDH-based BGMs at simulated high altitude in all tested glucose ranges. At true high altitude within-meter variation and accuracy compared to standard reference glucose solutions was better in the GDH based BGMs, and best in the Accu-Chek Compact Plus BGM, independent from glucose solutions used. In general, most BGMs tended to overestimate blood glucose at high altitude. g g g Research of BGMs at simulated high altitude up to 4000 m by Gautier et al. has reported underestimation up to 228.9% using venous blood samples of glucose ranges from 1.5 up to 26.3 mmol/L. However, this study did not differentiate between GOX or GDH based BGMs and BGMs used are outdated today. [4] Furthermore, O¨ berg and O¨ stenson tested four GDH and one GOX based BGM at simulated and true high altitude up to 4500 and 5895 m respectively. Their study reported overestimation ranging from 6.5 up to 15% of the GOX based BGM at normo- and hyperglycemic glucose levels at 4500 and 2500 m. GDH based BGMs performed better and overestimated true glucose levels by 0.8 to 6.5% [1]. Field experiment GOX and GDH based BGMs at Mount Kilimanjaro at different altitudes tested with standard glucose solutions. Table 3. GOX and GDH based BGMs at Mount Kilimanjaro at different altitudes tested with standard glucose solutions. Table 3. GOX and GDH based BGMs at Mount Kilimanjaro at different altitudes tested with standard glucose solutions. Altitude Control solution GDH BGMs GOX BGMs (m) Glucose (mmol/L) Glucose (mmol/L) Glucose (mmol/L) Accu-chek Compact Plus Contour Klinion Glucocard 1300 2.4–4.1 3.9 2.7 4.6 1.9 4.4–6.7 5.6 5.5 6.0 4.7 14.0–17.9 14.7 15.7 17.5 14.7 3000 2.4–4.1 3.9 3.0 3.0 1.6 4.4–6.7 5.9 5.6 5.6 4.3 14.0–17.9 14.9 16.6 15.5 14.2 3700 2.4–4.1 3.9 2.5 3.2 1.8 4.4–6.7 5.4 5.7 5.9 4.4 14.0–17.9 14.5 15.5 16.4 13.8 4600 2.4–4.1 3.8 3.1 2.6 1.4 4.4–6.7 6.0 5.9 4.9 4.2 14.0–17.9 14.9 16.2 13.9 14.0 doi:10.1371/journal.pone.0015485.t003 doi:10.1371/journal.pone.0015485.t003 standard reference glucose solutions, it constantly showed underestimation of the lowest glucose reference sample and was at the low end of the other reference samples. This is not well explained by chemical technique used as the other GOX based BGM (Klinion) did not show this phenomenon. A possible explanation might be that substances in the reference glucose solution or errors in test strips interfered with glucose measure- ment in this meter. [9] criteria primarily observed in the 3.6 to 5.2 mmol/L mean glucose range, and specifically due to an overestimation (Fig 3). In comparison to the reference BGM, the Glucocard, a GOX based BGM, performed better than the Contour (GDH) and the Klinion (GOX) BGMs (3 vs. 10 vs. 8% of bias outside performance criteria respectively). Most BGM performance criteria violations were observed at altitudes of 3000–3900 m. PLoS ONE \| www.plosone.org Field experiment when analyzing BGM accuracy in relation to the glucose solutions reference measurements one GOX based BGM (Glucocard) consistently underestimated true glucose measurement especially in the lower glucose range (2.4 to 4.4 mmol/L). These results were in accordance with results previously obtained in the hypobaric chamber. Three standard glucose control solutions were used as a crude reference method of BGM performance at Mt Kilimanjaro at altitudes of 1300 m, 3000 m, 3700 m and 4600 m. Also one single sample of capillary blood of eight subjects with type 1 diabetes mellitus and eight healthy control subjects was used to test BGMs at altitudes of 1300 m, 3000 m, 3770 m, 3900 m and 4600 m altitude. When comparing BGMs on Mount Kilimanjaro with the best tested BGM in the hypobaric chamber (Accuchek Compact Plus; a GDH based BGM) for capillary glucose samples of subjects with type 1 diabetes mellitus and healthy controls at altitudes ranging 1300–4600 m, a total of 228 paired glucose measurements were analyzed. Of these, 47 (21%) were outside of set performance Measurements with the standard glucose control solutions demonstrated consistent results of all BGMs except for the Klinion GOX based BGM which showed a large within BGM variance at different altitudes in all glucose ranges tested (Table 3). However, Table 2. Relative bias (%) of GOX based BGMs compared to the Hexokinase laboratory reference method at different simulated altitudes (10 mmol/L glucose sample). GOX Hexokinase * Statstrip Klinion Glucocard Altitude(m) Glucos(mmol/L) normo(%) hypo(%) normo(%) hypo(%) normo(%) hypo(%) 0 10.05 12.6 9.5 25.8 5.2 2.9 5.2 2000 9.73 14.7 10.8 27.5 0.8 3.7 3.7 3000 9.38 11.1 5.8 22.5 3.4 25.3 4.8 4000 8.93 4.5 6.1 20.8 0.8 3.5 6.1 5000 8.50 2.3 5.6 22.4 0.0 4.5 6.6 Note: normobaric BGMs stay at sea level. (normo = bias under normobaric conditions (sea level); hypo = bias under hypobaric conditions (simulated altitude), * = reference method). doi:10.1371/journal.pone.0015485.t002 sed BGMs compared to the Hexokinase laboratory reference method at different simulated Table 2. Relative bias (%) of GOX based BGMs compared to the Hexokinase laboratory reference method at different simulated altitudes (10 mmol/L glucose sample). Table 2. Relative bias (%) of GOX based BGMs compared to the Hexokinase laboratory reference method at different simulated altitudes (10 mmol/L glucose sample). November 2010 \| Volume 5 \| Issue 11 \| e15485 Measuring Blood Glucose at High Altitude Table 3. Discussion Black diamonds = Contour; White squares = Klinion; White diamonds = Glucocard; Dashed line = Accu Chek Compact Plus BGM as a reference glucose. doi:10.1371/journal.pone.0015485.g003 tested at true high altitude a wide range of bias was reported and no reference method was used. but not on true bias due to the lack of a laboratory reference method at high altitude. Also, the reference solution used was not compared to a laboratory reference method upon return from altitude. However, BGMs showed consistent findings that were in good accordance with results obtained with the hypobaric chamber experiment. Our results differ from the findings in previous research. First, we did not find a significant method-related difference between GOX and GDH based BGMs at simulated high altitude. Second, although not compared to a laboratory reference method, we did not find a wide range of bias in GOX and GDH based BGMs up to 4600 m at true high altitude. Second, on Mount Kilimanjaro blood glucose levels of tested BGMs were compared to the best performing BGM as a reference method based on results in a hypobaric chamber under constant environmental conditions. One could speculate that alternating temperatures and humidity on a mountain might influence the accuracy of tested BGMs and the accuracy of the reference BGM in particular, thereby compromising results. However, when testing with standard glucose solutions, the reference BGM showed best accuracy and within meter variation (Table 3). Moreover, in our study we tested BGMs at temperatures ranging from 10 to 28uC and these are well within the range of temperatures stated by BGM manufacturers to provide reliable glucose measurements. A possible explanation for the first difference observed in contrast to the study of Gautier et al. [4] might be improved oxygen-based technique and sensitivity of the reagent for oxygen. This is reflected by the fact that it takes less time to read out glucose test strips results now than it took thirteen years ago, suggesting that GOX based BGMs need less oxygen in order to show a proper reaction. In our study, BGMs were tested at simulated high altitude under constant temperature and humidity, using a laboratory reference method. Previous research reported low temperatures and humidity influence accuracy of BGMs independent of altitude [1–2]. This might have compromised field research at true high altitude and explained the wide range of bias in cold conditions at 5895 m in the other studies. Discussion Interestingly, two similar GDH based BGMs were used and bias observed did not correspond with our study; the Contour (GDH based) BGM in their study showed an underestimation (21.9 to 24.2%) and the Freestyle (GDH based) BGM only showed a bias of +0.8% of both ,5.8 and ,16.5 mmol/L glucose samples. When GDH based BGMs were At simulated high altitude one GDH based BGM performed better under hypobaric conditions (Precision Xceed), possibly due to in-meter failure of the normobaric BGM. This is suggested by the fact that under normobaric conditions (sea level) this BGM showed constant biases close to 15% in the 5, 10 and 15 mmol/L glucose samples. Furthermore, the most inaccurate BGM in all glucose ranges was the GDH based Freestyle Mini (28% of total glucose readings . +15% bias) without an apparent effect of altitude. Lastly, there was a time dependent glucose lowering effect seen in the reference samples (GHex) possibly due to cellular uptake despite perchloric acid preparation of samples. (Table 1, 2) As expected, a parallel decrement in glucose was observed when testing all BGMs. At true high altitude, the GOX based Glucocard BGM had good within-meter variation and was reasonably accurate com- pared to the reference BGM. However, when testing with November 2010 \| Volume 5 \| Issue 11 \| e15485 5 Measuring Blood Glucose at High Altitude Figure 3. Error grid analysis of glucose samples of all (healthy controls and diabetes type 1) subjects measured by GOX- and GDH- BGMs (y-axis) at different altitudes on Mount Kilimanjaro (range 1300–4600 m) compared to Accuchek BGM glucose as a reference method (x-axis); (mmol/L). Note: leftward deviation from dashed line indicates overestimation of glucose measured by BGM and rightward deviation means underestimation by BGM compared to Accu-Chek BGM. Black diamonds = Contour; White squares = Klinion; White diamonds = Glucocard; Dashed line = Accu Chek Compact Plus BGM as a reference glucose. doi:10.1371/journal.pone.0015485.g003 Figure 3. Error grid analysis of glucose samples of all (healthy controls and diabetes type 1) subjects measured by GOX- and GDH- BGMs (y-axis) at different altitudes on Mount Kilimanjaro (range 1300–4600 m) compared to Accuchek BGM glucose as a reference method (x-axis); (mmol/L). Note: leftward deviation from dashed line indicates overestimation of glucose measured by BGM and rightward deviation means underestimation by BGM compared to Accu-Chek BGM. Discussion [1,2]. Since at 4600 m we tested BGM’s at 28uC, the difference in temperature might explain this contrasting result. Based on the tests in the hypobaric chamber and at true high altitude we concluded that the Accu-check Compact Plus GDH based BGM was most accurate at simulated altitude and most precise at true high altitude. PLoS ONE \| www.plosone.org Acknowledgments We would like to thank the Centre for Human Aviation and the Dutch Airforce in Soesterberg, the Netherlands, for providing the hypobaric chamber testing facilities and their assistance. If a similar effect of overestimation is present at true high altitude this could have dangerous consequences in the normo- and hypoglycemic range. Therefore, true high altitude studies with approved laboratory reference methods with measurement of humidity and temperature as possible confounders are warranted. Conclusion This study has limitations that might influence its results. No differences were observed between GDH and GOX based BGMs at simulated altitude up to 5000 m. All of the tested BGMs, First, on Mount Kilimanjaro we could only test BGM’s on within meter variation and relative bias compared to the best tested BGM, PLoS ONE \| www.plosone.org November 2010 \| Volume 5 \| Issue 11 \| e15485 6 Measuring Blood Glucose at High Altitude except for two GDH based BGMs, performed within defined criteria for acceptable accuracy at simulated altitude. In general, at simulated high altitude BGMs tend to overestimate true glucose levels. At true high altitude GDH based BGMs performed better in relation to within-meter variation and accuracy. 9. Ginsberg BH (2009) Factors affecting blood glucose monitoring: sources of errors in measurement. J Diabetes Sci Technol 3: 903–913. Author Contributions Conceived and designed the experiments: PdM HGK STdV MJF HJGB BDD. Performed the experiments: PdM KMB MJF RR. Analyzed the data: HGK PdM. Contributed reagents/materials/analysis tools: MJF HGK BDD. Wrote the paper: HGK PdM. Permitted the use of hypobaric chamber of the royal airforce: RR. Conceived and designed the experiments: PdM HGK STdV MJF HJGB BDD. Performed the experiments: PdM KMB MJF RR. Analyzed the data: HGK PdM. Contributed reagents/materials/analysis tools: MJF HGK BDD. Wrote the paper: HGK PdM. Permitted the use of hypobaric chamber of the royal airforce: RR. Conceived and designed the experiments: PdM HGK STdV MJF HJGB BDD. Performed the experiments: PdM KMB MJF RR. Analyzed the data: HGK PdM. Contributed reagents/materials/analysis tools: MJF HGK BDD. Wrote the paper: HGK PdM. Permitted the use of hypobaric chamber of the royal airforce: RR. p g 8. Clarke WL, Cox D, Gonder-Frederick LA, Carter W, Pohl SL (1987) Evaluating clinical accuracy of systems for self-monitoring of blood glucose. Diabetes Care 10: 622–628. 4. Gaultier JF, Bigard AX, Douce P, Duvallet A, Cathelineau G (1996) Influence of simulated high altitude on performance of five blood glucose meters. Diabetes Care 19: 1430–1433. 7. Chan AY, Swaminathan R, Cockram CS (1989) Effectiveness of sodium fluoride as a preservative of glucose in blood. Clin Chem 35: 315–317. 3. Fink KS, Christensen DB, Ellsworth A (2002) Effect of high altitude on blood glucose meter performance. Diabetes Technol Ther 4: 627–635. y y g J g 6. Pecchio O, Maule S, Migliardi M, Trento M, Veglio M (2000) Effects of exposure at an altitude of 3,000 m on performance of glucose meters (Letter). Diabetes Care 23: 129–131. 1. O¨ berg D, O¨ stenson CG (2005) Performance of Glucose Dehydrogenase- and Glucose Oxidase-Based Blood Glucose Meters at High Altitude and Low Temperature (Letter). Diabetes Care 28: 1261. 2. Moore K, Vizzard N, Coleman C, McMahon J, Hayes R, Thompson CJ (2001) Extreme altitude mountaineering and type 1 diabetes: the Diabetes Federation of Ireland Kilimanjaro Expedition. Diabet Med 18: 749–755. 3. Fink KS, Christensen DB, Ellsworth A (2002) Effect of high altitude on blood glucose meter performance. Diabetes Technol Ther 4: 627–635. 4. Gaultier JF, Bigard AX, Douce P, Duvallet A, Cathelineau G (1996) Influence of simulated high altitude on performance of five blood glucose meters. Diabetes Care 19: 1430–1433. PLoS ONE \| www.plosone.org 1. O¨ berg D, O¨ stenson CG (2005) Performance of Glucose Dehydrogenase- and Glucose Oxidase-Based Blood Glucose Meters at High Altitude and Low Temperature (Letter). Diabetes Care 28: 1261. p ( ) 2. Moore K, Vizzard N, Coleman C, McMahon J, Hayes R, Thompson CJ (2001) Extreme altitude mountaineering and type 1 diabetes: the Diabetes Federation of Ireland Kilimanjaro Expedition. Diabet Med 18: 749–755. 5. Bilen H, Kilicaslan A, Akcay G, Albayrak F (2007) Performance of glucose dehydrogenase (GDH) based and glucose oxidase (GOX) based blood glucose meters systems at moderately high altitude. J Med Eng Technol 31: 152–156. 5. Bilen H, Kilicaslan A, Akcay G, Albayrak F (2007) Performance of glucose dehydrogenase (GDH) based and glucose oxidase (GOX) based blood glucose meters systems at moderately high altitude. J Med Eng Technol 31: 152–156. 6. Pecchio O, Maule S, Migliardi M, Trento M, Veglio M (2000) Effects of exposure at an altitude of 3,000 m on performance of glucose meters (Letter). Diabetes Care 23: 129–131. 7. Chan AY, Swaminathan R, Cockram CS (1989) Effectiveness of sodium fluoride as a preservative of glucose in blood. Clin Chem 35: 315–317. 8. Clarke WL, Cox D, Gonder-Frederick LA, Carter W, Pohl SL (1987) Evaluating clinical accuracy of systems for self-monitoring of blood glucose. Diabetes Care 10: 622–628. 9. Ginsberg BH (2009) Factors affecting blood glucose monitoring: sources of errors in measurement. J Diabetes Sci Technol 3: 903–913. References 5. Bilen H, Kilicaslan A, Akcay G, Albayrak F (2007) Performance of glucose dehydrogenase (GDH) based and glucose oxidase (GOX) based blood glucose meters systems at moderately high altitude. J Med Eng Technol 31: 152–156. PLoS ONE \| www.plosone.org November 2010 \| Volume \| Issue \| e15485 5 11 7 7
https://openalex.org/W3087118494	https://link.springer.com/content/pdf/10.1007/s10157-020-01969-7.pdf	English	null	Initial responsiveness to darbepoetin alfa and its contributing factors in non-dialysis chronic kidney disease patients in Japan	Clinical and experimental nephrology	2,020	cc-by	7,504	Abstract Background Hyporesponsiveness to erythropoiesis-stimulating agents (ESAs) is associated with cardiovascular events and poor renal outcome in patients with chronic kidney disease (CKD). This study aimed to investigate the initial responsiveness to darbepoetin alfa (DA) and its contributing factors using the data from the BRIGHTEN. g yp p y p g g ( ) poor renal outcome in patients with chronic kidney disease (CKD). This study aimed to investigate the initial responsiveness to darbepoetin alfa (DA) and its contributing factors using the data from the BRIGHTEN. Methods Of 1980 patients enrolled at 168 facilities, 1695 were included in this analysis [285 patients were excluded mainly due to lack of hemoglobin (Hb) values]. The initial ESA response index (iEResI) was defined as a ratio of Hb changes over 12 weeks after DA administration per weight-adjusted total DA dose and contributing factors to iEResI were analyzed. Results The mean age was 70 ± 12 years (male 58.8%; diabetic nephropathy 27.6%). The median creatinine and mean Hb levels at DA initiation were 2.62 mg/dL and 9.8 g/dL, respectively. The most frequent number of DA administration dur- ing 12 weeks was 3 times (41.1%), followed by 4 (15.6%) times with a wide distribution of the total DA dose (15–900 μg). Remarkably, 225 patients (13.3%) did not respond to DA. Multivariate analysis showed that male gender, hypoglycemic agent use, iron supplementation, high eGFR, low Hb, low CRP, low NT-proBNP, and low urinary protein–creatinine ratio were independently associated with better initial response to DA (P = < 0.0001, 0.0108, < 0.0001, 0.0476, < 0.0001, 0.0004, 0.0435, and 0.0009, respectively). Conclusions Non-responder to DA accounted for 13.3% of patients with non-dialysis CKD. Iron supplementation, low CRP, low NT-proBNP, and less proteinuria were predictive and modifiable factors associated with better initial response to DA. Keywords Erythropoiesis-stimulating agents · Hyporesponsiveness · Darbepoetin alfa · Chronic kidney disease · Pre- dialysis Keywords Erythropoiesis-stimulating agents · Hyporesponsiveness · Darbepoetin alfa · Chronic kidney disease · Pre- dialysis Terumasa Hayashi1 · Hideki Kato2 · Kenichiro Tanabe3 · Masaomi Nangaku2 · Hideki Hirakata4 · Takashi Wada5 · Hiroshi Sato6 · Yasushi Yamazaki7 · Takao Masaki8 · Tatsuo Kagimura3 · Hiroyasu Yamamoto9 · Hiroki Hase10 · Masahiro Kamouchi11 · Enyu Imai12 · Kyoichi Mizuno13 · Manabu Iwasaki14 · Tadao Akizawa15 · Yoshiharu Tsubakihara16 · Shoichi Maruyama17 · Ichiei Narita18 Terumasa Hayashi1 · Hideki Kato2 · Kenichiro Tanabe3 · Masaomi Nangaku2 · Hideki Hirakata4 · Takashi Wada5 · Hiroshi Sato6 · Yasushi Yamazaki7 · Takao Masaki8 · Tatsuo Kagimura3 · Hiroyasu Yamamoto9 · Hiroki Hase10 · Masahiro Kamouchi11 · Enyu Imai12 · Kyoichi Mizuno13 · Manabu Iwasaki14 · Tadao Akizawa15 · Yoshiharu Tsubakihara16 · Shoichi Maruyama17 · Ichiei Narita18 Received: 9 January 2020 / Accepted: 4 August 2020 / Published online: 19 September 2020 © The Author(s) 2020 Clinical and Experimental Nephrology (2021) 25:110–119 https://doi.org/10.1007/s10157-020-01969-7 Clinical and Experimental Nephrology (2021) 25:110–119 https://doi.org/10.1007/s10157-020-01969-7 ORIGINAL ARTICLE Initial responsiveness to darbepoetin alfa and its contributing factors in non‑dialysis chronic kidney disease patients in Japan Terumasa Hayashi1 · Hideki Kato2 · Kenichiro Tanabe3 · Masaomi Nangaku2 · Hideki Hirakata4 · Takashi Wada5 · Hiroshi Sato6 · Yasushi Yamazaki7 · Takao Masaki8 · Tatsuo Kagimura3 · Hiroyasu Yamamoto9 · Hiroki Hase10 · Masahiro Kamouchi11 · Enyu Imai12 · Kyoichi Mizuno13 · Manabu Iwasaki14 · Tadao Akizawa15 · Yoshiharu Tsubakihara16 · Shoichi Maruyama17 · Ichiei Narita18 * Ichiei Narita naritai@med.niigata‑u.ac.jp DA administration DA was administered within 8 weeks after the registration along its product information (30 μg every 2 weeks for the initial dose, subcutaneously or intravenously, and the dosage and duration should be adjusted thereafter to maintain Hb levels at ≥ 11 g/dL); however, the dose adjustment was actu- ally entrusted to the physicians’ discretion in each facility as the BRIGHTEN was conducted in a real-world clinical setting. Study population incidence of CVD events, especially stroke, and mortality. Furthermore, post hoc analyses of CHOIR and TREAT tri- als suggested that poor outcomes might not have resulted from achieving higher Hb level itself but resulted from toxicities associated with high-dose ESAs, patient-related factors promoting ESA hyporesponsiveness, or a combina- tion of both [16, 17]. Thus, in the management of anemia for CKD patients, ESA hyporesponsiveness should be pre- dicted and appropriate dose of ESA should be adminis- tered to patients with ESA hyporesponsiveness. However, clinically relevant definition of ESA hyporesponsiveness is not available at present because its definition should be based on the index associated with poor renal, cardiovas- cular, or overall outcome [18]. Patients aged ≥ 20 years with estimated glomerular filtra- tion rate (eGFR) of < 60 mL/min/1.73 m2 (calculated with the Japanese equation [20]) who presented renal anemia (Hb < 11 g/dL) were enrolled from June 2014 to Sep- tember 2016. Patients scheduled to initiate maintenance dialysis or undergo kidney transplantation until 24 weeks after registration; those with history of ESA treatment (administered with ESA temporarily and > 12 weeks before registration were eligible); and those with malig- nant tumors under treatment, hematologic diseases, or hemorrhagic diseases were excluded. Of 1980 patients enrolled in 168 facilities, 285 patients were excluded mainly due to the lack of Hb values at 0 and 12 weeks (84 ± 14 days). Finally, 1695 patients were included in the data analysis (Fig. 1). Patients were observed for 96 weeks since DA administration. The “oBservational clinical Research In chronic kid- ney disease patients with renal anemia: renal proGnosis in patients with Hyporesponsive anemia To Erythropoiesis- stimulating agents, darbepoetiN alfa (BRIGHTEN)” is a multicenter prospective observational study conducted in a real-world clinical setting in Japan to explore the preva- lence of hyporesponsiveness to darbepoetin alfa (DA) and to establish an appropriate index of hyporesponsiveness to DA associated with poor renal outcome and CVD events in non-dialysis CKD patients. Patient enrollment started in June 2014, and the observation period ended in September 2018. Therefore, this study aimed to investigate the initial responsiveness to DA and its contributing factors using the data from the BRIGHTEN. Introduction mortality, and kidney function as well as quality of life in this patient population [6–11]. Furthermore, the ben- eficial effect of ESAs was greater with high hemoglobin (Hb) levels; however, the optimal target Hb level in ESA treatment remains controversial. A series of large RCTs comparing the effect of normalizing Hb level (> 13 g/ dL) with conservative Hb level (10–11 g/dL) on mortality and cardiovascular disease (CVD) events was conducted over the past two decades and showed a consistent result of harm with normalizing Hb level [12–15]. The Correc- tion of Hemoglobin and Outcomes in Renal Insufficiency (CHOIR) trial [14] and the Trial to Reduce Cardiovascular Events with Aranesp Therapy (TREAT) [15] conducted on pre-dialysis CKD patients showed that targeting higher Hb level compared with targeting lower Hb level increased the Anemia in patients with chronic kidney disease (CKD) is associated with poor renal, cardiovascular, and overall outcomes [1–5]. Observational studies and small rand- omized controlled trials (RCTs) showed that correction of anemia with erythropoiesis-stimulating agents (ESAs) was associated with beneficial outcomes in morbidity, Electronic supplementary material The online version of this article (https://doi.org/10.1007/s10157-020-01969-7) contains supplementary material, which is available to authorized users. * Ichiei Narita naritai@med.niigata‑u.ac.jp Extended author information available on the last page of the article Electronic supplementary material The online version of this article (https://doi.org/10.1007/s10157-020-01969-7) contains supplementary material, which is available to authorized users. * Ichiei Narita naritai@med.niigata‑u.ac.jp Extended author information available on the last page of the article Vol:1 .(123456789 3 Clinical and Experimental Nephrology (2021) 25:110–119 111 incidence of CVD events, especially stroke, and mortality. Furthermore, post hoc analyses of CHOIR and TREAT tri- als suggested that poor outcomes might not have resulted from achieving higher Hb level itself but resulted from toxicities associated with high-dose ESAs, patient-related factors promoting ESA hyporesponsiveness, or a combina- tion of both [16, 17]. Thus, in the management of anemia for CKD patients, ESA hyporesponsiveness should be pre- dicted and appropriate dose of ESA should be adminis- tered to patients with ESA hyporesponsiveness. However, clinically relevant definition of ESA hyporesponsiveness is not available at present because its definition should be based on the index associated with poor renal, cardiovas- cular, or overall outcome [18]. Materials and methods Patient baseline characteristics (age, sex, etiology of CKD, smoking status, medical history, comorbidities, hypogly- cemic agent use, renin–angiotensin system inhibitor use, iron supplementation, body mass index, and blood pres- sure) were collected at the study registration. The clear Study design The study design and other details of the study protocol have been published elsewhere [19]. The protocol was approved by the main institutional review board (Nagoya University; No. 2014-0027) and then by each participating center. The research was conducted under the health insurance system of Japan and in accordance with the principles of the Decla- ration of Helsinki and Ethical Guidelines on Clinical Stud- ies of the Ministry of Health, Labor, and Welfare of Japan. Written informed consent was provided by each participant. The research was designed, implemented, and overseen by the BRIGHTEN Executive Committee, together with rep- resentatives of Translational Research Center for Medical Innovation, Kobe, Japan, a third-party organization inde- pendent of the investigators’ institutions and responsible for data collection and analysis. The manuscript was prepared by one of the authors and subsequently revised and edited by all authors. The study was registered to ClinicalTrials. gov (NCT02136563) and UMIN-CTR (UMIN000013464). 1 3 1980 enrolled 1695 included in the analysis 285 excluded (there was some overlapping) 39 withdrawal of consent 11 conﬂicted with inclusion or exclusion criteria 32 withdrew before 12 weeks 97 had no data of DA administraon 244 had no available hemoglobin value at 0 and 12 weeks Fig. 1 Screening of patient. DA darbepoetin alfa Fig. 1 Screening of patient. DA darbepoetin alfa 112 Clinical and Experimental Nephrology (2021) 25:110–119 were analyzed using the general linear model adjusted by gender as a factor. We included all baseline demographic and clinical variables into univariate analysis; then variables significantly associated with iEResI (P < 0.05) were incorpo- rated into multivariate analysis. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA), and P values of < 0.05 were considered significant. definition of diabetes and dyslipidemia was not adopted for the BRIGHTEN, while the diagnosis was entrusted to inves- tigators’ judgement. Complete blood count including Hb, serum creatinine, albumin, iron, ferritin, total iron-binding capacity (TIBC), hemoglobin A1c and urinary protein–creatinine ratio (PCR) were measured at each facility laboratory at the beginning of the study and on week 12. In addition, high-sensitivity C-reactive protein (CRP), folic acid, vitamin B12, aminoter- minal pro-brain natriuretic peptide (NT-pro BNP), iron, fer- ritin, and TIBC were measured at the clinical laboratory company (SRL, Tokyo, Japan). Results The mean patient age was 70 ± 12 years (male 58.8%). Dia- betic nephropathy was the leading cause of CKD (27.6%) followed by nephrosclerosis (23.5%) and chronic glomeru- lonephritis (22.5%). The prevalence of coronary artery dis- ease, heart failure, stroke, and peripheral artery disease was 15. 8%, 7.0%, 11.7%, and 11.0%, respectively. Creatinine and Hb levels at DA initiation were 2.62 (1.89, 3.61) mg/dL Response index to DA For the assessment of ESA hyporesponsiveness, the use of the following formulae (ESA resistance index; ERI [21]) were originally planned. ERI −1A = Dose of DA at 12 weeks (휇g) Concentration of Hb (g∕dL) at 12 weeks × body weight (kg), Dose of DA at 12 weeks (휇g) 1A = Dose of DA at 12 weeks (휇g) ERI −1A = Dose of DA at 12 weeks (휇g) Concentration of Hb (g∕dL) at 12 weeks × body weight (kg), ERI −1A = (휇g) Concentration of Hb (g∕dL) at 12 weeks × body weight (kg), ERI −1B = Dose of DA at 12 weeks (휇g) Concentration of Hb (g∕dL) at 12 weeks, ERI −2A = Total dose of DA during 12 weeks (휇g) ΔHb 0−12(g∕dL) × body weight (kg) , ERI −2B = Total dose of DA during 12 weeks (휇g) ΔHb 0−12(g∕dL) , ERI −1B = Dose of DA at 12 weeks (휇g) Concentration of Hb (g∕dL) at 12 weeks, and 9.8 ± 0.9 g/dL, respectively (Table1). The prevalence of patients with serum ferritin level < 50 μg and TSAT < 20% was 27.2% and 22.4%, respectively (Supplementary Fig. 1). The most frequent number of DA administration dur- ing 12 weeks was 3 times (n = 697, 41.1%), followed by 4 (n = 265, 15.6%) and 2 (n = 248, 14.6%) times with a wide distribution of the total DA dose ranging from 15 μg to 900 μg (Figs. 2, 3). The total DA dose and hemoglobin lev- els increased in relation to the number of DA administration for 12 weeks (Fig. 3). and 9.8 ± 0.9 g/dL, respectively (Table1). The prevalence of patients with serum ferritin level < 50 μg and TSAT < 20% was 27.2% and 22.4%, respectively (Supplementary Fig. 1). The most frequent number of DA administration dur- ing 12 weeks was 3 times (n = 697, 41.1%), followed by 4 (n = 265, 15.6%) and 2 (n = 248, 14.6%) times with a wide distribution of the total DA dose ranging from 15 μg to 900 μg (Figs. 2, 3). The total DA dose and hemoglobin lev- els increased in relation to the number of DA administration for 12 weeks (Fig. 3). Response index to DA ERI −2A = Total dose of DA during 12 weeks (휇g) ΔHb 0−12(g∕dL) × body weight (kg) , ERI −2B = Total dose of DA during 12 weeks (휇g) ΔHb 0−12(g∕dL) , where ΔHb 0–12 (g/dL) = Hb (g/dL) at 12 weeks—Hb (g/ dL) before DA administration. where ΔHb 0–12 (g/dL) = Hb (g/dL) at 12 weeks—Hb (g/ dL) before DA administration. iEResI increased with the number of DA administration, especially in the total DA dose between 90 μg and 180 μg; whereas, it was not associated with the total DA dose and even lower in the total DA dose of > 180 μg (Table 2 and Fig. 4). However, some patients showed decreased or no changes in Hb level (ΔHb 0–12 < 0 or = 0) during 12 weeks after DA administration; thus, these aforementioned formulae were not used in the data analysis. Instead, initial ESA response index (iEResI) was defined as a reciprocal of the ERI-2A. Notably, 225 patients (13.3%) had no response or even had reduced Hb level at week 12 (DA hyporesponsiveness group). A similar pattern of histograms of total DA dose during 12 weeks in two groups with and without DA hypore- sponsiveness was observed; however, there was a statisti- cally significant difference in the median cumulative DA dose between two groups with [100 (60, 180) μg] and with- out DA hyporesponsiveness [140 (90, 180) μg] (P < 0.0001) (Supplementary Fig. 2). iEResI = ΔHb 0−12(g∕dL) × body weight (kg) Total dose of DA during 12 weeks (휇g). Statistical analysis Baseline characteristics are reported as means ± stand- ard deviation (SD), median [interquartile range (IQR)], or number (percentage). The Wilcoxon rank sum test was used to compare two groups. As regards the predictors of initial response to DA, the association of iEResI with the num- ber of DA administration and total doses of DA was firstly investigated. Statistical analysis Furthermore, contributing factors to iEResI Multivariate analysis showed that male gender, hypo- glycemic agent use, iron supplementation, not serum fer- ritin or TSAT level, and high eGFR were independently and significantly associated with better initial response to DA (P = < 0.0001, 0.0108, < 0.0001, and 0.0476, respectively); whereas, high Hb, high CRP, high NT-proBNP, and high 1 3 3 Clinical and Experimental Nephrology (2021) 25:110–119 113 Table 1 Baseline characteristics Number of patients Prevalence, SD standard deviation, IQR interquartile range, CKD chronic kidney disease, RAS renin-angiotensin sys- tem, NT-proBNP aminoterminal pro-brain natriuretic peptide ( Q ) Age years 1695 70 ± 12 Male gender n (%) – 997 (58.8) Etiology of CKD Diabetic nephropathy n (%) – 467 (27.6) Chronic glomerulonephritis n (%) – 381 (22.5) Nephrosclerosis n (%) – 398 (23.5) Polycystic kidney disease n (%) – 95 (5.6) Other n (%) – 354 (20.9) Smoking status Current n (%) – 183 (10.8) Ever n (%) – 617 (36.4) Diabetes n (%) – 722 (42.6) Malignancy (past history) n (%) – 208 (12.3) Cardiovascular disease Coronary artery disease n (%) – 268 (15.8) Heart failure n (%) – 118 (7.0) Stroke n (%) – 198 (11.7) Peripheral artery disease n (%) – 187 (11.0) RAS inhibitor use Angiotensin II receptor blocker n (%) – 972 (57.3) Angiotensin converting enzyme inhibitor n (%) – 179 (10.6) Hypoglycemic agent use Dipeptidyl peptidase-4 inhibitor n (%) – 362 (21.4) Insulin n (%) – 193 (11.4) Iron supplementation n (%) – 252 (14.9) Body mass index (kg/m2) 1552 23.2 ± 4.0 Systolic arterial pressure (mmHg) 1603 134.3 ± 19.0 Diastolic arterial pressure (mmHg) 1601 71.3 ± 12.3 Creatinine (mg/dl) 1695 2.62 (1.89, 3.61) Estimated glomerular filtration rate (ml/min/1.73m2) 1695 18.0 (12.9, 25.3) Hemoglobin (g/dl) 1695 9.8 ± 0.9 Albumin (g/dl) 1652 3.7 ± 0.5 Ferritin (ng/ml) 1644 96.4 (46.4, 175.0) Transferrin saturation (%) 1645 26.2 (20.6, 31.9) High sensitive C-reactive protein (ng/dl) 1649 575 (219, 1810) Folic acid (ng/ml) 1604 7.2 (5.5, 9.8) Vitamin B12 (pg/ml) 1578 354 (258, 498) NT-proBNP* (pg/ml) 1648 516.0 (238.5, 1160.0) HbA1c (%) 1012 6.1 ± 0.9 Urinary protein-creatinine ratio (g/gCr) 1526 1.4 (0.5, 3.1) SD standard deviation, IQR interquartile range, CKD chronic kidney disease, RAS renin-angiotensin sys- tem, NT-proBNP aminoterminal pro-brain natriuretic peptide Discussion urinary PCR levels were significantly associated with poor initial response to DA (P = < 0.0001, 0.0004, 0.0435, and 0.0009, respectively) (Table 3). urinary PCR levels were significantly associated with poor initial response to DA (P = < 0.0001, 0.0004, 0.0435, and 0.0009, respectively) (Table 3). ESA hyporesponsiveness is a strong predictor of mortality and CVD events as well as poor renal survival in patients with CKD [16, 17, 21–25]. Although the definition of ESA hyporesponsiveness has been suggested in previous clinical 1 3 Clinical and Experimental Nephrology (2021) 25:110–119 114 1 2 3 4 5 6 7 8 Number of DA administraon during 12 weeks Frequency 600 400 200 0 1 2 3 4 5 6 7 8 Number of DA administraon during 12 weeks Frequency 600 400 200 0 Fig. 2 Number of DA administration during 12 weeks after DA ini- tiation. DA darbepoetin alfa the DA dose used was quite different from those used in our clinical practice setting in Japan [15, 25]. Therefore, extrapolating this definition into clinical practice in Japan seems difficult. In this context, an evidence-based index of ESA hyporesponsiveness and clinical guidelines for anemia management in CKD patients with ESA hyporesponsiveness should be established in Japan. In the current analysis, 225 patients (13.3%) showed no increase or even had decreased Hb level at 12 weeks after DA administration. This prevalence was similar to those reported in previous studies regardless of different defini- tions of ESA hyporesponsiveness adopted [29, 30]. Fur- thermore, the new index of initial responsiveness to DA (iEResI) was defined. Male gender, hypoglycemic agent use, iron supplementation, and high eGFR were positively associated with iEResI; whereas, high Hb, high CRP, high NT-proBNP levels, and high urinary PCR were negatively associated with iEResI. Previous studies have already shown that iron deficiency, decreased renal function, and inflam- mation were strong predictors for ESA hyporesponsiveness; however, reports regarding the association between gender difference or hypoglycemic agent use (proxy for diabetes) and ESA hyporesponsiveness are conflicting [23, 24, 29, 31]. In our study, iron supplementation, not serum ferritin or TSAT level, was associated with iEResI on the univari- ate and multivariate analysis. As shown in Supplementary Fig. 1, the prevalence of study subjects with iron deficiency (e.g., serum ferritin < 50 μg and TSAT < 20%) seems to be lower than that previously reported in dialysis-dependent Fig. Discussion 2 Number of DA administration during 12 weeks after DA ini- tiation. DA darbepoetin alfa guidelines for the treatment of anemia in patients with CKD, they were defined arbitrarily and were not associated with any outcomes [26, 27]. Only the Kidney Disease: Improv- ing Global Outcomes (KDIGO) anemia guideline published in 2012 defined ESA hyporesponsiveness as “patients who have no increased Hb concentration from baseline after the first month of ESA treatment on appropriate weight-based dose [28]. This definition was based on the secondary analy- sis of the TREAT study, which revealed that diabetic pre- dialysis CKD patients with initial hyporesponsiveness to DA had poor cardiovascular outcome [17]. However, in the TREAT study, nearly half of study subjects had CVD and 1 2 3 4 5 6 7 8 overall Number of DA administraon during 12weeks ( s k e e w 2 1 g nir u d e s o d A D la t o T μg) Hemoglobin change during 12 weeks (g/dl) 1000 800 600 400 200 0 7 6 5 4 3 2 1 0 -1 -2 -3 -4 1 2 3 4 5 6 7 8 overall Number of DA administraon during 12weeks Fig. 3 Total DA dose and hemoglobin change during 12 weeks after DA initiation stratified by the number of DA administration. DA darbepo- etin alfa 1 2 3 4 5 6 7 8 overall Number of DA administraon during 12weeks ( s k e e w 2 1 g nir u d e s o d A D la t o T μg) 1000 800 600 400 200 0 Hemoglobin change during 12 weeks (g/dl) 7 6 5 4 3 2 1 0 -1 -2 -3 -4 1 2 3 4 5 6 7 8 overall Number of DA administraon during 12weeks Number of DA administraon during 12weeks Fig. 3 Total DA dose and hemoglobin change during 12 weeks after DA initiation stratified by the number of DA administration. DA darbepo- etin alfa Fig. 3 Total DA dose and hemoglobin change during 12 weeks after DA initiation stratified by the number of DA administration. DA darbepo- etin alfa Fig. 3 Total DA dose and hemoglobin change during 12 weeks after DA initiation stratified by the number of DA administration. Discussion DA darbepo- etin alfa Clinical and Experimental Nephrology (2021) 25:110–119 Clinical and Experimental Nephrology (2021) 25:110–119 Clinical and Experimental Nephrology (2021) 25:110–119 115 Table 2 iEResI stratified by cumulative DA doses and number of DA administration Cumulative DA dose Number of DA administration Number of patients Mean Standard deviation ≤ 90 μg 1 54 1.4120 2.6105 2 135 0.5363 1.1022 3 293 0.6012 0.6158 4 10 1.0538 0.5824 5 1 0.9949 – 6 3 0.7087 0.1833 Overall 496 0.6823 1.1659 > 90 μg and ≤ 120 μg 1 4 0.4073 0.3080 2 31 0.4675 0.4046 3 125 0.4212 0.5637 4 80 0.6923 0.4602 5 6 0.7412 0.5405 6 7 1.2840 0.5611 Overall 253 0.5438 0.5369 > 120 μg and ≤ 180 μg 1 1 0.7467 – 2 10 0.3996 0.4987 3 142 0.4237 0.3584 4 95 0.5354 0.3770 5 96 0.7171 0.3853 6 130 0.6380 0.3328 7 1 0.6204 – Overall 475 0.5645 0.3794 > 180 μg 1 0 – – 2 16 0.2714 0.2156 3 48 0.3005 0.2492 4 60 0.3238 0.3194 5 40 0.3780 0.3925 6 99 0.3343 0.2603 7 6 0.3473 0.2317 8 1 0.7055 – Overall 270 0.3304 0.2917 CKD patients [32], which might make difficult to show sig- nificant and apparent relationship between these parameters of iron status and iEResI. The association between male gender and better initial responsiveness to DA may be a mathematical artifact. Namely, DA is administered through prefilled syringe at a dose of 15, 30, 40, 60, 120, or 180 μg for pre-dialysis CKD patients in Japan, not according to the body weight-adjusted dose for each patient. As the iEResI adopted in this study had a figure of “body weight” in the numerator, the iEResI value may be relatively higher in men than in women. Some studies reported that diabetes or insulin resistance was associated with ESA hyporespon- siveness in pre-dialysis CKD patients [29, 33]; whereas, others did not show any association between diabetes and ESA hyporesponsiveness in dialysis patients [23, 31]. In this study, patients with hypoglycemic agents use had higher prevalence of dyslipidemia than those without hypoglycemic agent use (67.6% vs. 49.7%; P < 0.0001). Statin use has also been reported to improve ESA responsiveness in dialysis and pre-dialysis CKD patients due to its anti-inflammatory effect [34, 35]. Although the prevalence of statin use was not obtained in this study, such a confounding factor may affect this association. Discussion 2 1 0 -1 -2 1 2 3 4 5 6 90<DA≤120μg 1 0.5 0 -0.5 -1.0 2 3 4 5 6 7 8 DA>180μg iEResI iEResI P<0.0001 P=0.2308 Number of DA administraon during 12 weeks Number of DA administraon during 12 weeks the total DA dose and even lower in the total DA dose of > 180 μg. DA Darbepoetin alfa, iEResI initial response index to erythropoiesis- stimulating agents. Note: The scale of vertical axes in each graph is not unified. 1 2 3 4 5 6 DA≤90μg 15 10 5 0 -5 Number of DA administraon during 12 weeks iEResI P=0.0062 2 1 0 -1 -2 1 2 3 4 5 6 90<DA≤120μg iEResI P<0.0001 Number of DA administraon during 12 weeks iEResI Number of DA administraon during 12 weeks Number of DA administraon during 12 weeks Number of DA administraon during 12 weeks 1 0.5 0 -0.5 -1.0 2 3 4 5 6 7 8 DA>180μg iEResI P=0.2308 Number of DA administraon during 12 weeks 1 2 3 4 5 6 7 2 1 0 -1 120<DA≤180μg iEResI P<0.0001 Number of DA administraon during 12 weeks 1 iEResI Number of DA administraon during 12 weeks Number of DA administraon during 12 weeks Fig. 4 Association between iEResI and the number of DA administra- tion stratified by the total DA dose until 12 weeks. iEResI increased with the number of DA administration, especially in the total DA dose between 90 μg and 180 μg; whereas, it was not associated with Fig. 4 Association between iEResI and the number of DA administra- tion stratified by the total DA dose until 12 weeks. iEResI increased with the number of DA administration, especially in the total DA dose between 90 μg and 180 μg; whereas, it was not associated with the total DA dose and even lower in the total DA dose of > 180 μg. DA Darbepoetin alfa, iEResI initial response index to erythropoiesis- stimulating agents. Note: The scale of vertical axes in each graph is not unified. associated with the iEResI. As the aforementioned study showed that patients with history of heart failure not of coro- nary artery disease showed ESA hyporesponsiveness [23], history of CVD which has been usually used for adjustment of confounders in previous studies has a crucial week point, which means that the history of CVD can not represent the severity of CVD. Discussion Our results also showed that high urinary PCR was associated with ESA hyporesponsiveness, which is consistent with a previous report [24]. Although the exact mechanism by which proteinuria could affect the ESA responsiveness has not been clarified, proteinuria is a proxy for inflammation or histological damage of the kidney and for progressive kidney function decline which could affect endogenous erythropoietin production or response to ESA. Of interest, history of CVD was not associated with the iEResI in our study; whereas, NT-proBNP was significantly 1 3 Clinical and Experimental Nephrology (2021) 25:110–119 116 1 2 3 4 5 6 DA≤90μg 15 10 5 0 -5 1 2 3 4 5 6 7 2 1 0 -1 120<DA≤180μg 2 1 0 -1 -2 1 2 3 4 5 6 90<DA≤120μg 1 0.5 0 -0.5 -1.0 2 3 4 5 6 7 8 DA>180μg Number of DA administraon during 12 weeks iEResI iEResI iEResI iEResI P=0.0062 P<0.0001 P<0.0001 P=0.2308 Number of DA administraon during 12 weeks Number of DA administraon during 12 weeks Number of DA administraon during 12 weeks Fig. 4 Association between iEResI and the number of DA administra- tion stratified by the total DA dose until 12 weeks. iEResI increased with the number of DA administration, especially in the total DA dose between 90 μg and 180 μg; whereas, it was not associated with the total DA dose and even lower in the total DA dose of > 180 μg. DA Darbepoetin alfa, iEResI initial response index to erythropoiesis- stimulating agents. Note: The scale of vertical axes in each graph is not unified. 4 5 6 4 5 6 7 2 1 0 -1 -2 1 2 3 4 5 6 90<DA≤120μg 1 0.5 0 -0.5 -1.0 2 3 4 5 6 7 8 DA>180μg nistraon during 12 weeks iEResI iEResI P=0.0062 P<0.0001 P<0.0001 P=0.2308 Number of DA administraon during 12 weeks nistraon during 12 weeks Number of DA administraon during 12 weeks e number of DA administra- 12 weeks. iEResI increased especially in the total DA s, it was not associated with the total DA dose and even lower in the total DA dose of > 180 μg. DA Darbepoetin alfa, iEResI initial response index to erythropoiesis- stimulating agents. Note: The scale of vertical axes in each graph is not unified. Discussion On the other hand, BNP or NT-proBNP can represent the severity of CVD. Patients with severe CVD usually have more serious inflammation, or take more medication, which could be associated with anemia or poor response to ESA, compared to those with less severe CVD. Thus, we think that it is quite reasonable that NT-proBNP, not the history of CVD, was associated with the respon- siveness to DA. Low Hb level before ESA treatment was reported as a predictor of ESA hyporesponsiveness [23, 29] because it means the presence of comorbidity or inflamma- tion that could be directly associated with ESA hyporespon- siveness. Our study result showing the opposite association may be interpreted as patients with higher Hb level before DA initiation may be less dependent on erythropoietin. Age, malnutrition, and RAS inhibitor were also reported as pre- dictive factors of ESA hyporesponsiveness [29, 36]; how- ever, this study did not show any association between these factors and iEResI, although precise nutritional assessment other than serum albumin level was not performed for the BRIGHTEN. Although the BRIGHTEN is nationwide prospective study that can closely monitor patients, several limita- tions should be considered. This study aimed to examine the initial responsiveness to DA in principally ESA-naïve patients. As ESA responsiveness is likely to change over time, 12 weeks may be relatively short to evaluate ESA responsiveness. As it is an observational study conducted in a real-world clinical setting, the dose and frequency of DA administration were not unified because they were determined based on the physician’s discretion and patients’ preference, which may affect the iEResI. Discussion Further- more, there were many missing data about iron supple- mentation (type of agents, dose, and route of administra- tion), so we could not analyze the relationship between the 1 3 3 Clinical and Experimental Nephrology (2021) 25:110–119 Clinical and Experimental Nephrology (2021) 25:110–119 117 CKD chronic kidney disease, RAS renin-angiotensin system, NT-proBNP aminoterminal pro-brain natriu- retic peptide a Adjusted by gender as a factor Univariatea Multivariateb Variables β coefficient P value β coefficient P value Age − 0.003 0.1187 Male gender (reference, female) 0.165 < 0.0001 0.234 < 0.0001 Etiology of CKD 0.9418 Diabetic nephropathy 0.030 Chronic glomerulonephritis 0.051 Nephrosclerosis 0.015 Polycystic kidney disease − 0.030 Smoking status, never (reference, current) − 0.107 0.2969 Diabetes, yes (reference, no) 0.082 0.0396 Past history of malignancy, yes (reference, no) − 0.097 0.1032 Cardiovascular disease (reference, no) Coronary artery disease, yes 0.043 0.4191 Heart failure, yes − 0.007 0.9228 Stroke, yes − 0.003 0.9543 Peripheral artery disease, yes 0.019 0.7633 RAS inhibitor use, yes (reference, no) − 0.016 0.7340 Hypoglycemic agent use, yes (reference, no) 0.085 0.0415 0.112 0.0108 Iron supplementation, yes (reference, no) 0.223 < 0.0001 0.240 < 0.0001 Body mass index 0.003 0.4960 Systolic arterial pressure − 0.001 0.2430 Diastolic arterial pressure 0.001 0.4385 Log (Estimated glomerular filtration rate) 0.133 0.0014 0.095 0.0476 Hemoglobin − 0.111 < 0.0001 − 0.123 < 0.0001 Albumin − 0.017 0.6501 Log (Ferritin) 0.014 0.5018 Log (Transferrin saturation) − 0.090 0.0703 Log (C-reactive protein) − 0.031 0.0198 − 0.048 0.0004 Log (Folic acid) 0.005 0.9112 Log (Vitamin B12) -0.008 0.8511 Log (NT-proBNP) − 0.039 0.0189 − 0.039 0.0435 HbA1c 0.046 0.1418 Log (Urinary protein–creatinine ratio) − 0.051 0.0005 − 0.053 0.0009 Table 3 Predictors associated with iEResI proteinuria were independently and significantly associated with better initial response to DA. In future analyses, we are planning to investigate the association between the iEResI as well as other ERIs and patient outcomes. dose of iron supplementation and responsiveness to DA. Factors associated with ESA hyporesponsiveness could highly depend on patient characteristics studied, comorbid disease, and treatment for anemia itself (e.g., ESA dose); thus, our results could not be extrapolated to CKD patients in Western countries and to those undergoing chronic dial- ysis treatment. Acknowledgements We would like to express our deepest gratitude to the patients, as well as the investigators and staff at the study sites (presented in the appendix) for their contribution to the study. Compliance with ethical standards Conflict of interest This trial is funded by KKC. The KKC proposed the concept of this trial and provided information, but do not take any part in collecting, analyzing, or interpreting the results. The Transla- tional Research Center for Medical Innovation, a third-party organiza- tion that is independent of the primary investigator’s institution, was contracted to perform the trial and manage the data. 12. Besarab A, Bolton WK, Browne JK, et al. The effects of normal as compared with low hematocrit values in patients with cardiac disease who are receiving hemodialysis and epoetin. N Engl J Med. 1998;339:584–90. 13. Drüeke TB, Locatelli F, Clyne N, et al. Normalization of hemo- globin level in patients with chronic kidney disease and anemia. N Engl J Med. 2006;355:2071–84. Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. 14. Singh AK, Szczech L, Tang KL, et al. Correction of anemia with epoetin alfa in chronic kidney disease. N Engl J Med. 2006;355:2085–98.f 15. Pfeffer MA, Burdmann EA, Chen CY, et al. A trial of darbepo- etin alfa in type 2 diabetes and chronic kidney disease. N Engl J Med. 2009;361:2019–32. 16. Szczech LA, Barnhart HX, Inrig JK, et al. Secondary analysis of the CHOIR trial epoetin-alpha dose and achieved hemoglobin outcomes. Kidney Int. 2008;74:791–8. y 17. Solomon SD, Uno H, Lewis EF, et al. Erythropoietic response and outcomes in kidney disease and type 2 diabetes. N Engl J Med. 2010;363:1146–55. 18. Yamamoto H, Nishi S, Tomo T, et al. 2015 Japanese society for dialysis therapy: guideline for renal anemia in chronic kidney disease. Ren Replace Ther. 2017;3:36. 19. Kato H, Nangaku M, Hirakata H, et al. References 1. Keane WF, Brenner BM, de Zeeuw D, et al. The risk of develop- ing end-stage renal disease in patients with type 2 diabetes and nephropathy: the RENAAL study. Kidney Int. 2003;63:1499–507. 20. Matsuo S, Imai E, Horio M, et al. Revised equations for esti- mated GFR from serum creatinine in Japan. Am J Kidney Dis. 2009;53:982–92. 2. Kovesdy CP, Trivedi BK, Kalantar-Zadeh K, et al. Association of anemia with outcomes in men with moderate and severe chronic kidney disease. Kidney Int. 2006;69:560–4. 21. Kilpatrick RD, Critchlow CW, Fishbane S, et al. Greater epo- etin alfa responsiveness is associated with improved survival in hemodialysis patients. Clin J Am Soc Nephrol. 2008;3:1077–83. 3. Levin A, Thompson CR, Ethier J, et al. Left ventricular mass index increase in early renal disease: impact of decline in hemo- globin. Am J Kidney Dis. 1999;34:125–34. 22. Fukuma S, Yamaguchi T, Hashimoto S, et al. Erythropoiesis- stimulating agent responsiveness and mortality in hemodialysis patients: results from a cohort study from the dialysis registry in Japan. Am J Kidney Dis. 2012;59:108–16. 4. Walker AM, Schneider G, Yeaw J, et al. Anemia as a predictor of cardiovascular events in patients with elevated serum creatinine. J Am Soc Nephrol. 2006;17:2293–8. 5. Pisoni RL, Bragg-Gresham JL, Young EW, et al. Anemia man- agement and outcomes from 12 countries in the Dialysis Out- comes and Practice Patterns Study (DOPPS). Am J Kidney Dis. 2004;44:94–111.f 23. López-Gómez JM, Portolés JM, Aljama P. Factors that conditon the responce to erythropoietin in patients on hemodialysis and their relation to mortality. Kidney Int. 2008;74(Suppl):S75–81. 24. Minutolo R, Conte G, Cianciaruso B, et al. Hyporespon- siveness to erythropoiesis-stimulating agents and renal sur- vival in non-dialysis CKD patients. Nephrol Dial Transplant. 2012;27:2880–6. 6. Hayashi T, Suzuki A, Shoji T, et al. Cardiovascular effect of nor- malizing the hematocrit level during erythropoietin therapy in predialysis patients with chronic renal failure. Am J Kidney Dis. 2000;35:250–6.f 25. Tanaka T, Nangaku M, Imai E, et al. Safety and effectiveness of long-term use of darbepoetin alfa in non-dialysis patients with chronic kidney disease: a post-marketing surveillance study in Japan. Clin Exp Nephrol. 2019;23:231–43. 7. Ofsthun N, Labrecque J, Lacson E, et al. The effects of higher hemoglobin levels on mortality and hospitalization in hemodi- alysis patients. Kidney Int. 2003;63:1908–14. 8. De Nicola L, Conte G, Chiodini P, et al. Compliance with ethical standards Rationale and design of observational clinical research in chronic kidney disease patients with renal anemia: renal prognosis in patients with hyporesponsive anemia to erythropoiesis-stimulating agents, darbepoetin alfa (BRIGHTEN Trial). Clin Exp Nephrol. 2018;22:78–84. Discussion In conclusion, in the current analysis of the BRIGHTEN, 13.3% of patients showed no increase or even decreased Hb level at 12 weeks after DA administration. Furthermore, male gender, hypoglycemic agent use, iron supplementation, high eGFR, low Hb, low CRP, low NT-proBNP, and low Funding The authors reported the following financial supports: Masaomi Nangaku received honoraria, manuscript fees, and subsi- dies or donations from Kyowa Kirin Co. Ltd. (KKC). Takashi Wada received honoraria and subsidies or donations from KKC. Terumasa 1 3 Clinical and Experimental Nephrology (2021) 25:110–119 118 patients with chronic kidney disease. Clin J Am Soc Nephrol. 2007;2:938–46. Hayashi received honoraria from KKC. Takao Masaki received hono- raria and subsidies or donations from KKC. Hiroyasu Yamamoto received honoraria and manuscript fees from KKC. Hiroki Hase received honoraria from KKC. Masahiro Kamouchi received travel fees, gifts, and others from KKC. Tadao Akizawa is in Employment/ Leadership position/Advisory role in KKC and received honoraria and manuscript fees from KKC. Yoshiharu Tsubakihara received honoraria from KKC. Shoichi Maruyama received honoraria and subsidies or donations from KKC. Ichiei Narita received honoraria and subsidies or donations from KKC. 9. Kuriyama S, Tomonari H, Yoshida H, et al. Reversal of anemia by erythropoietin therapy retards the progression of chronic renal failure, especially in nondiabetic patients. Nephron. 1997;77:176–85. 10. Tsubakihara Y, Gejyo F, Nishi S, et al. High target hemoglobin with erythropoiesis-stimulating agents has advantages in the renal function of non-dialysis chronic kidney disease patients. Ther Apher Dial. 2012;16:529–40. 11. Revicki DA, Brown RE, Feeny DH, et al. Health-related quality of life associated with recombinant human erythropoietin therapy for predialysis chronic renal disease patients. Am J Kidney Dis. 1995;25:548–54. References Stability of target hemoglobin levels during the first year of epoetin treatment in 26. Locatelli F, Aljama P, Bárány P, et al. Revised European best prac- tice guidelines for the management of anaemia in patients with 1 3 Clinical and Experimental Nephrology (2021) 25:110–119 119 chronic renal failure. Nephrol Dial Transplant. 2004;19(Suppl 2):ii1–47. 33. Abe M, Okada K, Soma M, et al. Relationship between insulin resistance and erythropoietin responsiveness in hemodialysis patients. Clin Nephrol. 2011;75:49–58. ) 27. KDOQI. KDOQI Clinical Practice Guideline and Clinical Practice Recommendations for anemia in chronic kidney disease: 2007 update of hemoglobin target. Am J Kidney Dis. 2007;50:471–530. 34. Sirken G, Kung SC, Raja R. Decreased erythropoietin require- ments in maintenance hemodialysis patients with statin therapy. ASAIO J. 2003;49:422–5. p g g y 28. Kidney disease: improving global outcomes (KDIGO) Anemia Work Group. KDIGO clinical practice guideline for anemia in chronic kidney disease. Kidney Int. 2012;2(Suppl):S279–335. 35. Chiang CK, Yang SY, Peng YS, et al. Atorvastatin increases eryth- ropoietin-stimulating agent hyporesponsiveness in maintenance hemodialysis patients: role of anti-inflammation effects. Am J Nephrol. 2009;29:392–7. 29. Rossert J, Gassmann-Mayer C, Frei D, et al. Prevalence and pre- dictors of epoetin hyporesponsiveness in chronic kidney disease patients. Nephrol Dial Transplant. 2007;22:794–800. 36. Kanbay M, Perazella MA, Kasapoglu B, et al. Erythropoiesis stimulatory agent-resistant anemia in dialysis patients: review of causes and management. Blood Purif. 2010;29:1–12. 30. Besarab A, Yee J. Candidate biomarkers for erythropoietin response in end-stage renal disease. Kidney Int. 2011;79:488–90. 31. Ifudu O, Uribarri J, Rajwani I, et al. Gender modulates respon- siveness to recombinant erythropoietin. Am J Kidney Dis. 2001;38:518–22. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 32. Nakai S, Hanafusa N, Masakane I, et al. An overview of regular dialysis treatment in Japan (as of 31 December 2012). Ther Apher Dial. 2014;6:535–602. Affiliations Terumasa Hayashi1 · Hideki Kato2 · Kenichiro Tanabe3 · Masaomi Nangaku2 · Hideki Hirakata4 · Takashi Wada5 · Hiroshi Sato6 · Yasushi Yamazaki7 · Takao Masaki8 · Tatsuo Kagimura3 · Hiroyasu Yamamoto9 · Hiroki Hase10 · Masahiro Kamouchi11 · Enyu Imai12 · Kyoichi Mizuno13 · Manabu Iwasaki14 · Tadao Akizawa15 · Yoshiharu Tsubakihara16 · Shoichi Maruyama17 · Ichiei Narita18 Terumasa Hayashi1 · Hideki Kato2 · Kenichiro Tanabe3 · Masaomi Nangaku2 · Hideki Hirakata4 · Takashi Wada5 · Hiroshi Sato6 · Yasushi Yamazaki7 · Takao Masaki8 · Tatsuo Kagimura3 · Hiroyasu Yamamoto9 · Hiroki Hase10 · Masahiro Kamouchi11 · Enyu Imai12 · Kyoichi Mizuno13 · Manabu Iwasaki14 · Tadao Akizawa15 · Yoshiharu Tsubakihara16 · Shoichi Maruyama17 · Ichiei Narita18 10 Division of Nephrology, Toho University Ohashi Medical Center, Tokyo, Japan 1 Department of Kidney Disease and Hypertension, Osaka General Medical Center, Osaka, Japan 1 Department of Kidney Disease and Hypertension, Osaka General Medical Center, Osaka, Japan 1 Department of Kidney Disease and Hypertension, Osaka General Medical Center, Osaka, Japan 2 Division of Nephrology and Endocrinology, The University of Tokyo Graduate School of Medicine, Tokyo, Japan 3 Translational Research Center for Medical Innovation, Kobe, Japan 4 Fukuoka Renal Clinic, Fukuoka, Japan 5 Department of Nephrology and Laboratory Medicine, Faculty of Medicine, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kanazawa, Japan 6 Division of Clinical Pharmacology and Therapeutics, Tohoku University Graduate School of Pharmaceutical Sciences and Faculty of Pharmaceutical Sciences, Sendai, Japan 7 Department of Nephrology and Rheumatology, Kagawa Prefectural Central Hospital, Takamatsu, Japan 8 Department of Nephrology, Hiroshima University Hospital, Hiroshima, Japan 9 Division of Nephrology and Hypertension, Department of Internal Medicine, The Jikei University School of Medicine, Tokyo, Japan 2 Division of Nephrology and Endocrinology, The University of Tokyo Graduate School of Medicine, Tokyo, Japan 11 Department of Health Care Administration and Management, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan 3 Translational Research Center for Medical Innovation, Kobe, Japan 12 Nakayamadera Imai Clinic, Takarazuka, Japan 12 Nakayamadera Imai Clinic, Takarazuka, Japan 4 Fukuoka Renal Clinic, Fukuoka, Japan 13 Mitsukoshi Health and Welfare Foundation, Tokyo, Japan 4 Fukuoka Renal Clinic, Fukuoka, Japan 13 Mitsukoshi Health and Welfare Foundation, Tokyo, Japan 14 School of Data Science, Yokohama City University, Yokohama, Japan 14 School of Data Science, Yokohama City University, Yokohama, Japan 5 Department of Nephrology and Laboratory Medicine, Faculty of Medicine, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kanazawa, Japan 15 Division of Nephrology, Department of Medicine, Showa University School of Medicine, Tokyo, Japan 15 Division of Nephrology, Department of Medicine, Showa University School of Medicine, Tokyo, Japan 6 Division of Clinical Pharmacology and Therapeutics, Tohoku University Graduate School of Pharmaceutical Sciences and Faculty of Pharmaceutical Sciences, Sendai, Japan 16 Course of Safety Management in Health Care Sciences, Graduate School of Health Care Sciences, Jikei Institute, Osaka, Japan 7 Department of Nephrology and Rheumatology, Kagawa Prefectural Central Hospital, Takamatsu, Japan 17 Department of Nephrology, Nagoya University Graduate School of Medicine, Nagoya, Japan 17 Department of Nephrology, Nagoya University Graduate School of Medicine, Nagoya, Japan 8 Department of Nephrology, Hiroshima University Hospital, Hiroshima, Japan 18 Division of Clinical Nephrology and Rheumatology, Niigata University Graduate School of Medical and Dental Sciences, 757 Ichibancho Asahimachidori Chuo‑ku, Niigata 951‑8510, Japan 9 Division of Nephrology and Hypertension, Department of Internal Medicine, The Jikei University School of Medicine, Tokyo, Japan 1 1 3 3
https://openalex.org/W2075936524	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0052917&type=printable	English	null	Multiple Protein Biomarker Assessment for Recombinant Bovine Somatotropin (rbST) Abuse in Cattle	PloS one	2,012	cc-by	11,651	Abstract The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: susann.ludwig@wur.nl * E-mail: susann.ludwig@wur.nl Abstract Biomarker profiling, as a rapid screening approach for detection of hormone abuse, requires well selected candidate biomarkers and a thorough in vivo biomarker evaluation as previously done for detection of growth hormone doping in athletes. The bovine equivalent of growth hormone, called recombinant bovine somatotropin (rbST) is (il)legally administered to enhance milk production in dairy cows. In this study, first a generic sample pre-treatment and 4-plex flow cytometric immunoassay (FCIA) were developed for simultaneous measurement of four candidate biomarkers selected from literature: insulin-like growth factor 1 (IGF-1), its binding protein 2 (IGFBP2), osteocalcin and endogenously produced antibodies against rbST. Next, bovine serum samples from two extensive controlled rbST animal treatment studies were used for in vivo validation and biomarker evaluation. Finally, advanced statistic tools were tested for the assessment of biomarker combination quality aiming to correctly identify rbST-treated animals. The statistical prediction tool k-nearest neighbours using a combination of the biomarkers osteocalcin and endogenously produced antibodies against rbST proved to be very reliable and correctly predicted 95% of the treated samples starting from the second rbST injection until the end of the treatment period and even thereafter. With the same biomarker combination, only 12% of untreated animals appeared false-positive. This reliability meets the requirements of Commission Decision 2002/657/EC for screening methods in veterinary control. From the results of this multidisciplinary study, it is concluded that the osteocalcin – anti-rbST- antibodies combination represent fit-for-purpose biomarkers for screening of rbST abuse in dairy cattle and can be reliably measured in both the developed 4-plex FCIA as well as in a cost-effective 2-plex microsphere-based binding assay. This screening method can be incorporated in routine veterinary monitoring programmes: in the European Union for detection of rbST abuse and in the control of rbST-free dairy farms in the United States of America and other countries. Smits NGE, van der Veer G, Bremer MGEG, Nielen MWF (2012) Multiple Protein Biomarker Assessment for Recombinant Bovine se in Cattle. PLoS ONE 7(12): e52917. doi:10.1371/journal.pone.0052917 Received May 25, 2012; Accepted November 22, 2012; Published December 27, 2012 Copyright: 2012 Ludwig et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This project was financially supported by the Dutch Ministry of Economic Affairs, Agriculture and Innovation (http://www.rijksoverheid.nl/ministeries/ eleni project 72029.01). Susann K. J. Ludwig1, Nathalie G. E. Smits1, Grishja van der Veer1, Maria G. E. G. Bremer1, Michel W. F. Nielen1,2 Susann K. J. Ludwig1, Nathalie G. E. Smits1, Grishja van der Veer1, Maria G. E. G. Bremer1, Michel W. F. Nielen1,2 1 RIKILT-Institute of Food Safety, Wageningen UR, Wageningen, The Netherlands, 2 Laboratory of Organic Chemistry, Wageningen University, Wageningen, The Netherlands y, Wageningen UR, Wageningen, The Netherlands, 2 Laboratory of Organic Chemistry, Wageningen University, Wageningen, Th LT-Institute of Food Safety, Wageningen UR, Wageningen, The Netherlands, 2 Laboratory of Organic Chemistry, Wageningen U rlands Editor: John Matthew Koomen, Moffitt Cancer Center, United States of America Biomarker Assessment for rbST Abuse Detection Biomarker Assessment for rbST Abuse Detection methods, suffer from the short half-life of rbST. Although biweekly injections containing slow-release formulations are used to prolong the presence of rbST in the cows’ body, the protein levels in treated animals cannot be distinguished from the background level throughout the whole two-week inter-injection period and large inter-individual differences in blood rbST levels were reported [19,20,24–27]. Furthermore, rbST immunoassays were not capable to distinguish the almost identical recombinant and endogenous forms of bST [19,20,24,25] and mass spectrometry- based methods on the other hand required very tedious sample preparation procedures [26,27]. For screening of rbST in cattle, a few biomarker-based methods were developed, but focused on a single candidate biomarker only [4,9,28–30]. In a recent study, three candidate biomarkers were combined in one screening tool, but the ,5% false-compliant rate target could not be achieved [31]. Nevertheless, biomarker-based screening for rbST can be considered a very promising start for detecting rbST abuse in dairy cows. the inhibition format, high sample biomarker concentrations yield low fluorescence signals and vice versa. For the anti-rbST- antibodies, a direct assay format with an rbST-coupled colour- encoded microsphere set was used, where the anti-rbST-antibod- ies bind and can be detected by fluorescently labelled anti-bovine detection antibodies. Here, a high biomarker level leads to a high fluorescence signal. With this 4-plex FCIA, biomarker profiles were measured in serum samples. Based on the biomarker profiles of 67 untreated animals from different origins, we assessed the inter-individual and physiological variability of these biomarkers within dairy cattle and determined decision limits, beyond which a sample could be classified rbST-treated. Then, we used a large set of serum samples obtained from two independent controlled rbST animal treatment studies to evaluate the discriminative power of each candidate biomarker and of all combinations of biomarkers for distinguishing rbST-treated from untreated cows. Following thorough statistical evaluations, the value of individual and multiple biomarkers was assessed for the prediction of rbST abuse in dairy cows. Biomarkers indicative for ST abuse are described in detail in literature and several of them are listed and referenced in Table 1. From these, we selected four different candidate biomarkers. These included two biomarkers of the IGF-1 axis, which respond quickly upon rbST treatment, namely insulin-like growth factor-1 (IGF-1) and IGF binding protein 2 (IGFBP2). Biomarker Assessment for rbST Abuse Detection The other two biomarkers were expected to show a delayed but long-lasting response; these are osteocalcin (marker of bone turnover) and antibodies which are endogenously produced against rbST (anti- rbST-antibodies). The overall aim of the study was the development and validation of a chemical analytical method for rbST-dependent biomarker detection according to European legislation for screening methods [23] and a data analysis approach for identifying biomarker combinations, which can reliably predict rbST abuse. This aim was reached with the help of a statistical prediction model based on the biomarker combination endoge- nously produced antibodies against rbST and osteocalcin. To be able to screen for these four candidate biomarkers in serum, we developed a 4-plex flow cytometric immunoassay (FCIA) enabling parallel biomarker analysis in a single sample. For IGF-1, IGFBP2 and osteocalcin, a competitive inhibition assay format was chosen, where the respective candidate biomarker is covalently coupled to one set of colour-encoded microspheres. The different microsphere sets can be discriminated by a red laser (Figure 1). Biomarker-specific and generic fluorescent secondary antibodies are used for quantification with a green laser. Due to Introduction Extensive studies were done for the biomarker-based detection of recombinant somatotropin (ST; or growth hormone, GH) in sports doping, where ST is abused by athletes for their performance enhancement [14–18]. A similar screening approach can be chosen for the detection of recombinant bovine ST (rbST) abuse in dairy cattle, where the hormone is administered for enhanced milk production [19,20]. The administration to dairy cattle is approved by the U.S. Food and Drug Administration in the United States of America and allowed in several other countries [21]. But treating cows with rbST is forbidden in the European Union since 1999 because of animal health and welfare reasons [22]. By European regulation, screening and confirmatory methods should be available for the detection of (ab)used veterinary drugs, with for screening, a maximum false-compliant rate of 5% (ß error) [23]. In contrast to the well-established human biomarker-based screening approach, the issue of rbST-dependent biomarker detection is still in its infancy: actually, routine veterinary control for rbST abuse has not been implemented at all, despite the EU ban. So far developed methods which detect rbST directly, such as immunoassays or mass spectrometry-based Many different techniques are available for detection of hormone abuse in sports doping and veterinary control, which all have to fulfil the requirements to be reliable, comparably fast and affordable. Biomarker profiling was suggested as a rapid screening approach for detection of doping practices because of its many advantages over the direct detection of the particular abused substances [1]. Biomarker profiles are indicative for more than one administered agent as they reflect the physiological effect, hence, the abuse of unknown compounds can also be detected [1,2]. Furthermore, in many cases, the analysis of biomarker profiles enables the detection of abused substances for a longer time period, because the biological effect lasts longer than the abused substance itself can be detected in the body [3,4]. A lot of work was focused on the identification of indicative biomarkers and the development of assays for detection of those [2,5–10]. But the suitability and discriminative power of each biomarker has to be evaluated in controlled studies where a treated group is compared with an untreated one [11–13]. 1 December 2012 \| Volume 7 \| Issue 12 \| e52917 December 2012 \| Volume 7 \| Issue 12 \| e52917 PLOS ONE \| www.plosone.org Results and Discussion doi:10.1371/journal.pone.0052917.g001 examined thoroughly and we used the presence of the specific endogenous antibodies against rbST as a biomarker for its abuse [4,30,33,34]. Although PIIINP, a marker of collagen turnover, is known to show potential in human and bovine hormone abuse detection [10,18], it has not been included into our biomarker panel yet because of the lack of a suitable commercially available standard protein and antibody. were calculated by analysis of sera from 67 untreated dairy cows (see paragraph 9.1 in the materials and methods section). Compared to the number of tested athletes in human studies, the number of tested control animals may seem to be rather low, but the variation within the dairy population is expected to be much lower, because of several reasons: First, only female cows have to be taken into account. Second, milking only occurs after first calving (usually at 20–24 months of age), thus after puberty, in which levels of IGF-1, IGFBP2 and osteocalcin are mainly changed due to growth and are more stable thereafter [37–39]. Third, since in this region of Europe mainly Frisian Holstein cows are used for milk production, we focussed on this particular race for the development of the test. And fourth, we do not need to consider sick animals, since their milk will not be allowed for consumption due to the presence of veterinary drug residues and therefore treatment with rbST is useless for sick dairy cows. Thus, the overall relative variation expected in dairy cows is anyway much lower than in athletes, where gender, different ethnicities, the effect of sports discipline, injury and all age groups need to be considered. Development of a 4-plex flow cytometric immunoassay For the simultaneous detection of these four candidate biomarkers, we developed a generic sample pre-treatment and 4-plex flow cytometric immunoassay (FCIA). To this end, our previously reported 3-plex assay [31] was extended with the biomarker osteocalcin. Adding osteocalcin to the existing triplex FCIA did not result in major interferences of any of the assay components of the four combined biomarker assays (data not shown). IGF-1 and osteocalcin concentrations of tested serum samples were calculated based on the obtained standard curves in serum-matched buffer (Figure S1). Results and Discussion The 4-plex FCIA is capable of determining IGF-1 and osteocalcin concentrations in the relevant range in serum, namely 64–400 ng mL21 for IGF-1 and 32– 320 ng mL21 for osteocalcin (note that serum samples were diluted 80-times prior to analysis, thus the standard curves cover protein concentrations of 0.8–5 ng mL21 for IGF-1 and 0.4– 4 ng mL21 for osteocalcin). For IGFBP2, the standard protein could not completely inhibit the B0 signal; therefore, we decided to work with normalized responses (B/B0) for the data analysis. For the induced anti-rbST-antibodies, we worked with the responses normalized to a single standard serum (B/Bd). Decision limits were 216 ng mL21 for IGF-1, 0.52 B/B0 for IGFBP2, 1.62 B/Bd for anti-rbST-antibodies and 160 ng mL21 for osteocalcin and are shown as green horizontal lines in Figure 2. Results of samples exceeding this limit were considered positive. Then, biomarker profiles of the dairy cows from both animal studies were measured (Figure 2). Results of the cows from animal study I are shown in dotted lines whereas the results of animal study II are shown in solid lines. Note that the animals from animal study I received two additional weekly rbST injections after the biweekly treatment period (the treatment schedules of both animal studies are indicated by the black horizontal bars above the graphs and shown in Figure S2). The generic sample pre-treatment was necessary for releasing IGF-1 from its binding protein-complex and preventing non- specific binding in the detection of anti-rbST-antibodies. The rather harsh pre-treatment protocol did not affect the detection quality of osteocalcin, thus it could be adopted for the combined 4- plex FCIA. Note that adding IGF-2 in excess, as done in commercially available human IGF-1 immunoassays, improved neither the normalized standard curves nor the detection of biomarker level differences in between treated and untreated animals. The developed assay showed high reproducibility for all measured candidate biomarkers (Table 2) and a comparable sensitivity to previous single biomarker methods [35,36]. However, the newly developed 4-plex FCIA has several advantages, such as the simultaneous measurement of all four markers in one sample from one well of a microtiter plate, which saves sample material, work load and time. Additionally, only one washing step was required compared to an average of six washing steps in a conventional enzyme-linked immunosorbent assay, making the 4- plex FCIA much faster and easy-to-use. Results and Discussion Furthermore, only occasionally a value exceeded the decision limit. Only the results of one cow were clearly above the decision limit, but these values were observed already during the adaptation period. In humans and despite large IGFBP2 levels (Figure 2B) are expected to decrease upon rbST treatment [32,41]. The IGFBP2 assay is of an inhibition format, thus B/B0 levels are inversely correlated with the concentration. Hence, higher B/B0 levels are expected after rbST treatment. For some of the rbST-treated cows, a slight increase in B/B0 levels can be observed after treatment (Figure 2B.1) with a decrease to baseline before the next treatment. But this pattern is not as pronounced as for IGF-1. Furthermore, only occasionally a value exceeded the decision limit. Only the results of one cow were clearly above the decision limit, but these values were observed already during the adaptation period. In humans and despite large Results and Discussion For the prediction of rbST abuse in dairy cows, we selected candidate biomarkers based on information found in literature (Table 1). These were markers of the IGF-axis (such as IGF-1 and IGFBP2) and bone markers (such as osteocalcin), known to be influenced by somatotropin and previously examined by the GH- 2000 group for detecting somatotropin abuse in athletes [13,32]. Furthermore, the immune response of cows treated with rbST was Table 1. Candidate biomarkers for ST abuse and their expected response upon ST treatment in human and cows. Biomarkers response to ST described for reference Acid labile subunit (ALS) increase human [45] Anti-rbST-antibodies increase bovine [30,31,33,34] Apolipoprotein A-1 (APOA1) decrease human [8] C-terminal cross-linked telopeptide of collagen I (ICTP) increase human [13,45,46] C-terminal propeptide of procollagen I (PICP) increase human [13] Haemoglobin a-chain (HbA1) increase human [6] IGF binding protein 2 (IGFBP2) decrease bovine [31,47] IGF binding protein 3 (IGFBP3) increase human [45] Insulin-like growth factor (IGF-1) increase human [45,46] Inter-alpha-trypsin inhibitor heavy chain H4 (ITIH4) decrease human [8] Leucine-rich a-2-glycoprotein (LRG) increase human [9] N-terminal propeptide of procollagen I (PINP) increase human [12,45] N-terminal propeptide of procollagen III (PIIINP) increase human [12,13,45,46] Osteocalcin increase human [12,13] Transthyretin (TTR) increase human [8] a-1 antitrypsin (AAT) increase human [8] doi:10.1371/journal.pone.0052917.t001 biomarkers for ST abuse and their expected response upon ST treatment in human and cows. Table 1. Candidate biomarkers for ST abuse and their expected response upon ST treatment in hu December 2012 \| Volume 7 \| Issue 12 \| e52917 PLOS ONE \| www.plosone.org 2 Biomarker Assessment for rbST Abuse Detection PLOS ONE \| www plosone org 3 December 2012 \| Volume 7 \| Issue 12 \| e December 2012 \| Volume 7 \| Issue 12 \| e52917 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org 3 Biomarker Assessment for rbST Abuse Detection Figure 1. Work flow for serum preparation, generic serum pre-treatment and 4-plex FCIA for serum candidate biomarkers. A detailed description can be found in Materials S1. Abbreviations: h – hour, IGF-1 – insulin-like growth factor 1, IGFBP2 – IGF binding protein 2, GS I – glycine solution I, GS II glycine solution II, min – minutes, PBST – phosphate-buffered saline with 0.05% (v/v) Tween-20, PBSTB – 0.1% (m/v) BSA in PBST, PE – phycoerythrin fluorescent label, rbST – recombinant bovine somatotropin, RT – room temperature, sec – seconds. Results and Discussion The whole assay procedure, starting from a serum sample until the results from the flow cytometer for all four candidate biomarkers, takes 3.5 hours for a whole 96 well microtiter plate. This demonstrates that the 4-plex FCIA is a rapid and promising screening tool for the detection of the four candidate biomarkers in serum. IGF-1 levels were found to be elevated directly after rbST treatment (Figure 2A.1) and returned back to baseline before the next treatment. This short response time was observed before in human studies, where IGF-1 concentrations were back to baseline one week after termination of somatotropin treatment [32]. Nevertheless, in athletes, IGF-1 stayed elevated throughout the treatment period. This difference in IGF-1 response to somato- tropin treatment could be due to the fact, that athletes were injected daily and, although a slow-release formulation was used in the here presented study, the biweekly treatment schedule does not reflect the same situation of permanently present somatotropin in circulation. IGF-1 levels of untreated animals (Figure 2A.2) remained below the decision limit. The found IGF-1 concentra- tions are consistent with previously reported serum IGF-1 concentrations in dairy cows [31,40]. IGF-1 levels were found to be elevated directly after rbST treatment (Figure 2A.1) and returned back to baseline before the next treatment. This short response time was observed before in human studies, where IGF-1 concentrations were back to baseline one week after termination of somatotropin treatment [32]. Nevertheless, in athletes, IGF-1 stayed elevated throughout the treatment period. This difference in IGF-1 response to somato- tropin treatment could be due to the fact, that athletes were injected daily and, although a slow-release formulation was used in the here presented study, the biweekly treatment schedule does not reflect the same situation of permanently present somatotropin in circulation. IGF-1 levels of untreated animals (Figure 2A.2) remained below the decision limit. The found IGF-1 concentra- tions are consistent with previously reported serum IGF-1 concentrations in dairy cows [31,40]. IGFBP2 levels (Figure 2B) are expected to decrease upon rbST treatment [32,41]. The IGFBP2 assay is of an inhibition format, thus B/B0 levels are inversely correlated with the concentration. Hence, higher B/B0 levels are expected after rbST treatment. For some of the rbST-treated cows, a slight increase in B/B0 levels can be observed after treatment (Figure 2B.1) with a decrease to baseline before the next treatment. But this pattern is not as pronounced as for IGF-1. Single candidate biomarker profiles of untreated and of rbST-treated cows For all of the candidate biomarkers, false-positive rates were quite low, indicating a high specificity of all of the biomarkers towards rbST treatment. Nevertheless, none of the candidate biomarkers reached the targeted 95% true-prediction (,5% false-compliant) rate at any time point required for a screening method according to Commission Decision 2002/657/ EC [23]. inter-individual differences, mean IGFBP2 levels responded quite well upon ST treatment, but the athletes were treated daily on three subsequent days [32]. B/B0 levels of untreated animals (Figure 2B.2) remained below the decision limit at almost all times. For the antibodies, endogenously produced by the cow as an immunological response upon rbST treatment [30], a delayed increase in signal was observed (Figure 2C.1). Most of the cows developed antibodies approximately 2 weeks after the first rbST injection and a maximum in response could be seen around the third injection (four weeks after start of rbST treatment). Thereafter, the responses declined slowly. Zwickl et al. reported an increase of antibody formation within the first three months of rbST treatment and a decline thereafter, but the amount of rbST administered in their study was much higher than recommended by the manufacturer and applied here [34]. For the untreated cows in our studies (Figure 2C.2), only one result was found to be above the decision limit. For osteocalcin, a slow increase in concentration was observed after rbST treatment (Figure 2D.1) compared to the untreated cows where the concentrations remained below the decision limit at almost all times. A similar effect on osteocalcin levels was observed in the human GH-2000 study [13]. Osteocalcin concentrations in our studies increased consistently in the 8 week treatment period, no gradual decline was observed as for the anti- rbST-antibodies. A slow osteocalcin decrease was noticed after rbST withdrawal but values remained above the decision limit for some of the cows until the end of the animal study. For all of the candidate biomarkers large inter-individual physiological differences in biomarker levels were apparent as for example seen in the adaptation period of the treated animals. IGF-1, IGFBP2 and osteocalcin levels differed quite a lot between individual animals. Biomarker levels are known to be influenced by many factors such as age and state of lactation. Nevertheless, the expected variation is much smaller than in athletes tested for ST abuse as already discussed above. Single candidate biomarker profiles of untreated and of rbST-treated cows After successful development of the 4-plex flow cytometric immunoassay, decision limits for each single candidate biomarker December 2012 \| Volume 7 \| Issue 12 \| e52917 December 2012 \| Volume 7 \| Issue 12 \| e52917 PLOS ONE \| www.plosone.org 4 Biomarker Assessment for rbST Abuse Detection Table 2. 4-plex FCIA assay performance characteristics for the single candidate biomarkers. Candidate biomarkers Performance characteristics IGF-1 IGFBP2 Anti-rbST antibodies Osteocalcin IC50 1.5a 9.5a - 1.1a Inter-assay variation 15.7b 7.9b 22.3b 17.1b Intra-assay variation 6.4b 5.7b 9.4b 9.5b Decision limit 216a 0.52c 1.62d 160a Stability The 4-plex FCIA can be performed stably over several months by different staff. Specificity No unwanted interaction in between the assays (analytes and antibodies) observed. IC50 related to 80-times diluted samples. ain ng mL21. bin %. cof B/B0. d f B/Bd (Figure 3). False-positive rates were calculated from untreated cows during the whole animal experiment (adaptation period samples from all cows and all the samples from untreated cows). High true- positive rates were reached by IGF-1 already at the beginning of the treatment period. Similar response patterns were observed for both studies. Only the double injections in study I led to a changed IGF-1 pattern. Also for the anti-rbST-antibodies, high true- positive rates of 75% were seen after the second rbST injection. But the response was study-dependent: while the animals from study I (equal age of 5 years) were found positive after the second injection until the end of the study period, a gradual decrease of the number of positively found animals was observed in study II (age ranged from 2 to 8 years). This could be due to the different ages of the animals in study II. We saw that the antibody response tended to be higher in the older animals. Younger animals also showed antibody response, which declined more quickly than for the older animals. For osteocalcin, as already seen in Figure 2D.1, some of the rbST-treated cows did not show osteocalcin concentrations beyond the decision limit in both studies. The increase of true-positive found samples at the end of the treatment period in study I was due to the double rbST injections. As already expected from the biomarker profiles (Figure 2B.1), IGFBP2 did not show high true-positive rates, i.e., none of the animals from study I and only some animals in study II were found above the decision limit. Single candidate biomarker profiles of untreated and of rbST-treated cows Note that we accounted for the variation in our untreated reference population used to assess the decision limits. Also the response upon rbST treatment differed in every individual cow. Some cows showed a big increase in IGF-1 levels short after injection while others did not show any response above decision limit (non-responders). Also for osteocal- cin, some cows hardly showed any response after treatment. December 2012 \| Volume 7 \| Issue 12 \| e52917 Statistical multiple biomarker analysis Since the single biomarker analysis and additive biomarker analysis, which were both based on decision limits, did not deliver satisfying results for predicting rbST abuse, a different biomarker- combining approach was chosen for analysis of the data. K-nearest neighbours (kNN), a statistical prediction tool, was used to build a model from one group of data (Group A: all animals of animal study II and untreated animals from animal study I) and predict the results of Group B (rbST-treated cows of animal study I and 67 independent untreated cows) on basis of the built model. Eleven different models (one for every possible combination of two to four biomarkers) were evaluated to find the optimal biomarker combination for rbST abuse prediction. True-positive rates of Group B data were calculated for every biomarker combination over the time of the whole animal study and are shown in Figure 6 (Table S1 shows corresponding data). Six of the eleven different models yielded true-positive rates above the 95% true-positive rate required for a screening method at several time points. For the biomarker combinations IGF-1 - IGFBP2 - anti-rbST-antibodies (IBA) and IGFBP2 - anti-rbST-antibodies - osteocalcin (BAO), only one time point within the biweekly treatment period was above 95%. Note that in total, samples from eleven time points were obtained and analysed during the biweekly treatment period of animal study I. For the biomarker combinations IBAO and IA, four and six time points within the biweekly treatment period were above the 95% target respectively. Seven time points above the 95% target within the biweekly treatment period were reached by the prediction models based on the biomarker combinations IAO and AO. For the three best performing models (IA, AO and IAO), true-positive rates above 95% (,5% false-compliant) were reached following the second rbST injection. For IA, a true-prediction rate of almost 60% was observed already one week after the first rbST injection, whereas AO only showed 30%, which is in accordance with expectations since IGF-1 is a quick responding biomarker and osteocalcin has a delayed response time. Since all of the rbST- treated cows were detected by the three best performing models (IA, AO, IAO) at the end of the biweekly treatment period, no further increased prediction rate was observed due to the subsequent two weekly injections. Additive biomarker analysis Since no single candidate biomarker was capable of predicting 95% of the rbST-treated samples correctly, we tested different possibilities of combining biomarker results for improvement of the predictive power of our 4-plex FCIA. One approach to do this is the additive biomarker analysis. In Figure 4, the number of candidate biomarkers responding above decision limit per cow and per time point within the animal studies is shown. As already described in paragraph 2 of the results section, there were big inter-individual differences: some cows responded in many markers, others only in one or two for some time points. There were also two extreme cases: one cow responded in all four tested The predictive power of each candidate biomarker was assessed by calculating true-positive rates for all samples from rbST-treated cows in their treatment and withdrawal period December 2012 \| Volume 7 \| Issue 12 \| e52917 PLOS ONE \| www.plosone.org December 2012 \| Volume 7 \| Issue 12 \| e52917 5 Biomarker Assessment for rbST Ab Biomarker Assessment for rbST Abuse Detection PLOS ONE \| www.plosone.org 6 December 2012 \| Volume 7 \| Issue 12 PLOS ONE \| www.plosone.org 6 December 2012 \| Volume 7 \| Issue 12 \| e52917 December 2012 \| Volume 7 \| Issue 12 \| e52917 December 2012 \| Volume 7 \| Issue 12 \| e52917 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org Biomarker Assessment for rbST Abuse Detection Biomarker Assessment for rbST Abuse Detection Figure 2. Biomarker profiles of rbST-treated (left) and untreated (right) dairy cows. Profiles from animal study I (dotted lines) and animal study II (solid lines) are shown. Sera from adaption period (3 sera from every cow), treatment period (13 sera per cow from animal study I and 9 sera per cow from animal study II) and withdrawal period (5 sera per cow from animal study I and 6 sera per cow from animal study II) were measured in duplicate. Biomarkers shown are concentrations of IGF-1 (A), B/B0 levels of IGFBP2 (B), B/Bd levels of antibodies against rbST (C) and concentrations of osteocalcin (D). The rbST treatment schedules for both animal studies are indicated by two black horizontal bars and decision limit per biomarker by the green horizontal line. Note that cows from animal study II received two additional rbST injections after the biweekly treatment period. doi:10.1371/journal.pone.0052917.g002 frequency will not be found in real practice. Additive biomarker analysis Furthermore, also with the additive biomarker analysis, quite some false-positive results were obtained throughout the whole study. markers above decision limit at one time point and another rbST- treated cow did not show any response above decision limit at any day. On the other hand, there were untreated cows, which showed positive responses in one of the candidate biomarkers. Figure 5 shows the true-positive rate obtained for the rbST-treated cows of both animal studies considering a sample positive, when at least one biomarker reacted above the respective decision limit. Although the true-positive rate obtained with the additive biomarker analysis was much higher than for the single candidate biomarkers, the 95% true-positive rate required for a screening method was only reached at some time points in study I within the biweekly treatment period. After the double rbST injection in study I, all of the cows were found positive, but this treatment Statistical multiple biomarker analysis After withdrawal of rbST, the true-positive rate of the models based on IA, AO and IAO remained above 95% for two more weeks and then declined to 70% four weeks after withdrawal. Figure 3. Predictive power of each single candidate biomarker for indicating rbST abuse. True-positive rates were calculated for all samples from rbST-treated cows in their treatment and withdrawal periods of study I and II. False-positive rates were calculated for all samples from untreated cows from the two animal studies (adaptation period samples from all cows and all the samples from untreated cows). Samples were considered positive if their biomarker value exceeded the respective decision limit. The treatment schedules of the two controlled animal studies are indicated by black horizontal bars on top of the graph. The targeted 95% true-positive (,5% false-compliant) rate according to 2002/657/EC is indicated by the dotted horizontal line. doi:10.1371/journal.pone.0052917.g003 Figure 3. Predictive power of each single candidate biomarker for indicating rbST abuse. True-positive rates were calculated for all samples from rbST-treated cows in their treatment and withdrawal periods of study I and II. False-positive rates were calculated for all samples from untreated cows from the two animal studies (adaptation period samples from all cows and all the samples from untreated cows). Samples were considered positive if their biomarker value exceeded the respective decision limit. The treatment schedules of the two controlled animal studies are indicated by black horizontal bars on top of the graph. The targeted 95% true-positive (,5% false-compliant) rate according to 2002/657/EC is indicated by the dotted horizontal line. doi:10.1371/journal.pone.0052917.g003 Since we used all of the untreated animals of both animal studies for model building, false-positive rates for the eleven different models were calculated based on the results of the 67 independent untreated cows (Table S2). For the three best- performing prediction models IA, AO and IAO, false-positive rates ranged from 10.6% to 14.7%, which was quite acceptable, since samples screened positive must be analysed by a subsequent confirmatory analysis method according to Commission Decision 2002/657/EC anyway [23]. The confirmation method is based on the detection of an N-terminal peptide of somatotropin, which has December 2012 \| Volume 7 \| Issue 12 \| e52917 December 2012 \| Volume 7 \| Issue 12 \| e52917 PLOS ONE \| www.plosone.org 7 Biomarker Assessment for rbST Abuse Detection Figure 4. Number of biomarkers reacting above the respective decision limit. Statistical multiple biomarker analysis Results shown per cow (in animal studies I and II) and day of the controlled animal studies. Each row represents one individual cow. Vertical dotted lines indicate the treatment time points in both animal studies. doi:10.1371/journal.pone.0052917.g004 Figure 4. Number of biomarkers reacting above the respective decision limit. Results shown per cow (in animal studies I and II) and day of the controlled animal studies. Each row represents one individual cow. Vertical dotted lines indicate the treatment time points in both animal studies. doi:10.1371/journal.pone.0052917.g004 Figure 4. Number of biomarkers reacting above the respective decision limit. Results shown per cow (in animal studies I and II) and day of the controlled animal studies. Each row represents one individual cow. Vertical dotted lines indicate the treatment time points in both animal studies. doi:10.1371/journal.pone.0052917.g004 The results obtained proof that the developed 4-plex FCIA reduced to an AO biomarker combination 2-plex FCIA, applied to an in vivo evaluation and combined with a thorough statistical multiple biomarker analysis can detect more than 95% of the rbST-treated cows truly positive directly after the second rbST injection until the end of their treatment period and even thereafter. This meets the requirements of Commission Decision a different terminal amino acid in the recombinant form of the hormone [26]. We concluded from the results of the here presented studies that the AO biomarker combination is the preferred model for predicting rbST abuse. It yielded seven out of eleven time points above the 95% target and if two biomarkers are equally well-suited for prediction as three biomarkers, the simpler version is favoured. Figure 5. Predictive power (shown as true-positive and false-positive rates) of the additive biomarker analysis. True-positive rates were calculated for all samples from rbST-treated cows in their treatment and withdrawal periods of study I and II. False-positive rates were calculated for all samples from untreated cows from the two animal studies (adaptation period samples from all cows and all the samples from untreated cows). Samples were considered positive if one of the candidate biomarkers exceeded its respective decision limit. The treatment schedules of the two animal studies are indicated by black horizontal bars on top of the graph. The targeted 95% true-positive (,5% false-compliant) rate according to 2002/657/EC is indicated by the dotted horizontal line. doi:10.1371/journal.pone.0052917.g005 Figure 5. Predictive power (shown as true-positive and false-positive rates) of the additive biomarker analysis. Statistical multiple biomarker analysis True-positive rates were calculated for all samples from rbST-treated cows in their treatment and withdrawal periods of study I and II. False-positive rates were calculated for all samples from untreated cows from the two animal studies (adaptation period samples from all cows and all the samples from untreated cows). Samples were considered positive if one of the candidate biomarkers exceeded its respective decision limit. The treatment schedules of the two animal studies are indicated by black horizontal bars on top of the graph. The targeted 95% true-positive (,5% false-compliant) rate according to 2002/657/EC is indicated by the dotted horizontal line. doi:10.1371/journal.pone.0052917.g005 r (shown as true-positive and false-positive rates) Figure 5. Predictive power (shown as true-positive and false-positive rates) of the additive biomarker analysis. True-positive rates were calculated for all samples from rbST-treated cows in their treatment and withdrawal periods of study I and II. False-positive rates were calculated for all samples from untreated cows from the two animal studies (adaptation period samples from all cows and all the samples from untreated cows). Samples were considered positive if one of the candidate biomarkers exceeded its respective decision limit. The treatment schedules of the two animal studies are indicated by black horizontal bars on top of the graph. The targeted 95% true-positive (,5% false-compliant) rate according to 2002/657/EC is indicated by the dotted horizontal line. December 2012 \| Volume 7 \| Issue 12 \| e52917 December 2012 \| Volume 7 \| Issue 12 \| e52917 PLOS ONE \| www.plosone.org 8 Biomarker Assessment for rbST Abuse Detection Figure 6. True-positive rates following statistical multiple biomarker analysis. True-positive rates, obtained with the prediction models based on the eleven different biomarker combinations, were calculated for rbST-treated cows from animal study I in their treatment and withdrawal period. The treatment schedules of animal study I is indicated by black horizontal bars on top of the graphs. The targeted 95% true-positive rate according to 2002/657/EC is indicated by the dotted horizontal lines. doi:10.1371/journal.pone.0052917.g006 Biomarker Assessment for rbST Abuse Detection Figure 6. True-positive rates following statistical multiple biomarker analysis. True-positive rates, obtained with the prediction models based on the eleven different biomarker combinations, were calculated for rbST-treated cows from animal study I in their treatment and withdrawal period. The treatment schedules of animal study I is indicated by black horizontal bars on top of the graphs. Statistical multiple biomarker analysis The targeted 95% true-positive rate according to 2002/657/EC is indicated by the dotted horizontal lines. doi:10.1371/journal.pone.0052917.g006 2002/657/EC for a screening assay for the detection of banned veterinary drugs such as rbST [23]. requirements of Commission Decision 2002/657/EC for screen- ing assays [23]. Furthermore, the extensive in vivo validation with two independent rbST animal treatment studies followed by statistical analysis revealed that a combination of just two candidate biomarkers is actually adequate for detection of rbST treatment. Therefore, even a 2-plex version (namely the combi- nation of anti-rbST-antibodies and osteocalcin) of our assay would already be fit-for-purpose based on the data presented here. When comparing with previously reported results of a 3-plex FCIA combining IGF-1, IGFBP2 and anti-rbST-antibodies [31], the models presented here seemingly perform somewhat less, especially at the beginning of the rbST treatment but the new models are much more realistic: Note that here, two completely independent groups were used for model building (Group A) and prediction (Group B), whereas in the 3-plex experiments [31], sample data used for prediction were from the same cows as the data on which the model was built, leading to an overestimation of true-positive results in that work. Nevertheless some issues should be considered. First of all, for obvious ethical and cost reasons, the rbST treatment period was limited to 8 weeks in our animal studies, so we do not know yet how the prediction models would perform for long-term treated animals. As it can be seen in Figure 2C.1, the antibody biomarker response declined somewhat in course of the treatment period and we do not know whether this would influence the prediction quality in a prolonged treatment. Second, in the presented animal studies, cows were treated with rbST for the first time in their lives December 2012 \| Volume 7 \| Issue 12 \| e52917 Chemicals and instruments Ultrasonic bath, monosodium phosphate monohydrate (NaH2- PO46H2O), potassium dihydrogen phosphate (KH2PO4), sodium chloride (NaCl), sodium azide (NaN3) and Tween 20 were obtained from VWR International (Amsterdam, The Nether- lands). Microcentrifuge Model 16K was purchased from Bio-Rad (Veenendaal, The Netherlands). Protein LoBind Tubes, Safe Lock Tubes (amber) and Centrifuge 5810R were obtained from Eppendorf (Hamburg, Germany). 1-Ethyl-3-(3-dimethylaminopro- pyl)carbodiimide (EDC), 2-(N-morpholino)ethanesulfonic acid (MES) hydrate, ovalbumin and bovine serum albumin (BSA) were obtained from Sigma-Aldrich (St. Louis, MO, USA). MultiScreen HTS filter plates were purchased from Millipore (Billerica, MA, USA). Purified bovine osteocalcin and mouse anti-bovine osteo- calcin antibodies were obtained from Haematologic Technologies, Inc. (Essex Junction, VT, USA). Insulin-like growth factor-I (IGF- I; human recombinant) was purchased from Fitzgerald Industries International (North Acton, MA, USA). Insulin-like growth factor binding protein-2 (IGFBP-2; bovine recombinant, receptor grade) was purchased from IBT (Reutlingen, Germany). Mouse anti-IGF- 1 was supplied by LifeSpan BioSciences, Inc. (clone SPM406, Seattle, WA, USA) and the rabbit anti-IGFBP-2 was from United States Biological (Swampscott, MA, USA). Monsanto rbST standard was obtained from the National Hormone & Peptide Program (NHPP) of Dr Parlow (Torrance, CA, USA). R- Phycoerythrin (PE)-labelled goat anti-bovine immunoglobulins (GAB-PE) were from Santa Cruz Biotechnology (Santa Cruz, CA, USA) and R-Phycoerythrin (PE)-labelled goat anti-mouse immu- noglobulins (GAM-PE) and goat anti-rabbit immunoglobulins (GAR-PE) were purchased at Prozyme (San Leandro, CA, USA). Donor adult bovine serum was from HyClone (South Logan, UT, USA). Sodium hydroxide (NaOH), disodium hydrogen phosphate dihydrate (Na2HPO462 H2O) and hydrochloric acid (HCl) were purchased from Merck (Darmstadt, Germany). SeroMAP micro- spheres (microsphere sets 025, 050, 078 and 084) and sheath fluid were obtained from Luminex (Austin, TX, USA). The Luminex 100 IS 2.2 system consisting of a Luminex 100 analyser and a Luminex XY Platform was purchased from Applied Cytometry Systems (ACS, Dinnington, Sheffield, South Yorkshire, UK). Snijder Test tube rotator was from Omnilabo International (Breda, The Netherlands). 10 mL polypropylene tubes were obtained from Greiner Bio-One (Alphen aan de Rijn, The Netherlands). Glycine was purchased from Duchefa (Haarlem, The Netherlands) and sulfo-N-Hydroxysuccinimide (Sulfo-NHS) from Fluka (Buchs, Switzerland). Sodium dodecyl sulphate (SDS) was obtained from Serva (Heidelberg, Germany). The microtiter vari-shaker was purchased from Dynatech (Guernsey, UK). PosilacH (rbST) 500 mg single dose syringes and syringes with only the slow release formula were obtained from Monsanto Company (St. Louis, MO, USA) for animal study I and from Elanco Animal Health (Greenfield, IN, USA) for animal study II. Biomarker Assessment for rbST Abuse Detection and there are no data about biomarker levels during a second treatment period after calving. According to the manufacturers’ treatment schedule, dairy cows are treated starting from 9 weeks after calving until the end of the lactation (typically a biweekly treatment) and the following year again. Eppard et al. and Zwickl et al. reported that repeated treatment periods did not cause an immunological memory effect with enhanced antibody production in the second treatment period [33,34]. For both situations, long- term treatment and repeated treatment, the IA, AO and IAO biomarker combinations should be tested and possibly the inclusion of other biomarkers could be considered. Since blood sampling in routine veterinary monitoring programmes might not be justified in some countries, we suggest a tiered approach according to Commission Decision 2002/657/EC. This would consist of three steps: First, a fast screening for anti-rbST- antibodies in tank milk using our previously described milk FCIA [29]. Second, in case of suspicious findings, a more detailed individual bovine serum biomarker profiling using the IA, AO or IAO FCIA presented here will provide additional evidence, since they are based on more biomarkers and data for individual cows. Note that in practice, a whole stable and not an individual cow is treated with rbST, thus increasing the chance of detecting rbST use. And third, for final confirmation of rbST itself in serum samples of suspect individual cows, a highly sensitive mass spectrometric confirmatory method, which fulfils the 2002/657/ EC confirmatory method requirements, is to be used [26]. Conclusions In this study a multidisciplinary approach was used for the development of an in vivo validated screening assay for rbST abuse in dairy cows. Four candidate biomarkers for rbST abuse were assessed using a newly developed 4-plex flow cytometric immu- noassay, in vivo validation studies and advanced statistics. Biomarkers indicative for rbST administration were evaluated based on two extensive animal studies with rbST-treated and untreated animals and an additional untreated reference popula- tion. Different data evaluation approaches were tested. The prediction tool kNN using a biomarker combination endogenously produced antibodies against rbST and osteocalcin proved to be very reliable and correctly predicted 95% of the treated samples starting from the second rbST injection until the end of the treatment period and even thereafter. This reduced 2-plex FCIA method (consisting of biomarkers anti-rbST antibodies and osteocalcin) combined with the statistical analysis approach was shown to be a fast, reliable and cost-effective approach to screen for rbST abuse in dairy cattle. These methods and models can be included in routine veterinary control programmes in the European Union for detection of rbST abuse and also in the control of rbST-free dairy farms in the United States of America and other countries. Chemicals and instruments Buffers and solutions Discussion For the first time, to the best of our knowledge, a 4-plex biomarker assay development and data evaluation model is presented for the detection of rbST abuse, which fulfils the December 2012 \| Volume 7 \| Issue 12 \| e52917 PLOS ONE \| www.plosone.org 9 Biomarker Assessment for rbST Abuse Detection Sample materials p Samples from different sources were used for analysis. Serum samples from two independent controlled animal treatment studies were used. In animal study I, eight Holstein cows were selected. These cows were all about 5 years old, divided in two groups of 4 animals each and treated with 500 mg rbST in slow-release formula or slow-release formula only. After a two-week adaptation period, they received an injection every second week, in total 4 times in accordance with the suggested treatment schedule by the manufacturer (http://www.fda.gov/downloads/AnimalVeterinary/ Products/ApprovedAnimalDrugProducts/FOIADrugSummaries/ ucm050022.pdf; accessed 2012 Apr 4). Since we did not know for sure whether we would see any response, the cows were thereafter treated two times more but with a weekly interval, followed by a final 4-week withdrawal period. In animal study II, 10 Holstein dairy cows were divided in two groups. In contrast to animal study I, these cows were of different age (2–8 years). After a 2-week adaptation period, 8 cows were treated every second week with 500 mg rbST in a slow-release formula during 8 weeks and 2 control cows were treated with the slow- release formula only. The biweekly treatment period according to manufacturers’ guidelines was directly followed by a 4-week withdrawal period. In both studies, blood sampling was scheduled similarly: During the two week adaptation period, blood samples were collected weekly; during the treatment period, blood samples were collected a day before, a day after and a week after injection and during withdrawal, blood samples were collected weekly for four more weeks, which yielded 21 serum samples per cow in animal study I and 18 serum samples per cow in animal study II. The treatment schedule and blood sampling time points are shown in Figure S2. Unfortunately, one untreated cow died in the beginning of animal study I because of swollen hocks, which led to general inflammation and sepsis. Therefore in study I, results could be obtained for 4 rbST-treated and 3 untreated cows. Furthermore, one cow from animal study II fell sick (hock joint inflammation, lung infection and sepsis) in course of the experiment and its biomarker level results were excluded from statistical analysis. Sample pre-treatment A generic sample pre-treatment procedure which was crucial for removing non-specific interferences and making the candidate biomarkers accessible for detection was described previously [4,28,29], is depicted in Figure 1 and described in-depth in Materials S1. Additive biomarker analysis. After evaluating biomarker profiles and true-positive rates based on single biomarkers, an additive biomarker approach was tested. Here, a sample was considered as rbST-treated when at least one of the candidate Additive biomarker analysis. After evaluating biomarker profiles and true-positive rates based on single biomarkers, an additive biomarker approach was tested. Here, a sample was considered as rbST-treated when at least one of the candidate Biomarker Assessment for rbST Abuse Detection microspheres (sets 050, 025, 078 respectively) was described before [4,28,29]. Coupling 75 mg mL21 osteocalcin to micro- spheres (set 084) was done following the same procedure. Four-plex flow cytometric immunoassay procedure The assay procedure for detection of three biomarkers was described before [31] and is similar for four biomarkers in the present study and summarized in Figure 1. The samples were analysed in duplicate in the flow cytometer at 1 mL s21 until 50 microspheres per set were counted, up to a maximum of 50 mL per sample. A typical analysis of a full 96 well microtiter plate takes 3.5 hours starting from raw serum until the results are obtained. Sample materials For investigation of natural physiological variations in biomarker levels, sera from 67 healthy, lactating cows varying in the age of two to eleven years, from two different locations, in different stages of their lactating cycle were analysed, to reflect a normal population of untreated dairy cows. Based on the origin of these animals the assumption of being untreated with rbST was justified. Standard preparation Protein standards of IGF-1, IGFBP2 and osteocalcin, prepared in serum-matched buffer (80 mg mL21 BSA in PBS), were used for standard curves ranging from 0.08 to 20 ng mL21 for IGF-1 and osteocalcin and from 0.2 to 50 ng mL21 for IGFBP2. Also blank standard samples (80 mg mL21 BSA in PBS without any IGF-1, IGFBP2 and osteocalcin) were measured. Note that no standards are commercially available for anti-rbST-antibodies. Single biomarker analysis approach. Using a single biomarker for prediction of unknown samples as rbST-treated or untreated, the calculation of decision limits for each biomarker was necessary. These were based on the results obtained from a population of 67 untreated dairy cows being diverse in age, in lactation stage and in origin. For every biomarker, results were averaged and two-times the standard deviation was added to obtain the decision limit. Samples found to show concentrations (for IGF-1 and osteocalcin), B/B0 (for IGFBP2) or B/Bd (for anti- rbST antibodies) beyond the respective calculated decision limit, were considered as rbST-treated (positive). True-positive and false- positive rates could be calculated for every single biomarker from the results of the controlled animal studies. Single biomarker analysis approach. Using a single biomarker for prediction of unknown samples as rbST-treated or untreated, the calculation of decision limits for each biomarker was necessary. These were based on the results obtained from a population of 67 untreated dairy cows being diverse in age, in lactation stage and in origin. For every biomarker, results were averaged and two-times the standard deviation was added to obtain the decision limit. Samples found to show concentrations (for IGF-1 and osteocalcin), B/B0 (for IGFBP2) or B/Bd (for anti- rbST antibodies) beyond the respective calculated decision limit, were considered as rbST-treated (positive). True-positive and false- positive rates could be calculated for every single biomarker from the results of the controlled animal studies. Data analysis Raw median fluorescence intensities (MFIs) were measured by the flow cytometer for every single candidate biomarker. Every sample was measured in duplicate and MFIs were averaged before further analysis. For IGF-1, IGFBP2 and osteocalcin, B/B0 values were calculated per sample by dividing the measured MFI by the MFI of a blank biomarker-free standard. Then, concentrations were recalculated from standard curves (non-linear four-parameter curve fit) using GraphPad Prism program (GraphPad Software Inc., San Diego, USA) for IGF-1 and osteocalcin. For IGFBP2, no complete inhibition could be obtained with the available standard protein, therefore, no actual concentrations were determined and B/B0 values were simply used. For anti-rbST-antibodies, which are endogenously produced by the cow in response to rbST treatment, no standard was available. To be able to normalize, measured sample MFIs were divided by the MFI of one serum sample, which was measured every time (B/Bd). This serum was donor adult bovine serum which was a mixture of sera from many cows from one herd. Since this is produced in large amounts, it can be used for a long time with constant quality. To assess the 4-plex FCIA quality and compare it to other methods, assay performance characteristics were calculated, such as IC50, inter-assay and intra-assay variation (describing precision and ruggedness). For IGF-1, IGFBP2 and osteocalcin, IC50 was read from standard curves at 50% inhibition of the signal of the blank. For all candidate biomarkers, inter-assay variation was determined by measuring 8 different serum samples on 8 days. Mean of results (concentrations for IGF-1 and osteocalcin, B/B0 for IGFBP2 and B/Bd for anti-rbST-antibodies), standard deviation and percentaged standard deviation (%CV) were calculated for every serum. The average of the 8 percentaged standard deviations was the inter-assay deviation. Intra-assay variation was calculated the same way from 8 repetitions of 8 sera within one microtiter plate. Ethics statement Buffers and solutions were prepared as follows: phosphate- buffered saline (PBS; 154 mM NaCl, 5.39 mM Na2HPO4, 1.29 mM KH2PO4, pH 7.4), PBST (PBS, 0.05% v/v Tween- 20), PBSTB (0.1% w/v BSA in PBST), glycine solution I (GS I; 27.5 mM glycine, pH 0.5 adjusted with HCl), glycine solution II (GS II; 400 mM glycine, 0.3% w/v SDS, pH 10 adjusted with NaOH), MES buffer (50 mM, pH 5), blocking buffer (PBS, 0.1% w/v BSA, 0.02% v/v Tween-20, 0.05% w/v NaN3). Permission for animal study I (EC2007/71) was obtained from the Ethical Commission of the Faculty of Veterinary Medicine of Ghent University (Belgium) on basis of the Dutch law on animal studies (Wet op de Dierproeven). For animal study II, permission (EC2010-21) was obtained from the Ethical Commission of the Animal Science Group of Wageningen University and Research Centre in Lelystad (The Netherlands). December 2012 \| Volume 7 \| Issue 12 \| e52917 PLOS ONE \| www.plosone.org 10 Biomarker Assessment for rbST Abuse Detection Acknowledgments We kindly thank Dr Philippe Delahaut from CER groupe (Marloie, Belgium) for supplying reference serum samples from untreated cows, Johan van Hende from University of Ghent (Belgium) for performing the controlled animal treatment experiment and Dr Maria J. Groot and Jan S. Ossenkoppele for their help with collecting the sample material. ) (DOCX) Table S2 False-positive rates of the statistical multiple biomarker analysis. Results were calculated for 67 indepen- dent untreated cows predicted with the eleven different biomarker combination models. (DOCX) For model building of the Group A data, a training and test set were chosen by using a stratified repeated random sub-sampling approach, which means that 70% of the rbST-treated and 70% of the untreated samples were selected for the training set and the remaining 30% of both groups for the test set for internal validation, which is necessary to build a strong model. Subse- quently, concentrations, B/B0 and B/Bd values of the training set were auto-scaled and a kNN model was built on the training set data. The optimal number of k (1#k#10) was chosen based on the bootstrapping approach [44] leaving out 10% of the training data (randomly with replacement), which was repeated 10-times. The resulting model was validated with the test set data and thereafter used for predicting Group B data. To obtain an average performance of the model, this procedure was run 10,000 times; every time different randomly chosen training and test sets of Microsphere preparation Covalent coupling of 100 mg mL21 Monsanto rbST standard, 100 mg mL21 IGF-1 and 10 mg mL21 IGFBP2 to seroMAP PLOS ONE \| www.plosone.org December 2012 \| Volume 7 \| Issue 12 \| e52917 11 Biomarker Assessment for rbST Abuse Detection biomarkers reacted above decision limit and also here, true- positive and false-positive rates were calculated. Group A data were applied. Correctly and falsely predicted results were evaluated for Group B and a true-positive rate and false- positive rate could be calculated for every Group B sample. Multiple biomarker statistical approach. After evaluating single candidate biomarkers and testing the additive biomarker approach, we assessed how well a statistical combination of two to four markers was capable to predict rbST abuse. Therefore, a k- nearest neighbours prediction model (kNN) in the R environment [42] and functions available in R package e1071 [43] were used to evaluate all eleven theoretical combinations of two to four biomarkers. As in the single biomarker approach, recalculated concentrations for IGF-1 and osteocalcin and B/B0 signals for IGFBP2 as well as B/Bd signals for rbST-induced antibodies for every sample from the animal studies were included in the data analysis. For obvious ethical reasons, we had only a limited number of rbST-treated animals available. Therefore, all serum sample time points per cow (21 time points in the trial period of 14 weeks for animal study I and 18 time points in the trial period of 13 weeks for animal study II) were used for data analysis, despite the fact that these were not completely independent. However, only data from independent cows were used for model building and sample prediction. Author Contributions Conceived and designed the experiments: MB NS MN SL. Performed the experiments: NS SL. Analyzed the data: SL GvdV. Contributed reagents/ materials/analysis tools: GvdV MB SL NS. Wrote the paper: SL MN NS GvdV. Conceived and designed the experiments: MB NS MN SL. Performed the experiments: NS SL. Analyzed the data: SL GvdV. Contributed reagents/ materials/analysis tools: GvdV MB SL NS. Wrote the paper: SL MN NS GvdV. 9. Boateng J, Kay R, Lancashire L, Brown P, Velloso C, et al. (2009) A proteomic approach combining MS and bioinformatic analysis for the detection and identification of biomarkers of administration of exogenous human growth hormone in humans. Proteom Clin Appl 3: 912–922. 8. Ding J, Okada S, Jørgensen JOL, Kopchick JJ (2011) Novel serum protein biomarkers indicative of growth hormone doping in healthy human subjects. Proteomics 11: 3565–3571. 1. Teale P, Barton C, Driver PM, Kay RG (2009) Biomarkers: unrealized potential in sports doping analysis. Bioanalysis 1: 1103–1118. 3. Draisci R, Montesissa C, Santamaria B, D’Ambrosio C, Ferretti G, et al. (2007) Integrated analytical approach in veal calves administered the anabolic androgenic steroids boldenone and boldione: urine and plasma kinetic profile and changes in plasma protein expression. Proteomics 7: 3184–3193. 7. Ding J, List EO, Okada S, Kopchick JJ (2009) Perspective: Proteomic approach to detect biomarkers of human growth hormone. Growth Horm IGF Res 19: 399–407. Materials S1 Serum preparation, generic serum pre- treatment and 4-plex FCIA for serum candidate bio- markers. (DOCX) First, the whole data set was divided into two groups: Group A data were used to build the time-point-independent prediction model. Therefore and to use sufficient sample numbers for the model building, this group contained all data from animal study II (diverse population with biweekly treatment only). Furthermore, since two control animals were not enough to represent untreated cows, Group A also contained the data from the untreated animals of animal study I. In total, 98 samples from treated and 119 samples from untreated cows were used for model building. Group B data were used for prediction based on the Group A model. Group B contained the data from the rbST-treated cows of animal study I (uniform in age with biweekly treatment and two additional weekly injections) and the 67 untreated cows. Note that these are sample data independent from Group A data. Table S1 True-positive rates of the statistical multiple biomarker analysis. True-positive rates, obtained with the prediction models based on the eleven different biomarker combinations, were calculated for rbST-treated cows from animal study I in their treatment (day 16–71) and withdrawal period (day 72–99). 1. Teale P, Barton C, Driver PM, Kay RG (2009) Biomarkers: unrealized potential in sports doping analysis. Bioanalysis 1: 1103–1118. 2. Cacciatore G, Eisenberg SW, Situ C, Mooney MH, Delahaut P, et al. (2009) Effect of growth-promoting 17[beta]-estradiol, 19-nortestosterone and dexa- methasone on circulating levels of nine potential biomarker candidates in veal calves. Anal Chim Acta 637: 351–359. 3. Draisci R, Montesissa C, Santamaria B, D’Ambrosio C, Ferretti G, et al. (2007) Integrated analytical approach in veal calves administered the anabolic androgenic steroids boldenone and boldione: urine and plasma kinetic profile and changes in plasma protein expression. Proteomics 7: 3184–3193. 4. Smits NGE, Bremer MGEG, Ludwig SKJ, Nielen MWF (2012) Development of a flow cytometric immunoassay for recombinant bovine somatotropin-induced antibodies in serum of dairy cows. Drug Test Analysis 4: 362–367. 5. Chung L, Baxter RC (2009) Detection of growth hormone responsive proteins using SELDI–TOF mass spectrometry. Growth Horm IGF Res 19: 383–387. p p g y y 2. Cacciatore G, Eisenberg SW, Situ C, Mooney MH, Delahaut P, et al. (2009) Effect of growth-promoting 17[beta]-estradiol, 19-nortestosterone and dexa- methasone on circulating levels of nine potential biomarker candidates in veal calves. Anal Chim Acta 637: 351–359. 6. Chung L, Clifford D, Buckley M, Baxter RC (2006) Novel biomarkers of human growth hormone action from serum proteomic profiling using protein chip mass spectrometry. J Clin Endocrinol Metab 91: 671–677. 4. Smits NGE, Bremer MGEG, Ludwig SKJ, Nielen MWF (2012) Development of a flow cytometric immunoassay for recombinant bovine somatotropin-induced antibodies in serum of dairy cows. Drug Test Analysis 4: 362–367. 5. Chung L, Baxter RC (2009) Detection of growth hormone responsive proteins using SELDI–TOF mass spectrometry. Growth Horm IGF Res 19: 383–387. 1. Teale P, Barton C, Driver PM, Kay RG (2009) Biomarkers: unrealized potential in sports doping analysis. Bioanalysis 1: 1103–1118. Supporting Information Figure S1 Standard curves of the three rbST-dependent biomarkers IGF-1, IGFBP2 and osteocalcin. Each data point is the mean of 8 separate measurements in a serum-matched buffer (80 mg mL21 BSA in PBS solution). All curves relate to 80- times diluted sera. (TIF) Figure S2 Treatment schedule and sampling time points for animal studies I and II. Arrows indicate the treatment of the cows with rbST in slow-release formula or the slow-release formula only; bold vertical lines indicate blood sampling time points. Biomarker Assessment for rbST Abuse Detection 10. Mooney MH, Situ C, Cacciatore G, Hutchinson T, Elliott C, et al. (2008) Plasma biomarker profiling in the detection of growth promoter use in calves. Biomarkers 13: 246–256. 28. Bremer MGEG, Smits NGE, Haasnoot W, Nielen MWF (2010) Multiplex ready flow cytometric immunoassay for total insulin like growth factor 1 in serum of cattle. Analyst 135: 1147–1152. y 29. Ludwig SKJ, Smits NGE, Bremer MGEG, Nielen MWF (2012) Monitoring milk for antibodies against recombinant bovine somatotropin using a microsphere immunoassay-based biomarker approach. Food Control 26: 68–72. 11. Powrie JK, Bassett EE, Rosen T, Jørgensen JO, Napoli R, et al. (2007) Detection of growth hormone abuse in sport. Growth Horm IGF Res 17: 220–226. 12. Kniess A, Ziegler E, Kratzsch J, Thieme D, Mu¨ller R (2003) Potential parameters for the detection of hGH doping. Anal Bioanal Chem 376: 696–700. 30. Rochereau-Roulet S, Gaudin I, Che´reau S, Pre´vost S, Andre´-Fontaine G, et al. (2011) Development and validation of an enzyme-linked immunosorbent assay for the detection of circulating antibodies raised against growth hormone as a consequence of rbST treatment in cows. Anal Chim Acta 700: 189–193. 13. Longobardi S, Keay N, Ehrnborg C, Cittadini A, Rosen T, et al. (2000) Growth hormone (GH) effects on bone and collagen turnover in healthy adults and its potential as a marker of GH abuse in sports: A double blind, placebo-controlled study. J Clin Endocrinol Metab 85: 1505–1512. 31. Smits NGE, Ludwig SKJ, Van der Veer G, Bremer MGEG, Nielen MWF (2012) Serum biomarker profiling for detection of recombinant bovine somatotropin abuse using a multiplex flow cytometric immunoassay. Analyst: doi: 10.1039/c1032an35226e. 14. Dall R, Longobardi S, Ehrnborg C, Keay N, Rose´n T, et al. (2000) The effect of four weeks of supraphysiological growth hormone administration on the insulin- like growth factor axis in women and men J Clin Endocr Metab 85: 4193–4200 like growth factor axis in women and men. J Clin Endocr Metab 8 32. Kicman AT, Miell JP, Teale JD, Powrie J, Wood PJ, et al. (1997) Serum IGF-I and IGF binding proteins 2 and 3 as potential markers of doping with human GH. Clin Endocrinol 47: 43–50. 15. Erotokritou-Mulligan I, Eryl Bassett E, Cowan DA, Bartlett C, Milward P, et al. (2010) The use of growth hormone (GH)-dependent markers in the detection of GH abuse in sport: Physiological intra-individual variation of IGF-I, type 3 pro- collagen (P-III-P) and the GH-2000 detection score. Biomarker Assessment for rbST Abuse Detection Clin Endocrinol 72: 520– 526. 33. Eppard PJ, Rogan GJ, Boysen BG, Miller MA, Hintz RL, et al. (1992) Effect of high doses of a sustained-release bovine somatotropin on antibody formation in dairy cows. J Dairy Sci 75: 2959–2967. 16. Healy M-L, Dall R, Gibney J, Bassett E, Ehrnborg C, et al. (2005) Toward the development of a test for growth hormone (GH) abuse: A study of extreme physiological ranges of GH-dependent markers in 813 elite athletes in the postcompetition setting. J Clin Endocr Metab 90: 641–649. y J y 34. Zwickl CM, Smith HW, Tamura RN, Bick PH (1990) Somatotropin antibody formation in cows treated with a recombinant bovine somatotropin over two lactations. J Dairy Sci 73: 2888–2895. J y 35. Lee AJ, Hodges S, Eastell R (2000) Measurement of osteocalcin. Ann Clin Biochem 37: 432–446. 17. Sartorio A, Jubeau M, Agosti F, Marazzi N, Rigamonti A, et al. (2006) A follow- up of GH-dependent biomarkers during a 6-month period of the sporting season of male and female athletes. J Endocrinol Invest 29: 237–243. 36. Armstrong JD, Cohick WS, Harvey RW, Heimer EP, Campbell RM (1993) Effect of feed restriction on serum somatotropin, insulin-like growth factor-I- (IGF-I) and IGF binding proteins in cyclic heifers actively immunized against growth hormone releasing factor. Domest Anim Endocrin 10: 315–324. 18. Wallace JD, Cuneo RC, Lundberg PA, Rosen T, Jorgensen JOL, et al. (2000) Responses of markers of bone and collagen turnover to exercise, growth hormone (GH) administration, and GH withdrawal in trained adult males. J Clin Endocrinol Metab 85: 124–133. 37. Blum WF, Horn N, Kratzsch J, Jørgensen JO, Juul A, et al. (1993) Clinical studies of IGFBP-2 by radioimmunoassay. Growth Regulat 3: 100–104. 19. Castigliego L, Grifoni G, Rosati R, Iannone G, Armani A, et al. (2009) On the alterations in serum concentration of somatotropin and insuline-like growth factor 1 in lactating cows after the treatment with a little studied recombinant bovine somatotropin. Res Vet Sci 87: 29–35. 38. Friedrich N, Alte D, Vo¨lzke H, Spilcke-Liss E, Lu¨demann J, et al. (2008) Reference ranges of serum IGF-1 and IGFBP-3 levels in a general adult population: Results of the Study of Health in Pomerania (SHIP). Growth Horm IGF Res 18: 228–237. 39. Sato R, Onda K, Ochiai H, Iriki T, Yamazaki Y, et al. (2011) Serum osteocalcin in dairy cows: Age-related changes and periparturient variation. Res Vet Sci 91: 196–198. 20. Biomarker Assessment for rbST Abuse Detection Zhao X, McBride BW, Trouten-Radford LM, Golfman L, Burton JH (1994) Somatotropin and insulin-like growth factor-I concentrations in plasma and milk after daily or sustained-release exogenous somatotropin administrations. Domest Anim Endocrin 11: 209–216. 40. Kerr DE, Laarveld B, Fehr MI, Manns JG (1991) Profiles of serum IGF-I concentrations in calves from birth to eighteen months of age and in cows throughout the lactation cycle. Can J Anim Sci 71: 695–705. 21. Food and Drug Administration (1993) Animal Drugs, Feeds, and Related Products; Sterile Sometribove Zinc Suspension. 58 Federal Register 59946. 41. Sharma BK, Vandehaar MJ, Ames NK (1994) Expression of insulin-like growth factor-I in cows at different stages of lactation and in late lactation cows treated with somatotropin. J Dairy Sci 77: 2232–2241. 22. European Union (1999) Council Decision 1999/879/EC of 17 December 1999 concerning the placing on the market and administration of bovine somatotrophin (BST) and repealing Decision 90/218/EEC. Off J Eur Commun L 331: 71–72. 42. R Development Core Team (2011) R: A language and environment for statistical computing. 23. European Union (2002) Commission Decision 2002/657/EC of 12 August 2002 implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results. Off J Eur Commun L 221: 8–36. 43. Dimitriadou E, Hornik K, Leisch F, Meyer D, Weingessel A (2011) Misc functions of the department of statistics (e1071), TU Wien. 44. Efron B (1979) Bootstrap methods: Another look at the jackknife. Ann Stat 7: 1– 26. 24. Castigliego L, Iannone G, Grifoni G, Rosati R, Gianfaldoni D, et al. (2007) Natural and recombinant bovine somatotropin: immunodetection with a sandwich ELISA. J Dairy Res 74: 79–85. 45. Nelson AE, Ho KKY, Ghigo E, Lanfranco F, Strasburger CJ (2011) Detection of growth hormone doping in sport using growth hormone-responsive markers. In: Melmed S, editor. Hormone use and abuse by athletes: Springer US. pp. 139– 150. 25. Lucy MC, Byatt JC, Curran TL, Curran DF, Collier RJ (1994) Placental lactogen and somatotropin: hormone binding to the corpus luteum and effects on the growth and functions of the ovary in heifers. Biol Reprod 50: 1136–1144. 46. Di Luigi L, Rigamonti AE, Agosti F, Mencarelli M, Sgro` P, et al. (2009) Combined evaluation of resting IGF1, N-terminal propeptide of type III procollagen and C-terminal cross-linked telopeptide of type I collagen levels might be useful for detecting inappropriate GH administration in female athletes. Eur J Endocrinol 160: 753–758. 26. References 6. Chung L, Clifford D, Buckley M, Baxter RC (2006) Novel biomarkers of human growth hormone action from serum proteomic profiling using protein chip mass spectrometry. J Clin Endocrinol Metab 91: 671–677. 1. Teale P, Barton C, Driver PM, Kay RG (2009) Biomarkers: unrealized potential in sports doping analysis. Bioanalysis 1: 1103–1118. 7. Ding J, List EO, Okada S, Kopchick JJ (2009) Perspective: Proteomic approach to detect biomarkers of human growth hormone. Growth Horm IGF Res 19: 399–407. 4. Smits NGE, Bremer MGEG, Ludwig SKJ, Nielen MWF (2012) Development of a flow cytometric immunoassay for recombinant bovine somatotropin-induced antibodies in serum of dairy cows. Drug Test Analysis 4: 362–367. 5. Chung L, Baxter RC (2009) Detection of growth hormone responsive proteins using SELDI–TOF mass spectrometry. Growth Horm IGF Res 19: 383–387. December 2012 \| Volume 7 \| Issue 12 \| e52917 PLOS ONE \| www.plosone.org 12 Biomarker Assessment for rbST Abuse Detection Biomarker Assessment for rbST Abuse Detection Le Breton MH, Rochereau-Roulet S, Chereau S, Pinel G, Delatour T, et al. (2010) Identification of cows treated with recombinant bovine somatotropin. J Agr Food Chem 58: 729–733. 27. Le Breton MH, Rochereau-Roulet S, Pinel G, Cesbron N, Le Bizec B (2009) Elimination kinetic of recombinant somatotropin in bovine. Anal Chim Acta 637: 121–127. 47. McGuire MA, Vicini JL, Bauman DE, Veenhuizen JJ (1992) Insulin-like growth factors and binding proteins in ruminants and their nutritional regulation. J Anim Sci 70: 2901–2910. December 2012 \| Volume 7 \| Issue 12 \| e52917 PLOS ONE \| www.plosone.org 13
https://openalex.org/W3157390405	https://iris.unimore.it/bitstream/11380/1245621/2/life-11-00406.pdf	English	null	The Anti-Ageing and Whitening Potential of a Cosmetic Serum Containing 3-O-ethyl-l-ascorbic Acid	Life	2,021	cc-by	10,275	Keywords: melanin; skin ageing; reconstructed human epidermis; human keratinocytes; ﬁbroblasts; photoprotection; vitamin C Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. life life life Article The Anti-Ageing and Whitening Potential of a Cosmetic Serum Containing 3-O-ethyl-L-ascorbic Acid Nicola Zerbinati 1, Sabrina Sommatis 2, Cristina Maccario 2, Serena Di Francesco 2, Maria Chiara Capillo 2, Raffaele Rauso 3 , Martha Herrera 4, Pier Luca Bencini 5, Stefania Guida 6 and Roberto Mocchi 2,* 1 Department of Medicine and Surgery, University of Insubria, 21100 Varese, Italy; nicola.zerbinati@uninsubria.it 1 Department of Medicine and Surgery, University of Insubria, 21100 Varese, Italy; nicola.zerbinati@uninsubria.it 2 Department of Molecular Medicine, UB-CARE S.r.l.-Spin-Off University of Pavia, 27100 Pavia, Italy; sabrina.sommatis@ub-careitaly.it (S.S.); cristina.maccario@ub-careitaly.it (C.M.); serena.difrancesco@ub-careitaly.it (S.D.F.); mariachiara.capillo@ub-careitaly.it (M.C.C.) serena.difrancesco@ub careitaly.it (S.D.F.); mariachiara.capillo@ub careitaly.it (M.C.C.) 3 Maxillofacial Surgery Unit, University of Campania “Luigi Vanvitelli”, 80138 Naples, Italy; raffaele.rauso@unicampania.it g y , , ; 5 Istituto di Chirurgia e Laser-Chirurgia in Dermatologia (I.C.L.I.D.), 20121 Milan, Italy; pl.bencini@iclid.it 6 Dermatology Unit, Department of Surgical, Medical, Dental and Morphological Sciences Related to Transplant, Oncology and Regenerative Medicine, University of Modena and Reggio Emilia, 41124 Modena, Italy; stefania.guida@unimore.it * Correspondence: roberto.mocchi@ub-careitaly.it Abstract: Skin ageing has many manifestations such as wrinkles, dryness, hyperpigmentation, and uneven skin tone. Extrinsic and intrinsic factors, especially solar ultraviolet light (UVB), contribute to skin ageing; its main features are brown spots, alterations in melanin pigmentation, and a decrease in collagen and hyaluronic acid linked to oxidative stress. Several studies showed that topical products containing ingredients with antioxidant activity can reduce oxidative damage; to provide a maximum anti-ageing effect to the skin, topical products can combine various ingredients. C-SHOT SERUM contains a combination of two molecules with a proven anti-ageing activity: a high percentage (30%) of a more stable vitamin C derivative, 3-O-ethyl-L-ascorbic acid, and lactic acid (1%). The product showed a high biocompatibility, assessed through an MTT assay on keratinocytes and on Reconstructed Human Epidermis (RHE, SkinEthic); the anti-ageing activity was demonstrated on human dermal ﬁbroblasts and keratinocytes by a statistically signiﬁcant increase in collagen production and a reduction of a UVB-induced DNA damage marker (γ-H2AX histone), indicating DNA protection. Moreover, a depigmenting activity, shown by a highly signiﬁcant decrease in melanin content on treated Reconstructed Human Pigmented Epidermis (RHPE), was assessed. According to the data of our study, the tested product contrasts the effect of skin ageing and irregular pigmentation due to the physiological decline of the skin. Citation: Zerbinati, N.; Sommatis, S.; Maccario, C.; Di Francesco, S.; Capillo, M.C.; Rauso, R.; Herrera, M.; Bencini, P.L.; Guida, S.; Mocchi, R. The Anti-Ageing and Whitening Potential of a Cosmetic Serum Containing 3-O-ethyl-L-ascorbic Acid. Article The Anti-Ageing and Whitening Potential of a Cosmetic Serum Containing 3-O-ethyl-L-ascorbic Acid Life 2021, 11, 406. https://doi.org/ 10.3390/life11050406 Academic Editor: Rolland Gyulai Received: 15 April 2021 Accepted: 26 April 2021 Published: 29 April 2021 1. Introduction g y C-SHOT SERUM is a cosmetic product (Matex Lab SpA, Via Carlo Urbani 2, ang Enrico Fermi, Brindisi, Italy) containing a high concentration of vitamin C (30%), in particular the much more stable derivative 3-O-ethyl ascorbic acid (3OAA) and lactic acid (1%), which play a key role in many basic molecular anti-ageing-related processes like collagen stimulation, melanogenesis inhibition, and protection against photoageing and UV damage. The 3-O-ethyl ascorbic acid has an ethyl group at the third carbon position and this structural modiﬁcation protects the 3-OH group from ionization and, thus, the molecule from oxidation [10]. Vitamin C has shown its anti-ageing activity already from use at 5% in cosmetic for- mulations; furthermore, to be effective, topical application requires pH levels below four to allow the delivery across the epidermal layer to be mediated by speciﬁc sodium-dependent vitamin C transporters (SVCT1 and SVCT2). Both transporters are hydrophobic membrane proteins that co-transport sodium, driving the uptake of vitamin C; SVCT1 is a low afﬁnity transporter closed to epithelial cells in the small intestine and kidney, while SVCT2 is a high afﬁnity transporter ubiquitous in the cells of the body. The vitamin C delivery is ﬁnely regulated by the availability of the SVCTs on the plasma membrane and, following its uptake, it is involved in many molecular processes. Several evidence demonstrate the role of vitamin C as a co-factor for the regulatory proline and lysine hydroxylases that stabilize the collagen molecule tertiary structure through the hypoxia-inducible factor (HIF)-1 activation, controlling the gene expressions involved in collagen synthesis (Col-1 and Col-3) and tissue remodeling as collagenases [11–16]. Vitamin C is a water-soluble antioxidant that protects the skin from reactive oxygen species (ROS) by solar radiation or environmental factors such as pollution [14]. Other studies evidenced vitamin C’s ability to interfere with the action of tyrosinase, the rate-limiting enzyme in melanogenesis. Particularly, it reduces the ortho-quinones obtained from the dihydroxyphenylalanine (DOPA) oxidation derived from tyrosinase catalysis. Indeed, it has been used in cosmetic products also to make the skin more radiant and brighter because it is able to act against the hyperpigmentation [15–17]. However, the physicochemical properties of the compound, such as rapid oxidation in solutions upon exposure to air, the melting point (190–192 ◦C) partition coefﬁcient (log P(o/w) = −1.85) and dissociation constant (pKa = 4.25) reduce its delivery across the skin and its molecular efﬁcacy. 1. Introduction Skin ageing is characterized by intrinsic and extrinsic factors that cause impairment of skin integrity, structural degradation and alterations [1]. Intrinsic or chronological ageing is an uncontrollable process caused by a physiological decline due to passing years and is characterized by smooth, dry, pale, and ﬁnely wrinkled skin [2]. Extrinsic ageing is generally caused by environmental factors such as ultraviolet (UV) radiation from the sun (major cause of skin damage as photoageing), alcohol intake, poor nutrition, overeating, and environmental pollution [3,4]. UVA constitute 95% of the sun’s rays that reach the Earth, but exposure to UVB, the remaining 5%, has a much greater biological impact on the skin than exposure to UVA when comparing similar irradiation doses. A major part of UVB is Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). https://www.mdpi.com/journal/life Life 2021, 11, 406. https://doi.org/10.3390/life11050406 Life 2021, 11, 406 2 of 13 adsorbed in the epidermis by DNA, aromatic amino acids of proteins, nicotinamide adenine dinucleotide (NADH), and nicotinamide adenine dinucleotide phosphate (NADPH), and does not penetrate deeper into the skin [5]. Regarding aged skin, brown spots also are found often as a consequence of alterations in the melanin pigmentation, particularly visible on the face and back of the hands [6]. Photoageing leads to a drastic decrease in skin hydration and total levels of collagen, elastin, and hyaluronic acid, causing a series of intracellular responses such as the production of oxidative stress. Particularly, UVB also can directly affect the DNA of skin cells through the formation of cyclobutane pyrimidine dimers (CPDs) and another dipyrimidine lesion, the (6–4) photoproducts [7]. Constant use of skin care products with added active molecules or antioxidants from an early age can lead to signiﬁcant beneﬁts in counteracting the effects of skin damage. The antioxidants, such as vitamins, have a preventative effect on ageing and their use, in the cosmetic ﬁeld, is due to their ability to easily penetrate the skin thanks to their small molecular weight, their safeness, and their efﬁcacy often found in the short term [8,9]. 2. Materials and Methods 2.1. International Nomenclature of Cosmetic Ingredients (INCI) The functional classiﬁcation of ingredients contained in the tested serum is summa- rized in Table 1. Table 1. Functional classiﬁcation of ingredients contained in the serum used in the study. Table 1. Functional classiﬁcation of ingredients contained in the serum used in the study. Function Ingredient Skin conditioning 3-O-ethyl ascorbic acid, lactic acid Humectant propanediol, sodium lactate Buffering & Keratolytic sodium lactate Preservative phenoxyethanol, imidazolidinyl urea Thickener hydroxyethylcellulose Chelating agent disodium edta Table 1. Functional classiﬁcation of ingredients contained in the serum used in the study. Function Ingredient Skin conditioning 3-O-ethyl ascorbic acid, lactic acid Humectant propanediol, sodium lactate Buffering & Keratolytic sodium lactate Preservative phenoxyethanol, imidazolidinyl urea Thickener hydroxyethylcellulose Chelating agent disodium edta Function Skin conditioning Humectant Buffering & Keratolytic Preservative Thickener Chelating agent 1. Introduction To overcome these disadvantages and ensure the molecular efﬁcacy, several vitamin C derivatives were synthetized and evaluated for their potential as pro- ascorbic acid derivatives. Among these, 3OAA is more lipophilic than the original, hence exhibiting an improved permeation into the skin, demonstrating an efﬁcient transdermal activity, and allowing the molecular mechanisms associated to the original compound (vitamin C) previously described. Overcoming the absorption limits due to the topical application, the choice of using 3OAA within cosmetic formulations has proved to be a valid means to improve the effectiveness of the original functional compound [18,19]. Life 2021, 11, 406 3 of 13 Besides vitamin C and its derivatives, topical α-hydroxy acids (α-HA) also showed clinical efﬁcacy in the anti-ageing skin treatment. The Cosmetic Ingredient Review (CIR) Panel assessed the evidence that α-HA ingredients are non-toxic, mutagenic, or carcino- genic recommending a maximum concentration of 10% with a pH above 3.5 in a cosmetic formulation. Among α-HA, glycolic and lactic acid have been shown to be effective in gly- cosaminoglycan (GAG) and collagen stimulation but, also, in the treatment of various types of hyper-pigmentary lesions, such as photoageing, solar lentigines, and post-inﬂammatory hyperpigmentation. Speciﬁcally, the proposed molecular mechanism underlying the skin pigmentation resulting from UV exposure is the increased epidermal turnover generating skin exfoliation/peeling and the suppression of melanin synthesis by inhibiting tyrosinase activity resulting in shiny and taut skin [19–21]. During this study, we investigated the potential of a cosmetic product containing concentrated 3-O-ethyl ascorbic acid at 30% in combination with lactic acid at 1%, to contrast the effect of skin ageing and irregular pigmentation due to the normal physiological decline of the skin on both in vitro cell cultures and 3D reconstructed tissue models. Cell viability (%) = (OD570nm test product/OD570nm control) × 100 (1) Reduction of cell viability by more than 30% is considered a cytotoxic effect. 2.2. Cell Cultures Human keratinocytes (HaCaT, BS code CL 168), were provided by I.Z.L.E.R. (In- stitute Zooproﬁlattico della Lombardia e Emilia Romagna) while human dermal ﬁbrob- lasts (NHDF-Ad−Human Dermal Fibroblasts, Adult, code CC-2511) were provided by Lonza (Basel, Switzerland). Cell lines were grown in a complete medium constituted by Dulbecco’s Modiﬁed Eagle’s Medium (DMEM, Biowest, Nuaillé, France) High Glucose with 10% fetal bovine serum (FBS, Gibco-Fisher Scientiﬁc, Waltham, MA, USA), 1% of L-glutamine (Capricorn Scientiﬁc, Ebsdorfergrund, Germany), and antibiotics 100 U/mL penicillin and 100 µg/mL streptomycin (Capricorn Scientiﬁc, Ebsdorfergrund, Germany), in conditions of complete sterility and maintained in incubation at 37 ◦C with a 5% carbon dioxide (CO2) atmosphere. p The Reconstructed Human Epidermis (RHE) model (EPISKIN Laboratories, Lyon, France) was supplied by SkinEthic™laboratories; it is a reconstructed tissue from normal human keratinocytes grown for 17 days in a chemically deﬁned medium. The tissue model consists of a fully differentiated epidermis including a basal cell layer, stratum spinosum, stratum granulosum, and stratum corneum on a 0.5 cm2 surface of inert polycarbonate ﬁlter at the air–liquid interface [22–24]. The SkinEthicTM Reconstructed Human Pigmented Epidermis (RHPE) model (EPISKIN Laboratories, Lyon, France) consists of a stratiﬁed and differentiated epidermal layer con- taining human melanocytes and keratinocytes. The cell model used is made of inert polycarbonate inserts of 0.5 cm2 on which normal human cells differentiate to form a cellular multilayer. y Regarding both reconstructed models (RHE and RHPE), two speciﬁc media provided by EPISKIN were used: maintenance and growth media (Table 2). Life 2021, 11, 406 4 of 13 Table 2. Technical data and batch information of maintenance and growth media. Maintenance Media Growth Media RHE (19-RHE-036) 19SMM010 19SGM026 RHPE (19-RHPE-019) 19SMM019 19SGMRHPE013 Table 2. Technical data and batch information of maintenance and growth media. 2.3. Biocompatibility 2.3.1. Cell Viability (MTT Test) Cell viability was evaluated using the 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetra- zolium bromide assay, as previously described [25]. Brieﬂy, keratinocytes were homoge- neously seeded in 96-well plates at a density of 1.5 × 104 and incubated at 37 ◦C, with a 5% CO2 humidiﬁed atmosphere. After 24 h, cells were treated with the cosmetic prod- uct, starting at 40 mg/mL, following serial dilution (1:2) in cell medium (tested range 0.313–40 mg/mL). Untreated cells were used as the control (Ctrl). The test was conducted in three replicates for each dilution. After 24 h of treatment, cells were incubated with 1 mg/mL MTT solution (Merck, Darmstadt, Germany) at 37 ◦C for 2 h. After removing the medium from each well, isopropanol was added to dissolve formazan crystals and the absorbance was read at a 570 nm wavelength using a microplate reader (MultiSkan, Thermo Scientiﬁc, Waltham, MA, USA). Cell survival was calculated measuring the dif- ference in optical density (OD) of the tested cosmetic product with respect to the control (Equation (1)). (1) Cell viability (%) = (OD570nm test product/OD570nm control) × 100 (1) 2.3.2. Skin Irritation Test on 3D Model The day of receipt, RHE inserts were placed in the maintenance medium (6-well plate) under sterile conditions and stored in an incubator at 37 ◦C, at 5% CO2 overnight. After this pre-incubation step, the product was topically applied on the surface of the epithelium insert for 42 min at a 32 µL/cm2 concentration, thus entering into contact with the in vitro reconstructed epithelium directly. To parallel, the other inserts were treated with a negative control Dulbecco Phosphate Buffer Solution (DPBS, Merck, Darmstadt, Germany) and a positive control consisting of a 5% (w/v in water) solution of Sodium Dodecyl Sulfate (SDS, Merck, Darmstadt, Germany), representing the irritating treatment after a short exposure. Occurring at the end of the exposure step, RHE inserts were rinsed with DPBS and trans- ferred in 6-well plates for incubation at 37 ◦C, 5% CO2, 95% humidiﬁed atmosphere for 42 h. Tissue viability was assessed using an MTT test: tissues were incubated with an MTT solution (1 mg/mL) for 3 h, then the extraction with isopropanol was performed. After extraction, the OD of the samples was quantiﬁed by spectrophotometry at a 570 nm wavelength. Isopropanol was used as a blank. After subtracting the blank OD from all raw data, mean OD values and standard deviations (SD) were calculated. Viability of the epidermis treated with the product was calculated as a ratio of the corrected optical densities of the sample over the negative control (untreated sample) (Equation (2)). (2) Cell viability (%) = (OD570nm test product/OD570nm negative control) × 100 (2) Cell viability (%) = (OD570nm test product/OD570nm negative control) × 100 (2) When the cell viability value was ≤50%, the product was classiﬁed as an irritant. To better evaluate the skin irritation effect of the cosmetic product, as well as the viability resulting from the direct contact of the product on the inserts, the levels of inter- leukin (IL)-1α released after treatment also were measured after 42 h of recovery time by an Enzyme-linked Immunosorbent Assay (ELISA) kit (Diaclone, Besançon cedex, France) Life 2021, 11, 406 5 of 13 following the manufacturer’s protocol. Brieﬂy, supernatants were collected and used for the coating of a speciﬁcally pre-treated 96-well ELISA plate provided in the kit. The standards were used to perform the standard curve (3.9–250 pg/mL). Samples, blank, and standards were added to each well in duplicate and the assay was performed according to the supplier’s instructions. 2.4.2. Collagen Synthesis NHDF cells were homogeneously seeded in 24-well plates at a density of 6 × 104 cells per well and incubated at 37 ◦C, with a 5% CO2 humidiﬁed atmosphere. After 24 h, the medium was replaced by one with a low FBS concentration (0.5%) as recommended in the kit. After 24 h, the two concentrations of the product (1.25 and 2.5 mg/mL), demonstrated to be non-cytotoxic and having the best solubility from the preliminary MTT so were chosen to be used in this assay (data not shown). Occurring at the end of the treatment, the supernatant of each sample was collected into a sterile tube and incubated overnight with the Isolation and Concentration Reagent (polyethylene glycol TRIS-HCl buffer, pH 7.6). The day after, the measurement of collagen synthesis was performed using a commercial kit (Sircol, Soluble Collagen assay kit, Biocolor Life Science Assays, Carrickfergus, UK), according to the manufacturer’s instructions, and the absorbance of samples was measured at 555 nm. The concentration of collagen then was calculated using a standard curve. 2.4.1. Photoprotective Effect HaCaT cells were seeded (5 × 104) homogeneously in 22 × 22 mm slides placed in Petri dishes and subsequently treated with the product at the highest non-cytotoxic concentrations (5–10 mg/mL). After 24 h, treatment was removed and cells were exposed to ultraviolet (UV)B at 2.5 mJ/cm2 (CAMAG® UV Lamp 4, wavelength 302 nm, Muttenz, CH) as a dose capable to cause damage on a DNA level. The dose was measured us- ing a Spectroline DRC-100X digital radiometer (Spectronics Corporation Westbury, NY, USA). After irradiation, cells were maintained in a complete medium for a recovery time (24 h) at 37 ◦C before ﬁxation in methanol. The day after, samples were incubated with the anti-phospho-Histone γ-H2A.X (diluted 1:5000 in a DPBS buffer containing 1% BSA) primary antibody for 1 h. After washing, incubation with a secondary antibody diluted 1:200 was performed for 30 min; then, Hoechst 33258 staining to highlight the cells’ DNA followed. Images of ﬁxed cells were taken with a Nikon Eclipse E400 ﬂuorescence micro- scope equipped with a Canon Power Shot A590 IS digital camera. Regarding processing, approximately 200 total cells were counted for each slide (marked with Hoechst), and the positive ones for histone γ-H2AX were evaluated among these too. 2.5. Depigmenting Effect The day of receipt, RHPE inserts were placed in a maintenance medium (6-well plate) under sterile conditions, removing the excess of agar, and stored in an incubator at 37 ◦C, 5% CO2 overnight. After this pre-incubation step, the product was deposed daily on the surface of the stratum corneum of the tanned epidermal tissue (1 µL/insert) for four days, while the medium was replaced every day. DPBS was used as a negative control following the same protocol. Inserts were incubated at 37 ◦C, 5% CO2, 95% humidiﬁed atmosphere and, after 6 days (48 h after the last day of treatment), tissues were assessed for cell viability (MTT test) and melanin content. 2.3.2. Skin Irritation Test on 3D Model The absorbance was then measured at 450 nm using a mi- croplate reader (MultiSkan, Thermo Scientiﬁc, Waltham, MA, USA). Data were analyzed as mean ± SD. 2.5.2. Melanin Content by cutting out the polyca (Pe ki El e Waltha Regarding melanin content, epidermal tanned tissues were removed from the insert by cutting out the polycarbonate ﬁlter and then plunged into 400 µL of Solvable Solution (Perkin Elmer, Waltham, MA, USA). Then, samples were heated at 100 ◦C for 45 min, cen- trifuged, and melanin extract was measured by spectrophotometry at a 500 nm wavelength using a synthetic melanin calibration curve as reference. (Perkin Elmer, Waltham, MA, USA). Then, samples were heated at 100 C for 45 min, centrifuged, and melanin extract was measured by spectrophotometry at a 500 nm wave- length using a synthetic melanin calibration curve as reference. 2.6. Statistical Analysis All h i d d i i li h h i ifi d 2.6. Statistical Analysis p and the results were exp All the experiments were conducted in triplicate, except when otherwise speciﬁed, and the results were expressed as a mean ± SD. Statistical signiﬁcance was calculated using the One-Way analysis of variance (ANOVA) with Fisher’s Least Signiﬁcant Difference (LSD) multiple comparison as the post-test. Statistical analysis was performed using GraphPad Prism version 9.0.0. (GraphPad Software, Inc., San Diego, CA, USA). Only for the melanin amount evaluation was the statistical signiﬁcance calculated using the Student’s t-test. ing the One-Way analysis of variance (ANOVA) with Fisher’s Least Significant Difference (LSD) multiple comparison as the post-test. Statistical analysis was performed using GraphPad Prism version 9.0.0. (GraphPad Software, Inc., San Diego, CA, USA). Only for the melanin amount evaluation was the statistical significance calculated using the Stu- dent’s T-test. 3 R lt 2.5.1. Cytotoxicity on RHPE (MTT Test) RHPE tissue samples were incubated in an MTT solution (1 mg/mL) for 3 h at 37 ◦C, 5% CO2 and, subsequently, formazan crystals were dissolved in 1.5 mL isopropanol for 2 h at room temperature under agitation. Two thousand microliters were transferred Life 2021, 11, 406 6 of 13 n trip- u i g in triplicate into a 96-well plate and the concentration of formazan was quantiﬁed by measuring the OD through spectrophotometry at a 570 nm wavelength. Cellular viability was obtained by comparing the OD of the insert treated with the cosmetic product and the OD of the negative controls expressed as a percentage. tained by comparing the OD of the insert treated with the cosmetic product and the OD of the negative controls expressed as a percentage. 2.5.2. Melanin Content Regarding melanin content epidermal tanned tissues were removed from the insert 3.1. Evaluation of Cell Viability and Skin Irritation 3.1. Evaluation of Cell Viability and Skin Irritation The cosmetic product (Matex Lab SpA Via The cosmetic product (Matex Lab SpA, Via Carlo Urbani 2, ang Enrico Fermi, Brindisi, Italy), showed high biocompatibility on human skin. The cytotoxicity test was required to evaluate the effect of the product on the cellular viability and to identify that the appropriate concentrations did not cause a decrease in cell respiration exceeding 30%. Regarding Figure 1 and Table 3, results of the HaCaT viability after treatment with different concentrations of the tested product (24 h) are shown, expressed as a percentage compared to the control (untreated cells). The cosmetic product showed cytotoxic activity only after treatment with the highest tested concentration, equal to 40 mg/mL, providing a cell viability percentage of 52.18%; higher product concentrations were not tested, assuming a further decrease in cell viability. The cosmetic product (Matex Lab SpA, Via Carlo Urbani 2, ang Enrico Fermi, Brin disi, Italy), showed high biocompatibility on human skin. The cytotoxicity test was re- quired to evaluate the effect of the product on the cellular viability and to identify that the appropriate concentrations did not cause a decrease in cell respiration exceeding 30%. Re- garding Figure 1 and Table 3, results of the HaCaT viability after treatment with different concentrations of the tested product (24 h) are shown, expressed as a percentage com- pared to the control (untreated cells). The cosmetic product showed cytotoxic activity only after treatment with the highest tested concentration, equal to 40 mg/mL, providing a cell viability percentage of 52.18%; higher product concentrations were not tested, assuming a further decrease in cell viability. Fi u e 1 Cell iability afte t eat e t ith the odu t o HaCaT ell Cell e e t eated ith Figure 1. Cell viability after treatment with the product on HaCaT cells. Cells were treated with Figure 1. Cell viability after treatment with the product on HaCaT cells. Cells were treated with different concentrations (0.313–40 mg/mL) of the serum for 24 h, and cell viability was assessed hrough an MTT assay (n = 3, replicates = 3). Figure 1. Cell viability after treatment with the product on HaCaT cells. Cells were treated with different concentrations (0.313–40 mg/mL) of the serum for 24 h, and cell viability was assessed through an MTT assay (n = 3, replicates = 3). Figure 1. Cell viability after treatment with the product on HaCaT cells. 3.1. Evaluation of Cell Viability and Skin Irritation 3.1. Evaluation of Cell Viability and Skin Irritation The cosmetic product (Matex Lab SpA Via Reported in Figure 2 and in Table 4, the tested product did not show any irritant activity after a short direct exposure on the tested tissue with a viability greater than 50% (threshold used to identify a substance as not an irritant according to ISO 10993-10: 2010 [26]). Figure 2. Cell viability on Reconstructed Human Epidermis (RHE). The inserts were treated with negative (Ctrl −, DPBS), positive (Ctrl +, SDS 5%) controls, and with the serum for 42 min. After 42 h of recovery, cell viability was assessed through an MTT assay. Figure 2. Cell viability on Reconstructed Human Epidermis (RHE). The inserts were treated with negative (Ctrl −, DPBS), positive (Ctrl +, SDS 5%) controls, and with the serum for 42 min. After 42 h of recovery, cell viability was assessed through an MTT assay. Figure 2. Cell viability on Reconstructed Human Epidermis (RHE). The inserts were treated with negative (Ctrl −, DPBS), positive (Ctrl +, SDS 5%) controls, and with the serum for 42 min. After 42 h of recovery, cell viability was assessed through an MTT assay. Figure 2. Cell viability on Reconstructed Human Epidermis (RHE). The inserts were treated with negative (Ctrl −, DPBS), positive (Ctrl +, SDS 5%) controls, and with the serum for 42 min. After 42 h of recovery, cell viability was assessed through an MTT assay. Table 4. Percentage values in cell viability after treatment (42 min) and 42 h of recovery of the RHE inserts with negative control (Ctrl −), positive control (Ctrl +), and with the serum (n = 1, rep- licates = 4). Table 4. Percentage values in cell viability after treatment (42 min) and 42 h of recovery of the RHE inserts with negative control (Ctrl −), positive control (Ctrl +), and with the serum (n = 1, replicates = 4). 3.1. Evaluation of Cell Viability and Skin Irritation 3.1. Evaluation of Cell Viability and Skin Irritation The cosmetic product (Matex Lab SpA Via Cells were treated with different concentrations (0.313–40 mg/mL) of the serum for 24 h, and cell viability was assessed through an MTT assay (n = 3, replicates = 3). Figure 1. Cell viability after treatment with the product on HaCaT cells. Cells were treated with different concentrations (0.313–40 mg/mL) of the serum for 24 h, and cell viability was assessed through an MTT assay (n = 3, replicates = 3). Table 3. Percentage changes (mean ± SD) in cell viability after 24 h of treatment with the product (n = 3, replicates = 3). * approximate value. Table 3. Percentage changes (mean ± SD) in cell viability after 24 h of treatment with the product (n = 3, replicates = 3). * approximate value. Sample (mg/mL) 0 0.313 0.625 1.25 2.50 5 10 20 40 Cell Viability (%) 100 109.78 114.20 120.52 117.04 110.91 109.60 72.89 52.18 Standard Deviation (SD) 0 10.86 10.71 13.69 9.89 9.31 3.29 4.30 0.93 IC50 Value 0.93 * Life 2021, 11, 406 7 of 13 0.93 Skin irritation was investigated on 3D-reconstructed human epidermis, evaluating tissue viability by an MTT test and the interleukin (IL)-1α amount. Reported in Figure 2 and in Table 4, the tested product did not show any irritant activity after a short direct exposure on the tested tissue with a viability greater than 50% (threshold used to identify a substance as not an irritant according to ISO 10993-10:2010 [26]). Skin irritation was investigated on 3D-reconstructed human epidermis, evaluating tissue viability by an MTT test and the interleukin (IL)-1α amount. Reported in Figure 2 and in Table 4, the tested product did not show any irritant activity after a short direct exposure on the tested tissue with a viability greater than 50% (threshold used to identify a substance as not an irritant according to ISO 10993-10: 2010 [26]). Skin irritation was investigated on 3D-reconstructed human epidermis, evaluating tissue viability by an MTT test and the interleukin (IL)-1α amount. Reported in Figure 2 and in Table 4, the tested product did not show any irritant activity after a short direct exposure on the tested tissue with a viability greater than 50% (threshold used to identify a substance as not an irritant according to ISO 10993-10:2010 [26]). Skin irritation was investigated on 3D-reconstructed human epidermis, evaluating tissue viability by an MTT test and the interleukin (IL)-1α amount. 3.1. Evaluation of Cell Viability and Skin Irritation 3.1. Evaluation of Cell Viability and Skin Irritation The cosmetic product (Matex Lab SpA Via Sample Cell Viability (%) Ctrl − 100 ± 19.75 Ctrl + 3.17 ± 0.81 C-SHOT SERUM 114.34 ± 9.35 Sample Cell Viability (%) Ctrl − 100 ± 19.75 Ctrl + 3.17 ± 0.81 C-SHOT SERUM 114.34 ± 9.35 Sample Cell Viability (%) Ctrl − 100 ± 19.75 Ctrl + 3.17 ± 0.81 C-SHOT SERUM 114.34 ± 9.35 Sample Cell Viability (%) Ctrl − 100 ± 19.75 Ctrl + 3.17 ± 0.81 C-SHOT SERUM 114.34 ± 9.35 The amount of IL 1α released in the medium after treatment with the tested product was lower than the 9 International Unit (IU)/mL, threshold according to ISO 10993-10: 2010 [26] (Figure 3 and Table 5), indicating a good biocompatibility of the product. The amount of IL-1α released in the medium after treatment with the tested prod- uct was lower than the 9 International Unit (IU)/mL, threshold according to ISO 10993- 10:2010 [26] (Figure 3 and Table 5), indicating a good biocompatibility of the product. 8 of 14 Figure 3. Expression of IL-1α (IU/mL) after treatment and 42 h of recovery of the RHE inserts with the negative (Ctrl −), positive (Ctrl +) controls and the serum. (n = 1, replicates = 4) ** p values ≤ 0.0001 were considered statistically significant compared to Ctrl −. Table 5. Amount of IL-1α (IU/mL) after treatment (42 min) and 42 h of recovery of the RHE inserts Figure 3. Expression of IL-1α (IU/mL) after treatment and 42 h of recovery of the RHE inserts with the negative (Ctrl −), positive (Ctrl +) controls and the serum. (n = 1, replicates = 4) p values ≤0.0001 were considered statistically signiﬁcant compared to Ctrl −. Figure 3. Expression of IL-1α (IU/mL) after treatment and 42 h of recovery of the RHE inserts with the negative (Ctrl −), positive (Ctrl +) controls and the serum. (n = 1, replicates = 4) p values ≤ 0.0001 were considered statistically significant compared to Ctrl −. Figure 3. Expression of IL-1α (IU/mL) after treatment and 42 h of recovery of the RHE inserts with the negative (Ctrl −), positive (Ctrl +) controls and the serum. (n = 1, replicates = 4) ** p values ≤0.0001 were considered statistically signiﬁcant compared to Ctrl −. Life 2021, 11, 406 8 of 13 Table 5. 3.1. Evaluation of Cell Viability and Skin Irritation 3.1. Evaluation of Cell Viability and Skin Irritation The cosmetic product (Matex Lab SpA Via Amount of IL-1α (IU/mL) after treatment (42 min) and 42 h of recovery of the RHE inserts with negative control (Ctrl −), positive control (Ctrl +), and the serum (n = 1, replicates = 4). Sample IL-1α (IU/mL) Ctrl − 5.90 ± 0.65 Ctrl + 115.27 ± 0.83 C-SHOT SERUM 0.98 ± 0.17 Table 5. Amount of IL-1α (IU/mL) after treatment (42 min) and 42 h of recovery of the RHE inserts with negative control (Ctrl −), positive control (Ctrl +), and the serum (n = 1, replicates = 4). Table 5. Amount of IL-1α (IU/mL) after treatment (42 min) and 42 h of recovery of the RHE inserts with negative control (Ctrl −), positive control (Ctrl +), and the serum (n = 1, replicates = 4). 3.2. Protection Against Ultraviolet (UV)B-Induced DNA Damage 3.2. Protection Against Ultraviolet (UV)B-Induced DNA Damage The photo-protective activity of the product after UVB irradiation was investigated through the activation of phosphorylated H2AX histone (γ-H2AX). Considering this assay, the two highest concentrations with a cell viability greater than 80% were selected: 5 and 10 mg/mL. Figure 4 shows representative images of human keratinocytes (HaCaT) in ﬂuorescence microscopy (magniﬁcation 100×) after treatment with the tested product while, in Figure 5 and in Table 6, the quantiﬁcation of γ-H2AX positive cells is reported. 9 of 14 Figure 4. Representative fluorescence microscopy images (100×) of HaCaT cells after treatment with the product (5–10 mg/mL) and UVB 2.5 mJ/cm2 irradiation, the control sample (Ctrl, un- treated, and non-irradiated cells) and positive control sample (Ctrl+, untreated, and irradiated cells). DNA is marked with Hoechst 33258 (blue staining), while green staining evidenced the phosphorylated histone γH2AX. Figure 4. Representative ﬂuorescence microscopy images (100×) of HaCaT cells after treatment with the product (5–10 mg/mL) and UVB 2.5 mJ/cm2 irradiation, the control sample (Ctrl, untreated, and non-irradiated cells) and positive control sample (Ctrl+, untreated, and irradiated cells). DNA is marked with Hoechst 33258 (blue staining), while green staining evidenced the phosphorylated histone γH2AX. Table 6. γ-H2AX positive cells (± SD) expressed as a percentage in control and positive control samples and after treatment with the product (n = 2, replicates = 2). Figure 4. Representative fluorescence microscopy images (100×) of HaCaT cells after treatment with the product (5–10 mg/mL) and UVB 2.5 mJ/cm2 irradiation, the control sample (Ctrl, un- treated, and non-irradiated cells) and positive control sample (Ctrl+, untreated, and irradiated cells). DNA is marked with Hoechst 33258 (blue staining), while green staining evidenced the phosphorylated histone γH2AX. Figure 4. Representative ﬂuorescence microscopy images (100×) of HaCaT cells after treatment with the product (5–10 mg/mL) and UVB 2.5 mJ/cm2 irradiation, the control sample (Ctrl, untreated, and non-irradiated cells) and positive control sample (Ctrl+, untreated, and irradiated cells). DNA is marked with Hoechst 33258 (blue staining), while green staining evidenced the phosphorylated histone γH2AX. Table 6. γ-H2AX positive cells (± SD) expressed as a percentage in control and positive control samples and after treatment with the product (n = 2, replicates = 2). Sample γ-H2AX Positive Cells (%) Ctrl 6.15 ± 2.82 Ctrl + 65.70 ± 2.51 C-SHOT SERUM (5 mg/mL) 49.87 ± 5.36 C-SHOT SERUM (10 mg/mL) 52.83 ± 11.40 Table 6. 3.2. Protection Against Ultraviolet (UV)B-Induced DNA Damage γ-H2AX positive cells (± SD) expressed as a percentage in control and positive control samples and after treatment with the product (n = 2, replicates = 2). Sample γ-H2AX Positive Cells (%) Ctrl 6.15 ± 2.82 Ctrl + 65.70 ± 2.51 C-SHOT SERUM (5 mg/mL) 49.87 ± 5.36 C-SHOT SERUM (10 mg/mL) 52.83 ± 11.40 Table 6. γ-H2AX positive cells (± SD) expressed as a percentage in control and positive control samples and after treatment with the product (n = 2, replicates = 2). Table 6. γ-H2AX positive cells (± SD) expressed as a percentage in control and positive control samples and after treatment with the product (n = 2, replicates = 2). Sample γ-H2AX Positive Cells (%) Ctrl 6.15 ± 2.82 Ctrl + 65.70 ± 2.51 C-SHOT SERUM (5 mg/mL) 49.87 ± 5.36 C-SHOT SERUM (10 mg/mL) 52.83 ± 11.40 9 of 13 Life 2021, 11, 406 Figure 5. Quantification of γ-H2AX positive cells in control cells (Ctrl, untreated, and non-irradi- ated), positive control (Ctrl +, untreated, and irradiated cells) and in samples treated with the se- rum (5–10 mg/mL) (n = 2, replicates = 2) * p values ≤ 0.05 were considered statistically significant compared to positive control cells; **** p values ≤ 0.0001 were considered statistically significant compared to untreated cells. Figure 5. Quantiﬁcation of γ-H2AX positive cells in control cells (Ctrl, untreated, and non-irradiated), positive control (Ctrl +, untreated, and irradiated cells) and in samples treated with the serum (5–10 mg/mL) (n = 2, replicates = 2) * p values ≤0.05 were considered statistically signiﬁcant compared to positive control cells; **** p values ≤0.0001 were considered statistically signiﬁcant compared to untreated cells. 10 of 14 Figure 5. Quantification of γ-H2AX positive cells in control cells (Ctrl, untreated, and non-irradi- ated), positive control (Ctrl +, untreated, and irradiated cells) and in samples treated with the se- um (5–10 mg/mL) (n = 2, replicates = 2) * p values ≤ 0.05 were considered statistically significant ompared to positive control cells; **** p values ≤ 0.0001 were considered statistically significant ompared to untreated cells. Figure 5. 3.2. Protection Against Ultraviolet (UV)B-Induced DNA Damage Quantiﬁcation of γ-H2AX positive cells in control cells (Ctrl, untreated, and non-irradiated), positive control (Ctrl +, untreated, and irradiated cells) and in samples treated with the serum (5–10 mg/mL) (n = 2, replicates = 2) * p values ≤0.05 were considered statistically signiﬁcant compared to positive control cells; ** p values ≤0.0001 were considered statistically signiﬁcant compared to untreated cells. 10 of 14 Table 6. γ-H2AX positive cells (± SD) expressed as a percentage in control and positive control samples and after treatment with the product (n = 2, replicates = 2). Sample γ-H2AX Positive Cells (%) Ctrl 6.15 ± 2.82 Ctrl + 65.70 ± 2.51 C-SHOT SERUM (5 mg/mL) 49.87 ± 5.36 C-SHOT SERUM (10 mg/mL) 52.83 ± 11.40 The obtained data highlighted that UVB radiations are a good stimulus to induce the activation of H2AX histone: there was a significant difference between untreated and non- The obtained data highlighted that UVB radiations are a good stimulus to induce the activation of H2AX histone: there was a signiﬁcant difference between untreated and non-irradiated cells (Ctrl) and the positive control, untreated, and UV-irradiated cells (Ctrl +) (p value ** ≤0.0001). Particularly, in the control sample the percentage of γ- H2AX positive cells was 6.15%, while the positive control was equal to 65.70%, indicating the efﬁcacy of UVB stimulus. The lower tested concentration of the serum, 5 mg/mL, determined a statistically signiﬁcant reduction of γ-H2AX positive cells compared to the positive control (p value * ≤0.05), equal to 24%, suggesting a protective effect against UV-damage. There seems to be a reduction trend even with the higher concentration, 10 mg/mL, but it does not have a statistically signiﬁcant result. irradiated cells (Ctrl) and the positive control, untreated, and UV-irradiated cells (Ctrl +) (p value **** ≤ 0.0001). Particularly, in the control sample the percentage of γ-H2AX posi- tive cells was 6.15%, while the positive control was equal to 65.70%, indicating the efficacy of UVB stimulus. The lower tested concentration of the serum, 5 mg/mL, determined a statistically significant reduction of γ-H2AX positive cells compared to the positive con- trol (p value * ≤ 0.05), equal to 24%, suggesting a protective effect against UV-damage. There seems to be a reduction trend even with the higher concentration, 10 mg/mL, but it does not have a statistically significant result. 3.3. Modulation of Collagen Synthesis 3.3. Modulation of Collagen Synthesis C ll th i tifi Collagen synthesis was quantiﬁed to evaluate anti-ageing cosmetic potential. To perform this assay, human dermal ﬁbroblasts (NHDF) were used and 1.25 and 2.5 mg/mL were chosen as the highest non-cytotoxic concentrations with the best solubility in the medium. Considering Figure 6 and Table 7, graphic and data are reported as a concentration (µg/mL) with respect to untreated cells (Ctrl) on the NHDF cell line. A pre-treatment of 24 h with the tested serum induced a signiﬁcant increase in collagen production of more than 10 times with both tested concentrations, conﬁrming the excellent capacity of this cosmetic serum to contrast the natural and physiological skin-ageing effect. Collagen synthesis was quantified to evaluate anti-ageing cosmetic potential. To per- form this assay, human dermal fibroblasts (NHDF) were used and 1.25 and 2.5 mg/mL were chosen as the highest non-cytotoxic concentrations with the best solubility in the medium. Considering Figure 6 and Table 7, graphic and data are reported as a concentra- tion (μg/mL) with respect to untreated cells (Ctrl) on the NHDF cell line. A pre-treatment of 24 h with the tested serum induced a significant increase in collagen production of more than 10 times with both tested concentrations, confirming the excellent capacity of this cosmetic serum to contrast the natural and physiological skin-ageing effect. Figure 6. Collagen levels (μg/mL) in a sample treated with the product compared to the control (Ctrl, untreated cells) (n = 3, replicates = 2). * p values ≤ 0.001 were considered statistically signifi- cant compared to the control (untreated cells). Figure 6. Collagen levels (µg/mL) in a sample treated with the product compared to the control (Ctrl, untreated cells) (n = 3, replicates = 2). * p values ≤0.001 were considered statistically signiﬁcant compared to the control (untreated cells). Figure 6. Collagen levels (μg/mL) in a sample treated with the product compared to the control (Ctrl, untreated cells) (n = 3, replicates = 2). * p values ≤ 0.001 were considered statistically signifi- cant compared to the control (untreated cells). Figure 6. Collagen levels (µg/mL) in a sample treated with the product compared to the control (Ctrl, untreated cells) (n = 3, replicates = 2). * p values ≤0.001 were considered statistically signiﬁcant compared to the control (untreated cells). Life 2021, 11, 406 10 of 13 10 of 13 Table 7. 3.3. Modulation of Collagen Synthesis 3.3. Modulation of Collagen Synthesis C ll th i tifi Cell viability expressed as a percentage compared to negative control (Ctrl −, cells treated with DPBS) following daily treatment (4 days) of the RHPE inserts with the serum (n = 1, replicates = 4). with DPBS) following daily treatment (4 days) of the RHPE inserts with the serum (n = 1, repli- cates = 4). cates = 4). Melanin content was quantified and data are reported in Figure 8 and Table 8 as mel- anin content in µg/mL compared to the negative control (Ctrl −). The melanin amount in the control inserts was equal to 20.02 μg/mL, while in treated inserts it was reduced to 16 60 μg/mL with a highly significant decrease equal to 17 10% Statistical analysis was Melanin content was quantiﬁed and data are reported in Figure 8 and Table 8 as melanin content in µg/mL compared to the negative control (Ctrl −). The melanin amount in the control inserts was equal to 20.02 µg/mL, while in treated inserts it was reduced to 16.60 µg/mL, with a highly signiﬁcant decrease equal to 17.10%. Statistical analysis was calculated using the Student’s t-test. Melanin content was quantified and data are reported in Figure 8 and Table 8 as mel- anin content in µg/mL compared to the negative control (Ctrl −). The melanin amount in the control inserts was equal to 20.02 μg/mL, while in treated inserts it was reduced to 16.60 μg/mL, with a highly significant decrease equal to 17.10%. Statistical analysis was calculated using the Student’s t-test. 16.60 μg/mL, with a highly significant decrease equal to 17.10%. Statistical analysis was calculated using the Student’s t-test. Figure 8. Melanin content (μg/mL) evaluated after four days of serum treatment of the tanned epidermis inserts compared to the untreated ones (negative control Ctrl ) (n = 1 replicates = 4) Figure 8. Melanin content (μg/mL) evaluated after four days of serum treatment of the tanned epidermis inserts compared to the untreated ones (negative control, Ctrl −) (n = 1, replicates = 4). p values ≤ 0.01 were considered statistically significant compared to untreated cells. Figure 8. Melanin content (µg/mL) evaluated after four days of serum treatment of the tanned epidermis inserts compared to the untreated ones (negative control, Ctrl −) (n = 1, replicates = 4). p values ≤0.01 were considered statistically signiﬁcant compared to untreated cells. Figure 8. 3.3. Modulation of Collagen Synthesis 3.3. Modulation of Collagen Synthesis C ll th i tifi Collagen levels expressed as a concentration (± SD) measured after treatment with tested product (n = 3, replicates = 2). Sample Collagen Levels (µg/mL) Ctrl 0.08 ± 0.01 C-SHOT SERUM (1.25 mg/mL) 0.99 ± 0.09 C-SHOT SERUM (2.5 mg/mL) 1.24 ± 0.12 0.08 ± 0.01 0.99 ± 0.09 1.24 ± 0.12 3.4. Evaluation of Melanin Content on Reconstructed Human Pigmented Epidermis (RHPE) The depigmenting activity was tested on the RHPE model on which the viability and content of extracted melanin, after repeated treatment of 4 days with the tested product, was investigated. Considering the obtained results by the analysis of tissue viability, the serum did not show any negative effect on viability, with a percentage greater than 80% compared to the negative control (Ctrl −), treated with Dulbecco Phosphate Buffer Saline (DPBS) (Figure 7). Even in the case with the RHE, 50% was the threshold above which viability was acceptable, according to ISO 10993-10:2010 [26]. 11 of 14 11 of 14 Figure 7. Cell viability expressed as a percentage compared to negative control (Ctrl −, cells treated with DPBS) following daily treatment (4 days) of the RHPE inserts with the serum (n = 1, repli- Figure 7. Cell viability expressed as a percentage compared to negative control (Ctrl −, cells treated with DPBS) following daily treatment (4 days) of the RHPE inserts with the serum (n = 1, replicates = 4). Figure 7. Cell viability expressed as a percentage compared to negative control (Ctrl −, cells treated with DPBS) following daily treatment (4 days) of the RHPE inserts with the serum (n = 1, repli- cates = 4). Figure 7. Cell viability expressed as a percentage compared to negative control (Ctrl −, cells treated with DPBS) following daily treatment (4 days) of the RHPE inserts with the serum (n = 1, repli- Figure 7. Cell viability expressed as a percentage compared to negative control (Ctrl −, cells treated with DPBS) following daily treatment (4 days) of the RHPE inserts with the serum (n = 1, replicates = 4). with DPBS) following daily treatment (4 days) of the RHPE inserts with the serum (n = 1, repli- cates = 4). Figure 7. Cell viability expressed as a percentage compared to negative control (Ctrl −, cells treated with DPBS) following daily treatment (4 days) of the RHPE inserts with the serum (n = 1, repli- Figure 7. 4. Discussion Improvement in life expectancy results in increased functional cosmetic production, with the purpose of preventing the signs of ageing and keeping a healthy appearance. Skin-ageing is associated with deep wrinkles, loss of skin tone and elasticity, irregular pigmentation, and reduction in the thickness of the epidermis and dermis. Taking a molecular point of view, ageing is due to cumulative internal and external, structural and physiological alterations via genetics, light radiation exposure, marked loss of ﬁbrillin- positive structures, and a reduced content of collagen ﬁbers. Fibroblasts, which are the typical cells of the dermis, provide tensile strength and elasticity through the production and secretion of various components of the extracellular matrix (ECM). Collagen is one of these main components and is also the most abundant protein within the body, able to provide not only tensile strength but also cell adhesion and migration. The molecular knowledge of these biological events allows the formulation of functional innovative cosmetic products active on these, promoting an anti-ageing effect. p p g g g During this study, we investigated the cosmetic C-SHOT SERUM’s (Matex Lab SpA, Via Carlo Urbani 2, ang Enrico Fermi, Brindisi, Italy) ability to prevent skin-ageing through in vitro tests and 3D cell culture systems. Most in vitro models are based on a monolayer of one kind of cell line. This system shows some disadvantages because, in the in vivo environment, different cell types interact with each other within the extracellular matrix (ECM). It may be difﬁcult to assess the actual activity of a substance only based on a cell monolayer without considering the complexity of a biological model such as the human skin. However, combining different assays can improve the effectiveness of in vitro models, and they could be a good and cost-efﬁcient starting point. Moreover, to support the 2D model analyses performed, we decided to test the product also using an innovative system like 3D cell cultures, which represent a much more accurate microenvironment allowing us to mimic more closely the in vivo skin tissue [27,28]. The serum contains a high concentration of vitamin C, in particular its derivative 3-O-ethyl-L-ascorbic acid (3OAA) (30%) and lactic acid (1%), with a pH equal to 3.82 that allows good penetration and efﬁcacy of these functional ingredients [17,20]. 3.3. Modulation of Collagen Synthesis 3.3. Modulation of Collagen Synthesis C ll th i tifi Melanin content (μg/mL) evaluated after four days of serum treatment of the tanned Figure 8. Melanin content (μg/mL) evaluated after four days of serum treatment of the tanned epidermis inserts compared to the untreated ones (negative control, Ctrl −) (n = 1, replicates = 4). p values ≤ 0.01 were considered statistically significant compared to untreated cells. Figure 8. Melanin content (µg/mL) evaluated after four days of serum treatment of the tanned epidermis inserts compared to the untreated ones (negative control, Ctrl −) (n = 1, replicates = 4). ** p values ≤0.01 were considered statistically signiﬁcant compared to untreated cells. Life 2021, 11, 406 11 of 13 11 of 13 Table 8. Melanin quantification, expressed as µg/mL and percentage, after treatment (daily treatment for four days) of the RHPE inserts with the tested product. Negative control (Ctrl −) (n = 1, replicates = 4). Sample Melanin Content (µg/mL) Melanin Content (%) Reduction (%) Ctrl − 20.02 ± 1.43 100.00 ± 8.03 - C-SHOT SERUM 16.60 ± 1.92 82.89 ± 9.57 −17.10 4 Discussion Table 8. Melanin quantification, expressed as µg/mL and percentage, after treatment (daily treatment for four days) of the RHPE inserts with the tested product. Negative control (Ctrl −) (n = 1, replicates = 4). Sample Melanin Content (µg/mL) Melanin Content (%) Reduction (%) Ctrl − 20.02 ± 1.43 100.00 ± 8.03 - C-SHOT SERUM 16.60 ± 1.92 82.89 ± 9.57 −17.10 4. Discussion Since, with ageing, the pH of the skin increases to around 6.0, skin care products with a low pH in the range of 3.5–4.0 normalizes the epidermal barrier function [29]. Biocompatibility is the ﬁrst property of a cosmetic that must be evaluated. During our study, it was assessed through a conventional in vitro assay (MTT test) and with a 3D model to study the skin irritation potential on a model closer to in vivo skin tissue. The serum showed an excellent biocompatibility, the starting point for evaluating further beneﬁcial effects of the formulation. Vitamin C, contained in the product as a functional ingredient, is an antioxidant biomolecule that protects skin cells from oxidative damage after exposure to ultraviolet (UV) rays; it shows a proven skin anti-ageing effect for its ability to induce the production of collagen (Col)-1 and Col-3 (enhancer of collagen production), and the inhibition the collagenase 1 [13,16]. Extrinsic factors, especially UV radiation, are responsible for most age-related changes in skin appearance, producing free radicals, degrading collagen and elastin ﬁbers, and inducing DNA damage. Therefore, the photo-protective activity from UVB rays and collagen synthesis stimulation were analyzed. The effect of UVB rays on the DNA damage of skin cells was evaluated in association with a pre-treatment of the product to investigate a photoprotective effect. The phosphorylation of histone H2AX (γ-H2AX) is a direct index Life 2021, 11, 406 12 of 13 12 of 13 of DNA molecular lesion [30] and it plays an important role in the DNA repair process. Therefore, damage to the keratinocytes’ DNA was determined by an immunoﬂuorescence analysis of γ-H2AX positive cells. The results showed a potential photoprotective effect on skin cells using the lower tested concentration of 5 mg/mL: fewer γ-H2AX positive cells than the untreated control were counted through ﬂuorescence microscopy, with a reduction in the percentage equal to 24%. The collagen assay showed a promising anti-ageing effect of the cosmetic product. The ability to promote collagen production was evaluated and shown after 24 h of treatment, with both tested concentrations and a dose-dependent increase in collagen content equal to 1078% and 1115%, respectively, compared to untreated cells. g q p y p Along with wrinkles, pigmentary changes are associated with premature photo-ageing and are its evident cutaneous manifestations. 4. Discussion The whitening power of this serum, enriched with 30% of 3-O-ethyl-L-ascorbic acid in combination with 1% of lactic acid, were evaluated on reconstructed human pigmented epidermis (EPISKIN Laboratories, France). After repeated treatment for 4 days, there was a statistical reduction in the melanin content of 15.52% compared to a reference negative control. These results demonstrate another beneﬁcial effect of this cosmetic product in association with its high biocompatibility, photoprotection, and collagen enhancing production, making the formulation useful to prevent skin modiﬁcations, ageing-related, overall. p g g To conclude, our study represents a promising starting point for the evaluation of C-SHOT SERUM effectiveness using different analytical approaches based on monolayer and 3D cell models. Nevertheless, further in vivo investigations will be useful for a more suitable and careful evaluation of the biological serum activity to obtain a predictive index of the in vitro data respecting the in vivo real effect of the serum. Author Contributions: Conceptualization, S.S., C.M. and R.M.; methodology, S.S., C.M. and R.M.; formal analysis, M.C.C. and S.D.F.; investigation, S.S., S.D.F. and M.C.C.; data curation, S.S., C.M. and R.M.; writing—original draft preparation, S.D.F. and M.C.C.; writing—review and editing, S.S. and R.M.; visualization, R.R., M.H., P.L.B. and S.G.; supervision, N.Z.; project administration, N.Z. and R.M. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by Matex Lab S.p.A., Via C. Urbani 2, ang. Via E. F, Brindisi, Italy. Funding: This research was funded by Matex Lab S.p.A., Via C. Urbani 2, ang. Via E. F, Brindisi, Italy. Institutional Review Board Statement: Not applicable Funding: This research was funded by Matex Lab S.p.A., Via C. Urbani 2, ang. Via E. F, Brindisi, Italy. Institutional Review Board Statement: Not applicable. Institutional Review Board Statement: Not applicable. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Data are included in the text; raw data are available from the corre- sponding author. Data Availability Statement: Data are included in the text; raw data are available from the corre- sponding author. Conﬂicts of Interest: N.Z. is the Scientiﬁc Director of Matex Lab. The authors declare no other conﬂicts of interest. Conﬂicts of Interest: N.Z. is the Scientiﬁc Director of Matex Lab. The authors declare no other conﬂicts of interest. 6. Nisticò, S.P.; Tolone, M.; Zingoni, T.; Tamburi, F.; Scali, E.; Bennardo, L.; Cannarozzo, G. A New 675 nm Laser Device in the Treatment of Melasma: Results of a Prospective Observational Study. Photomed. Laser Surg. 2020, 38, 1–5. [CrossRef] 8. Vermeer, B.J.; Gilchrest, B.A. A proposal for rational deﬁnition, evaluation, and regulation. Arch Dermatol. 1996, 132, 337–340. [CrossRef] p y g 7. Christoph, L.; Sang-In, K.; Steven, E.B.; Arthur, X.L.; Xiwei, W.; Gerd, P.P. UVB irradiation does not d changes of DNA methylation in human keratinocytes. F1000Research 2013, 2, 45. meyer, A.; Luiten, R.M. Oxidation events and skin aging. Ageing Res. Rev. 2015, 21, 16–29. [CrossRef] Kohl, E.; Steinbauer, J.; Landthaler, M.; Szeimies, R.-M. Skin ageing. J. Eur. Acad. Dermatol. Venereol. 2011, 2 [PubMed] References Ganceviciene, R.; Liakou, A.I.; Theodoridis, A.; Makrantonaki, E.; Zouboulis, C.C. Skin anti-aging strategies. Dermatoendocrinology 2012, 4, 308–319. [CrossRef] 14. Stamford, N.P. Stability, transdermal penetration, and cutaneous effects of ascorbic acid and its derivatives. J. Cosmet. Dermatol. 2012, 11, 310–317. [CrossRef] [PubMed] 15. Caritá, A.C.; Fonseca-Santos, B.; Shultz, J.D.; Michniak-Kohn, B.; Chorilli, M.; Leonardi, G.R. Vitamin C uses. Advances for delivery, efﬁciency and stability. Nanomedicine 2020, 24, 102117. [CrossRef] a-Santos, B.; Shultz, J.D.; Michniak-Kohn, B.; Chorilli, M.; Leonardi, G.R. Vitamin C: One compound, several á, A.C.; Fonseca-Santos, B.; Shultz, J.D.; Michniak-Kohn, B.; Chorilli, M.; Leonardi, G.R. Vitamin C: One comp Advances for delivery efﬁciency and stability Nanomedicine 2020 24 102117 [CrossRef] Advances for delivery, efﬁciency and stability. Nanomedicine 2020, 24, 102117. [CrossRef] 16. Pullar, J.; Carr, A.; Vissers, M. The roles of vitamin C in skin health. Nutrients 2017, 9, 866. [CrossRef] 16. Pullar, J.; Carr, A.; Vissers, M. The roles of vitamin C in skin health. Nutrients 2017, 9, 866. [CrossRef] 17. De Dormael, R.; Bastien, P.; Sextius, P.; Gueniche, A.; Ye, D.; Tran, C.; Chevalier, V.; Gomes, C.; Souverain, L.; Tricaud, C. Vitamin C Prevents Ultraviolet-induced Pigmentation in Healthy Volunteers: Bayesian Meta-analysis Results from 31 Randomized Controlled versus Vehicle Clinical Studies. J. Clin. Aesthet. Dermatol. 2019, 12, E53–E59. 18. Tai, A.; Aburada, M.; Ito, H. A simple efﬁcient synthesis and biological evaluation of 3-O-ethylascorbic acid. Biosci. Biotechnol. Biochem. 2014, 78, 1984–1987. [CrossRef] g, J.-H. Dual Effects of Alpha-Hydroxy Acids on the Skin. Molecules 2018, 23, 863. [CrossRef] [PubMed] 19. Tang, S.-C.; Yang, J.-H. Dual Effects of Alpha-Hydroxy Acids on the Skin. Molecules 2018, 23, 863. [Cro 20. Sharad, J. Glycolic acid peel therapy—A current review. Clin. Cosmet. Investig. Dermatol. 2013, 6, 281–288. [CrossRef] 21. Kornhauser, A.; Coelho, S.G.; Hearing, V.J. Effects of cosmetic formulations containing hydroxyacids on sun-exposed skin: Current applications and future developments. Dermatol. Res. Pract. 2012, 2012, 710893. [CrossRef] [PubMed] , ; , ; g, J g y y p Current applications and future developments. Dermatol. Res. Pract. 2012, 2012, 710893. [CrossRef] [PubMed] g g y y urrent applications and future developments. Dermatol. Res. Pract. 2012, 2012, 710893. [CrossRef] [PubMed] vá, H.; Liebsch, M.; Schmidt, E.; Genschow, E.; Traue, D.; Spielmann, H.; Meyer, K.; Steinhoff, C.; Tornier, C 22. Kandárová, H.; Liebsch, M.; Schmidt, E.; Genschow, E.; Traue, D.; Spielmann, H.; Meyer, K.; Steinhoff, C.; Tornier, C.; de Wever, B.; et al. References 1. Cho-Rong, K.; Young-Min, K.; Min-Kyeong, L.; In-Hye, K.; Youn-Hee, C.; Taek-Jeong, N. Pyropia yezoensis peptide promotes collagen synthesis by activating the TGF-β/Smad signaling pathway in the human dermal fbroblast cell line Hs27. Int. J. Mol. Med. 2017, 39, 31–38. , , 2. Cannarozzo, G.; Fazia, G.; Bennardo, L.; Tamburi, F.; Amoruso, G.F.; Del Duca, E.; Nisticò, S.P. A New 675 nm Laser Device in the Treatment of Facial Aging: A Prospective Observational Study. Photomed. Laser Surg. 2021, 39, 118–122. [CrossRef] [PubMed] 3. Puizina-Ivi´c, N. Skin aging. Acta Derm. Alp. Pannonica Adriat. 2008, 17, 47–54. 3. Puizina-Ivi´c, N. Skin aging. Acta Derm. Alp. Pannonica Adriat. 2008, 17, 47–54 3. Puizina-Ivi´c, N. Skin aging. Acta Derm. Alp. Pannonica Adriat. 2008, 17, 47–54. 4. Kohl, E.; Steinbauer, J.; Landthaler, M.; Szeimies, R.-M. Skin ageing. J. Eur. Acad. Dermatol. Venereol. 201 [PubMed] 5. Kammeyer, A.; Luiten, R.M. Oxidation events and skin aging. Ageing Res. Rev. 2015, 21, 16–29. [CrossRef] 6. Nisticò, S.P.; Tolone, M.; Zingoni, T.; Tamburi, F.; Scali, E.; Bennardo, L.; Cannarozzo, G. A New 675 nm Laser Device in the Treatment of Melasma: Results of a Prospective Observational Study. Photomed. Laser Surg. 2020, 38, 1–5. [CrossRef] g y y 8. Vermeer, B.J.; Gilchrest, B.A. A proposal for rational deﬁnition, evaluation, and regulation. Arch Dermatol. 1996, 132, 337–340. [CrossRef] Life 2021, 11, 406 13 of 13 13 of 13 9. Buono, S.; Langellotti, A.L.; Martello, A.; Bimonte, M.; Tito, A.; Carola, A.; Apone, F.; Colucci, G.; Fogliano, V. Biological activities of dermatological interest by the water extract of the microalga Botryococcus braunii. Arch Dermatol Res. 2012, 304, 755–764. [CrossRef] 10. Iliopoulosa, F.; Silb, B.C.; Moorec, D.J.; Lucasc, R.A.; Lanea, M.E. 3-O-ethyl-L-ascorbic acid: Characterisa single solvent systems for delivery to the skin. Int. J. Pharm. X 2019, 1, 100025. [CrossRef] 11. Nusgens, B.V.; Humbert, P.; Rougier, A.; Colige, A.C.; Haftek, M.; Lambert, C.A.; Richard, A.; Creidi, P.; Lapière, C.M. Topically applied vitamin C enhances the mRNA level of collagens I and III, their processing enzymes and tissue inhibitor of matrix metalloproteinase 1 in the human dermis. J. Investig. Dermatol. 2001, 116, 853–859. [CrossRef] [PubMed] 12. Haftek, M.; Mac-Mary, S.; le Bitoux, M.; Creidi, P.; Seité, S.; Rougier, A.; Humbert, P. Clinical, biometric and structural evaluation of the long-term effects of a topical treatment with ascorbic acid and madecassoside in photoaged human skin. Exp. Dermatol. 2008, 17, 946–952. [CrossRef] 13. 30. Sun, Y.; Wang, P.; Li, H.; Dai, J. BMAL1 and CLOCK proteins in regulating UVB-induced apoptosis and DNA damage responses in human keratinocytes. J. Cell Physiol. 2018, 233, 9563–9574. [CrossRef] References Assessment of the skin irritation potential of chemicals by using the SkinEthic reconstructed human epidermal model and the common skin irritation protocol evaluated in the ECVAM skin irritation validation study. Altern. Lab. Anim. 2006, 34, 393–406. [CrossRef] [PubMed] 23. Kandárová, H.; Liebsch, M.; Spielmann, H.; Genschow, E.; Schmidt, E.; Traue, D.; Guest, R.; Whittingham, A.; Warren, N.; Gamer, A.O.; et al. Assessment of the human epidermis model SkinEthic RHE for in vitro skin corrosion testing of chemicals according to new OECD TG 431. Toxicol. In Vitro 2006, 20, 547–559. [CrossRef] [PubMed] 24. Rosdy, M.; Clauss, L.C. Terminal epidermal differentiation of human keratinocytes grown in chemically deﬁned medium on inert ﬁlter substrates at the air-liquid interface. J. Investig. Dermatol. 1990, 95, 409–414. [CrossRef] [PubMed] 25. Van Merloo, J.; Kaspers, G.J.; Cloos, J. Cell sensitivity assays: The MTT assay. Methods Mol. Biol. 2011, 731, 237–245. 26 ISO Bi l i l E l ti f M di l D i P t 10 T t f I it ti d Ski S iti ti (ISO 10993 10 2012) ISO L d 25. Van Merloo, J.; Kaspers, G.J.; Cloos, J. Cell sensitivity assays: The MTT assay. Methods Mol. Biol. 2011, 731, 237–245. 26. ISO. Biological Evaluation of Medical Devices–Part 10: Test for Irritation and Skin Sensitization (ISO 10993-10:2012); ISO: London, UK, 2012. 27. Ramata-Stunda, A.; Boroduskis, M.; Vorobjeva, V.; Ancans, J. Cell and tissue culture-based in vitro test systems for evaluation of natural skin care product ingredients. Environ. Exp. Biol. 2013, 11, 159–177. p g p 28. Edmondson, R.; Jenkins Broglie, J.; Adcock, A.F.; Yang, L. Three-dimensional cell culture systems and their applications in drug discovery and cell-based biosensors. Assay Drug Dev. Technol. 2014, 12, 207–218. [CrossRef] y y g 29. Sotiropoulou, G.; Zingkou, E.; Pampalakis, G. Redirecting drug repositioning to discover innovative cosmeceuticals. Exp. Dermatol. 2021, 30, 628–644. [CrossRef] 30. Sun, Y.; Wang, P.; Li, H.; Dai, J. BMAL1 and CLOCK proteins in regulating UVB-induced apoptosis and DNA damage responses in human keratinocytes. J. Cell Physiol. 2018, 233, 9563–9574. [CrossRef]
https://openalex.org/W3139138105	https://revistabioagro.mx/index.php/revista/article/download/259/239	es		Salsa a base de zarzamora (Rubus fruticosus) como potencial alimento funcional	Revista biológico agropecuaria Tuxpan	2,018	cc-by	2,116	Salsa a base de zarzamora (Rubus fruticosus) como potencial alimento funcional Blackberry sauce (Rubus fruticosus) like a potential functional food. Ceballos-Luna Oscar1, Morales-Sotelo Rogelio1, López-Márquez Alex Amir1, Grapain-López Orquídea1, GarcíaGutiérrez Emmanuel1, Sedeño-Mota Juan Carlos1, Álvarez Campos Edith2, Lozada García José Armando1 1 Facultad de Biología Xalapa, Universidad Veracruzana. Circuito Universitario Gonzalo Aguirre Beltrán s/n, Col. Zona Universitaria, Xalapa, Veracruz C.P. 91000 2 Unidad de Capacitación para el Desarrollo Rural No. 2 Tlanalapa s/n Col. Los Carriles, Coatepec, Veracruz, México. C.P. 91500  Autor de correspondencia: alozada@uv.mx Recibido: 13/07/2018 Aceptado: 12/09/2018 RESUMEN La zarzamora (Rubus fruticosus) es una rosácea comestible cultivada en la microrregión de Coyopolan municipio de Ixhuacán de los Reyes, Veracruz, México. Esta fruta es importante en esta localidad ya que es utilizada en la elaboración de una variedad de productos artesanales, entre los que destacan el licor de zarzamora y mermelada. Algunos de estos productos además de ser un aporte económico para los habitantes, proporcionan un beneficio a la salud por sus componentes como los compuestos fenólicos que actúan como antioxidantes naturales. En este sentido el presente trabajo tuvo como objetivo elaborar una salsa como potencial alimento funcional teniendo como ingrediente principal la zarzamora y una variedad local de chile, llamado “tabaquero”; a estas salsas se les evaluó su calidad nutricional, la presencia de microorganismos a través de pruebas como coliformes totales, conteo de bacterias aerobias en placa y conteo de mohos y levaduras en alimentos, posteriormente se cuantificaronlos flavonoides, fenoles y antocianinas totales, así como también su actividad antioxidante. Como resultado se obtuvieron tres tipos de salsa, en las que se determinó que no hay presencia de microorganismos, las pruebas de calidad nutrimental indicaron que tienen buenos valores y respecto a la funcionalidad se observó que, en comparación con el fruto fresco se mantuvo la actividad antioxidante aún después del proceso. Por lo que concluimos que los productos cuentan con valores apropiados para considerarlos como potenciales alimentos funcionales. Palabras clave: alimento funcional, antioxidante, zarzamora, flavonoides. ABSTRACT The blackberry (Rubus fruticosus) is an edible rosacea grown in the microregion of Coyopolan at Ixhuacán de los Reyes, Veracruz, México. This one is important in the region since it is used in the Revista Científica Biológico Agropecuaria Tuxpan 6 (Especial) ISSN: 2007-6940 72 Ceballos–Luna et al., 2018 elaboration of a big variety of products, among which stand out the liquor of blackberry and jam. Some products in addition to be the economic sustenance for the habitants of the region, has health benefits from some components such as phenolic compounds that act as natural antioxidants. In this job has the objective of develop a sauce as a functional food having as principal ingredient blackberrys, also using a local variety of chili, called "tabaquero chili", and this sauce was evaluated their nutritional quality, the presence of total coliforms, quantification of flavonoids, phenols, anthocyanins and antioxidant activity. Three types of sauce were obtained, from which it was determined that there is not presence of total coliforms in the sauces, the nutritional quality tests indicate that they have a low contribution in high density lipids (bad cholesterol) and the functionality is observed that the product and the three sauces in comparison maintains an activity antioxidant. So we conclude that the product has the appropriate values to consider it a functional food. Keywords: functional food, antioxidant, blackberry, flavonoids. INTRODUCCIÓN MATERIALES Y MÉTODOS La tendencia mundial busca en los alimentos la contribución de ciertos bioactivos que sean capaces de ejercer alguna acción beneficiosa al consumidor; estos aspectos son los que se buscan en los llamados “alimentos funcionales”, productos nutracéuticos o farmalimentos (Bello, 1995; Aranceta et al., 2003). En los últimos años el interés por estos productos ha incrementado su demanda, principalmente por las propiedades antioxidantes con las que cuentan, entre éstos se encuentran los frutos rojos o “berrys” (Coronado et al., 2014). Se elaboraron tres salsas a base de Zarzamora (Rubus fruticosus) y una variedad de chile local conocida como “Tabaquero”, la primera de estas era dulce, la segunda salada y la tercera salada condimentada con hoja de acuyo (Piper auritum). Todas las salsas fueron pasteurizadas y guardadas al vacío. Uno de los frutos rojos que contienen altos niveles de antioxidantes, flavonoides, antocianinas y taninos son las especies de Rubus (De Carvalho et al., 2010), es por ello que el objetivo del presente estudio fue incorporar estos metabolitos secundarios con actividad biológica en productos innovadores con características de alimento funcional, con la finalidad de apoyar a los productores de la microrregión de Coyopolan, municipio de Ixhuacán de los Reyes, Veracruz, México. Una vez obtenida la salsa base que contenía 30 gr de azúcar y 20 gr de sal, se pasteurizó y se dividió en tres porciones iguales, las cuales correspondierona cada uno de los tres tipos de salsas, la primera porción se le agregó más azúcar adicionándole 110 gr, a la segunda se le agregó 10 gr de sal y finalmente a la tercera se le incorporó un triturado de una hoja asada de Piper auritum pesando aproximadamente 25 gr, además de 5 gr de sal. Calidad Alimenticia y Microbiológica Se evaluaron parámetros de calidad nutrimental con pruebas como contenido de grasas, humedad, pH, acidez, cenizas y azúcares bajo los estándares internacionles de la AOAC (Asociación de Químicos Analíticos Oficiales, 2015). La cantidad de proteína total se determinó Revista Científica Biológico Agropecuaria Tuxpan 6 (Especial) ISSN: 2007-6940 73 Ceballos–Luna et al., 2018 mediante el método de Biuret (Coutiño et al., 2015 Horwitz, 1990). Para el caso de los parámetros microbiológicos se realizazon con base en las Normas Oficiales Mexicanas:NOM113-SSA1-1994, NOM-111-SSA1-1994 y NOM-092-SSA1-1994. Determinacion de flavonoides, fenoles, antocianinas totales y actividad antioxidante Se realizaron pruebas para determinar el contenido de estos compuestos en frutos frescos y en las salsas, mediante técnicas espectrofotométricas como reducción de AlCl3, Folin & Ciocalteun y pH diferencial (PyrkoszBiardzka et al., 2014, Palomino et al., 2009). La actividad antioxidante se calculó mediante el método de reducción de Molibdato (Moncada et al., 2012). RESULTADOS Y DISCUSIÓN Como se observa en la tabla 1, las salsas mostraron tener baja cantidad de grasa por cada muestra de 100 gr, la humedad contenida es de aproximadamente 50 a 60 %, las proteínas totales oscilan entre el 7 y 9%. Los azúcares totales variaron en las salsas, siendo la salsa dulce la de mayor cantidad en porcentaje. En lo que respecta a la acidez del producto, fue menor al 1%, lo que indica que es ligeramente acida al paladar. Finalmente, para el caso del pH, el promedio fue de tres. Tabla 1. Información nutrimental y microbiológica de las salsas de Zarzamora. Dulce Salada Salsa con Acuyo Grasas 0.4313gr 0.1896gr 0.0993gr Humedad 59% 51% 49% pH 3.02 2.8 2.89 Acidez 1% 1% 1% Cenizas 38% 18% 15% Proteínas 8% 9% 7% Azúcares 50% 36% 40% NOM-092-SSA1-1994 2 UFC 2 UFC 5 UFC NOM-113-SSA1-1994 0 UFC 0 UFC 0 UFC NOM-111-SSA1-1994 0 UFM 0 UFM 0 UFM En cuanto al contenido de microorganismos anaerobios se encontró mayor presencia en la salsa de acuyo con un promedio de 5 UFC, de acuerdo a la NOM-092-SSA1-1994, estos se encuentran dentro de los valores de aceptación, para coliformes totales no hubo presencia a las 24 h de incubación, en el caso de mohos y levaduras se encontraron por debajo de los parámetros mínimos permicibles, por lo cual se Revista Científica Biológico Agropecuaria Tuxpan 6 (Especial) ISSN: 2007-6940 74 Ceballos–Luna et al., 2018 consideran libres de estos microorganismos (NOM-092, NOM-111, NOM-113, 1994). Actividad antioxidante y compuestos fenólicos En cuanto a los flavonoides totales, el fruto fresco y la salsa salada, presentaron una concentración promedio de 0.41 mg EQ (Equivalentes de Quercetina), lo cual sugiere que éstos permanecen después del proceso de elaboración. La salsa dulce presentó mayor contenido de flavonoides totales, y triplicó el contenido de estos polifenoles en comparación con el fruto fresco. La presencia de flavonoides como la quercetina sugirió que puede tener beneficios a la salud previniendo diversos tipos de enfermedades crónico-degenerativas por su actividad antioxidante (Knekt et al., 2002; Geleijnse et al., 2002). En cuanto a las antocianinas totales (fruto fresco, salsas salada, dulce y con acuyo) presentaron un bajo contenido entre 0.004-0.009 de mg equivalentes Cianidina3-O-glucósido. Los fenoles fueron los únicos compuestos que no se pudieron determinar en ninguna de las muestras, esto se debe a que se encuentran en cantidades mínimas. Respecto a la actividad antioxidante que hayentre el fruto fresco, las salsas dulce y salada, los resultados muestran que esta característica se mantiene aún después del proceso de elaboración (Tabla 2). Por otra parte, cabe resaltar que la salsa con acuyo tuvo mayor actividad antioxidante, esto puede deberse a que esta planta aporta moléculas con esta característica, mostrando un alto contenido en compuestos fenólicos (García et al., 2002). Tabla 2. Contenido de compuestos fenólicos y actividad antioxidante. Compuestos Fenólicos cuantificados /Prueba Antocianinas Flavonoides Fenoles Reducción de ClAl3 Folin&Ciocalteun (mg EQ) (mg EAT) (mg = cyanidin 3-Oglucoside Fruto fresco 0.36 0 0.004 1.62 Salsa dulce 1.31 0 0.009 1.64 Salsa salada .0.47 0 0.009 1.39 Salsa con acuyo 0.21 0 0.006 5.59 Muestra pH diferencial Actividad antioxidante (reducción de molibdato) EQ: Equivalentes a Quercetina, EAT= Equivalentes a Ácido Tánico CONCLUSIÓN Se puede inferir que las salsas son un producto innovador ya que pueden ser consideradas un potencial alimento funcional, además de contar con valores nutrimentales aceptables y estándares microbiológicos dentro de las normas oficiales mexicanas. Por otra parte se le incorporaron compuestos fenólicos de la Revista Científica Biológico Agropecuaria Tuxpan 6 (Especial) ISSN: 2007-6940 75 Ceballos–Luna et al., 2018 zarzamora, como son los flavonoides y las antocianinas los cuales son benefiosos para la salud humana, además de que muestran actividad antioxidante significativa, en especial la salsa que contiene acuyo. LITERATURA CITADA Aranceta, J. Y Serra, L. 2003. Guía de alimentos funcionales. Sociedad Española de Nutrición Comunitaria (SENC). 105. Bello, J. 1995. Los alimentos funcionales o nutracéuticos. Nueva gama de productos en la industria alimentaria. Alimentaria 265: 25-29. Coronado, G. M. A., García, P. M., Santiago, H. V., García, P. M., Santiago, H. V. G., Córdova,Y. A. y Vásquez, N. R. 2014. La zarzamora, un mercado potencial para los productores agropecuarios de la Sierra de Sonora. Sociedad Mexicana de Administración Agropecuaria A.C. Torreón, México.784-794. Coutiño, R., Fernández, S y Palmeros, B. 2015. Manual de la experiencia educativa biología celular. Universidad Veracruzana. 132. https://doi.org/10.25009/uv.2018.37 De Carvalho, T., Thomsen, M. R. y Clark, J. R. 2010. Commercial fresh blackberry shipping market growth and price trends in the United States. Small Fruit News.102: 11. García, L., Rojo, D. M. D., García, L. V. G. y Hernández, M. A. 2002. Plantas con propiedades antiinflamatorias. Revista Cubana de Investigaciones Biomédicas 21(3): 214-216. Geleijnse, J. M., Launer, L. J, Van der Kuip, D. A. M, Hofman, A. y Witteman, J. C. M. 2002. Inverse association of tea and flavonoid intakes with incident myocardial infection: the Rotterdam study. The American Journal of Clinical Nutrition 75, 880-886. https://doi.org/10.1093/ajcn/75.5.880 Horwitz, W.1990. Official methods of analysis Washington, DC: Association of Official Analytical Chemists. 54: 171. Knet, P., Järvinen, R., Reunanen, A. y Maatela, J. 1996. Flavonoid intake and coronary mortality in Finland: A cohort Study. British. Medical Journal 312: 478-481. https://doi.org/10.1136/bmj.312.7029.47 8 Martínez, N. S. 2012. Producción de antocianinas por cultivo “in vitro” de Rubusa denotrichus Schltdl (zarzamora). Tesis de doctorado. División de estudios de posgrado. Facultad de ciencias biológicas. Universidad Autónoma de Nuevo León. Monterrey, Nuevo León, México. 85. Moncada, M. V., Giraldo, A.M. y Landazuri, P. 2012. Actividad antioxidante de los extractos etanólicos y acuosos de la cáscara y la semilla de Annona muricata y de las hojas brownea ariza. Revista de la Asociación Colombiana de Ciencias Biológicas 24: 143-151. Palomino, G. L. R., García, C. M. P., Gil, J. H., Rojano, G. B. A. y Durango, D. L. 2009. Determinación del contenido de fenoles y evaluación de la actividad antioxidante de propóleos recolectados en el departamento de Antioquia (Colombia). Vitae 16 (3): 388-395. Secretaria de salud. Norma Oficial Mexicana NOM-092-SSA1-1994, bienes y servicios. Métodos para la cuenta de bacterias aerobias en placa. Recuperado de: Revista Científica Biológico Agropecuaria Tuxpan 6 (Especial) ISSN: 2007-6940 76 Ceballos–Luna et al., 2018 http://www.ordenjuridico.gob.mx/Docu mentos/Federal/wo69532.pdf Secretaria de salud. Norma Oficial Mexicana NOM-113-SSA1-1994, bienes y servicios. Método para la cuenta de microorganismos coliformes totales en placa. Recuperado de: http://www.dof.gob.mx/normasOficiales /3924/stps1/stps1.htm Secretaria de salud. Norma Oficial Mexicana, NOM-111-SSA1-1994, bienes y servicios. Método para la cuenta de mohos y levaduras en alimentos. Recuperado de: http://www.cofepris.gob.mx/MJ/Docum ents/Normas/111ssa1.pdf Pyrkosz-Biardzka, K., Kucharska, A. Z, SokółŁętowska, A., Strugala, P. y Gabrielska, J. 2014. A Comprehensive Study on Antioxidant Properties of Crude extracts from Fruits of Berberis vulgaris L., Cornus mas L. and Mahonia aquifolium. Polish Journal Food and Nutrition Science 64(2): 91-99. https://doi.org/10.2478/v10222-012-009 7-x Revista Científica Biológico Agropecuaria Tuxpan 6 (Especial) ISSN: 2007-6940 77
https://openalex.org/W4309603931	https://hal.science/hal-04131965/file/fspas-09-1063849.pdf	English	null	STELLA—Potential European contributions to a NASA-led interstellar probe	Frontiers in astronomy and space sciences	2,022	cc-by	9,062	To cite this version: Robert F. Wimmer-Schweingruber, Nicolas André, Stanislav Barabash, Pontus C Brandt, Timothy S Horbury, et al.. STELLA-Potential European contributions to a NASA-led interstellar probe. Frontiers in Astronomy and Space Sciences, 2022, 9, pp.1063849. 10.3389/fspas.2022.1063849. hal- 04131965 STELLA-Potential European contributions to a NASA-led interstellar probe Robert F. Wimmer-Schweingruber, Nicolas André, Stanislav Barabash, Pontus C Brandt, Timothy S Horbury, Luciano Iess, Benoit Lavraud, Ralph L Mcnutt, Elena A Provornikova, Eric Quémerais, et al. To cite this version: Robert F. Wimmer-Schweingruber, Nicolas André, Stanislav Barabash, Pontus C Brandt, Timothy S Horbury, et al.. STELLA-Potential European contributions to a NASA-led interstellar probe. Frontiers in Astronomy and Space Sciences, 2022, 9, pp.1063849. 10.3389/fspas.2022.1063849. hal- 04131965 Robert F. Wimmer-Schweingruber, Nicolas André, Stanislav Barabash, Pontus C Brandt, Timothy S Horbury, Luciano Iess, Benoit Lavraud, Ralph L Mcnutt, Elena A Provornikova, Eric Quémerais, et al. Robert F. Wimmer-Schweingruber1, Nicolas André2, Stanislav Barabash3, Pontus C. Brandt4, Timothy S. Horbury5, Luciano Iess6, Benoit Lavraud7, Ralph L. McNutt Jr4, Elena A. Provornikova4, Eric Quémerais8, Robert Wicks9, Martin Wieser3 and Peter Wurz10 1Extraterrestrial Physics, Institute of Experimental & Applied Physics, Christian-Albrechts-University Kiel, Kiel, Germany, 2Institut de Recherche en Astrophysique et Planétologie, UPS, CNES, Université de Toulouse, Toulouse, France, 3Swedish Institute od Space Phyics, Kiruna, Sweden, 4Applied Physics Laboratory, Johns Hopkins University, Laurel, MD, United States, 5Department of Physics, Faculty of Natural Sciences, Space and Atmospheric Physics Group, Imperial College, London, United Kingdom, 6Department of Mechanical and Aerospace Engineering, Sapienza University of Rome, Rome, Italy, 7Laboratoire d’astrophysique de Bordeaux, University Bordeaux, Pessac, France, 8Héliosphère, Exosphéres Planétaires, Plasmas et Interfaces, Laboratoire Atmosphères, Observations Spatiales, Guyancourt, France, 9Department of Mathematics, Physics and Electrical Engineering, Northumbria University, Newcastle upon Tyne, United Kingdom, 10Space Physics and Planetary Sciences, Physics Institute, Faculty of Science, University of Bern, Bern, Switzerland Wimmer-Schweingruber RF, André N, Barabash S, Brandt PC, Horbury TS, Iess L, Lavraud B, McNutt RL Jr, Provornikova EA, Quémerais E, Wicks R, Wieser M and Wurz P (2022), STELLA—Potential European contributions to a NASA-led interstellar probe. The discovery of a myriad of exoplanets in the past decades has revolutionized the understanding of our place in the Universe. How different are exoplants and do some of them harbor life, just like Earth? To do so, their parent stars must drive a stellar wind and carve what we call astrospheres into the surrounding interstellar medium. Astrospheres are ubiquitous in our immediate neighborhood and show similar structure to our heliosphere. Voyager 1 and 2, Ulysses, Cassini, and IBEX have shown that the interaction between interstellar medium and solar wind is much more complex and involved than previously believed. This stellar-interstellar interaction is key to understand astrospheres and the shielding they provide to the planetary systems they harbor. This article summarizes a whitepaper that was submitted to NASA’s 2023/2024 decadal survey which is being conducted by the US National Academies. It is based in parts on a proposal submitted to the European Space Agency (ESA) in response to its 2021 call for medium-class mission proposals. The whitepaper and this article propose to study the interaction described above in situ at the heliospheric boundaries and to explore the very local interstellar medium beyond. Furthermore, they highlight possible European contributions to a NASA-led Interstellar Probe (ISP). Front. Astron. Space Sci. 9:1063849. doi: 10.3389/fspas.2022.1063849 © 2022 Wimmer-Schweingruber, André, Barabash, Brandt, Horbury, Iess, Lavraud, McNutt, Provornikova, Quémerais, Wicks, Wieser and Wurz. HAL Id: hal-04131965 https://hal.science/hal-04131965v1 Submitted on 17 Jun 2023 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. TYPE Methods PUBLISHED 17 November 2022 DOI 10.3389/fspas.2022.1063849 mission proposal, interstellar probe, space science, solar system, very local interstellar medium, fundamental physics STELLA—Potential European contributions to a NASA-led interstellar probe OPEN ACCESS EDITED BY Francesco Malara, University of Calabria, Italy REVIEWED BY Federico Fraternale, University of Alabama in Huntsville, United States Lennard Fisk, University of Michigan, United States CORRESPONDENCE Robert F. Wimmer-Schweingruber, wimmer@physik.uni-kiel.de SPECIALTY SECTION This article was submitted to Space Physics, a section of the journal Frontiers in Astronomy and Space Sciences RECEIVED 07 October 2022 ACCEPTED 26 October 2022 PUBLISHED 17 November 2022 CITATION Wimmer-Schweingruber RF, André N, Barabash S, Brandt PC, Horbury TS, Iess L, Lavraud B, McNutt RL Jr, Provornikova EA, Quémerais E, Wicks R, Wieser M and Wurz P (2022), STELLA—Potential European contributions to a NASA-led interstellar probe. OPEN ACCESS EDITED BY Francesco Malara, University of Calabria, Italy REVIEWED BY Federico Fraternale, University of Alabama in Huntsville, United States Lennard Fisk, University of Michigan, United States CORRESPONDENCE Robert F. Wimmer-Schweingruber, wimmer@physik.uni-kiel.de SPECIALTY SECTION This article was submitted to Space Physics, a section of the journal Frontiers in Astronomy and Space Sciences Robert F. Wimmer-Schweingruber1, Nicolas André2, Stanislav Barabash3, Pontus C. Brandt4, Timothy S. Horbury5, Luciano Iess6, Benoit Lavraud7, Ralph L. McNutt Jr4, Elena A. Provornikova4, Eric Quémerais8, Robert Wicks9, Martin Wieser3 and Peter Wurz10 Frontiers in Astronomy and Space Sciences 1 Introduction A substantial amount of work has been performed in the past decades that has shown that an interstellar probe (ISP) is today a feasible, timely, and compelling mission that would enable humanity’s journey to interstellar space. A 498-page study led by the Johns Hopkins University (JHU)/Applied Physics Laboratory (APL) provides a detailed account of the science goals and questions, the mission concept, scientific and technical implementation, as well as programmatics and concepts for the managements of such a 50-year mission (McNutt et al., 2021). The white paper reported here draws heavily from that study, as well as from white papers submitted previously to the European Space Agency (ESA) (Wimmer-Schweingruber et al., 2009; Wimmer-Schweingruber, 2013; Wimmer-Schweingruber, 2019) in response to its Cosmic Vision and Voyage 2050 calls for proposals, and the STELLA proposal submitted to ESA’s 13 December 2021 call for a medium-size mission opportunity (Barabash and Wimmer-Schweingruber, 2022). In line with the JHU/APL study, STELLA is a proposal for a substantial European contribtion to a NASA-led Interstellar Probe (ISP) that would reach 350 au after 50 years (Brandt et al., 2022). Here we highlight possible European contributions to such a NASA-led mission as well as the science questions that could be addressed by such a contribution. Today there is a growing awareness of the significance of the physics of the outer heliosphere. Its understanding helps to clarify the structure of our immediate interstellar neighborhood, contributes to the clarification of fundamental astrophysical processes like the acceleration of charged particles at a stellar wind termination shock and beyond, and also sheds light on the question to what extent interstellar-terrestrial relations are important for the environment of Earth and exoplanets. Even with the growing knowledge about the heliosphere provided by space missions, its global shape and dimensions remain a highly contested topic (Dialynas et al., 2017; Pogorelov et al., 2017; Schwadron and Bzowski, 2018). Most critically, the total plasma pressure in the heliosheath as a function of time and the relative importance of the various plasma populations (in particular at low energies < 200 eV) are only poorly constrained with present-day observations (Galli et al., 2017; Zirnstein et al., 2018). Attempts to explain consistently the known observations are currently roiled in controversy, with no clear trend towards a consensus (McNutt et al., 2021). Robert F. Wimmer-Schweingruber1*, Nicolas André2, Stanislav Barabash3, Pontus C. Brandt4, Timothy S. Horbury5, Luciano Iess6, Benoit Lavraud7, Ralph L. McNutt Jr4, Elena A. Provornikova4, Eric Quémerais8, Robert Wicks9, Martin Wieser3 and Peter Wurz10 This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. KEYWORDS 01 frontiersin.org Frontiers in Astronomy and Space Sciences Wimmer-Schweingruber et al. 10.3389/fspas.2022.1063849 10.3389/fspas.2022.1063849 TQ1: How is our heliosphere upheld by the physical processes from the Sun to the Very Local Interstellar Medium (VLISM)? TQ2: How do the Sun’s activity as well as the interstellar medium and its possible inhomogeneity influence the dynamics and evolution of the global heliosphere? TQ3: How do the current VLISM properties inform our understanding of the evolutionary path of the heliosphere? TABLE 1 ISP top-level science questions (McNutt et al., 2021). Scientific Requirements (Section 2.3) Stella Objective: Determine the (isotopic) composition of the GCR; determine Cosmic Ray Spectrometer (C Cosmic Ray Spectrometer (CRS): Large geometry factor (> 2 cm2 sr); mass resolution ∼0.1 amu; multiple viewing directions; measurements of the magnetic field European heritage: RAD on Mars Science Laboratory, EPD HET on Solar Orbiter, AHEPaM on Athena Cosmic Ray Spectrometer (CRS): Large geometry factor (> 2 cm2 sr); mass resolution ∼0.1 amu; multiple viewing directions; measurements of the magnetic field E h i RAD M S i L b EPD HET i European heritage: RAD on Mars Science Laboratory, EPD HET on Solar Orbiter, AHEPaM on Athena P Science Objective: Understand the filtration as the VLISM passes the heliospheric boundaries and processes responsible for structuring the VLISM he heliospheric interface (the Hydrogen wall) and beyond ISP Science Objective: Understand the filtration as the VLISM passes the heliospheric boundaries and processes responsible for structuring the VLISM at the heliospheric interface (the Hydrogen wall) and beyond Stella Science Requirement Q4: How does the local interstellar medium become structured when it meets the heliosphere? (Section 2.4) Stella Objective: Determine hydrogen populations properties, number density and velocity distribution, in the inner heliosheath, the hydrogen wall, and the VLISM Ly-α Spectrometer: Directional, high-resolution (< 20 km/s Doppler shift) Lyman-α spectra and their variation with distance from the Sun European heritage: Imaging UltraViolet Spectrograph on MAVEN. ISP Science Objective: (New and unique for the European contribution). Explore the gravity law on the scale of VLISM Stella Science Requirement Q5: Are there any deviations from the gravity law on the scale of the VLISM? (Section 2.5) Stella Objective: Determine the spacecraft acceleration Radio Science: Measure the two-way range to an accuracy of 20 cm @ 100 s integration time European heritage: BepiColombo; VERITAS. ( y g ) y ent Q4: How does the local interstellar medium become structured when it meets the heliosphere? (Section 2.4) Ly-α Spectrometer: Directional, high-resolution (< 20 km/s Doppler shift) Lyman-α spectra and their variation with distance from the Sun European heritage: Imaging UltraViolet Spectrograph on MAVEN. ISP Science Objective: (New and unique for the European contribution). Explore the gravity law on the scale of VLISM Stella Science Requirement Q5: Are there any deviations from the gravity law on the scale of the VLISM? (Section 2.5) Stella Objective: Determine the spacecraft acceleration Radio Science: Mea 20 cm @ 100 s integr heliosheath (Richardson et al., 2022). 1 Introduction nal European contributions to the ISP science questions (McNutt et al., 2021), their measurement requirements, and European heritage. arabash and Wimmer-Schweingruber (2022). 1 Introduction The many and diverse in situ and remote-sensing observations obtained to date clearly emphasize the need for a new generation of more comprehensive measurements that are required to understand the global nature of our Sun’s interaction with the local galactic environment. Science requirements informed by the now available observations drive the measurement requirements of an ISP’s in situ and remote-sensing capabilities that would allow to answer the open questions highlighted in Tables 1, 2. The latter exhibits notional, possible European contributions to the ISP science objectives and payload, but such contributions would, of course, need to undergo the usual competititve selection process. For the time being, they may be considered as an incomplete illustration of European capabilities. Other examples of European contributions are a magnetometer or a radio- plasma-wave instrument. All aforementioned ISP mission proposals aim at answering fundamental questions about heliophysics which can be summarized as shown in Table 1. In addition to answering te questions posed in Table 1, an ISP would enable planetary science with its fly-by of Jupiter and its moons after the era of the ESA’s Jupiter Icy Moons Explorer (JUICE) mission and to fly by a Kuiper-Belt Object (KBO), very similar to the successful New Horizons example. On its way to the Very Local Interstellar Medium (VLISM), ISP would allow bonus science goals from a variety of scientific disciplines (Section 2.6) and perform a fundamental measurement to test the 1/r2 law of gravitation at a hitherto unverified distance scale (Section 2.5). Thus, an ISP would serve a large scientific community while addressing the top-level questions given in Table 1. In Section 2 we discuss in more detail the science questions which could be addressed by possible European payload contributions to an ISP, some of which are shown in Table 2. After the exciting in situ observations of the termination shock, heliosheath, and heliopause, both Voyagers have now entered “interstellar space” (Stone et al., 2019). “Interstellar Space” is shown in quotation marks because it is not the pristine VLISM, but rather this space continues to be influenced by the Sun and heliosphere. Because of the various competing 02 frontiersin.org Frontiers in Astronomy and Space Sciences Frontiers in Astronomy and Space Sciences 10.3389/fspas.2022.1063849 Wimmer-Schweingruber et al. TABLE 2 Notional European contributions to the ISP science questions (McNutt et al., 2021), their measurement requirements, and European heritage. Adapted from Barabash and Wimmer-Schweingruber (2022). Scientific Requirements ISP Science Objective: Advance knowledge and understanding of the galactic neighborhood’s origin and evolution ISP Science Objective: Advance knowledge and understanding of the galactic neighborhood Stella Science Requirement Q1: What is the composition of the VLISM gas? (Section 2.1) Stella Objective: Determine the chemical and isotopic composition of the VLISM gas of all major species entering the heliosphere: H, He, O, Ar ISP Science Objective: Advance knowledge and understanding of the galactic neighborhood’s origin and evolution Stella Science Requirement Q1: What is the composition of the VLISM gas? (Section 2.1) Stella Objective: Determine the chemical and isotopic composition of the VLISM gas of all major species entering the heliosphere: H, He, O, Ar Neutral Gas Mass Spectrometer (NGMS): Mass-spectrometric measurements of thermal gas with m/Δm = 300 and sensitivitie at 1#/m3 ISP Science Objective: Advance knowledge and understanding of the galactic neighborhood’s origin and evolution Stella Science Requirement Q1: What is the composition of the VLISM gas? (Section 2.1) Stella Objective: Determine the chemical and isotopic composition of the VLISM gas of all major species entering the heliosphere: H, He, O, Ar Neutral Gas Ma measurements o / 3 Neutral Gas Mass Spectrometer (NGMS): Mass-spectrometric measurements of thermal gas with m/Δm = 300 and sensitivities at 1#/m3 European heritage: NGMS on Luna-Resurs, NIM/PEP on JUICE. ISP Science Objective: Resolve the birth and evolution of interstellar and inner-source PUIs; Understand acceleration processes and pressure balance in the solar wind and at the heliospheric boundaries; Establish the nature and structure of the heliopause, the ribbon, and the ring; Investigate how th heliospheric boundary is modified by solar dynamics; Determine the extent and impact of solar disturbances in the VLISM. ISP Science Objective: Resolve the birth and evolution of interstellar and inner-source PUIs; Understand acceleration processes and pressure balances in the solar wind and at the heliospheric boundaries; Establish the nature and structure of the heliopause, the ribbon, and the ring; Investigate how the heliospheric boundary is modified by solar dynamics; Determine the extent and impact of solar disturbances in the VLISM. Stella Science Requirement Q2: How is our dynamical heliosphere upheld and how does it change from the Sun to the VLISM? Frontiers in Astronomy and Space Sciences Scientific Requirements (Section 2.2) Stella Objective: Determine the 3D velocity distribution functions of ions and electrons Plasma Science System (PSS): Measurements of differ of ions and electrons in the energy range <3 eV to <2 Heritage: PAS on Solar Orbiter, SWEA on STEREO an p yi y y ; p Stella Science Requirement Q2: How is our dynamical heliosphere upheld and how does it change from the Sun to the VLISM? (Section 2.2) Plasma Science System (PSS): Measurements of differential fluxes of ions and electrons in the energy range <3 eV to <20 keV Heritage: PAS on Solar Orbiter, SWEA on STEREO and MAVEN, HIA and CODIF on Cluster, PLS on Voyager, PIMS on Europa Clipper Plasma Science System (PSS): Measurements of differential fluxes of ions and electrons in the energy range <3 eV to <20 keV Heritage: PAS on Solar Orbiter, SWEA on STEREO and MAVEN, HIA and CODIF on Cluster, PLS on Voyager, PIMS on Europa Clipper ISP Science Objective: Understand the nature and detailed origin of GCRs; Establish the exact nature of heliospheric shielding against GCRs; Establish how far solar disturbances reach into the VLISM how far solar disturbances reach into the VLISM. Stella Science Requirement Q3: What is the origin and role of galactic cosmic rays in the solar system and beyond? (Section 2.3) Stella Objective: Determine the (isotopic) composition of the GCR; determine the abundance of Li-B-Be; determine GCR pitch-angle distributions Cosmic Ray Spectrometer (CRS): Large geom cm2 sr); mass resolution ∼0.1 amu; multiple vi measurements of the magnetic field European heritage: RAD on Mars Science Labo how far solar disturbances reach into the VLISM. Stella Science Requirement Q3: What is the origin and role of galactic cosmic rays in the solar system and beyond? (Section 2.3) Stella Objective: Determine the (isotopic) composition of the GCR; determine the abundance of Li-B-Be; determine GCR pitch-angle distributions Cosmic Ray Spectrometer (CRS): Large geometry factor (> 2 cm2 sr); mass resolution ∼0.1 amu; multiple viewing directions; measurements of the magnetic field European heritage: RAD on Mars Science Laboratory, EPD HET on Solar Orbiter, AHEPaM on Athena Stella Science Requirement Q3: What is the origin and role of galactic cosmic rays in the solar system and beyond? 2.1 Stella Q1: What is the composition of the VLISM gas? Early determinations of the VLISM composition were from pickup ions in the solar wind, from anomalous cosmic rays, and from spectroscopic observations towards nearby stars. All these methods are indirect and depend on modelling for the determination of the composition of the VLISM neutral gas. Interstellar Boundary Explorer (IBEX) provided us with first direct measurements of VLISM H, He, O (Möbius et al., 2009), and Ne (Bochsler et al., 2012). However, the absolute values of their abundances in the VLISM are very uncertain because of the strong and variable photoionization in the inner Solar System (Bzowski et al., 2013) and filtration at the heliospheric boundary. Key ratios for planetary evolution studies, e.g., Ne/O and D/H, are not known with sufficient accuracy, and are limited by their small fluxes at 1 AU. Other VLISM species (C, N, Na, Mg, Al, Si, P, S, Ar, Ca, Fe) as well as the isotopic composition of He, O, Ne, have not been measured directly at all. For example, C is not able to penetrate the heliopause because it is fully ionized in the interstellar medium. This also prevents the use of indirect methods. ISP would for the first time directly measure all major neutral gas species in the VLISM. While the physics of PUIs within the inner heliosphere has been previously addressed with the Solar Wind Ion Composition Spectrometer (SWICS) on Ulysses and the Advanced Composition Explorer (ACE) (Gloeckler et al., 1992; Geiss et al., 1995; Schwadron et al., 2000; Wimmer- Schweingruber and Bochsler, 2003; Allegrini et al., 2005), the small geometric factor of SWICS did not allow to reconstruct the full 3D velocity distribution function of PUIs which is critical to understand PUI evolution through acceleration and transport processes in the heliosphere. The more modern PLAsma and SupraThermal Ion Composition [PLASTIC, Galvin et al. (2008)] instrument on the Solar TERrestrial Observatory (STEREO) mission or the Proton-Alpha Sensor (PAS) and Heavy Ion Sensor (HIS) on ESA’s Solar Orbiter (Owen et al., 2020) are examples of the kind of required instruments. The abundances of elements and their isotopes in the Solar System present a snapshot in time of the composition of the interstellar medium from which the Sun formed 4.57 Gyr ago. 2.2 Stella Q2: How is our dynamical heliosphere upheld and how does it change from the sun to the VLISM? In the following Sections 2.1, 2.2, 2.3, 2.4, 2.5 we highlight the science questions which could be addressed by a notional European payload (see Table 2) on which the white paper reported here was based. We stress that these notional contributions by no means cover all required measurements for an ISP, nor do they encompas all European capabilities. The solar wind expands from the Sun and interacts with the neutral gas in the VLISM to form the heliosphere. In situ observations by the Voyagers and New Horizon, and remote neutral atom measurements by IBEX strongly suggest that turbulence is one of the critical processes acting at the heliospheric interface (Fraternale et al., 2022). This interaction also serves as a model for how stellar winds interact with their VLISM to form astrospheres, but is further complicated by the inflow of neutral VLISM gas. This gas is ionized in the inner heliosphere, thus creating a population of pickup ions (PUIs) which is highly non-thermal and defines the pressure balances in the heliosheath (HS) and at the heliopause (HP) against the flow of the VLISM (Rankin et al., 2019). This crucial role of PUIs in the global solar wind-VLISM interaction can not be studied with Voyager 1 and 2 because they were not equipped to measure PUIs. Scientific Requirements While New Horizons is expected to reach the termination shock Stern et al. (2018), it does not carry a magnetometer (Weaver et al., 2008), which is a critical measurement to characterize the termination shock, heliosheath, and VLISM. definitions of VLISM we here clarify that we use the words “pristine VLISM” in the sense of the very local ISM in the vicinity of the local cloud or clouds and reserve the expressions LISM and ISM for what lies beyond. We stress that we do not expect an Interstellar Probe to reach the pristine VLISM in its lifetime, the relevant scales are simply too large even when compared with the distance that could be covered by this mission in its lifetime. We do not expect the Voyagers to last long enough to enter the undisturbed, pristine VLISM which is likely situated way beyond their current locations and remaining lifetime. Nevertheless, the Voyager results allow us to design an ISP to perform measurements of all quantities relevant to characterize and understand the VLISM. Moreover, observations by the Voyagers have underlined the importance of understanding the termination shock and how it influences the structure of the Although anchored in heliophysics, ISP, on its journey out of the Solar System, also offers other ground-breaking discoveries across astrophysics and planetary science. Any outward trajectory, dictated by the heliophysics objectives, offers flybys of at least one of the 130 dwarf planets or 4,000 KBOs, to understand the almost unexplored world of the outer Solar System. Remote and in situ dust observations would provide an unprecedented view of the circumsolar dust disk, critical for understanding the formation and evolution not only of our Solar System, but also to understand planetary formation in Frontiers in Astronomy and Space Sciences frontiersin.org 03 Wimmer-Schweingruber et al. 10.3389/fspas.2022.1063849 A Neutral Gas Mass Spectrometer (NGMS, Table 2) would provide the required composition measurements of the VLISM gas, as well as gradients due to the different ionization and neutralization processes. surprisingly young dust disks. Beyond the zodiacal cloud, the infrared Universe opens up to measure the poorly constrained extragalactic background light providing insight into galactic and stellar evolution since 200 My after the Big Bang. Further examples of such science opportunities are mentioned in Section 2.6. Frontiers in Astronomy and Space Sciences 2.1 Stella Q1: What is the composition of the VLISM gas? The evolution of interstellar matter that our Solar System encountered on its journey through the Galaxy can be inferred from the present state of the VLISM by measuring these elements and their isotope ratios. Measurements of the isotopic ratios of interstellar neutrals H/D, 3He/4He, and 20Ne/22Ne with high accuracy will constrain our understanding of the chemical evolution of our Galaxy. With the recent realization that the ISM is highly structured and that we are entering a new interstellar cloud now (Linsky et al., 2019), neutral gas measurements at 100s au scales will be critical for assessing the interstellar environment that the heliosphere will encounter on its interstellar trajectory. PUIs experiencing a secondary charge exchange process with interstellar neutrals become Energetic Neutral Atoms (ENAs). Knowledge of PUI properties is crucial for understanding ENA images of the global heliospheric interaction. IBEX has imaged heliospheric ENAs over a solar cycle which show substantial variations and demonstrate that solar activity drives the global response of the entire heliosphere and its interaction with the VLISM (McComas et al., 2020). But how multi-scale solar wind structures propagate and evolve in the outer heliosphere; how they influence suprathermal and energetic particles; what plasma flows they cause in the heliosheath; and how locations Frontiers in Astronomy and Space Sciences 04 frontiersin.org Wimmer-Schweingruber et al. 10.3389/fspas.2022.1063849 Li-Be-B can offer the capability to distinguish solar from interstellar from mixed plasmas in the HS, HP, and boundary layer(s), and VLISM. Studies of the unobserved species of GCRs in the VLISM are of importance not only to heliophysics and the nature of particle acceleration and consequences of GCRs in the heliosphere, but also to astrophysics and the nature of the Universe itself. Observations of particularly rare GCR isotopes, GCR electrons, and antimatter in the VLISM can even shed light on and further constrain models of the nature of the Big Bang and dark energy. However, because of heliospheric shielding of lower-energy GCRs, the critical observations required to answer such open questions rely on observations of these GCR species in the comparitavely unperturbed VLISM by an ISP. of boundaries change because of pressure pulses, shocks, and waves in the solar wind is not known. Voyager 1 detected shocks and pressure waves beyond the heliosphere which were likely driven by transient solar events (Gurnett et al., 2013). 2.1 Stella Q1: What is the composition of the VLISM gas? Our understanding of the multi-component, multi-scale plasma physics in the outer boundary layers of the heliosphere and its coupling to the neutral component of the VLISM is extremely limited (See, e.g., Mostafavi et al., 2022, for a review.). IBEX discovered the ENA ribbon, an enigmatic band of intense ENA fluxes in the sky. The ribbon is an “imprint” of the complex physical processes between charged and neutral particles controlled by the interstellar magnetic field. Understanding its origin will tell us about the fundamental physics of the heliosphere’s boundary, and thus about other astrospheres. The Interstellar Mapping and Acceleration Probe (IMAP, launch 2025) will explore ENAs in much more detail but it will not be able to investigate the micro-physical processes in this complex plasma. A launch of ISP in late summer of 2036 would let it traverse the ribbon and perform in situ particle measurements to establish its origin. A Cosmic Ray Spectrometer (CRS, Table 2) together with the PSS would provide the required measurements. 2.4 Stella Q4: How does the local interstellar medium become structured when it meets the heliosphere? A Plasma Science System (PSS, Table 2) together with NGMS would provide the required measurements of the VLISM plasma and neutral gas populations. The Local Interstellar Cloud surrounding the heliosphere primarily consists of atomic hydrogen (H). Approaching the heliosphere, interstellar H atoms pile up and form a Hydrogen Wall, found to be a common feature of astrosphere- ISM interaction (Baranov and Malama, 1993; Linsky and Wood, 1996). The heliospheric Hydrogen Wall was detected remotely, however, it was never explored with in situ measurements leaving open questions about its extent, peak density, and 3D structure. Despite being critical for the global heliospheric structure, properties of interstellar H atoms far away from the heliosphere in the LIC and their variations due to ion-neutral coupling at the heliospheric boundary remain largely unknown. Measuring solar Lyman-a (121.567 nm) emission backscattered from interstellar H atoms is a powerful technique to probe interstellar H atoms. The spectral shape of the Lyman-α emission line holds key information on the spatial and velocity distribution of interstellar H and enables to infer momentum exchange between hydrogen and plasma. The study of the Lyman-α line profile in the outer heliosphere allows characterizing the distribution (bulk velocity and temperature) of the hot hydrogen population, which is created by charge exchange in the inner heliosheath. After crossing the heliopause, these same measurements will provide the hydrogen atom distribution in the outer heliosheath and the Hydrogen Wall. As the probe moves away to larger distances the pristine interstellar H population will contribute increasing amounts to the Ly-α line profile thus allowing to derive (e.g., by extrapolation) the true number density, bulk velocity, and temperature in the Local Interstellar Cloud. The combination of in situ measurements of the H and proton density with other species with different ionization states will provide key Frontiers in Astronomy and Space Sciences 2.3 Stella Q3: What is the origin and role of galactic cosmic rays in the solar system and beyond? Galactic Cosmic Ray (GCR) anisotropies are sensitive to remote field variations and can therefore be used as an effective remote diagnostic of the field configuration of the heliosphere, and once beyond the heliopause, they provide insight into how solar disturbances can propagate deep into the VLISM (Krimigis et al., 2013; Gurnett et al., 2015; Rankin et al., 2019; Hill et al., 2020). The Voyagers’ cosmic ray instrument had limited look directions, and confounding, species-dependent anisotropies in GCR angular distributions have been reported (Rankin et al., 2019). There is no current consensus on what causes those anisotropies. With its improved instrumentation, ISP would offer a new opportunity to study the nature of GCR anisotropies, GCR shielding by the heliosphere, and the properties of the unshielded GCR spectra in the VLISM, including rare species and isotopes, that could not be observed by the Voyagers. Measurements of rare and unstable cosmic ray isotopes can be used to determine GCR source regions via spallation and direct acceleration, galactic escape rates, and solar modulation (Wiedenbeck et al., 2007). For instance, the abundance of Li, Be, and B at cosmic ray energies is comparable (same order of magnitude) to that of C, N, and O, and thus more than four to six orders of magnitude higher than their relative abundance in the Solar System. Observations of Frontiers in Astronomy and Space Sciences 05 frontiersin.org 10.3389/fspas.2022.1063849 Wimmer-Schweingruber et al. information about the ionization state and processes in the LISM. A Ly-α spectrometer (Table 2) would provide the required measurements. B3: Acquire Ly-α absorption spectra against known stellar neighbors to map out the neutral hydrogen density structure in the local neighbourhood. The increasing distance of an interstellar probe would gradually build up this 3D information. information about the ionization state and processes in the LISM. A Ly-α spectrometer (Table 2) would provide the required measurements. B4: The Solar System does not only include the Sun, planets, and asteroids, but also a large number of dwarf planets (a status to which Pluto has been demoted), Kuiper-Belt Objects (KBOs), and Trans-Neptunian Objects (TNOs). It thus comes as no surprise that there are many opportunities for a fly by of such an object. This would dramatically increase our understanding of these objects and of their diversity. 3 Mission description Apart from describing important science goals of an ISP, Table 2 also hints at possible European provisions of payload elements (instruments). Obviously, this would have to be subject to further refinement by a joint NASA - ESA science definition team and a competitive process. The instruments mentioned are given to illustrate European expertise and possible contributions. While these instruments have sufficient heritage and already have or will have TRL 6 by 2026, they would have to be qualified for a 50 years (or longer) mission duration. NASA and ESA would have to play a key role in this process. This science question would not require any dedicated payload, but the radio system would be required to provide precise range and range rate measurements (Table 2). 2.5 Stella Q5: Is the 1/r2 law of gravitation valid at scales of hundreds of au? The laws of gravity are unevenly tested at different spatial scales [Fundamental Physics Roadmap Advisory Team (FPR- AT), 2010]. The most precise knowledge is obtained from Solar System dynamics, where space probes allowed the most accurate tests of general relativity to date. Violations of the 1/r2 law, associated with the Compton wavelength of the graviton, have also been tested precisely in the Solar System. But at short spatial scales (at the atomic level) or beyond the Solar System (>50 au) gravity has never been accurately tested. ISP offers the opportunity to carry out an accurate test of the 1/r2 law at distances beyond 50 au and up to 350 au, by precise range and range rate measurements of the probe. In addition, ISP could determine or set up an upper limit on the density of dark matter in a sphere of radius approximately equal to the heliocentric distance of the probe. Gravitational signals from TNOs could also be detected providing a new “window” for discovering these objects. B5: As the ISP retreats from the Solar System its back-looking view of the Solar System would offer a remote observation of a known planetary system as a model for a habitable exo-planetary system. Frontiers in Astronomy and Space Sciences 2.6 Stella bonus science opportunities B1–B5 p One of the key European industrial and programmatic contributions proposed in the STELLA proposal to ESA is an upgrade of the European deep space communication facility that would allow the precise range and range-rate measurements of the probe to address STELLA science goal Q5 but would also provide additional downlink of ISP data and thus increase the ISP science return. The facility would be a critical augmentation of the European Deep Space Antennas (DSA) not only for ISP but also for other planned missions, e.g., to the icy giants. Less than a decade after its launch, returning data from ISP would require a receiving station with the equivalent of 4 35-m dishes or more than 16 18-m dishes, i.e., the equivalent of a 75-m dish. The architecture of the array should be studied during an early phase of the mission (0/A). European industries are among the world leaders in the field. mtex antenna technology (Germany) is the sole prime to develop a production-ready design and produce a prototype 18-m antenna for the US National Research Observatory (NRAO) Very Large Array (ngVLA) facility. Thales/Alenia (France/Italy), Schwartz Hautmont (Spain) are heavily involved in the development of the new 35-m DSA antenna. The five science questions discussed above are examples of heliophysics and fundamental physics questions that would be addressed by an ISP. We highlight five additional, bonus science “teasers” (B1–B5) to demonstrate the breadth of ISP. Indeed, this mission would also contribute to astrophysics and planetology. B1: Infrared emission from dust in the Solar System (Zodiacal light) obscures the extragalactic background light which is the red-shifted remnant of the emission of galaxies formed only a few hundred million years after the Big Bang (Cooray, 2016). Observing this light would constrain models of Galaxy formation, the evolution of cosmic structure, the re- ionization epoch, and allow for searches beyond the standard model (Tyson, 1995). B2: As the solar wind interacts with planetary magnetospheres, atmospheres, or surfaces, it emits characteristic X-ray emission. An X-ray instrument would thus allow to determine the soft X-ray background in the heliosphere and solar-wind planet interactions. frontiersin.org 06 Wimmer-Schweingruber et al. 2.6 Stella bonus science opportunities B1–B5 10.3389/fspas.2022.1063849 At the other end of the communication system, i.e., on the ISP spacecraft itself, Europe could also provide the High Gain Antenna (HGA) and communication system which should supply, together with the ground segment, a telemetry rate of at least 200 bps at 350 au and would allow to address STELLA science goal Q5. It would have to be a fully redundant X- band system; optical communication would very likely pose too stringent pointing requirements on the spacecraft. Hardware for this subsystem is mature technology with ample deep space heritage. The antenna complement would include a 5-m High Gain Antenna (HGA), a 0.4-m Medium Gain Antenna (MGA) coaligned with the HGA, and fore and aft Low Gain Antennas (LGAs). succession. ISP as the anchor would enable an entire half-century of scientists to gain mission experience. 4 Technology developments Of course, this 50-year (or longer) mission also requires significant, although entirely manageable, technology developments, the most noteable of which are highlighted in the following paragraphs. A more complete listing is given in McNutt et al. (2021). Frontiers in Astronomy and Space Sciences 3.1.1 Mission operations We envision that NASA would assume the main responsibilities for ISP flight dynamics and operations, and that ESA’s primary responsibility would be to operate the European payloads, and provide a dedicated link to ISP to drastically increase ISP’s science return. ESA would also perform the special operations required by the European Radio Science experiment and could provide additional uplink. The contribution envisaged in the STELLA proposal (Barabash and Wimmer-Schweingruber, 2022) only covers operations for the first 10-year segment of the 50-year nominal life time because it is foreseen to divide the ISP mission up into 10-year segments which would be reviewed individually. 3.1.2 Knowledge transfer The very long duration of the mission would necessitate the broad involvement of young and mid-career scientists as well as encouraging scientists who were not necessarily involved in the hardware phase to join the mission. We envision that PIs would change for the different phases of the mission. For each 10-year segment of the mission, all project and instrument teams would be requested to include three categories of their team members: 3.1 Maintaining a long-term mission Previous proposals have included solar sails, nuclear ion propulsion, or electric sails, which are technologically not mature for a flight in the 2030s and put complex constraints on spacecraft implementation. With the increasing availability of Super-Heavy Lift Launch Vehicles (SHLLVs), such as NASA Space Launch Systems (SLS), SpaceX Starship, or BlueOrigin vehicles, higher solar-system escape speeds are now becoming realistic using conventional chemical propulsion. For example, a launch with an SLS Block 2 followed directly by an Atlas Centaur upper stage and a Star-48 BV, all directly after launch, would put an 860- kg spacecraft on a 6-month trajectory to Jupiter. Here, a passive gravity assist would allow an escape speed twice that of the fastest solar-system escaping spacecraft, Voyager 1, of up to 7.2 au/year (McNutt et al., 2022). This 50-year (or longer) mission will require careful handling of mission operations and knowledge transfer: 4.2 Power Nuclear power would be unavoidable. NASA is currently developing the Next-Generation Radioisotope Thermoelectric Generator (NextGen RTG) through the Radioisotope Power Systems (RPS) Program at Glenn Research Center. The RPS Program is on track to provide the first flight-qualified NextGen RTG by 2028, well before 2036 when Interstellar Probe would require units. Power supply electronics and distribution units could be reused from New Horizons and Parker Solar Probe (PSP). See Section 5.9 in (McNutt et al., 2021) for more details. Frontiers in Astronomy and Space Sciences 4.3 Reliability • Senior members with extensive experience. These would typically be mentors of mid-career and junior scientists and potential PIs. Strategies to the ensure reliabilty of ISP for 50 years (or longer) of operations in space will need to be developed. Our knowledge of how to assess reliability has greatly improved over the past decades and has resulted in missions which have significantly exceeded their planned lifetimes, as is exemplified by the Voyagers, SOHO, Ulysses, and other missions. For instance, studies have shown that parts failures can be more accurately associated with inadequacies in design • Mid-career scientists who lead smaller teams and are in charge of important subsystems or projects. • Junior scientists who lead individual ISP science investigations. Through the course of the mission, scientists would flow through these categories to ensure knowledge transfer and 07 frontiersin.org 10.3389/fspas.2022.1063849 Wimmer-Schweingruber et al. or improper handling of components or workmanship (see McNutt et al., 2021, Sec. F.2). would also consider the use of the ESA facility for other missions to the icy giants and outer Solar System. The required TDAs should be kicked off as early as possible to ensure completion by the end of phase 0/A of an ISP mission. 4.5 Communication While communication across hundreds of au may be viewed as a technological challenge, the Voyagers have shown that this can be done. Current technology on Earth, paired with spacecraft communication equipment from the 1970s already allow this reliably across more than 150 au. The STELLA proposal to ESA (Barabash and Wimmer-Schweingruber, 2022) and the JHU/APL study (McNutt et al., 2021), as well as the first paragraphs of Section 3 give more details about possible improvements. Nevertheless, investments in communications would improve science return, and also benefit other outer-solar- system missions (e.g., to the giant planets and icy moons). If an ISP is recommended in first rank order of Large Strategic Missions and/or US Congress demands such a mission, then it would be prudent for NASA to establish a Science and Technology Definition Team (STDT) in consultation with ESA. Investments in the fields highlighted in the previous subsections would significantly enhance humanity’s space-faring capabilities and benefit (Solar System) space science. 5 International context An ISP would also pose a management challenge by necessitating a well-organized and continuous transfer of knowledge from one generation of engineers, scientists, and managers to the next as discussed in Section 3.1. An Interstellar Probe with varying names has been proposed to different space agencies. Apart from NASA (McNutt et al., 2021), ESA has commissioned such studies (Lyngvi et al., 2007), and China is also studying an interstellar probe which would reach 100 AU by 2049 to celebrate the 100-th anniversary of modern China. It was discussed in the 639th Academic Symposium of the renowned Xiangshan Science Conference series in October 2018. An international meeting on this topic was held at ISSI in Beijing in November 2019. Numerous mission white papers have been submitted to ESA (Wimmer-Schweingruber et al., 2009; Wimmer- Schweingruber, 2013; Wimmer-Schweingruber, 2019) and laid the ground work for the STELLA mission proposal (Barabash and Wimmer-Schweingruber, 2022). The necessarily long duration and ensuing cost of such a mission strongly suggest a coordinated NASA-ESA effort. Data availability statement Apart from obvious technology developments such as the development of the power sources and ensuring the powerful launcher needed for ISP mentioned above, the following technology development activities (TDAs) would be needed for an ISP: The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author. References Allegrini, F., Schwadron, N. A., McComas, D. J., Gloeckler, G., and Geiss, J. (2005). Stability of the inner source pickup ions over the solar cycle. J. Geophys. Res. 110, A05105. doi:10.1029/2004JA010847 Allegrini, F., Schwadron, N. A., McComas, D. J., Gloeckler, G., and Geiss, J. (2005). Stability of the inner source pickup ions over the solar cycle. J. Geophys. Res. 110, A05105. doi:10.1029/2004JA010847 Geiss, J., Gloeckler, G., Fisk, L. A., and von Steiger, R. (1995). C+ pickup ions in the heliosphere and their origin. J. Geophys. Res. 100, 23373–23378. doi:10.1029/95JA03051 Barabash, S., and Wimmer-Schweingruber, R. F. (2022). Stella – europe’s contribution to interstellar probe – humanity’s journey to interstellar space. Tech. rep. Kiruna, Sweden: Swedish Institute of Space Physics. Gloeckler, G., Geiss, J., Balsiger, H., Bedini, P., Cain, J. C., Fischer, J., et al. (1992). The solar wind ion composition spectrometer. Astron.& Astrophys. Suppl. 92, 267–289. Baranov, V. B., and Malama, Y. G. (1993). Model of the solar wind interaction with the local interstellar medium numerical solution of self-consistent problem. J. Geophys. Res. 98, 15157–15164. doi:10.1029/93JA01171 Gurnett, D. A., Kurth, W. S., Burlaga, L. F., and Ness, N. F. (2013). In situ observations of interstellar plasma with voyager 1. Science 341, 1489–1492. doi:10.1126/science.1241681 Bochsler, P., Petersen, L., Möbius, E., Schwadron, N. A., Wurz, P., Scheer, J. A., et al. (2012). Estimation of the neon/oxygen abundance ratio at the heliospheric termination shock and in the local interstellar medium from IBEX observations. Astrophys. J. Suppl. Ser. 198, 13. doi:10.1088/0067-0049/198/2/13 Gurnett, D. A., Kurth, W. S., Stone, E. C., Cummings, A. C., Krimigis, S. M., Decker, R. B., et al. (2015). Precursors to interstellar shocks of solar origin. Astrophys. J. 809, 121. doi:10.1088/0004-637X/809/2/121 Hill, M. E., Allen, R. C., Kollmann, P., Brown, L. E., Decker, R. B., McNutt, J., et al. (2020). Influence of solar disturbances on galactic cosmic rays in the solar wind, heliosheath, and local interstellar medium: Advanced composition explorer, new Horizons, and voyager observations. Astrophys. J. 905, 69. doi:10.3847/1538- 4357/abb408 Brandt, P. C., Provornikova, E. A., Cocoros, A., Turner, D., DeMajistre, R., Runyon, K., et al. (2022). Interstellar probe: Humanity’s exploration of the galaxy begins. Acta Astronaut. 199, 364–373. doi:10.1016/j.actaastro.2022. 07.011 Bzowski, M., Sokół, J. M., Kubiak, M. A., and Kucharek, H. (2013). Modulation of neutral interstellar He, Ne, O in the heliosphere. Survival probabilities and abundances at IBEX. Astron. Astrophys. 557, A50. doi:10.1051/0004-6361/201321700 Krimigis, S. M., Decker, R. References B., Roelof, E. C., Hill, M. E., Armstrong, T. P., Gloeckler, G., et al. (2013). Search for the exit: Voyager 1 at heliosphere’s border with the galaxy. Science 341, 144–147. doi:10.1126/science.1235721 Linsky, J. L., Redfield, S., and Tilipman, D. (2019). The interface between the outer heliosphere and the inner local ISM: Morphology of the local interstellar cloud, its hydrogen hole, strömgren shells, and 60Fe accretion. Astrophys. J. 886, 41. doi:10.3847/1538-4357/ab498a Cooray, A. (2016). Extragalactic background light measurements and applications, 3. London, UK: Royal Society Open Science, 150555. doi:10.1098/rsos. 150555 Dialynas, K., Krimigis, S. M., Mitchell, D. G., Decker, R. B., and Roelof, E. C. (2017). The bubble-like shape of the heliosphere observed by Voyager and Cassini. Nat. Astron. 1, 0115. doi:10.1038/s41550-017-0115 Linsky, J. L., and Wood, B. E. (1996). The α centauri line of sight: D/H ratio, physical properties of local interstellar gas, and measurement of heated hydrogen (the “hydrogen wall”) near the heliopause. Astrophys. J. 463, 254–270. doi:10.1086/177238 Fraternale, F., Adhikari, L., Fichtner, H., Kim, T. K., Kleimann, J., Oughton, S., et al. (2022). Turbulence in the outer heliosphere. Space Sci. Rev. 218, 50. doi:10.1007/s11214-022-00914-2 Lyngvi, A. E., van den Berg, M. L., and Falkner, P. (2007). Study overview of the interstellar heliopause probe. Noordwijk, Netherlands: European Space Agency. Tech. Rep. SCI-A/2006114/IHP, issue 3, rev. 4. Available at: https://sci.esa.int/ science-e/www/object/doc.cfm?fobjectid=40926. Fundamental Physics Roadmap Advisory Team (FPR-AT) (2010). A Roadmap for fundamental physics in space. Tech. rep. Noordwijk, Netherlands: European Space Agency, ESA. Available at: https://sci.esa.int/science-e/www/object/ doc.cfm?fobjectid=47597. McComas, D. J., Bzowski, M., Dayeh, M. A., DeMajistre, R., Funsten, H. O., Janzen, P. H., et al. (2020). Solar cycle of imaging the global heliosphere: Interstellar boundary explorer (IBEX) observations from 2009-2019. Astrophys. J. Suppl. Ser. 248, 26. doi:10.3847/1538-4365/ab8dc2 Galli, A., Wurz, P., Schwadron, N. A., Kucharek, H., Möbius, E., Bzowski, M., et al. (2017). The downwind hemisphere of the heliosphere: Eight years of IBEX-lo observations. Astrophys. J. 851, 2. doi:10.3847/1538-4357/ aa988f McNutt, R. L., Paul, M. V., Brandt, P. C., and Kinnison, J. D. (2021). Interstellar probe. Tech. rep. Laurel, MD: Applied Physics Laboratory, Johns Hopkins University. Galvin, A. B., Kistler, L. M., Popecki, M. A., Farrugia, C. J., Simunac, K. D. C., Ellis, L., et al. (2008). The plasma and suprathermal ion composition (PLASTIC) investigation on the STEREO observatories. Space Sci. Rev. 136, 437–486. doi:10.1007/s11214-007-9296-x McNutt, R. L., Wimmer-Schweingruber, R. F., Gruntman, M., Krimigis, S. M., Roelof, E. Conflict of interest The STELLA proposal (Barabash and Wimmer- Schweingruber, 2022) and the white paper reported here were developed using information from the JHU/APL ISP study (McNutt et al., 2021) which was supported by Task Order 80MSFC18F0139 under NASA Contract NNN06AA01C. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Acknowledgments All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. We thank the countless individuals who have contributed to our understanding of the heliosphere and the various studies and proposals of such a mission to explore our cosmic neighborhood. Author contributions Development and implementation of qualification procedures for missions with nominal lifetimes of 50 years and beyond. This would provide the community with knowledge of designing long-lived space equipment and be helpful for other programs such as Artemis. RW-S wrote the manuscript based on the inputs of his co- authors and provided information about the energetic particle environment, NA contributed information about the plasma state of the VLISM, SB led the STELLA proposal submitted to ESA in February 2022, PB provided input on US and international efforts for an ISP as well as the plasma and neutral gas measurements, TH contributed overall scientific guidance, especially about the magnetic structure of the heliosphere, LI contributed information about verifying the 1/r2 law of gravitation, BL contributed information about the plasma state of the VLISM, RM provided input on US and international efforts for an ISP, EP, and EQ contributed to the Ly-α questions, RW contributed overall scientific guidance, MW and PW contributed to the neutral gas questions. Studies of using integrated X/Ka deep space transponders (IDSTs) on an ISP mission. IDSTs will be flown on NASA’s Venus Emissivity, Radio Science, InSAR, Topography, and Spectroscopy (VERITAS) mission as a contribution from the Italian Space Agency and are already at high TRL level. The possible use of Ka- band relies on the pointing accuracy of the spinning spacecraft and requires dedicated studies. Studies of the deep space communication facility architectures, and of an upgrade of the DSA to provide a dedicated link for ISP: 4 × 35-m dishes or more than 16 18- m dishes, antenna’s location, numbers in each site. The study frontiersin.org 08 Wimmer-Schweingruber et al. 10.3389/fspas.2022.1063849 Frontiers in Astronomy and Space Sciences References C., Brandt, P. C., et al. (2022). Interstellar probe - destination: Universe. Acta Astronaut. 196, 13–28. doi:10.1016/j.actaastro.2022.04.001 09 Frontiers in Astronomy and Space Sciences frontiersin.org Wimmer-Schweingruber et al. 10.3389/fspas.2022.1063849 Möbius, E., Bochsler, P., Bzowski, M., Crew, G. B., Funsten, H. O., Fuselier, S. A., et al. (2009). Direct observations of interstellar H, He, and O by the interstellar boundary explorer. Science 326, 969–971. doi:10.1126/science.1180971 Stone, E. C., Cummings, A. C., Heikkila, B. C., and Lal, N. (2019). Cosmic ray measurements from Voyager 2 as it crossed into interstellar space. Nat. Astron. 3, 1013–1018. doi:10.1038/s41550-019-0928-3 Mostafavi, P., Burlaga, L. F., Cairns, I. H., Fuselier, S. A., Fraternale, F., Gurnett, D. A., et al. (2022). Shocks in the very local interstellar medium. Space Sci. Rev. 218, 27. doi:10.1007/s11214-022-00893-4 Tyson, J. A. (1995). “The optical extragalactic background radiation,” in Extragalactic background radiation meeting. Editors D. Calzetti, M. Livio, and C. Madau (Cambridge, NY: Cambridge University Press), 103–133. Weaver, H. A., Gibson, W. C., Tapley, M. B., Young, L. A., and Stern, S. A. (2008). Overview of the new Horizons science payload. Space Sci. Rev. 140, 75–91. doi:10.1007/s11214-008-9376-6 Owen, C. J., Bruno, R., Livi, S., Louarn, P., Al Janabi, K., Allegrini, F., et al. (2020). The solar orbiter solar wind analyser (SWA) suite. Astron. Astrophys. 642, A16. doi:10.1051/0004-6361/201937259 Wiedenbeck, M. E., Binns, W. R., Cummings, A. C., Davis, A. J., de Nolfo, G. A., Israel, M. H., et al. (2007). An overview of the origin of galactic cosmic rays as inferred from observations of Heavy ion composition and spectra. Space Sci. Rev. 130, 415–429. doi:10.1007/s11214-007-9198-y Pogorelov, N. V., Fichtner, H., Czechowski, A., Lazarian, A., Lembege, B., le Roux, J. A., et al. (2017). Heliosheath processes and the structure of the heliopause: Modeling energetic particles, cosmic rays, and magnetic fields. Space Sci. Rev. 212, 193–248. doi:10.1007/s11214-017-0354-8 Rankin, J. S., McComas, D. J., Richardson, J. D., and Schwadron, N. A. (2019). Heliosheath properties measured from a voyager 2 to voyager 1 transient. Astrophys. J. 883, 101. doi:10.3847/1538-4357/ab3d9d Wimmer-Schweingruber, R. F., and Bochsler, P. (2003). On the origin of inner-source pickup ions. Geophys. Res. Lett. 30, 1077–1080. doi:10.1029/ 2002GL015218 Richardson, J. D., Burlaga, L. F., Elliott, H., Kurth, W. S., Liu, Y. D., and von Steiger, R. (2022). Observations of the outer heliosphere, heliosheath, and interstellar medium. Space Sci. Rev. 218, 35. doi:10.1007/s11214-022-00899-y Wimmer-Schweingruber, R. F. (2013). In situ investigations of the local interstellar medium. Tech. rep. Zirnstein, E. J., Kumar, R., Heerikhuisen, J., McComas, D. J., and Galli, A. (2018). Stochastic acceleration of ∼0.1–5 keV pickup ions in the heliotail. Astrophys. J. 860, 170. doi:10.3847/1538-4357/aac3de References Kiel, Germany: Christian-Albrechts-University. Wimmer-Schweingruber, R. F. (2013). In situ investigations of the local interstellar medium. Tech. rep. Kiel, Germany: Christian-Albrechts-University. Wimmer-Schweingruber, R. F. (2019). In situ investigations of the local interstellar medium. Tech. rep. Kiel, Germany: Christian-Albrechts-University. Wimmer-Schweingruber, R. F. (2019). In situ investigations of the local interstellar medium. Tech. rep. Kiel, Germany: Christian-Albrechts-University. Schwadron, N. A., Geiss, J., Fisk, L. A., Gloeckler, G., Zurbuchen, T. H., and von Steiger, R. (2000). Inner source distributions: Theoretical interpretation, implications, and evidence for inner source protons. J. Geophys. Res. 105, 7465–7472. doi:10.1029/1999JA000225 Wimmer-Schweingruber, R. F., McNutt, R., Schwadron, N. A., Frisch, P. C., Gruntman, M., Wurz, P., et al. (2009). Interstellar heliospheric probe/heliospheric boundary explorer mission—A mission to the outermost boundaries of the solar system. Exp. Astron. (Dordr). 24, 9–46. doi:10.1007/s10686-008- 9134-5 Schwadron, N. A., and Bzowski, M. (2018). The heliosphere is not round. Astrophys. J. 862, 11. doi:10.3847/1538-4357/aacbcf Zirnstein, E. J., Kumar, R., Heerikhuisen, J., McComas, D. J., and Galli, A. (2018). Stochastic acceleration of ∼0.1–5 keV pickup ions in the heliotail. Astrophys. J. 860, 170. doi:10.3847/1538-4357/aac3de Stern, S. A., Weaver, H. A., Spencer, J. R., and Elliott, H. A. (2018). The new Horizons kuiper belt extended mission. Space Sci. Rev. 214, 77. doi:10.1007/s11214- 018-0507-4 frontiersin.org 10 Frontiers in Astronomy and Space Sciences
https://openalex.org/W3029711209	https://built-heritage.springeropen.com/track/pdf/10.1186/BF03545743	English	null	Shanghai and New York: Mid-Century Urban Avant-Gardes	Built heritage	2,019	cc-by	7,323	Shanghai and New York: Mid-Century Urban Avant-Gardes Rosemary Wakeman Rosemary Wakeman History Department & Urban Studies Program, Fordham University, New York, United States Email: rwakeman@fordham.edu History Department & Urban Studies Program, Fordham University, New York, United States Email: rwakeman@fordham.edu ABSTRACT This article examines the cultural and architectural exchanges between Shanghai and New York in the mid-20 th century and their iconic roles as avant-garde global capitals. It considers the cultural and architectural free-thinkers and the groundbreaking movements they led, as well as their influence on the character and identity of both cities. Art Deco was the emblematic symbol of these modern forces. The article analyses the new technolo gies and theatrical perspective, the performative and spectacular qualities of Art Deco as the signature styles and cultural heritage of both cities. KEYWORDS Art Deco, New York, avant-garde, Radio City Music Hall, Park Hotel, lighting design, technology, Ca thay Cinema, cityscape, film Received May 12, 2019; accepted August 15, 2019. Received May 12, 2019; accepted August 15, 2019. Received May 12, 2019; accepted August 15, 2019. Much has already been written about New York and Shanghai in the 1930s. The scholarship on ‘Art Deco New York’ and ‘Art Deco Shanghai’ is multitudinous. This article focuses more narrowly on specific facets of this heritage. It broadens out the imaginary of built heritage to include the composite of cosmopolitan in fluences that shaped the urban fabric, and that made these cities into global crossroads. While the influence of New York and the American architectural idiom on Shanghai is taken fully into account, the article argues that similar avant-garde influences inspired architects in both cities and predisposed the way their built en vironments were produced. It takes a closer look at ar chitecture as urban theatre and redefines architectural space to include what are often considered ephemeral or inconsequential elements, and yet were decisive in making architecture into avant-garde production. Spe cifically, new building technologies and theatre design were both at the nexus of perceptions about modernism and urban form. Technology was the guiding leitmotif of the 20th century. Technology along with architecture in the service of mass culture and consumption, popu lar entertainment and urban spectacle opens a rich perspective on built heritage. It ties it to the myriad transformations that made New York and Shanghai into modern urban icons. Shanghai and New York: Mid-Century Urban Avant-Gardes In the 1920s and 1930s, New York and Shanghai were bustling cosmopolitan metropolises, melting pots of nationalities and by extension architectural styles from around the world. New York was already a metropolis of some 7 million people and just behind London as the largest city in the world. With 3.2 million inhabitants in 1932, Shanghai was the fifth largest city in the world. Its population increased to 5 million by the end of the 1930s. Both cities were teaming ports and places of en counter with a flux-and-flow of peoples moving across borders and frontiers. They were shaped by exchanges and the mingling of publics, ideas, trade. This circulation of peoples and cultures, the collision of all these ideas and ways of seeing was a boiling cauldron of creativity. They made New York and Shanghai into global pivots for the production and consumption of urban modernity. The modern temperament, the famous New York and Shanghai ‘styles’, was defined by skyscrapers, department stores, and their glitzy ballrooms and jazz dance clubs. Their built heritage comprised manifold cultures. Both cities were the sites of competing architectural styles- ne oclassical and pompous Beaux-Arts buildings sat along side Art Deco towers that exemplified the new age. The aesthetic tensions that existed in New York and Shanghai created mesmerising cityscapes that was already legend ary in the 1930s. 49 R. Wakeman capitalism, while in Shanghai, it was a visible sign of the colonial presence and a mesmerising newness invented in the West. Alluding to the lesser role that Shanghai often played in relation to the great western capitals, the pub lisher of The China Journal argued that Shanghai ‘must show that it is one of the “Great Three” along with London and New York … Shanghai is the London, New York and Paris of the East, all rolled into one’ (Sowerby 1928, 114). Shanghai was becoming ‘an apartment city just as are New York and London’. At the same time, the cosmopolitan qualities of Art Deco were directly connected to Eurocen trism. A plea for Chinese expertise in modern architecture asserted that ‘We want variety in the “brave new world”. We most emphatically do not want an exact reproduction of Europe or America in Asia.’ (North China Herald 1929, 1934) Yet in this dialogue between east and west, the re semblances between the cities were as provocative as their differences. Art Deco as Future City It was the Art Deco skyscrapers that crowned their sky lines in the 1920s and 1930s that made New York and Shanghai into unabashed icons of the future. Rockefel ler Center, the Chrysler Building, and the Empire State Building in New York were glorious signatures of urban modernism as were Shanghai’s Park Hotel, the Sassoon House, and Broadway Mansions. Of course, not everyone was convinced that the skyscraper was the future. Ven erable American urban theorist Lewis Mumford called New York’s skyscrapers ‘a blind alley and an insupportable luxury’ (Mumford 1933). But for the majority gazing up at the soaring towers, Art Deco was the vision of a futuristic city. It formed part of a cosmopolitan imagination ema nating from the West. Architects consciously abandoned the historical forms that dominated architecture up to the 1920s as stogy and unsuitable to the modern age. Instead, they found inspiration from a wild variety of sources in what amounted to a global intermixing of styles. Art Deco exuded optimism and aesthetic exuberance. It was a spec tacular statement of ‘newness’, of a future filled with pos sibilities. It was an immediate international sensation, in fusing cities across the globe with a scintillating verve, an architectural jazz that matched the modern age. Although profoundly eclectic, it was the technology that pervaded modern life that most infused the new Art Deco aesthetic. Architects and designers embraced airships and cruise ships, the whirring cogs and wheels of the factory. They rendered the machine aesthetic in linear and geometric forms and cubist volumes, in sumptuous metals, zigzag and streamlined configurations that evoked the energy and power of technology. They explored the interplay of planes, the juxtaposition of patterns, colours and textures. In this regard, the impact of the 1925 Paris Exposition des Arts Décoratifs et Industriels Modernes on New York as well as on Shanghai was profound and rapid. The Exposi tion graced the Seine River in the west of Paris from April to October that year with an official attendance of over 15 million people. The exposition was itself a place of fantasy and wonder. Under the banner of ‘Art Deco’, the Exposi tion challenged participants with an avant-garde credo: ‘Works admitted to the Exposition must show new inspi ration and real originality.’ It was this experimental aesthetic temperament that made Art Deco so adaptable across cultures. Shanghai and New York: Mid-Century Urban Avant-Gardes In both cities, Art Deco symbolised money and power and took on a hegemonic caste (Lee 2001, 92– 95). It is easy to interpret this heritage as a capitalist logic exulting in the flashy facades. It was the rich and famous who inhabited the luxurious Art Deco skyscrapers rising on the skylines, and their glamorous lives became the stuff of Hollywood films shown in the cinemas of Shanghai and New York. The Bund and Fifth Avenue were the frontiers of global capitalism. Art Deco as Future City Rather than If architecture exuded wealth and privilege, it was also a product of the crossroads positions of both cities, a kind of migratory or melting pot modernism that scaled ter ritorial frontiers and wandered through urban cultures, imbibing local design practices, picking up fragments and adapting to local aesthetics. British sociologist Gerard Delanty argues that there are moments of world open ness in which the encounter between the local and the global shaped cultural modes of mediation (Delanty 2006, 27). The avant-garde production of the 1920s and 1930s was indeed such a moment. In both cities, the genealogi cal roots that made up this cultural mélange were many, and the meaning of this modernist heritage goes beyond the hypnotic decorative spell that Art Deco caste. It was a form of global cultural knowledge situated in the local (Windover 2009, 24). The geometric forms and motifs of East Asian decorative art added to the mix as did Futurism and Constructivism. Central European avant-garde design was a significant ingredient to this stew of global influ ences, as were avant-garde theories of democratic theatre and mass society, along with the technological wizardry of engineering and communication media, and dreams of cities made up of spiny-needle skyscrapers. In the hands It was this experimental aesthetic temperament that made Art Deco so adaptable across cultures. Rather than simply emulate Paris, when it was transplanted inter nationally, the design took on different meanings and fashioned distinct New York and Shanghai ‘styles’. In New York, Art Deco was the signature of a triumphant 50 BUILT HERITAGE 2019 / 3 Figure 1 Walter Gropius (1883– 1969), 1927. Berlin: Total Thea tre: Plan deep stage in position (Source: ARTSTOR). 31 Figure 1 Walter Gropius (1883– 1969), 1927. Berlin: Total Thea tre: Plan deep stage in position (Source: ARTSTOR). 31 design of the playhouse itself (Figure 1). As part of the broad social theory shared by these cosmopolitan elites, the modern theatre was imagined as a mass democratic medium (Gropius and Schlemmer, 1924). The focus was on dramatic technique and technics. Revolutionary uses of space, form, colour, sound, motion would dissolve into the shifting, illusionary arena of the imagination. The New York architects eagerly adapted these European expres sionist and modernist configurations to the city’s building boom. Art Deco as Future City The emotional and decorative effects of colour and pattern, and illumination were in full view in Rockefeller Center (which received official landmark status in 1978). of the avant-garde, these appealed to the emotions and the imagination. They made up a broad modernist heritage in which the ‘irrational fantasies’, as Rem Koolhaas called them, about the future city came to full light (Koolhaas 1997, 173). Art Deco as Aesthetic Mélange Edward Durell Stone, principal architect; Donald Deskey, interior designer (Source: Wikipedia Commons). 2 were an immediate sensation and spurred Deskey to join with Hood, Harrison, and Viennese designer Paul Frankl in promoting modern design and architecture. Deskey in troduced mass-produced tubular metal furniture based on designs by Marcel Breuer and Ludwig Mies van der Rohe he had seen at the Bauhaus. But European modernism was not the only source of inspiration for American design ers. They turned to the burgeoning American consumer industries as mechanisms for a revolution in everyday life. It was evident in Deskey’s 20-year association with the Procter & Gamble Company during which he designed dozens of domestic household products. for carrying out this vision and reconceptualising spatial context and texture. Colour and artificial lighting were an integral component of this theatrical built environment. Journalists remarked on Radio City Music Hall’s scale, its showmanship, with its interior ‘flooded with enough light, mostly coloured light, to illuminate the city of Minneapo lis’. Lighting was treated by modernists as pure colour and form. The illuminating experts ‘can weave together effects like aurora borealis, sunrise at sea, shooting-star showers, burning cities …’. The architects used tricks of perspective to shorten distances with the walls and ceilings curving into the stage (New Yorker Magazine 1932, 48–50). Like Art Deco itself, illumination was exuberantly decorative and ostentatiously ornamental. It was the moderne style that dazzled and thrilled the crowds, and augured the urban future. Deskey’s design of Radio City Music Hall promised avant-garde architecture of the 1930s as visual theatre. Deskey worked with theatrical set designer Edmond Jones to fashion the interior of Radio City Music Hall as a bold spectacle of cubist forms, rich materials, glowing metal surfaces, with a sumptuous Art Deco golden crown en veloping the stage (Okrent 2003, 222–230). Marble and gold foil mixed with aluminium and cork in a blaze of decorative light fixtures and enormous murals (Figure 2). The North China Herald pronounced it ‘the largest and best theatre in the world’ (29 March 1939). The theatre revealed the lighting effects that were essential to creat ing mid-century architecture and design as theatre. In his Event Cities, Bernard Tschumi argued that architecture is both a performance and an event. It is a mise-en-scene where vectors, voids, solids, envelopes and flow negoti ate the delicate relationship between space, movement and action (Tschumi 1994, 13). Art Deco as Aesthetic Mélange A host of official American delegates from organisations such as the American Institute of Architects wandered through the 1925 Paris Exposition, stunned at the radi cally new design in architecture, furniture, textiles and graphics. In February the next year, New York’s Metropoli tan Museum of Art welcomed visitors to their miniature version of the Paris Exposition with ‘A Selected Collection of Objects’. New York’s most prominent architects of mid- century modernism, individuals such as Harvey Wiley Corbett, Raymond Hood, Wallace Harrison, and Edmund Durell Stone, had all been trained in Europe, primarily at the Ecole des Beaux Arts, and were immediately inspired by the designs they viewed at the 1925 Exposition. These were the designers behind Rockefeller Center and Radio City Music Hall, perhaps the penultimate examples of New York’s avant-garde Art Deco architecture. They set off on a European grand tour, searching out avant-garde theatre designs and meeting up with German architects such as Expressionist Erich Mendelsohn to see his stream lined Universum Cinema and indulging in discussions with Peter Behrens and Walter Gropius. Gropius had ap plied his concept of ‘total theatre’ to the architecture and fi Although the European avant-garde was a touchstone, New York’s version of Art Deco was more commercial, more geared to the entertainment industry and certainly more flamboyant. Donald Deskey was characteristic of mid-century modernists such as Wallace Harrison and Raymond Hood, as well as Edward Durell Stone. Schooled at the Ecole des Beaux Arts, Deskey made a second trip to Paris in 1925 to attend the Exposition des Arts Déco ratifs. The spectacle of chrome and steel, the plastic, the straight lines and sharp edges were a revelation for him. He continued his sojourn in Europe with a pilgrimage to the Bauhaus in Dessau. On returning to New York, he was hired to decorate the windows of department stores along New York’s famed Fifth Avenue. Like many cutting-edge American designers of this period, he worked in advertis ing and sallied back-and-forth between theatre design, marketing, and industrial design. The avant-garde deco rative drama he concocted for department store displays 51 R. Wakeman Figure 2 Radio City Music Hall, Rockefeller Center, interior, audi torium with view toward Great Stage. Edward Durell Stone, principal architect; Donald Deskey, interior designer (Source: Wikipedia Commons). 2 Figure 2 Radio City Music Hall, Rockefeller Center, interior, audi torium with view toward Great Stage. Art Deco as Aesthetic Mélange Art Deco was a formula Like their New York counterparts, Chinese architects and artists such as Liu Jipiao and Lin Fengmian made the pilgrimage to Paris, enrolled in the Ecole des Beaux Arts and toured the 1925 Paris Exposition. Liu Jipiao de signed the interior of the Exposition’s Chinese Pavilion in a flamboyant moderne style. Art Deco arrived in Shang hai through a pair of young French architects, Alexandre Leonard and Paul Veysseyre. They began work on the Cercle Sportif Français, a lavish country club in the west ern part of the French Concession. Opened in 1926, the French Club, as expatriates called it, presented a conserva tive Beaux-Arts exterior fronting its style moderne interior in imitation of that showcased at the 1925 Paris Exposi tion. Leonard and Veysseyre went on to complete over a 52 BUILT HERITAGE 2019 / 3 Figure 3a Park Hotel, Laszlo Hudec architect, 1934 (Source: Virtual Shanghai.net). Figure 3b American Radiator Building, Raymond Hood architect (Source: Irma and Paul Milstein Di vision of United States History, The New York Public Library. Accessed June 27, 2019). 3a 3b Figure 3a Park Hotel, Laszlo Hudec architect, 1934 (Source: Virtual Shanghai.net). Figure 3b American Radiator Building, Raymond Hood architect (Source: Irma and Paul Milstein Di vision of United States History, The New York Public Library. Accessed June 27, 2019). 3a Hudec ual or chitect ein Di ory, The essed 3a 3b Figure 3a Park Hotel, Laszlo Hudec architect, 1934 (Source: Virtual Shanghai.net). Figure 3b American Radiator Building, Raymond Hood architect (Source: Irma and Paul Milstein Di vision of United States History, The New York Public Library. Accessed June 27, 2019). Figure 3a Park Hotel, Laszlo Hudec architect, 1934 (Source: Virtual Shanghai.net). Figure 3b American Radiator Building, Raymond Hood architect (Source: Irma and Paul Milstein Di vision of United States History, The New York Public Library. Accessed June 27, 2019). Figure 3a Park Hotel, Laszlo Hudec architect, 1934 (Source: Virtual Shanghai.net). Figure 3b American Radiator Building, Raymond Hood architect (Source: Irma and Paul Milstein Di vision of United States History, The New York Public Library. Accessed June 27, 2019). 3a monument of American business and enterprise, and an American visitor to Shanghai remarked that the skyscrap ers along the Bund were ‘monuments to the prosperity of the owners and to the possibilities of Eastern trade’ (Frankl 1928, 61; New York Times, 25 February 1923). Art Deco as Aesthetic Mélange They were the world of big business and international commerce. The advent of the New York skyscraper as ideal of the future meant an Americanisation of architectural pro duction and a rising demand for American materials and architectural know-how. Surveying the dizzying height of New York’s skyscrapers (Figure 4), a Shanghai journal ist remarked ‘There literally seems to be no limit … The modern skyscraper is gradually producing a new species of human beings.’ Architect George Wilson reported in The China Journal on how ‘the pleasing effect of the modern American skyscraper’ in New York came about (Wilson 1932, 238). The American Institute of Architecture, the United States government, as well as American educa tional and philanthropic organisations championed the growing presence of American architects and builders in Shanghai. Foreign businesses confronted an increasingly powerful Chinese nationalist movement that targeted the treaty system and foreign privilege. Yet even in the chaotic Republican era, city-building functioned around avant- garde international networks. American engineering firms were involved in the construction of new electric power plants and substations in Shanghai as well as the expan sion of telephone service. Architectural schools such as the University of Pennsylvania and Pratt Institute in New York were opening their doors to Chinese students such as hundred commissions, most of them in the Art Deco style. But they were not the only champions of the eclectic mod ernism of Shanghai. There were also Hungarian-Slovak László Hudec, Englishman George ‘Tug’ Wilson, Russian Alexander Yaron, Spaniard Abelardo Lafuente (see the article by Álvaro Leonardo Pérez in this issue), and an ex panding roster of foreign-trained Chinese architects such as Robert Fan and Dong Dayou. Following the pattern seen elsewhere around the world, the city’s architects com bined a potpourri of architectural traditions with the more ‘universalist’ imported elements of Art Deco. László Hudec represented this fusion of patterns. Like many of the refu gees fleeing the former Austro-Hungarian Empire, Hudec was early on influenced by the atmospheric qualities of German Expressionism, especially that of Erich Mendel sohn. By dint of refugee status, Hudec along with other expatriate architects became ‘citizens of the world’. He maintained contact with the European architectural avant- garde throughout his career. Hudec initially worked for the American architectural firm of Roland Curry in Shanghai. Figure 3a Park Hotel, Laszlo Hudec architect, 1934 (Source: Virtual Shanghai.net). Figure 3b American Radiator Building, Raymond Hood architect (Source: Irma and Paul Milstein Di vision of United States History, The New York Public Library. Accessed June 27, 2019). Art Deco as Aesthetic Mélange A trip to America to study New York’s skyscrapers pro foundly shaped his ideas, especially the verticality of build ings by architect Raymond Hood (Figures 3a, Figure 3b). Hudec’s landmark Park Hotel in Shanghai was based on his sketches of Hood’s American Radiator Building (dedicated to the new heating technology) at New York’s Bryant Park (Poncellini 2010). The Park Hotel was Shanghai’s first sky scraper and long the city’s tallest building. In his book New Dimensions (1928), New York-based Viennese designer Paul Frankl defined skyscrapers as the 53 R. Wakeman BUILT HERITAGE 2019 / 3 4 5 Figure 4 Skyscrapers of New York City: aerial view of Battery Park and the skyscrapers of lower Manhattan, looking north wards, 1929 (Source: New York Public Library Digital Collections. Accessed June 26, 2019). Figure 5 Nanjing Road (Source: Japanese book Nostalgia for Shanghai published by Kokusho-kankoukai). 4 5 Figure 4 Skyscrapers of New York City: aerial view of Battery Park and the skyscrapers of lower Manhattan, looking north wards, 1929 (Source: New York Public Library Digital Collections. Accessed June 26, 2019). Figure 5 Nanjing Road (Source: Japanese book Nostalgia for Shanghai published by Kokusho-kankoukai). 4 Figure 4 Skyscrapers of New York City: aerial view of Battery Park and the skyscrapers of lower Manhattan, looking north wards, 1929 (Source: New York Public Library Digital Collections. Accessed June 26, 2019). Figure 5 Nanjing Road (Source: Japanese book Nostalgia for Shanghai published by Kokusho-kankoukai). 54 BUILT HERITAGE 2019 / 3 Figure 6 Wing On Department Store, Art Deco Tower, architects Elliot Hazard and Edward Phillips (Source: Peter Hack on abc.net.au). 6 Figure 6 Wing On Department Store, Art Deco Tower, architects Elliot Hazard and Edward Phillips (Source: Peter Hack on abc.net.au). 6 Liang Sicheng, Yang Tingbao, Chen Zhi, Zhao Shen, and Lin Huiyin. They were instrumental in bringing Ameri can practices and standards to China and merging them with local aesthetic traditions and heritage. With architec tural offices in Shanghai, the New York firm of Henry K. Murphy employed Art Deco trendsetters Elliot Hazzard and Poy Gum Lee as well as a number of Chinese archi tects. Cathay Mansions was likened to ‘… one of the more magnificent apartment houses in New York City …’ (North China Daily News, 14 November 1932 and North China Herald, 19 May 1932). Art Deco as Aesthetic Mélange The richness of display on Nanjing and Bubbling Well roads made them into Shanghai’s ‘Great White Way’ in imitation of Broadway (Figure 5). Art Deco building mirrored construction in New York, where caste aluminium spandrels were also being used, for example, in Rockefeller Center. Concrete, steel fram ing and aluminium were then adapted to the structure of Broadway Mansions in Shanghai by Scottish architect Bright Fraser. Shanghai’s Art Deco buildings introduced the automatically-controlled Otis elevators and escalators, the air-conditioning and heating, the automatic sprinklers and fire alarm systems, that made skyscrapers practical and also made them technological marvels. Even the rein forced pilings that made raising the towers possible in the Shanghai mud were based on new engineering technolo gies. Architectural modernism was tightly bound with this machine culture. The majority of these engineering and construction feats were done by American and German companies although Chinese firms were also subcontract ed for building construction. Art Deco as Technics These references to an American-style Deco design evi denced the increasingly influential role New York played in the Shanghai imagination. New York architects Elliot Hazard and Edward Phillips designed the 26-storey Art Deco tower extension of the Wing On Department Store (Figure 6). It was a mark of the intimacy between Art Deco design, advertising and commercialism. The tower rose over Nanjing Road across from the Sincere Depart ment Store and fashioned the intersection into what was dubbed Shanghai’s ‘Times Square’ (Denison 2017, 277). Hazzard also designed the Shanghai Power Com pany Building on Nanjing Road. The reinforced concrete The Art Deco buildings rising along Nanjing Road and on the Bund were conjured from this technological wiz ardry. They emerged alongside the development of pho tography and cinematography, radiography, telephony, electricity, and engineering. ‘Modern genies who, on a flat mud bank, rear cement and concrete palaces to the sky …’ reported the North China Daily News (14 Novem ber 1932). Technology made possible the functionality, the movement and circulation that were the unquestion able basis of modernist architecture as well as modern life. Technology was not merely contextual, but instead 55 R. Wakeman Figure 7 Sassoon House, ’Radio Central of China’, Far Eastern Review (March 1931): 182. 7 Central Radio Office, installed in Shanghai a. Perforating apparatus; b. Automatic transmitter; c. Undulator; d. Direct printing machines. Figure 7 Sassoon House, ’Radio Central of China’, Far Eastern Review (March 1931): 182. Raymond Hood searched for an aesthetic style in keeping with avant-garde media that were ‘entirely new’. Hood was among the most celebrated architects of the age and had absorbed a riot of European influences, especially from Finish architect Eliel Saarinen as well as Walter Gropius and Adolf Meyer. Hood ‘resorted to light’ and ‘developed an architectural effect that depended almost entirely on changing lights. Sheets of light, changing in tone and colour, stream up the wall … Other lights glow from behind great pilasters … The fixtures themselves, gigantic for lighting fixtures, have inside of each one an almost complete stage electrical equipment … At the state end, a glowing screen was ingeniously devised that has almost the qualities of a colour organ. Practically, we adapted all of the tricks of stage lighting to the architec ture of a room’ (Hood 1928, 6). Art Deco as Technics The result was Art Deco as techno-theatre for a new media that drew three-quar ters of a million people to the live broadcasts in its first year of operation. was constitutive of modernist aesthetics (Danius 2002, 2–3) and urban heritage. Passing through their doors into sumptuously decorated interiors, the public found telegraph and telephone exchanges, and ticker tapes for real-time information on markets. The buildings were al legories to global communication. They were the nodal points that connected New York and Shanghai to the world and the mechanisms for global urban pretensions. It was precisely this euphoria about the machine age that Art Deco aesthetics attempted to capture. The ‘Radio Cen tral of China’ (Figure 7) was located in Shanghai’s iconic Art Deco citadel, the Sassoon House. The ground and first floors were jam-packed with transmission equip ment and offices where ‘messages received from all parts of the world are made ready for delivery’. It was ‘in direct and permanent day and night communication with Paris’ (Shecklen 1930, 290; Pavlovsky 1931, 182). This broader understanding of the heritage of modernism that em braces the technologies of movement and communication goes beyond decorative appearances to the complex fabric of structures and spaces that impacted people’s lives. Her itage becomes a pathway to understanding the media rev olution and future building design. In 1926, architect and stage designer Joseph Urban designed the New York headquarters of the Hearst Pub lishing empire to honour his hero, the distinguished Viennese architect Otto Wagner. High-fluted columns topped with urns reproduced the imagery of the 1925 Paris Exposition, while Art Deco colossal statues at the base symbolised the publishing industry. The build ing exuded Urban’s expertise as a theatrical stage de signer. Urban was trained in Vienna and took part in the Jugendstil movement that applied modernist prin ciples to architecture and theatre design. He was heavily Art Deco as Theatre Theatrical tactics and stage machinery were just as essen tial to modernist perception. In designing the thirty-five radio and television studios for the National Broadcast ing Company at Rockefeller Center (Figure 8), architect 56 BUILT HERITAGE 2019 / 3 ckefeller Center, dio, architect Ray ource: Museum New York). h Urban, Ziegfeld Front, Watercol 1926–1927 h Urban Collec k & Manuscript bia University). ph Urban, Vene Ziegfeld Follies e: Joseph Urban ook & Manuscript bia University s III. New York). 8 9 10 8 8 8 9 10 Figure 8 30 Rockefeller Center, Television Studio, architect Ray mond Hood (Source: Museum of the City of New York). Figure 9 Joseph Urban, Ziegfeld Theater, Street Front, Watercol our, New York, 1926–1927 (Source: Joseph Urban Collec tion, Rare Book & Manuscript Library, Columbia University). Figure 10 Joseph Urban, Vene tian Stage Set, Ziegfeld Follies of 1921 (Source: Joseph Urban Papers, Rare Book & Manuscript Library, Columbia University Libraries. Series III. New York). Figure 8 30 Rockefeller Center, Television Studio, architect Ray mond Hood (Source: Museum of the City of New York). 10 influenced by the Wiener Werkstätte. On immigrating to the United States, he opened a Wiener Werkstätte shop on Fifth Avenue to advertise their sleek crafts. In Urban's work, the theatre became an implied metaphor (Figure 9). He designed striking sets for the Ziegfeld Follies in New York, for the Metropolitan Opera and the Broad way stage. Urban argued that the inspiration for modern theatre design must come from the architect rather than the dramatist (Figure 10). Openness, elasticity of func tion, enveloping the audience in the production through the use of colour and lighting effects were essential to the democratic theatre (Urban 1929). Urban’s buildings exuded the same drama as his theatrical stage design. They were a scenographic blend of European modernism and American commercial design. His innovative use of colour and lighting, his radical forms made his designs among the most avant-garde in New York. He employed them as well at the 1933 Century of Progress Exposition in Chicago, where he concocted an elaborate colour and lighting scheme that bathed the buildings in a ‘colour ge ography’ and unified the fairground design around the paths of sun and shade. Art Deco as Colour and Light In explaining his own techniques of theatrical design, American architect Claude Bragdon found parallels with architecture as a ‘permanent stage set’ of ‘dream pictures, 57 R. Wakeman Figure 11 Claude Fayette Bragdon, Coloured Decorative Design (Source: The Miriam and Ira D. Wallach Division of Art, Prints and Photographs: Art & Architecture Collection, The New York Public Library. ‘Coloured decorative design’ New York Public Li brary Digital Collections. Accessed June 27, 2019). Figure 12 The Great White Way: Broadway north from 45 th Street, 1923 (Source: The Miriam and Ira D. Wallach Division of Art, Prints and Photographs: Photography Collection, The New York Public Library. ‘Broadway north from 45th St.’ New York Public Library Digital Collections. Accessed June 27, 2019) . 11 12 11 11 Figure 11 Claude Fayette Bragdon, Coloured Decorative Design (Source: The Miriam and Ira D. Wallach Division of Art, Prints and Photographs: Art & Architecture Collection, The New York Public Library. ‘Coloured decorative design’ New York Public Li brary Digital Collections. Accessed June 27, 2019). Figure 12 The Great White Way: Broadway north from 45 th Street, 1923 (Source: The Miriam and Ira D. Wallach Division of Art, Prints and Photographs: Photography Collection, The New York Public Library. ‘Broadway north from 45th St.’ New York Public Library Digital Collections. Accessed June 27, 2019) . adapted to advertising and became integral to architectur al aesthetics. Focusing on lighting underscores building form as a dreamscape of colour, light, and glass. It inter twines the heritage of Art Deco with the legacy of the city as theatre. In a Shanghai review of avant-garde cinema design in Europe and the United States, the techno-drama of the Théâtre Pigalle in Paris is a model in which light and machinery create magical staging. For its architect, ‘the façade of a theatre is only a curtain-wall, the function of which is to serve as a background for a blazon of luminous letters ...’. The exterior lighting sets the dramatic atmos phere for the theatre interior conceived as ‘pure architec ture—the proportion of masses in voids and solids, light and shade …’ (North China Daily, 14 November 1932). These were the touchstones for aesthetic gratification and experiential authenticity. Theatrical machinery combined the sensory with architectural aesthetics. as if seen in the depths of some great crystal sphere’ (Figure 11). The effect was produced through colour and lighting (Bragdon 1929, 109 and 114). Art Deco as Colour and Light The taste for spec tacular staging effects was derived from the emergence of modernity in the 19th century as well as with the ‘ar chitecture of light’ of European modernists. Illusions to the ‘architecture of light’ and to glass crystals were widely shared among avant-garde architects of this period, par ticularly by figures such as Walter Gropius, László Mo holy-Nagy, and Bruno Taut in their writings of the 1920s and 1930s. Architect George Wilson famously designed Shanghai’s Sassoon House around the Lalique glass he had admired in Paris. When the building opened in 1929, ‘… one almost loses interest in everything except the wonderful Lalique lighting, which is a riot of beauty … one colour melting into another, giving the effect of sun rise and sunset, blazing day and moonlight’ (North China Herald, 2 August 1929). Architectural historian Cary Y. Liu has argued that ‘ar chitecture’ comprises both material and immaterial con nections, as well as aesthetic concepts and standards of taste (Liu 2014, 118). Imagining built heritage as a broad phenomenon across the technological and theatrical land scape shifts our gaze to the cityscape as magical theatre in places such as the Bund, Nanjing Road, or Times Square. First introduced in Times Square, dazzling displays of moving and flashing neon lights attracted huge crowds of people (Figure 12). The content of the ads was overshad owed by the electric display. The spectacle changed the Most emblematic of the vintage glow of urban modern ism was neon, first unveiled by French inventor Georges Claude at the Paris Motor Show in December 1910. The new methods of electrical illumination developed in the 1920s and 1930s such as neon gas-filled tubes, indirect lighting, luminous coloured lighting techniques made the built form and the spaces of New York and Shanghai into wildly aesthetic wonderlands. In Paris, officials quickly imposed legal restrictions on the use of electric advertis ing signage. In New York and Shanghai, neon was quickly 58 BUILT HERITAGE 2019 / 3 13 Figure 13 Night view of the Shang hai department stores, 1937. Claude Berruyer photographer (Source: Virtual Shanghai.net). Figure 13 Night view of the Shang hai department stores, 1937. Claude Berruyer photographer (Source: Virtual Shanghai.net). 13 commercial signage became a new kind of globalised urban semiotics and a new kind of heritage. Mao Tun cap tured the immediacy of this visual theatre in Midnight, his seminal 1933 novel of Shanghai. Art Deco as Colour and Light Looking out over the cityscape ‘one saw with a shock of wonder on the roof of a building a gigantic neon sign in flaming red and phospho rescent green: LIGHT, HEAT, POWER!’ The ‘kaleidoscope of red, yellow, green, black, shiny, square, cylindrical, leap ing, dancing shapes’ overwhelm his protagonists (Tun 2001, 1, 8). Some of the sky-signage, such as the glitzy Wrigley Spearmint Chewing Gum illumination in New York, became crowd favourites and was reproduced end lessly in photography and film as the signature of Times Square (Taylor 1991, 240–242). Others were temporary, perishable, remembered only by people who enjoyed their fleeting exoticism. In Shanghai Fox-Trot, neo-sensationalist writer Mu Shiying depicted the ‘Neon light stretches out a coloured finger and writes a large character in the black- ink night. An English gentleman stands in front, wearing a red swallow-tail coat, gripping a cane, vigorously strid ing along. Below his feet is written: “Johnny Walker: Still Going Strong”’ (Mu 1934). This kind of urban spectacle was by its very nature an ephemeral art. In celebration of the 1937 coronation of King George VI, ‘Shanghai burst into a joyous, display of myriad coloured lights from one end of the Bund to the other skyscraper after skyscraper blazed forth brilliantly in the night sky, making the whole waterfront a mass of twinkling light …’ (North China Herald, 19 May 1937) The riot of colour and light on the marquees in the theatre and cinema district along Broad way in New York, and the skyscrapers and cinema palaces of Shanghai made the city into illusionistic performance. way people experienced space and made their engagement with the city a form of ‘cinematic spectatorship’ (Cordon 2010, 65). The cityscape was conceived as a virtual stage set through which people moved, mesmerised by the il lusionary marvels. Electric theatres such as this revealed the potential of urban space as radical avant-garde tactic. The built form of the city became sensuous display. The urban panorama was to be looked at as an image or object of aesthetic pleasure, a position increasingly reproduced in photography and film of the 1920s and 1930s. Art Deco as Colour and Light For ex ample, the 1933 Soviet-style avant-garde montage film City Nights by director Fei Mu, captures the carnivalesque blaze of light along the Bund, the city’s signature ‘poster buildings’ outlined in luminous neon, the dazzling ad vertising displays on the Wing On, the Sincere, and Sun department stores along Nanjing Road. More than merely architectural ornament, the lavish illumination dominated cultural identity and transformed Shanghai and New York into global mass media (Figure 13). While architecture itself moved increasing toward an abstracted, minimalist modern style, the urban panorama remained ornamental, sheathed in light, ever more evocative of urban identity and experience. In good part, it was the ultimate mark of the ‘society of spectacle’ and demonstrated the phantas magoria of commodification that haunted people's daily lives and, according to the urban theorists of the Frankfurt School, ultimately subjugated them. Art Deco as Commerce There was an intimate relationship between the avant- garde and the sorcery of commercial advertising. Glittering 59 R. Wakeman 14 Figure 14 Cathay Theatre, 1949 (Source: Virtual Shanghai.net). Figure 14 Cathay Theatre, 1949 (Source: Virtual Shanghai.net). Figure 14 Cathay Theatre, 1949 (Source: Virtual Shanghai.net). This urban spectatorship was mirrored in the Para mount Ballroom and Cathay Cinema, the Grand Theater, the cabarets and film palaces, bathed in light and decora tive splendour. Along with theatre, film was also a source of inspiration, at times directly overlapping with archi tectural design. Joseph Urban for example designed film sets and directed numerous films for William Randolph Hearst’s Cosmopolitan Productions. Cinema was the most democratic, universalist form of mass entertainment. The introduction of sound brought an entirely different qual ity to the illusion provided by motion pictures. Sumptuous movie theatres such as the Rialto and Roxy in New York and the Paramount and Cathay in Shanghai belonged to an era of fantasy and mirage, one that Art Deco design indulged in with abandon. The Paramount’s glass tower sheathed in red and green lights could be seen for miles away (see the article by Andrew Field in this issue). The Cathay Cinema opened on January 1, 1932 with an Ameri can movie called Free Soul. Until 1949 it played English and American Films. It was designed by Czech architect Charles Henry Gonda, who also designed Shanghai’s Capi tol Theater and the Sun Department Store. Central Euro peans added to the jumble of nationalities in Shanghai. Gonda’s theatrical imagination was sharpened in Vienna and Paris where he mixed with the European avant-garde movements that were transforming architecture and design. Conda used lighting to extraordinary effect in the Cathay. The 1000-seat auditorium, done in ‘American style’, bathed the glass ceiling beams, the bronze and gold colour scheme in indirect lighting. The Cathay logo shot up into the sky along the building’s central spire (Figure 14). The North China Herald reported that it was ‘the first time in the Far East a lighting scheme was created which architec turally constitutes a perfect unit with the decoration of the Auditorium’ (5 January 1932). It was a frank assertion that a building façade was a form of advertising—a marquee. With their neon-lite spires and marble facades, the movie palaces became iconic fixtures in the city. Art Deco as Commerce A moviegoer ef fused, ‘When night falls, the foreign names in neon lights on their fronts: Grand Theatre or Cathay … glow in their alluring and glittering colours’ (Poshek 2003, 34). Conclusion Art Deco played a significant role in the dialogue be tween local place and globalisation. It was a signature avant-garde production that embraced theatrics and technological wizardry and spectacularised the built en vironment. Both Shanghai and New York were captivated by this thrilling visualisation of the modern city. The influences that made up their urban ‘styles’ in the 1920s and 1930s were many. Similar avant-garde influences in spired architects in both cities. They shared inspiration from the European experiments, particularly as it was interpreted in Paris, Berlin, and more generally in central Europe. They shared in the excitement of new technolo gies and the brilliance of a nighttime panorama illumi nated by electric stars. These facets that were so emblem atic of modern urban life open a rich perspective on built heritage as do mass culture and consumption, popular entertainment and urban spectacle. Often considered ephemeral or inconsequential, these were decisive to the 60 BUILT HERITAGE 2019 / 3 Paul T.Frankl. 1928. New Dimensions; the Decorative Arts of Today in Words & Pictures. New York: Payson & Clarke. production of the urban fabric and key to what made Shanghai and New York distinctive. Not only are they an expressive fragment of the cultural legacy of the past that deserves acknowledgement. They have also inspired a discourse about both local identity and the nature of urbanity that remains to the present day. In part this nar rative is nostalgic, in part it references the globalising influences and creative forces that produced the iconic New York and Shanghai ‘styles’. Pavlovsky, M., and H. Sauve. 1931. “French Equipped International Wireless Station in Shanghai.” The Far Eastern Review, March. Poncellini, Luca. 2010. Hudec László. Budapest: Holnap Kiadó. Poshek, Fu. 2003. Between Shanghai and Hong Kong: The Politics of Chinese Cinemas. Stanford, CA: Stanford University Press. Shecklen, George F. 1930. “Shanghai, the Radio Central of China.” Far Eastern Review, June. References Bragdon, Claude. 1929. “The Technique of Theatrical Pro duction.” The Architectural Record 66 (2). Sowerby, Arthur. 1928. “Shanghai’s Position in the World.” The China Journal 3(8). Cordon, Eric. 2010. The Urban Spectator: American Con cept-Cities from Kodak to Google. Lebanon, NH: Dart mouth. Taylor, William R. 1991. Inventing Times Square: Com merce and Culture at the Crossroads of the World. Balti more & London: John Hopkins University Press. Danius, Sara. 2002. The Senses of Modernism: Technology, Perception, and Aesthetics. Ithaca, NY: Cornell Univer sity Press. Tschumi, Bernard. 1994. Event Cities. Cambridge, MA: MIT Press. Tun, Mao. 2001. Midnight. Amsterdam: Fredonia Books. Delanty, Gerard. 2006. “The Cosmopolitan Imagination: Critical Cosmopolitanism and Social Theory.” The British Journal of Sociology 57 (1): 25–47. Denison, Edward. 2017. Architecture and the Landscape of Modernity in China before 1949. London & New York: Routledge. Gropius, Walter, and Oskar Schlemmer, eds. (1924) 2014. The Theater of the Bauhaus. Reprint, Middletown, CT: Wesleyan University Press. Hood, Raymond. 1928. “The National Broadcasting Stu dios New York.” The Architectural Record 64 (1). Koolhaas, Rem. 1997. Delirious New York: A Retroactive Manifesto for Manhattan. New York: Monacelli Press. Lee, Ou-fan Leo. “Shanghai Modern: Reflections on Urban Culture in China in the 1930s.” In Alternative Moder nities, edited by Gaonkar Dilip Parameshwar, 86–122. Durham and London: Duke University Press. Liu, Cary Y. 2014. “Encountering the Dilemma of Change in the Architectural and Urban History of Shanghai.” Journal of the Society of Architectural Historians 73 (1): 118–136. Mu, Shiying. 1934. “Shanghai Fox-Trot.” [Shanghai de hubuwu.] Xiandai 2 (1): 4. Translated by Andrew Field at http://shanghaisojourns.net/blog/2011/3/20/shang hai-fox-trot.html Mumford, Lewis. 1933 “The Skyline: Skyscrapers and Ten ements.” The New Yorker, June 3. Okrent, Daniel. 2003. Great Fortune: The Epic of Rockefel ler Center. New York: Viking. 61 R. Wakeman R. Wakeman
https://openalex.org/W3042464714	https://journals.us.edu.pl/index.php/RSL/article/download/9458/7338	Russian	null	Когнитивнао-нарративная реализация концептов «совесть» и «добро» в дискурсе летературных журналов	Rusycystyczne Studia Literaturoznawcze	2,019	cc-by-sa	7,552	COGNITIVE AND NARRATIVE IMPLEMENTATION OF THE CONCEPTS OF “CONSCIENCE” AND “GOOD” IN THE DISCOURSE OF LITERARY JOURNALS The article deals with the problem of value transformations in modern Russian society, the transla- tor of which are “thick” literary and social and political journals . The authors explore this process by analyzing the cognitive and narrative mechanisms implementing the concepts of “conscience” and “good” in magazines. g g Keywords: axiology, narrative, cognitive script, literary journals, Russian society. КОГНИТИВНО-НАРРАТИВНАЯ РЕАЛИЗАЦИЯ КОНЦЕПТОВ «СОВЕСТЬ» И «ДОБРО» В ДУСКУРСЕ ЛИТЕРАТУРНЫХ ЖУРНАЛОВ Rusycystyczne Studia Literaturoznaw Вера Антропова Члябнскй гсударстнный унрстт ORCID https://orcid.org/0000-0003-3421-9978 Василий Федоров Члябнскй гсударстнный унрстт ORCID https://orcid.org/0000-0001-6295-5054 КОГНИТИВНО-НАРРАТИВНАЯ РЕАЛИЗАЦИЯ КОНЦЕПТОВ «СОВЕСТЬ» И «ДОБРО» В ДУСКУРСЕ ЛИТЕРАТУРНЫХ ЖУРНАЛОВ COGNITIVE AND NARRATIVE IMPLEMENTATION OF THE CONCEPTS OF “CONSCIENCE” AND “GOOD” IN THE DISCOURSE OF LITERARY JOURNALS Rusycystyczne Studia Literaturoznawcze nr 29 (2019) ISSN 2353-9674 DOI https://doi.org/10.31261/RSL.2019.29.03 Rusycystyczne Studia Literaturoznawcze nr 29 (2019) ISSN 2353-9674 DOI https://doi.org/10.31261/RSL.2019.29.03 e Studia Literaturoznawcze 2019 (29) POZNAWCZO-NARRACYJNA REALIZACJA POJĘĆ „SUMIENIE” I „DOBRO” W DYSKURSIE CZASOPISM LITERACKICH Artykuł porusza problem przemian wartości we współczesnym społeczeństwie rosyjskim, których no- śnikami stają się literackie, artystyczne, społeczne i polityczne „tołstyje żurnały”. Autorzy badają ten proces, analizując poznawczo-narracyjne mechanizmy realizacji pojęć „sumienie” i „dobro”, przejawia- jących się w dyskursie tychże czasopism. ją y ę y y p Słowa kluczowe: aksjologia, narracja, skrypt poznawczy, czasopisma literackie, społeczeństwo rosyjskie. ją y ę y y p Słowa kluczowe: aksjologia, narracja, skrypt poznawczy, czasopisma literackie, społeczeństwo rosyjskie. Акслгчскй крзс сх сфрах жзн рссйскг бщста прл к цннстнй днант научных зыска- 54 нй бласт гуантарстк. Флсфы, культурлг, сцлг цнают срнную стуацю как «цннст- ный акуу», «цннстн-нратную нпрдлннсть», «цннстный рлятз нглз», бусллнны «б- щ снжн цннстнг статуса рал, нрастн- нст, разытстью рнтр пнан дбра зла», «упадк сцальнг атртта традцнных цннстй», «прфанзацй д лчнй ттстннст лчнг ыбра»1. Днака ж така, чт Рсс эт прцссы «дут напралн услня дффрнцац плюрал- зац нратн-цннстных сст»2. Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Журналстка как сцальный нсттут журналстск трчст как д кратнй дятльнст п пррд сй был стаются цннстн аркранны, с- гдня раках журналстскй наук, п нню нктрых спцалст, знкла даж тдльная учбная дсцпл- на научная парадга — акслгя журналстк (Вктр Алксандрч Сдр, Калла Рнатна гатуллна, Сргй Сргч Ильчнк), прзанная стуац культур- нг нгбразя тразть цннстный кнфлкт дйнй ральнст. Вра Антрпа, Васлй Фдр Когн Ункальны прдукт срннй дакультуры я- ляются так назыаы «тлсты» тчстнны журналы тпа «Октябрь», «зда», «Дружба нард», «а», «- ый р», «ная», «аш срннк» нктры друг, схдящ к лтратурны, сатрчск энцклпд- чск журнала XVIII–XIX к, а пзж — к стск «тлсты» журнала. 1 Т.В. Тлнна, Причины кризиса ценностных ориентаций в современных российских семьях // «Тря практка бщстннг разтя» 2012, № 6, http://teoria-practica.ru/rus/files/arhiv_zhurnala/2012/6/scilgiyа/tili- nina.pdf (5.01.2019). 2 С.В. Мара, Динамика норм и ценностей россиян // «Сцс» 2013, № 7, с. 120. 3 И.В. Шкарская, Становление и развитие советских «толстых» жур- развитие советских «толстых» жур- развитие советских «толстых» жур- налов в первой половине 20-х годов: автореф. дис. канд. филол. наук, М- ска 1999, http://cheloveknauka.com/stanovlenie-i-razvitie-sovetskih-tols tyh- zhurnalov-v-pervoy-polovine-20-h-godov (5.01.2019). 4 Офцальный сайт журнала «аш срннк»: http://www.nash-sovre- mennik.ru (5.01.2019). POZNAWCZO-NARRACYJNA REALIZACJA POJĘĆ „SUMIENIE” I „DOBRO” W DYSKURSIE CZASOPISM LITERACKICH Стск «тлсты» журналы задуы- алсь, с днй стрны, как прнк дрлюцнных журнал, с другй — эт был прнцпальн н куль- 55 турн-эсттчск ялн журнальн пл стскй пчат, пскльку н знкл нн 20–30- гды ХХ ка как «сздатл ыразтл гсударстннй плт- к бласт культуры», прзанны «набл эффктн здйстать на бщстнн сзнан»3. И нстря на т, чт пдбны жсячны лтратурн-худжстнны бщстнн-плтчск журналы прдсталяют сбй сгдня тдльный — элтарный — сгнт журнальнй п- рдк Рсс, пртстящй друг нфрацнны птка, даж г структур наблюдатся длгчская культурн-эсттчская дскртнсть разбщннсть. Так, с днй стрны, нахдятся радкальны, ппзцн- ны зданя, прдящ длгю нацнальнг кнср- атза, пддржающ традцнны патрархальны цннст («аш срннк»), с другй стрны, журналы, рнтранны на западны лбральн-дкратчск далы («ная»). «аш срннк», прначальн п- чаташй прдстатлй «дрнскй прзы» — Фдра Абраа, Васля Бла, Сргя алыгна, Валнтна Рас- путна, Владра Слухна, Васля Шукшна — бъ- ял сбя «трбунй днйшх плтк патртч- скг напралня», ппзцй прзападны устанка, «дущ к разрушню Рсс гбл русскг нарда»4. Журнал «ная», задуанный как здан лтратурнг бъдння Краснй Ар Флта, с 1934 п 1990 гды был рган Сюза псатлй СССР, гд, наряду с псатля- сцралза, пчаталсь такж Андрй Платн, Юрй Тынян, Алксандр Тардскй, Вктр крас, Юрй Казак, Кнстантн Сн, Юрй Трфн, Вктр Астафь, Варла Шала, пзж — Булат Окуджаа, Фа- турн-эсттчск ялн журнальн пл стскй пчат, пскльку н знкл нн 20–30- гды ХХ ка как «сздатл ыразтл гсударстннй плт- к бласт культуры», прзанны «набл эффктн здйстать на бщстнн сзнан»3. 5 Офцальный сайт журнала «ная»: http://znamlit.ru (5.01.2019). 6 «Знамя»: очерк истории, http://znamlit.ru/history.html (5.01.2019). 7 .Ф. Алфрнк, Поэтическая энергия слова. Синергетика языка, созна- ния и культуры, Academia, Мска 2002, с. 17. POZNAWCZO-NARRACYJNA REALIZACJA POJĘĆ „SUMIENIE” I „DOBRO” W DYSKURSIE CZASOPISM LITERACKICH И нстря на т, чт пдбны жсячны лтратурн-худжстнны бщстнн-плтчск журналы прдсталяют сбй сгдня тдльный — элтарный — сгнт журнальнй п- рдк Рсс, пртстящй друг нфрацнны птка, даж г структур наблюдатся длгчская культурн-эсттчская дскртнсть разбщннсть. Так, с днй стрны, нахдятся радкальны, ппзцн- ны зданя, прдящ длгю нацнальнг кнср- атза, пддржающ традцнны патрархальны цннст («аш срннк»), с другй стрны, журналы, рнтранны на западны лбральн-дкратчск далы («ная»). «аш срннк», прначальн п- чаташй прдстатлй «дрнскй прзы» — Фдра Абраа, Васля Бла, Сргя алыгна, Валнтна Рас- путна, Владра Слухна, Васля Шукшна — бъ- ял сбя «трбунй днйшх плтк патртч- скг напралня», ппзцй прзападны устанка, «дущ к разрушню Рсс гбл русскг нарда»4. Журнал «ная», задуанный как здан лтратурнг бъдння Краснй Ар Флта, с 1934 п 1990 гды был рган Сюза псатлй СССР, гд, наряду с псатля- сцралза, пчаталсь такж Андрй Платн, Юрй Тынян, Алксандр Тардскй, Вктр крас, Юрй Казак, Кнстантн Сн, Юрй Трфн, Вктр Астафь, Варла Шала, пзж — Булат Окуджаа, Фа- Rusycystyczne Studia Literaturozn Rusycystyczne Studia Litera 56 зль Искандр, Людла Птрушская, Владр Маканн, Гргй Влад, Васлй Аксн, Владр Вйнч. В 1990 гду СМИ «бъял сбя нзасы, тстя с праа плутралтнй судбнй тяжб с Сюз псатлй СССР»5; пстпнн прратлсь «здан, пслда- тльн утрждающ далы дкрат тчстннг лбралза»6. Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Так браз, на прдсталятся цлсбразны рас- стрть кгнтны нарратны ханзы плщ- ня нктрых базых цннстй ткстах дух журнал, сстящх ппзтных тншнях. В качст базых братся к ду спряжнны цннстя этчскг тл- ка: «дбр» «ссть». 8 К.С. Грбачч, Словарь эпитетов русского литературного языка, - рнт, Санкт-Птрбург 2002; М.Р. Ль, Словарь антонимов русского языка: более 2000 антоним. пар, Русскй язык, Мска 1984; В.А. Мас- ла, Лингвокультурология, Издатльскй цнтр «Акадя», Мска 2001; С.И. Ожг, .Ю. Шда, Толковый словарь русского языка: 80 000 слов и фразеол. выражений, Азбукнк, Мска 1999; Ю.. Караул, Г.А. Чркаса, .В. Уфца, Ю.А. Сркн, Е.Ф. Та рас, Русский ас- социативный словарь: в 2-х т., АСТ–Астрль, Мска 2002; Л.В. Грхн- Грхн- Грхн- а, Т.М. Гршка, Е.Р. Прслгна, Л.И. Ручна, А.В. Снла, Словарь концептов русской народной сказки: в 2-х ч., Изд- жгрд. ун-та, жнй грд 2011, 2015; А.П. Егньа (рд.), Словарь русского язы- ка: в 4-х т., Плграфрсурсы, Мска 1999; тй ж (рд.), Словарь сино- нимов русского языка, АСТ, Мска 2003; П.. Днса, В.В. Мрккн, Словарь сочетаемости слов русского языка, АСТ, Мска 2005; М. Фас- Фас- Фас- р, Этимологический словарь русского языка, пр. О.. Турбач, АСТ– Астрль, Мска 2007. POZNAWCZO-NARRACYJNA REALIZACJA POJĘĆ „SUMIENIE” I „DOBRO” W DYSKURSIE CZASOPISM LITERACKICH В такй сязк ы х рассатра слу х кплтн-нструнтальных (т сть дплн- тльн-спгатльных) тншнй: рл «дбра» я- лятся т, чт н пртрчт «сст», ны сла, «ссть» стантся этчск снан зрня, п- наня «дбра», разлчня «дбра» «зла». р р р Мтдлгчск нструнт зучня цннстй ткстах ыбранных пчатных СМИ стал кнцпт — кгн- тная (ыслтльная) катгря, пратная днца «па- ят культуры», кант знаня; слжн, нжстк структу- рранн сысл бразан псатльн-бразнг цннстн-рнтраннг характра7. Кнцпт тд- лгчск рлантн псаню цннстных трансфрацй, пскльку т ядрн-прфрйную рганзацю: ядр структуры — эт цнтральны, нарантны сыслы к- пннты, характрзующся сантчскй стабльнстью, а элнты прфр бладают слаблннй, псрдан- нй (част ассцатнй, тафрчскй, этлгчскй т.д.) сантчскй сязью с цнтр. Прнцп плг 57 структурраня кнцпта пддржатся абслютны бльшнст рссйскх лнгкнцптлг (нада Данлна Ппа, Исф Абрач Стрнн, Сргй Грг- рьч Вркач, клай Фдрч Алфрнк, Владр Ильч Караск, Гннадй Гннадьч Слышкн, Гргрй Валрьч Ткар). Для фксац прдлнных цннстн-сыслых транс- фрацй сначала кнструруют бщязыкую, узуальную (т сть сстанлнную п стрчск срнны сларя) дль кнцпт, зат — дскурсны дл с цлью дальнйшг срання прдлня сыслй д- нак. e Studia Literaturoznawcze 2019 (29) Итак, узуальная дль кнцпта «ссть», п данны лк- скграфчскх стчнк8, ыглядят слдующ браз: usycystyczne Studia Literaturoznawcze 2019 ( — ядрная зна: 1. «чуст сзнан ральнй тт- стннст за с пдн пступк прд са сбй, прд кружающ людь, бщст»; 2. «нрастнны прнцпы, згляды, убждня»; — ядрная зна: 1. «чуст сзнан ральнй тт- стннст за с пдн пступк прд са сбй, прд кружающ людь, бщст»; 2. «нрастнны прнцпы, згляды, убждня»; Rusycystyczne Studia Literaturozn — клядрная зна: «срдст аплляц к нрастнн- этчск нра с цлью ыять нсттст эт нр- а ызать чуст стыда». POZNAWCZO-NARRACYJNA REALIZACJA POJĘĆ „SUMIENIE” I „DOBRO” W DYSKURSIE CZASOPISM LITERACKICH Очдн, чт ядрны кн- цптуальны прзнак псыают нрастнн-этчск 58 прнцпы лчнст с тчк зрня нддуальнг кнтр- ля (члк са кнтрлрут с пдн, сущстлят рфлксю сбстнных дйстй), а прядрны — с тчк зрня сцальнг кнтрля (кружающ цнают п- ступк ндда, бщст пказыат у пральный а- рант). Пр эт налч сст тсутст языкй картн ра днзначн стнсны с бнарны ппз- ця «хрш/плх» «пральн/нпральн»; Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… р р р — прфрйная зна (прналсь нан данны ассцатных, этлгчскх сларй, сларй эпт- т): 1. «нчт, дсталяющ фзчскй л псхлгч- скй дскфрт» (больная, взбаламученная, возмущенная, неспокойная, потревоженная, растревоженная, щекотли- вая, мучает, грызет, гложет, жгучая); 2. «чт-т, сязанн с спрят чстты, ста» (кристальная, кристально- чистая, незапятнанная, нечистая, черная, чистая, непогре- шимая, свет); 3. «нчт, чт т фзчск характр- стк» (глубокая, глухая, растяжимая, бездонная, глиняная, чугунная, резиновая); 4. «сязанн с прфсснальны длг» (врачебная, общественная, партийная, политиче- ская, юридическая, чекиста); 5. «сязанн с сстян сна паят» (дремлющая, пробудившаяся, проснувшаяся, за- бывчивая, спит); 6. «лцр, нскрннсть слдст утраты» (продажная, сговорчивая, потерянная, проданная, змеиная); 7. «т, чт ялятся пржтк пршлг, чт уж н стрбан; нку ннужный дал» (в шкафу на ве- шалке, мечта идиота, советского человека, фантом); 8. «т, чт т рнную характрстку» (эпохи, нетленная, давно); 9. «сязанн с ладн нфрацй» (знать, ве- дать). Так браз, прфрйная зна снат нзыбл- ую пратнсть кнстантнсть «сст», даннй част пля эта цннсть прдстална тнстльн: ссть н сгда нужна, на тпральн бусллна, на ызыат блзннны щущня, жн пйт на сдлку с сбстн- нй сстью, эт срдст ст заблуждн члка т.д. 59 П таку ж прнцпу характрзу бщязыкую - дль кнцпта «дбр»: — ядрная зна: «с плжтльн, хрш, т, чт пр- тпсталн злу»; — ядрная зна: «с плжтльн, хрш, т, чт пр- тпсталн злу»; — клядрная зна: 1. «т, чт прнст пльзу»; 2. «щ, ущст»; 3. POZNAWCZO-NARRACYJNA REALIZACJA POJĘĆ „SUMIENIE” I „DOBRO” W DYSKURSIE CZASOPISM LITERACKICH «сглас на чт-лб, дбрн пступка»; 4. «чнь хрш качст». Так браз, сл ядрная сантка т с ж абстрактный характр кнцп- рут «дбр» с тчк зрня «тлчннй» этк, т к- лядрны кпннты страают «дбр» псдн- н-практчскую бласть дятльнст члка, сязанную с прагаткй г пступк, сцальнй кункацй, атральны дстатк т.д.; — зна прфр: 1. «жл, ласк, пртл тншн к ку-лб; благжлатльнсть п тншню к ку-лб» (добро пожаловать, мама, улыбка, нежность, ра- дость); 2. «с, чт сязан с тчн рн, пршлы, па- ятью» (детство, старость, на век, вечно, память); 3. «т, чт дфцт» (редко, редкость, кончается); 4. «т, чт дста- лят эсттчск удльст, на чт прятн стрть» (цветы, сон, красота); 5. «снатльнсть рзультат труда» (кузнец, ремесленник, художник); 6. «т, чт дсталят фз- чск удльст» (выпивка, в штанах). — зна прфр: 1. «жл, ласк, пртл тншн к ку-лб; благжлатльнсть п тншню к ку-лб» (добро пожаловать, мама, улыбка, нежность, ра- дость); 2. «с, чт сязан с тчн рн, пршлы, па- ятью» (детство, старость, на век, вечно, память); 3. «т, чт дфцт» (редко, редкость, кончается); 4. «т, чт дста- лят эсттчск удльст, на чт прятн стрть» (цветы, сон, красота); 5. «снатльнсть рзультат труда» (кузнец, ремесленник, художник); 6. «т, чт дсталят фз- чск удльст» (выпивка, в штанах). Rusycystyczne Studia Literaturozn Rusycystyczne Studia Liter Тпрь прйд к дскурснй ралзац кнцпт (для аналза был зят сь асс публцстчскх ткст за 2016 гд з назанных журнал, рбалзующх данны кнцпты). POZNAWCZO-NARRACYJNA REALIZACJA POJĘĆ „SUMIENIE” I „DOBRO” W DYSKURSIE CZASOPISM LITERACKICH Для начала прдста клчстнны характ- рстк прзнак кнцпт «ссть» «дбр»: Параметр «Знамя» «Наш современник» Клчст рпрзнтант (рбалзатр, бъктатр) кнцпта «ссть» кнцпта «дбр» 10 23 3 13 Параметр «Наш современник» «Наш современник» «Знамя» 10 23 60 ивная реализация… Клчст кнткст (случа ралзац, уптрблнй) кнцпта «ссть» кнцпта «дбр» 43 126 8 31 Клчст ыялнных сантчскх прзнак (групп, сысл) кнцпта «ссть» кнцпта «дбр» 6 10 4 6 ализация… Клчст кнткст (случа ралзац, уптрблнй) кнцпта «ссть» кнцпта «дбр» 43 126 8 31 Клчст ыялнных Клчст кнткст (случа ралзац, уптрблнй) кнцпта «ссть» кнцпта «дбр» Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… ивная реали Клчст ыялнных сантчскх прзнак (групп, сысл) кнцпта «ссть» кнцпта «дбр» 6 10 4 6 Клчст ыялнных сантчскх прзнак (групп, сысл) кнцпта «ссть» кнцпта «дбр» Вплн чдна бщая тнднця кнцптуалзац ука- занных нтальных кнструкт: так назыаый кнцпт- «срдст» («ссть») прдсталн гразд ньшй стпн, ч кнцпт-«слдст» («дбр»), чт грт н- сфрулраннст, разытст, ннятнст сах з- ртльных срдст, тд, спсб, с пщью ктрых пстгаются цннст, ющ бл тлчнный характр. Вра Антрпа, Васлй Фдр Когн Чт ж касатся качстнных пказатлй, т ркнструк- ця дл кнцпта «ссть» дскурс журнала «ная» пказала слдующ: прзшл спадн ядрнг дскур- снг сысла «нрастнный закн, ргулрующй пд- н члка» (14 кнткст) с днтчны узус (т сть языкй картн ра), чт сдтльстут б унрсаль- нст рассатраг кнцпта ст с т — нзк птнцал бразаня ных сысл. Однак бнаружн дн ный клядрный кпннт «фра рганзац дятльнст ыскг качста» (11 ткстых фрагнт), н стртшйся ран псан узуальнй дл. Эт пзлят сдлать прдплжн трансфрац сысл, тнднц прхда сантк з сфры «тлчннй» нра- стннст прдтную плсксть пказатля качста дя- тльнст кг-лб. Так браз, качст кгнтнй стратг цннстнй трансфрац ы тча кннта- тнсть, цнчнсть «сст», так как этк-нрастнная сантка уступат ст сантк цнк качста, блада- ющй счслястью, прагатчскй знястью. POZNAWCZO-NARRACYJNA REALIZACJA POJĘĆ „SUMIENIE” I „DOBRO” W DYSKURSIE CZASOPISM LITERACKICH 61 Журнал «аш срннк» гразд ньшй стпн апллрут к кнцпту «ссть» рассатрат г с- ключтльн раках традцннй кгнтнй атрцы, гд прсутстуют с бщязыкы сантчск элн- ты: «нрастнный гарант спрадлст прадст» (6 кнткст), «чстнсть, прядчнсть» (3), «ттстн- нсть за прнят ршня» (3). Объктаця кнцпта «дбр» журнал «ная» г- рт б тсутст какх-лб зннй ядр прядрнй част, так как цнтральный сысл «чт-лб плжтльн, пртпсталнн злу» с ркрдн часттны пказатл (81 кнткст) заплнл практчск с сантчск пр- странст, «пглт» даж клядрную часть, журнал экс- траплрал этт ажный днстнный сысл на бль- шй фрагнт пля. Эт сдтльстут т, чт кнцпт «дбр» даж дскурс элтарнг журнала «ная» «закн- сррался», ных тнднцй разт г цннстнг птнцала здан пка н дт. дт, как на прд- сталятся, п нскльк прчна: 1. са сдржан этй цннст сгдня скльк-нбудь нятн н сфрулран, пэту журнал аксальн стржн, слдуя акслг- чск традця русскй культуры, ртранслрут сант- ку «дбра» т саы пддржат бщязыкую картну ра; 2. нн «дбр», как нкакая другая цннсть, сгд- ня сбнн затн пдргатся акслгчску ыхла- щаню рлятзу, сяз с ч здан пшл п пут рпрдукц, упрщня закрплня цннстнг сдр- жаня (дн базый сысл заплнл днрнн ядрную прядрную зны). Rusycystyczne Studia Literaturoznawc Аналгчную картну, слдуя качстнны пказатля, жн наблюдать тншн этг ж кнцпта дскур- с «ашг срннка»: с бльш прс лдрут сыслй кпннт «чт-лб плжтльн, пртп- сталнн злу» (16 кнткст). Клчстнны ж данны сдтльстуют т, чт здан гразд ньшй стп- н актуалзрут этт кнцпт сзнан чтатлй. 62 Так браз, цннстн-трансфрацнны прцссы п-разну плтлсь дскурсных практках «на- н» «ашг срннка». 9 Д.В. Барышнка, Когнитивный поворот в постклассической нарра то- логии (Обзор новых англоязычных книг) // « лтратурн бзрн» 2013, № 119, www.magazines.russ.ru/nlo/2013/119/b32.html (5.01.2019). POZNAWCZO-NARRACYJNA REALIZACJA POJĘĆ „SUMIENIE” I „DOBRO” W DYSKURSIE CZASOPISM LITERACKICH «аш срннк» птрл языкую картну ра схдя з закна экн рчых услй, т саы пдтрд статус кнсратнг з- даня п фр, п сдржаню. «ная», транслруя сдржан кнцпта «дбр», пддржал бщязыкую картну ра, н пддааясь акслгчску рлятз- у, а структур кнцпта «ссть» актуалзрал н значн плжтльнй цнк качста. С днй стрны, жн грть б акслгчск ыхлащан сн- жн статуса кнцпта, чт сттстут бщкультурнй днак, с другй — ппытк элтарнг зданя как- т браз актуалзрать, ранрать, «бнть» на- хдящся глубк крзс этчск цннст. Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Матралы, публканны «тлстых» лтратурных журналах, прдсталяют сбй арант нарратнг дскур- са. арратзацю жн рассатрать как спсб рал- зац цннстных сысл фр устйчых пста- тльных структур. В эт тншн тксты эксплцруют кгнтны (нтальны) дл. Изучн нарратных ханз началсь трй п- лн XX ка, нарратная тря пршла элюцю т структуралстскг тлканя, рассатрающг п- стан как структуру, набр тдльных элнт, д кг- нтных сслданй, ктрых зучаются нтальны, псхлгчск культурлгчск снаня прзд- ста наррат9. Испльзуя трю Тна ан Дйка стншн кгнт- нй (стуацннй) дл дскурса, ы плуч тд- лгчскй аппарат для псаня нарратных стратгй р- алзац кнцпт «ссть» «дбр». Учный тчат, чт «ажнйш кпннт прцс- с пстрня спрятя ткст ялятся сыслн 9 Д.В. Барышнка, Когнитивный поворот в постклассической нарра то- логии (Обзор новых англоязычных книг) // « лтратурн бзрн» 2013, № 119, www.magazines.russ.ru/nlo/2013/119/b32.html (5.01.2019). 63 стящх за н сцальных стуацй х кгнтная рпрзнтаця»10. Так браз, знкат ппзця: р- альная сцальная стуаця (фрагнт) vs. 10 Т.А. ан Дйк, Язык. Познание. Коммуникация: сборник работ, рд. пр. В.И. Грас, сст. В.В. Птр, Пргрсс, Мска, с. 161. 11 Та ж, с. 164. 12 W. Labov, Oral narratives of personal experience, http://ling.upenn.edu/~wla- bov/Papers/FebOralNarPE.pdf (5.01.2019). 13 D. Schiffrin, Redoing and replaying // тй ж, In Other Words: Variation in Reference and Narrative, Cambridge University Press, Cambridge 2006, с. 321. 12 W. Labov, Oral narratives of personal experience, http://ling.upenn.edu/~wla- bov/Papers/FebOralNarPE.pdf (5.01.2019). 10 Т.А. ан Дйк, Язык. Познание. Коммуникация: сборник работ, рд. пр. В.И. Грас, сст. В.В. Птр, Пргрсс, Мска, с. 161. 11 Та ж, с. 164. 10 Т.А. ан Дйк, Язык. Познание. Коммуникация: сборник работ, рд. пр. В.И. Грас, сст. В.В. Птр, Пргрсс, Мска, с. 161. 13 D. Schiffrin, Redoing and replaying // тй ж, In Other Words: Variation in Reference and Narrative, Cambridge University Press, Cambridge 2006, с. 321. 12 W. Labov, Oral narratives of personal experience, http://ling.upenn.edu/~wla- bov/Papers/FebOralNarPE.pdf (5.01.2019). 13 D. Schiffrin, Redoing and replaying // тй ж, In Other Words: Variation in Reference and Narrative, Cambridge University Press, Cambridge 2006, с. 321. POZNAWCZO-NARRACYJNA REALIZACJA POJĘĆ „SUMIENIE” I „DOBRO” W DYSKURSIE CZASOPISM LITERACKICH прдсталн нй, ктрая сущстут д сцнаря, л фрйа, пр- бртннг рзультат сцкультурнг пыта, н ы- ступат н тльк шаблн для прздста наррата, н сг рда нтрпртацннй ракй. Втрй элнт прдлят пстрн дскурса (чт как ы буд г- рть, сталкаясь ральнй с стуацй л псыая рч): «Кгнтны дл ссталяют рфрнцальный базс для нтрпртац дскурса, а н ‘фрагнт ральн- г ра’, л стуацй»11. Эт пзлят на ыять п- сать н прст набры ткст, а слжшся данн тп кункац пстатльны структуры, яляющся атрцй для пстрня кнкртных ткст транслрую- щ гты нтальны цннстны стртпы. usycystyczne Studia Literaturoznawcze 2019 ( Спсбнсть пстатльных длй ыступать нс- тл культурнй нфрац пыта пдтрждат У- лья Лаб, сслдашй устны рассказы англязычных рспнднт: «These forms of communication may draw upon the fundamental human capacity to transfer experience from one person to another through oral narratives of personal experience»12. Rusycystyczne Studia Literaturozn Rusycystyczne Studia Litera Дбра Шффрн прдлжат сслдан наррат, указыая на разрабтанны Улья Лаб тпчны сцлнгстчск тды. Атр тчат: «Narrative is a form discourse trough which we reconstruct and represent past experience both for ourselves and for others»13. Мжн грть т, чт знкл кгнтн напралн зучн наррат. В цнтр сслданй казыатся засязь 64 сцкультурнг пыта члка ханз г рчй рпрзнтац. сцкультурнг пыта члка ханз г рчй рпрзнтац. Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Для нас ажн тзс б зрфнст кгнтных ха- нз, тчающх за пстрн устных пстанй лчн пыт людй, длй наррата дскурс «тл- стых» лтратурных журнал. Так браз, дскурсная ралзаця кнцпт «с- сть» «дбр» кнструрутся псрдст нарратных структур, ктры задйстуют разлчны сыслы кп- ннты кнцпт. 14 И.T. Янн, Черноморская Пальмира // «аш срннк» 2016, № 9, http://nash-sovremennik.ru/archive/2016/n9/1609-14.pdf (5.01.2019). Вс ы- POZNAWCZO-NARRACYJNA REALIZACJA POJĘĆ „SUMIENIE” I „DOBRO” W DYSKURSIE CZASOPISM LITERACKICH Сущстан такх наррат прд- лятся налч тпчных стрй, л «сюжт», ыступа- ющх культурны днца кллктнг пыта, т сть прдсталяют сыслы цннст д рассказ, ктры ют структурн-сдржатльн тждст. аша задача сстт т, чтбы ыять псать такй набр нарра- т дскурс «тлстых» лтратурных журнал, фр- ру публцстчск, кртчск атрала, прдлть, как браз н рпрзнтруют цннстны трансфрац срнн рссйск бщст. Опраясь на сцкультурную длгчскую характ- рстку журнал «ная» «аш срннк», ы ж грть разлчных арантах нарратзац кнцпт «ссть» «дбр». В журнал «аш срннк» кнцпт «ссть» ралзу- тся нарратнй дл, ктрая пстрна на пртст- ян стлкнн, ральных прдплагаых. С днй стрны, эт нутрннй кнфлкт, плщающй ядрный сысл «чуст сзнан ральнй ттстннст за с пдн пступк прд са сбй, прд кружа- ющ людь, бщст». апрр: Кпнсацнны лспсадк Хках крстнстях благплучн прнялсь бщают ыраст лучш пржнх. А чт ж блудная дчь Х- кнскг лса? В 2015 гду на ст с сй сьй прхала Эст- ню на пстянн ст жтльста. Пчу? Ссть зала? пхж14. 14 И.T. Янн, Черноморская Пальмира // «аш срннк» 2016, № 9, http://nash-sovremennik.ru/archive/2016/n9/1609-14.pdf (5.01.2019). Вс ы- 65 С другй стрны, ншнй кнфлкт, бъктрующй клядрный кпннт: «срдст аплляц к нрастн- н-этчск нра с цлью ыять нсттст эт нра ызать чуст стыда», — ыплняющй функц сцальнг кнтрля: «нат, чт ны члны Ерсюза ыпрашают грны срдста, бз зазрня сст крча б угрз»15. Пр эт, бльшнст уптрблнй кнфлкт этт тльк прдплагатся, н тг н сущстлятся. Та- кй устйчый сыслй кпннт («нг», «нсст- яшгся» кнфлкта) пдчркат рнтацю журнала, дскурс ктрг срнн бщст тдльный н- дд бльшнст с н страаются традцнны ральн-этчск прнцпы, с «н» (сбнн запад- н) н т сй цннстнй рарх сст. 15 Та ж. д л ня ткстых фрагнт пдчркнуты шрфт прнадлжат атра стать. 15 Та ж. 16 Е.П. Щгла, Вера, истина, справедливость — и дочка Маша // «ная» 2016, № 4, http://znamlit.ru/publication.php?id=6242 (5.01.2019). 17 Та ж. POZNAWCZO-NARRACYJNA REALIZACJA POJĘĆ „SUMIENIE” I „DOBRO” W DYSKURSIE CZASOPISM LITERACKICH Пказат- лн эт сысл прый прр, ктр спльзан б- блйскй нтрткст (блудная дчь), днстррующй ухд т сст, птрю. Са кнфлкт, «наррат пртстя- ня», рпрзнтрут пзцю журнала как «фрпста» с- ст хатчнй дйсттльнст, пстянную г брьбу за ссть. Пэту журнал «аш срннк» прсхдт аплляця к кллктнй сст, на эт указыат ыская часттнсть фразлгчскх датчскх днц (за- учла ссть, бз зазрня сст). Rusycystyczne Studia Literaturozn В дскурс журнала «ная» нарратная рпрзнтаця кнцпта «ссть» бнаружат другую закнрнсть. «Кнфлктный сюжт» трансфррутся сюжт нра- стнн ыбр тдльнг члка л гря худжстн- нг прздня. Ины сла, ссть прдстат как нддуальный ыбр, как ршн «Я» сблюдать прнц- пы л тказаться т нх: В дскурс журнала «ная» нарратная рпрзнтаця кнцпта «ссть» бнаружат другую закнрнсть. «Кнфлктный сюжт» трансфррутся сюжт нра- стнн ыбр тдльнг члка л гря худжстн- нг прздня. Ины сла, ссть прдстат как нддуальный ыбр, как ршн «Я» сблюдать прнц- пы л тказаться т нх: И пчу члчская пррда устрна так, чт саы гршны чустуют сбя с н т, ку эт прстал? Пчу стыд грызт саых д л ня ткстых фрагнт пдчркнуты шрфт прнадлжат атра стать. 15 Та ж 66 17 Та ж. прядчных дстйных, т ря как бссстны прбыают п- к благплуч?16 прядчных дстйных, т ря как бссстны прбыают п- к благплуч?16 Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Я тчаянн, страшн нат, даж н прд чтатл, а пржд сг прд сбстннй сстью за лжь сх пстях, ктрую так длг ыдаал за праду, — слн бы грт са дух этй кнг Л. Пан- тл, а частнст, «сй рукй ыл этт пдзаглк (т.. чт п- сть «атбграфчская», как напсан сх зданях). И т тпрь расплачаюсь17. Пказатльна лксчская счтасть (ссть: атр, я, сбстнная). В такй нддуальнй ттстнн- ст прд сстью знкат дстанця прд са - прат сст, абстрактны пнят, н случайн ссть публкацях журнала «ная» ыступат как т- длнная т члка ысшая сла. Так браз, нарра- тная ралзаця кнцпта «ссть» спадат с узуаль- нй длью (бщязыкй), днак здсь ы наблюда нтллктуальн-рацналстчск пнан сст, аналтчск тдляющ нру, ктрая сущстут нза- с т л тдльнг члка, слдан этй нр- как д. Важн ттть щ дн ктр нарратзац кнцпта «ссть» дскурс журнала «ная». Так, ы уж бращал нан на пялн нг, прагатчск рнтран- нг сыслг кпннта «фра рганзац дятльн- ст ыскг качста». Вслдст этг пялся наррат дбрсстн ыплнн сх бязаннстй («дбр- сстн»), качстнн, а н клчстнн стт- ст нра сст. Мжн сказать, чт, с днй стрны, дбрсстнсть стантся нструнт плщня с- ст псдннст, сзнан дятльнст ндда. С другй стрны, стуац цннстных трансфрацй дбрсстнсть ыступат сязующ зн жду лч- 67 нстны пыт абстрактнй катгрй, устаналат сыслы тншня. Так браз, «тлсты» лтратурны журналы «аш срннк» «ная», актзруют сршнн разны кгнтны дл (Т. ан Дйк) кнцпта «ссть». Пр здан прдлагат брьбу за ссть, грт дградац этй цннст срнн р, спльзуя наррат брь- бы стлкння дух сл. 18 Т.Л. Мрна, Разговор по-русски — разговор по душам // «аш ср- ннк» 2016, № 8, http://www.nash-sovremennik.ru/archive/2016/n8/1608- 11.pdf (5.01.2019). Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… 19 И.Т. Янн, Блестящий сочинский гамбит // «аш срннк» 2016, № 11, http://www.nash-sovremennik.ru/archive/2016/n11/1611-14.pdf (5.01.2019). прядчных дстйных, т ря как бссстны прбыают п- к благплуч?16 Втр здан рпрзнтрут ссть как лчнстную катгрю длат ппытку найт сязь жду частны, нддуальны пыт дй «с- ст», фрруя наррат «сбднг ыбра». у арратная ралзаця кнцпта «дбр» актн спль- зутся дскурс журнала «аш срннк». В цл, на сттстут узуальнй дл (бщязыкй), н транс- фррут ядрный сыслй кпннт «с плжтль- н, хрш, т, чт пртпсталн злу». Так, ыстраа- тся нарратная дль, ктрй «дбр» сязан с паятью пршлы, чт пзлят грть пртпсталн дбрых/ндбрых (дй, качст, людй), эт прждат сю- жт «сх» «чужх». Ины сла, с дбр сязан с, чт сттстут цннстя нра, рззрню журнала «аш срннк», «дбр» прдлят пра на сущстан, пскльку эт ысшая катгря, н над бъяснять. дсь ажн пдчркнуть, чт дскурс журнала «аш срннк» рактуалзруются дрнйш сыслы кнцпта «дбр», так как «хрш сдланный, праль- ный, сттстующй нр», чт пзлят прнять этт кнцпт для устанлня лгтнст пральнст чг-лб. Так браз, фррутся наррат, пстрн- ный на ппзц «ы»–«н», арантах «сй»–«чужй», «закнный»–«нзакнный»: Rusycystyczne Studia Literaturozn эт ж крн, сптанн нас «люд пгут!» джт на, кгда на случатся прхдть на пщь друг, даж сйчас, па- скудн ря, кгда цлы сндкаты, эксплуатруя налд, калк, аллтк, пстал на службу сг крыстнг бгащня наш д- бры скрнн чуста, люд, зная, чт ппршайк, убг, «нал- 68 ды», «люд нстны» — нструнт нажы руках прступнг ра, с ран тянутся к кшлька, пту чт пгать — н сзнан на- ш, а характр, кр18. ды», «люд нстны» — нструнт нажы руках прступнг ра, с ран тянутся к кшлька, пту чт пгать — н сзнан на- ш, а характр, кр18. Сгдня Рсс прхдтся фактчск зан сстаналать утра- чнн прздст, ытснять с тчстннг аптчнг рынка нд- брсстных нстранных прздтлй19. Сгдня Рсс прхдтся фактчск зан сстаналать утра- чнн прздст, ытснять с тчстннг аптчнг рынка нд- брсстных нстранных прздтлй19. В дскурс журнала «ная» ы бнаружл сущста- н нсклькх нарратных длй, пш саы частт- ны з нх. В-прых, эт наррат, ктр дбр рас- крыатся как нрастнн-этчская катгря (дбрый, дбрпрядчный, дбрсрдчный т. п.), качст чл- ка. В такх ткстах рчь дт зжнст сттстать члка гуанстчск цннстя. В-трых, стр дбр как нкй нр сцальнй рганзац члч- скй жзн (распрстраннн спльзан н сущ- сттльнг «дбр» г лксчскх фр). В-тртьх, аплляця к дбру как услю прзнаня чг-лб, т сть «дбрн». Сущстан этх трх нарратных аран- т нь тсылат нас к «эгцнтрчскй» пзц, кт- рй члк са тчат за с пступк, рганзут сй лчный пыт. В эт сысл дбр как нкая нра рпр- знтрут абстрактн-флсфскую катгрю, тдлнную т псднных практк прстг члка, н прсутстут как бъктная дя. Тгда дбрн/ндбрн стан- ятся прялн л, актнг начала. Пказатльн, чт наррат, гд сыслй кпннт «дбрн» грат ажную рль, пялятся кртчск-рнчск тнш- н к субъкту, пряшу сю лю: Вра Антрпа, Васлй Фдр Когн 69 Эт антзападн пклн, сющся над западнка-лбрала- дбряющ прсналстскй рж Путна, п сут дла, пр састятльн нзас пклн с плн западны быты прычка даж труды наыка. Он хтят уют зарабаты- ать дньг с пщью напряжннг калфцраннг труда20. Ины сла, дбрн — дйст, бнаружающ лю члка г спсбнсть сттстать дя д- бра, спрадлст. Так браз, нарратная ралзаця кнцпта «д- бр» журналах «аш срннк» «ная» эксплцру- ют слдующ кгнтны дл: дбр как с закнн, пральн, тлчающ сх т чужх, дбр как нра- стнн-этчская катгря, как ыбр нддуальнг с- знаня, прялн г л. Сгдня Рсс прхдтся фактчск зан сстаналать утра- чнн прздст, ытснять с тчстннг аптчнг рынка нд- брсстных нстранных прздтлй19. usycystyczne Studia Literaturoznawcze 2019 ( Итак, дскурс «тлстых» лтратурных журнал стант- ся транслятр цннстных трансфрацй рссйск бщст. а прр кгнтнй нарратнй рал- зац кнцпт «ссть» «дбр» жн сдлать ыд, с днй стрны, сущстан стуац культурнг нгбразя срнн русскязычн дапрстран- ст, с другй стрны, стлкнн лбральн рнт- раннй кнсратнй сст цннстй. Rusycystyczne Studia Literaturoz Rusycystyczne Studia Litera Исслдан ыплнн за счт гранта Рссй- скг научнг фнда (пркт № 16-18-02032). 20 С.В. Црль, Что нас ждет? // «ная» 2016, № 9, http://magazines.russ. ru/znamia/2016/9/chto-nas-zhdet.html (5.01.2019). REFERENCES Alefirenko, Nikolay F. Poeticheskaya energiya slova. Sinergetika yazyka, soznaniya. Moskva: Academia, 2002 [Алфрнк, клай Ф. Поэтическая энер- энер- энер- гия слова. Синергетика языка, сознания и культуры. Мска: Academia, 2002]. Baryshnikova, Dar’ya V. “Kognitivnyy povorot v postklassicheskoy narratologii (Obzor novykh angloyazychnykh knig).” Novoye literaturnoye obozreniye, no. 119, 2013, www.magazines.russ.ru/nlo/2013/119/b32.html. Accessed 5 Ja- . 119, 2013, www.magazines.russ.ru/nlo/2013/119/b32.html. Accessed 5 Ja- 119, 2013, www.magazines.russ.ru/nlo/2013/119/b32.html. Accessed 5 Ja- 119, 2013, www.magazines.russ.ru/nlo/2013/119/b32.html. Accessed 5 Ja- www.magazines.russ.ru/nlo/2013/119/b32.html. Accessed 5 Ja- .magazines.russ.ru/nlo/2013/119/b32.html. Accessed 5 Ja- magazines.russ.ru/nlo/2013/119/b32.html. Accessed 5 Ja- .russ.ru/nlo/2013/119/b32.html. Accessed 5 Ja- russ.ru/nlo/2013/119/b32.html. Accessed 5 Ja- .ru/nlo/2013/119/b32.html. Accessed 5 Ja- ru/nlo/2013/119/b32.html. Accessed 5 Ja- /nlo/2013/119/b32.html. Accessed 5 Ja- nlo/2013/119/b32.html. Accessed 5 Ja- /2013/119/b32.html. Accessed 5 Ja- b32.html. Accessed 5 Ja- 32.html. Accessed 5 Ja- html. Accessed 5 Ja- . Accessed 5 Ja- Accessed 5 Ja- 5 Ja- nuary 2019 [Барышнка, Дарья В. “Кгнтный прт пстклас- Барышнка, Дарья В. “Кгнтный прт пстклас- , Дарья В. “Кгнтный прт пстклас- Дарья В. “Кгнтный прт пстклас- . “Кгнтный прт пстклас- Кгнтный прт пстклас- прт пстклас- прт пстклас- пстклас- пстклас- 20 С.В. Црль, Что нас ждет? // «ная» 2016, № 9, http://magazines.russ. ru/znamia/2016/9/chto-nas-zhdet.html (5.01.2019). 70 счскй нарратлг (Обзр ных англязычных кнг).” Новое ли- тературное обозрение, № 119, 2013, www.magazines.russ.ru/nlo/2013/119/ b32.html. Дата бращня: 5 янаря 2019]. счскй нарратлг (Обзр ных англязычных кнг).” Новое ли- тературное обозрение, № 119, 2013, www.magazines.russ.ru/nlo/2013/119/ b32.html. Дата бращня: 5 янаря 2019]. Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Вра Антрпа, Васлй Фдр Когнитивно-нарративная реализация… Van Dijk, Teun A. Yazyk. Poznaniye. Kommunikatsiya: sbornik rabot. Ed. Transl. Gerasimov, Vladimir I. Ed. Petrov, Vasiliy V. Moskva: Progress, 1986. [Ван Дйк, Тн А. Язык. Познание. Коммуникация: сборник работ. Рд. пр. Г- рас, Владр И. Сст. Птр, Васлй В. Мска: Пргрсс, 1989]. Fasmer, Maks. Etimologicheskiy slovar’ russkogo yazyka. Transl. Turbachev, Oleg N. Moskva: AST–Astrel’, 2007 [Фаср, Макс. Этимологический словарь рус- ского языка. Пр. Турбач, Олг ., Мска: АСТ–Астрль, 2007]. g y g y y g Moskva: AST–Astrel’, 2007 [Фаср, Макс. Этимологический словарь рус- ского языка. Пр. Турбач, Олг ., Мска: АСТ–Астрль, 2007]. Gorbachevich, Kirill S. Slovar’ epitetov russkogo literaturnogo yazyka. Sankt- Peterburg: Norint, 2002 [Грбачч, Крлл С. Словарь эпитетов русско- го литературного языка. Санкт-Птрбург: рнт, 2002]. REFERENCES Grekhneva, Larisa V., et al. Slovar’ kontseptov russkoy narodnoy skazki: v 2-kh ch. Nizhniy Novgorod: Izd-vo Nizhegorod. un-ta 2011, 2015 [Грхна, Лар- са В. др. Словарь концептов русской народной сказки: в 2-х ч. жнй грд: Изд- жгрд. ун-та, 2011, 2015]. Karaulov, Yuriy N., et al. Russkiy assotsiativnyy slovar’: v 2-kh t. Moskva: AST– Astrel’, 2002 [Караул, Юрй . др. Русский ассоциативный словарь: в 2-х т. Мска: АСТ–Астрль, 2002]. Labov, William. Oral narratives of personal experience, http://ling.upenn edu/~wlabov/Papers/FebOralNarPE.pdf. Accessed 5 January 2019. L’vov, Mikhail R. Slovar’ antonimov russkogoyazyka: boleye 2000 antonim. par. Moskva: Russkiy yazyk, 2001 [Ль, Мхал Р. Словарь антонимов рус- ского языка: более 2000 антоним. пар. Мска: Русскй язык, 1984]. Mareyeva, Svetlana V. “Dinamika norm i tsennosteyrossiyan.” Sotsis, no. 7, 2013. [Мара, Стлана В. “Днака нр цннстй рссян.” Социс, № 7, 2013]. Maslova, Valentina A. Lingvokul’turologiya. Moskva: Izdatel’skiy tsentr “Akademiya,” 2001 [Масла, Валнтна А. Лингвокультурология. М- ска: Издатльскй цнтр “Акадя,” 2001]. Mironova, Tat’yana L. “Razgovor po-russki — razgovor po dusham.” Nash sovremennik, no. 8, 2016, http://www.nash-sovremennik.ru/archive/2016/ n8/1608-11.pdf. Accessed 5 January 2019 [Мрна, Татьяна Л. “Разгр п-русск — разгр п душа.” Наш современник, № 8, 2016, http:// www.nash-sovremennik.ru/archive/2016/n8/1608-11.pdf. Дата бращня: 5 янаря 2019]. Ozhegov, Sergey I., Shvedova, Nataliya Yu. Tolkovyy slovar’ russkogo yazyka: 80 000 slov i frazeol. vyrazheniy. Moskva: Azbukovnik, 1999 [Ожг, Сргй И., Шда, аталя Ю. Толковыий словарь русского языка: 80 000 слов и фразеол. выражений. Мска: Азбукнк, 1999]. Schiffrin, Deborah. “Redoing and replaying.” In Other Words: Variation in Reference and Narrative. Cambridge: Cambridge University Press, 2006. Shcheglova, Yevgeniya P. “Vera, istina, spravedlivost’ — i dochka Masha.” Znamya, no. 4, 2016, http://znamlit.ru/publication.php?id=6242. Accessed 5 January 71 2019 [Щгла, Егня П. “Вра, стна, спрадлсть — дчка Маша.” Знамя, № 4, 2016, http://znamlit.ru/publication.php?id=6242. Дата бращня: 5 янаря 2019]. Shkarovskaya, Iren V., Stanovleniye i razvitiye sovet·skikh “tolstykh” zhurnalov v pervoy polovine 20-kh godov: avtoref. dis. kand. filol. nauk, Moskva 1999, http://cheloveknauka.com/stanovlenie-i-razvitie-sovetskih-tolstyh-zhur- nalov-v-pervoy-polovine-20-h-godov. Accessed 5 January 2019 [Шкар- ская, Ирэн В. Становление и развитие советских “толстых” журналов в первой половине 20-х годов: автореф. дис. канд. филол. наук. Мска 1999, http://cheloveknauka.com/stanovlenie-i-razvitie-sovetskih-tolstyh-zhur na- lov-v-pervoy-polovine-20-h-godov. Дата бращня: 5 янаря 2019]. Slovar’ russkogo yazyka: v 4-х t. Ed. Yevgen’yeva, Anastasiya P. Moskva: Russkiy yazyk, Poligrafresursy, 1999 [Словарь русского языка: в 4-х т. Рд. Егнь- а, Анастася П. REFERENCES Мска: Русскй язык, Плграфрсурсы, 1999]. Rusycystyczne Studia Literaturoznawcze 2019 (29) Slovar’ sinonimov russkogoyazyka. Ed. Yevgen’yeva, Anastasiya P. Moskva: AST, 2003 [Словарь синонимов русского языка. Рд. Егньа, Анастася П. Мска: АСТ, 2003]. Slovar’ sochetayemosti slov russkogo yazyka. Ed. Denisov, Petr N., Morkovin, Valeriy V. Moskva: AST, 2005 [Словарь сочетаемости слов русского языка. Рд. Днс, Птр ., Мркн, Валрй В. Мска: АСТ, 2005]. Rusycystyczne Studia Literaturoznawc Tilinina, Tat’yana V. “Prichiny krizisa tsennostnykh oriyentatsiy v sovremennykh rossiyskikh sem’yakh.” Teoriya i praktika obshchestvennogo razvitiya, no. 6, 2012, http://teoria-practica.ru/rus/files/arhiv_zhurnala/2012/6/scilgiyа/ tilinina.pdf. Accessed 5 January 2019 [Тлнна, Татьяна В. “Прчны кр- зса цннстных рнтацй срнных рссйскх сьях.” Теория и практика общественного развития, № 6, 2012, http://teoria-practica.ru/ rus/files/arhiv_zhurnala/2012/6/scilgiyа/tilinina.pdf. Дата бращня: 5 янаря 2019]. Tsirel’, Sergey V. “Chto nas zhdet?” Znamya, no 9, 2016, http://magazines.russ. ru/znamia/2016/9/chto-nas-zhdet.html. Accessed 5 January 2019 [Црль, Сргй В. “Чт нас ждт?” Знамя, № 9, 2016, http://magazines.russ.ru/ znamia/2016/9/chto-nas-zhdet.html. Дата бращня: 5 янаря 2019]. Yanin, Igor’ T. “Blestyashchiy sochinskiy gambit.” Nash sovremennik, no. 11, 2016, http://www.nash-sovremennik.ru/archive/2016/n11/1611-14.pdf. Accessed 5 January 2019 [Янн, Игрь Т., “Блстящй счнскй габт.” Наш современник, № 11, 2016, http://www.nash-sovremennik.ru/archive/2016/ n11/1611-14.pdf. Дата бращня: 5 янаря 2019]. Yanin, Igor’ T. “Chernomorskaya Pal’mira.” Nash sovremennik, no. 9, 2016, http:// nash-sovremennik.ru/archive/2016/n9/1609-14.pdf. Accessed 5 January 2019 [Янн, Игрь Т. “Чрнрская Пальра.” Наш современник, № 9, 2016, http://nash-sovremennik.ru/archive/2016/n9/1609-14.pdf. Дата бращня: 5 янаря 2019]. “Znamya:” ocherk istorii, http://znamlit.ru/history.html. Accessed 5 January 2019 [“Знамя:”очерк истории, http://znamlit.ru/history.html. Дата бращня: 5 янаря 2019].
https://openalex.org/W3099320484	https://pure.rug.nl/ws/files/195836987/1_s2.0_S0927025620306200_main.pdf	English	null	Effective mobility of BCC dislocations in two-dimensional discrete dislocation plasticity	Computational materials science	2,021	cc-by	14,026	Effective mobility of BCC dislocations in two-dimensional discrete dislocation plasticity Katiyar, T.; Van der Giessen, E. Published in: Computational Materials Science IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Katiyar, T., & Van der Giessen, E. (2021). Effective mobility of BCC dislocations in two-dimensional discrete dislocation plasticity. Computational Materials Science, 187, Article 110129. https://doi.org/10.1016/j.commatsci.2020.110129 Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). University of Groningen University of Groningen Effective mobility of BCC dislocations in two-dimensional discrete dislocation plasticity Katiyar, T.; Van der Giessen, E. Effective mobility of BCC dislocations in two-dimensional discrete dislocation plasticity T. Katiyar , E. Van der Giessen * Zernike Institute for Advanced Materials, University of Groningen, Groningen 9747AG, The Netherlands A R T I C L E I N F O Keywords: Discrete dislocation plasticity BCC crystal Dislocation mobility Strain rate sensitivity Two-dimensional discrete dislocation plasticity (2D-DDP) has shown to be a powerful tool for studying micro- plasticity problems such as size effects in single crystals, fracture of bimaterial interfaces, delamination of thin films, fatigue crack growth etc. The power of 2D-DDP lies in the application of edge dislocation dipoles as the vehicle for plastic slip: the loss of accuracy in the description of dislocation structures is counter-balanced by its simplicity and the possibility to reach larger plastic strains. The constitutive rules for dislocation evolution in 2D- DDP used so far are tuned to FCC crystals and need to be modified to be used for BCC materials. One of the key challenges in extending the method to BCC materials is that, contrary to FCC, the mobilities of edge and screw dislocations in BCC crystals differ vastly from each other, so that the screw mobility will be rate limiting the plastic slip. Thus, a method is required to map the edge and screw mobilities of dislocation loops into an effective mobility to be used in 2D. To do so, we here propose a 3D-to-2D procedure that is based on the notion of conservation of in-plane plastic strain rate. The consequence of this approach is that the effective 2D mobility for FCC crystals is not simply equal to the uniform mobility of a dislocation loop, as has been assumed by all 2D models to date, but also on the size of edge dipoles. In order to assess the consequences of this departure from the current literature, we considered a few key problems involving plasticity size effects and crack growth, and compared the predictions assuming constant mobility versus the proposed effective mobility. After observing that, overall, the predictions do not deviate substantially, we proceed with application of the 3D-to-2D procedure to compute the effective 2D mobility for BCC materials based on their edge and screw mobilities. The validation of the approach is done by comparison of the predicted rate sensitivity of polycrystalline iron with the experi mental rate sensitivity at room temperature, which are found to be in fairly good agreement. Copyright strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of th or copyright holder(s), unless the work is under an open content license (like Creative Commons). The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date: 24-10-2024 Computational Materials Science 187 (2021) 110129 * Corresponding author. E-mail address: E.van.der.Giessen@rug.nl (E. Van der Giessen). Available online 12 November 2020 0927-0256/© 2020 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1016/j.commatsci.2020.110129 Received 6 March 2020; Received in revised form 2 October 2020; Accepted 4 October 2020 Available online 12 November 2020 0927-0256/© 2020 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). * Corresponding author. E-mail address: E.van.der.Giessen@rug.nl (E. Van der Giessen). https://doi.org/10.1016/j.commatsci.2020.110129 Received 6 March 2020; Received in revised form 2 October 2020; Accepted 4 October 2020 * Corresponding author. E-mail address: E.van.de 1. Introduction Rice [20] showed that this can be reconciled with the 12 slip systems of type {111}〈110〉of FCC crystals by considering edge dislocations to glide on three effective slip systems, as illustrated in Fig. 1; for example, by assuming equal activity in the [110] and [011] directions on the (111) plane yields a slip plane for 2D-DDP in the [121] direction. A similar construction leads to three effective slip planes for 2D-DDP of BBC crystals having 12 slip systems. The objective of the present work is to devise and establish a meth odology to incorporate the widely different edge and screw mobilities into an effective mobility Meff that can represent BCC dislocations in 2D- DDP. To this aim, and after a recap of the objectives and limitations of 2D-DDP, we propose a 3D-to-2D procedure in Section 3 which projects the expansion of a 3D dislocation loop to 2D dipoles based on the con servation of in-plane plastic shear rate. The resulting effective mobility of 2D dislocations in FCC crystals is found not to be constant, as assumed till date in 2D-DDP computations, but to be a function of loop size. To asses the influence of this discrepancy on past results obtained by simply assuming a constant effective mobility, a number of key 2D-DDP pre dictions published in the literature are re-explored in Section 4. After observing that, overall, the predictions do not deviate substantially, we proceed in Section 5 by applying the 3D-to-2D procedure to compute the effective 2D mobility for BCC crystals based on their edge and screw mobilities. In order to validate the obtained effective mobility, 2D-DDP predictions of the rate sensitivity of the yield strength in α-iron at room temperature are compared against experimental data available in the literature. Section 6 describes the employed 2D-DDP framework for the prediction of the rate sensitivity and the validation approach. The rate- sensitivity calculations are then presented in Section 7 and the findings are discussed in Section 8. Finally, the important conclusions and im plications of the study are shortly summarized in Section 9. Plastic deformation can be either computed explicitly as the average plastic strain rate over small samples of volumes (see, e.g., [21]) and passed-on to a crystal plasticity model or determined implicitly by keeping track of the motion of individual dislocations while solving a boundary-value problem [6]. 1. Introduction microplasticity and the associated dislocation microstructure but despite recent accelerations [3] they suffer from the high computational burden associated with the complicated geometry of dislocation loops and their complex interactions in 3D [4]. Simplification of the loop shape, as proposed very recently in [5], indeed drastically reduces computing times, but the implications of the assumptions made have not been fully assessed. Therefore, the use of two-dimensional plane strain DDP models (2D-DDP) has been popular during the past two decades, especially after the inception of a versatile superposition method for solving complex boundary value problems by Van der Giessen and Needleman [6]. Even though it cannot capture detailed aspects of 3D dislocation microstructures, 2D-DDP has shown to be a powerful tool in understanding many aspects of dislocation physics. It has been suc cessfully applied to study some key micro-plasticity problems such as size effects in single crystals [7,8] and polycrystals [9], bending [10], Dislocations are the fundamental carriers of plastic flow in ductile crystals. The discreteness of dislocations plays an important role at the micron/sub-micron scale where several features of deformation are directly affected by the motion and interaction of individual disloca tions. While the dislocations are averaged in a continuum sense in traditional continuum plasticity models, atomistic models cannot address large numbers of dislocations. This has led to the emergence of the discrete dislocation plasticity (DDP) methodology which, in terms of length and time scales, sits between continuum and atomistics plasticity; that is, it adopts a continuum description of the crystal lattice while retaining dislocations as discrete entities (see Po et al. [1] and El-Awady et al. [2] for comprehensive reviews). Fully three-dimensional disloca tion dynamics simulations have been extensively used to investigate Computational Materials Science 187 (2021) 110129 Fig. 1. Effective slip systems in FCC and BCC crystals allowing for plane strain plastic deformation in the respective x1–x2 planes according to Rice [20]. 3D slip systems are shown dashed, while effective 2D slip planes are indicated with thick lines. T. Katiyar and E. Van der Giessen Computational Materials Science 187 (2021) 110129 T. Katiyar and E. Van der Giessen Fig. 1. Effective slip systems in FCC and BCC crystals allowing for plane strain plastic deformation in the respective x1– systems are shown dashed, while effective 2D slip planes are indicated with thick lines. Fig. 1. 1. Introduction In either approach, and following [18,6], the long-range interactions between dislocations are incorporated directly through the continuum elastic fields, while short-range in teractions are formulated as distinct constitutive rules. These constitu tive rules for edge dislocations are designed to be simple, yet without compromising the essential physics. For this reason, in 2D we use dislocation dipoles, where the sign of the Burgers vector magnitude reflects the line direction (into or out of the plane of consideration). Such a dipole can be regarded as a sideview or a projection of a dislocation loop in a direction parallel to the loop, as illustrated in Fig. 2. In this way, the displacement discontinuity is contained on the line between the two edge dislocations, just like the displacement discontinuity in the 3D view is bounded by the dislocation contour. 1. Introduction Effective slip systems in FCC and BCC crystals allowing for plane strain plastic deformation in the respective x1–x2 planes according to Rice [20]. 3D slip systems are shown dashed, while effective 2D slip planes are indicated with thick lines. Fig. 2. (a) Top view of dislocation loop on slip plane and (b) side view of edge dipole with the same Burgers vector as a 2D projection. fatigue crack growth [11], fracture of bimaterial interfaces [12] and plastic deformation in thin films [13,14]. Though widely applied, the existing 2D-DDP framework is limited to face-centered cubic (FCC) crystals. In order to study microplasticity in body-centered cubic (BCC) crystals, the framework needs to account for important aspects of plasticity in BCC crystals that are distinctly different from FCC crystals. One of these key differences, apart from the different slip systems, is that screw and edge dislocations in BCC crystals have vastly different mobilities: the screw segments are much less mo bile than non-screw segments owing to a high lattice friction experi enced during their motion. This is due to the non-planar cores of the screws which makes it difficult for them to move, especially at low to moderate stress levels and temperatures (see [15,16]). As a conse quence, during plastic deformation, the loops become elongated due to the fast-moving edge segments, resulting in a high density of slow moving screw dislocations. Therefore, screw dislocations are established to be the dominant defects governing the plastic deformation in BCC crystals with their mobility controlling the rate of the deformation. However, in 2D-DDP only edge dislocations exist. To simply assign these edge dislocations either the mobility of pure edge or pure screw seg ments would be an inaccurate representation of the BCC dislocations in two-dimensions. Fig. 2. (a) Top view of dislocation loop on slip plane and (b) side view of edge dipole with the same Burgers vector as a 2D projection. Kubin et al. pioneered with simulating the dynamics of systems of par allel edge dislocations under the influence of long-range interactions between dislocations, e.g. [17–19]. In a second stage, it was realized that moving edge dislocations can be used to describe mesoscopic plasticity, as a phenomenon emerging from the continuously evolving slipped distance. By construction, the evolution of parallel edge dislocations generates plane strain plastic deformation. 2. Two-dimensional DDP Frank-Read sources emerge naturally during 3D dislocation dynamics simulations, but in most 2D simulations they are assumed to be present from the start of the simulation and interpreted as remainders from previous plastic deformation. The strength of a Frank–Read source is determined by the length of the generating segment [22], but in 2D the latter information is not available. The majority of 2D-DDP studies therefore simply assumed a Gaussian dis tribution of source strengths. Some years ago, however, Shishvan and Van der Giessen [14] returned to the relation between Frank–Read length and the critical stress τnuc and proposed a log-normal distribution of Frank-Read lengths that is restricted by the size of the grain. The resulting distribution of source strengths is nearly log-normal too; the lowest strength is determined by the largest grain and the strength is bounded at high values by the theoretical shear strength of the material. In order to account for the fact that bowing-out of sources may be constrained by various obstacles, the log-normal source strength in the model of Shishvan and Van der Giessen [14], τLN nuc , is taken to be moderated according to 2D-DDP, obviously, cannot account for the detailed short-range in teractions between dislocation loops leading to, for instance, locks which may subsequently act as obstacles or evolve into Frank-Read sources. The objective of DDP with straight edge dislocations is to pro vide a platform to describe plastic deformation taking into account the effects of long-range dislocation interactions and source-limited plas ticity. Despite its simplicity, DDP studies of FCC single crystals and polycrystals in 2D have shown to be able to capture size effects in, for example, tension [8], bending[10] and indentation [23]; to predict the Hall–Petch effect [9] and to provide insight in the dual role of disloca tions in crack propagation [24]. While these studies aimed primarily at qualitative results, it has also been possible to quantitatively predict size effects in thin films [13,14] and micro cantilevers [25] using dedicated models with appropriate fitting to a single set of experimental results. The dimensions of the regions analyzed in these cited works and the strain levels achieved are feasible thanks to the simplicity of the 2D model; similar studies in 3D would require dramatically more compu tational resources and more refined initial conditions. As a compromise, Benzerga et al. 2. Two-dimensional DDP This projection from 3D loops to 2D dipoles has been used explicitly to define the Frank-Read source in 2D-DDP. In 3D, a Frank-Read source The field of discrete dislocation plasticity was initiated by a long standing interest in dislocation pattern formation. Ghoniem et al. and 2 Computational Materials Science 187 (2021) 110129 Fig. 3. (a) Elliptical dislocation loop on x-z plane with Burgers vector b and major and minor axes, 2p and 2q, respectively. t and n are the unit line vector and unit normal vector of a dislocation segment, respectively. N is the unit normal to the x-z plane. Ms and Me are the mobilities of pure screw and edge segments, respectively. (b) Edge dislocation–dipole on the x-z plane, denoted by ‘T’ symbols with the as yet unknown effective mobility Meff. T. Katiyar and E. Van der Giessen Computational Materials Science 187 (202 Computational Materials Science 187 (2021) 110129 T. Katiyar and E. Van der Giessen Fig. 3. (a) Elliptical dislocation loop on x-z plane with Burgers vector b and major and minor axes, 2p and 2q, respectively. t and n are the unit line vector and unit normal vector of a dislocation segment, respectively. N is the unit normal to the x-z plane. Ms and Me are the mobilities of pure screw and edge segments, respectively. (b) Edge dislocation–dipole on the x-z plane, denoted by ‘T’ symbols with the as yet unknown effective mobility Meff. stress acting on the leading dislocation in a pile-up exceeds the obstacle strength τobs, it is released to move past the obstacle. is a pinned dislocation segment that bows out in the direction of its Burgers vector under the influence of an applied shear stress; when the stress is sufficiently high, the bowed-out configuration becomes unsta ble, the loop bows around the pinning points until segments with opposite line direction meet, annihilate and leave a free dislocation loop and a copy of itself. In 2D, the Frank-Read segment is projected onto a point, which generates a dislocation dipole when the shear stress has maintained the critical value τnuc (representing the stress for which the bowed-out segment becomes unstable) for a sufficiently long time (to represent the time between the unstable configuration at τnuc and the complete loop) [6,22]. 2. Two-dimensional DDP [26] proposed an extension of the set of constitutive rules summarized above which incorporate some short-range 3D effects. In this paper, we will stay with the standard 2D formulation and make a first step towards a DDP formulation for BCC crystals. 4.1. Background and introduction where Fn is the magnitude of the Peach-Koehler force resolved in the glide plane and normal to the segment. Under an applied shear stress σxy, Fn is given by Fn = σxyb.The plastic strain rate tensor ˙ε induced by the expanding loop in a certain volume V is given by As summarized in Section 2, 2D-DDP has been used extensively in studying the size-dependent plastic behaviour under tension of FCC specimens with dimensions below tens of microns. These studies aimed at revealing what the key physical mechanisms are behind, for instance, the Hall–Petch effect in bulk polycrystals [9] and to either falsify or substantiate strain gradient plasticity theories, see e.g. [35]. In view of the relevance for the behaviour of thin films, a significant fraction of these studies were performed for thin films with columnar grains, e.g. [13,36,14], a configuration that was investigated under plane strain conditions in experiments carried out by Xiang and Vlassak [37]. ˙ε = 1 2V ∮ vLP dl(N ⊗b + b ⊗N), (3) (3) where the integral is the rate of change of the area swept by the loop and N is the unit normal to the slip plane. For the N and b as described by the loop in Fig. 3(a) and with the aid of Eq. (2), the only non-zero compo nent of ˙ε is given by A second group of studies has been concerned with the effect of limited plasticity on crack propagation, either under monotonic or cyclic loading [24,11,12]. A two-dimensional analysis of such problems is justified as plastic deformation near the crack tip is predominantly plane strain under mode-I loading. While the studies cited above assumed isotropic elasticity, Shishvan and Van der Giessen [38] took the cubic symmetries of single crystals into account. ˙εLP xy = b 2V ∮ vLP dl = b2 2Vσxy ∮ M ( θ ) dl. (4) (4) (where the superscript LP is used to indicate quantities for a loop). The corresponding quantity for an edge dislocation dipole lying on the x-z plane, Fig. 3(b), with out-of-plane length L is given by These two characteristic classes of problems are revisited in the subsequent subsections by considering the mobility of the edge dislo cations to be governed by the effective mobility according to Eq. (8). 3. Projection of dislocation mobility to 2D In FCC crystals, the mobility is uniform along a dislocation loop and, therefore, its value has been simply adopted in 2D discrete dislocation plasticity. However, in BCC crystals, the mobility along a loop depends on the nature of the dislocation segment—i.e., edge versus screw—with the screw segments generally being much slower. It is not obvious what mobility to use for dislocations in 2D-DDP for BCC crystals, but it is intuitive that the mobility of edge dislocations in 2D (or 2D dislocations) needs to reflect the slow-moving screw segments in terms of an effective mobility Meff. This idea is illustrated in Fig. 3. To achieve a functional form of Meff in terms of the mobility of a loop in 3D, we propose a projection method that is rooted in equating the plastic strain rate of a dislocation loop in 3D to that of a 2D dipole. Before computing the plastic strain rate of a loop, we first recapitulate some basic notions of plastic deformation produced by an expanding dislocation loop. τnuc = τLN nuc + τ0 nuc, (1) (1) τnuc = τLN nuc + τ0 nuc, with a material dependent parameter τ0 nuc that is determined by fitting the yield stress to experimental data. After having done so for free- standing copper micrometer thin films comprising columnar grains of a particular size and film thickness, Shishvan and Van der Giessen [14] were able to predict the size dependent response of films with other dimensions at remarkable quantitative precision. In materials that have not been produced by vapor deposition like thin films, but whose manufacturing has involved forming and heat treatments, there is a variety of obstacles such as inclusions, precipitates, forest dislocations on intersecting slip planes etc., that can hinder the glide of dislocations. These are explicitly modeled in 2D-DDP as point obstacles that are randomly distributed over the slip planes with a density ρobs. A dislocation arriving at such an obstacle gets pinned, possibly leading to the development of a dislocation pile-up. When the An elliptical dislocation loop is considered to be representative of dislocations in both FCC and BCC metals at room temperature. Pio neering theoretical work by DeWit and Koehler [27] on the critical equilibrium shapes of dislocations in anisotropic crystals and subsequent 3 T. Katiyar and E. 4. Effect of effective versus fixed mobility in FCC materials (2) vLP = M ( θ ) Fn, (2) 3. Projection of dislocation mobility to 2D Van der Giessen Computational Materials Science 187 (2021) 110129 may seem to be a pathological consequence of the approach but can readily be eliminated by the requirement that the slipped areas are the same in 3D and in 2D, consistent with the equality of the plastic strain rates. This means that πD2/4 = sL, where s denotes the distance between the two edge dislocations. Hence L is at any instant given by refinements by Kovacs [28], Foreman [29] and Fitzgerald [30] predict the shape of dislocations to be elliptical with the major axis aligned with the Burgers vector due to the lower energy of pure screws compared to edges. Even though these studies did incorporate the crystallographic anisotropy, the conclusion relies on the assumption of isotropic mobility. With a strong anisotropy in mobilities of edges and screws in BCC crystals, the dislocation loops can become more elongated along the major axis. In-situ observations of dislocation kinetics in pure iron at room temperature by Caillard [31,32] indeed report the shapes to be approximately elliptical. There are studies such as [33,34] that predict the emergence of sharp corners in the equilibrium shapes; however, this occurs at very high temperature (̃1180 K) and hence, are ignored in our analysis. It is noted that by assuming an elliptical shape, factors like line tension and self energy that help to determine this shape are incorpo rated implicitly. L = πD2 4s , (7) and therefore Meff = 2s D M. (8) L = πD2 4s , (7) L = πD2 4s , (7) and therefore and therefore Meff = 2s D M. (8) Meff = 2s D M. (8) The key observation to make here is that the effective mobility for edges in an FCC crystal is, in general, not equal to the mobility of the loop. This is something that has not been incorporated in any 2D-DDP study known to the authors. Granted the principle of equality of plastic strain, this observation thus raises the question how large the effect of the discrepancy is on the results obtained by simply assuming the effective mobility to be equal to M. We will first explore this for a number of key 2D-DDP predictions published in the literature in the subsequent section. 3. Projection of dislocation mobility to 2D In order to determine the plastic strain rate caused by the expansion of an elliptical loop, we consider the x-z plane to be the glide plane of a dislocation loop with Burgers vector b (magnitude b) along the x-axis, as shown in Fig. 3(a). The semi-major and semi-minor axes of the loop are denoted by p and q, respectively. Since the glide mobility of the loop M depends on the nature of a dislocation segment locally, it is taken to be a function of the angle θ between b and the line direction t i.e., M ≡M(θ). This mobility determines the glide velocity vLP of a segment according to 4.1. Background and introduction In order to compute this effective mobility at every step, we not only need to keep track of the size of every dipole through ˙εDP xy = bvDP( 2L ) 2V = Lb2Meffσxy V , (5) (5) where vDP is the glide speed of each of the dislocations (though in opposite directions) determined by vDP = Meffbσxy in terms of their effective mobility Meff. ˙s = 2vDP, ˙s = 2vDP, As mentioned previously, the projection of the mobility of a loop in an FCC crystal to a 2D description is facilitated by the fact that the mobility is uniform. Thus, an initially circular loop will remain circular, with major and minor axes equal to the diameter of the loop, D. In that case, with M(θ) ≡M, Eq. (4) yields but also the corresponding loop diameter D in 3D. Following [6], the size s0 of a freshly generated dipole in 2D is determined by their mutual attractive force and the strength of source strength τnuc through s0 = μb 2π(1 −ν)τnuc s0 = μb 2π(1 −ν)τnuc ˙εLP xy = πb2D 2V Mσxy. (6) ˙εLP xy = πb2D 2V Mσxy. (6) with μ the elastic shear modulus and ν Poisson’s ratio. In 3D, when assuming that the critical shape of a bowed-out Frank–Read segment is semi-circular, the classical line-tension argument [29] yields that the critical diameter D0 for the same source strength is given by [26,14]. The requirement that the plastic strain rate ˙εDP xy produced by the extending edge dipole in 2D be equal to the strain rate corresponding to the expanding loop, ˙εLP xy according to (6), leads to the expression D0 = βnuc μb τnuc , with βnuc = 1 2π ( ln D0 b + 0.5 ) . Meff = πD 2L M. Meff = πD 2L M. Assuming that the initial diameter D0 and the initial dipole size s0 are Assuming that the initial diameter D0 and the initial dipole size s0 are The emergence of the out-of-plane length L of the edge segments in 2D Computational Materials Science 187 (2021) 110129 T. Katiyar and E. Van der Giessen Fig. 4. Unit cell of a free-stranding thin film with thickness h comprising columnar grains having width d. The film is subject to an overall uniaxial strain ε under plane strain conditions perpendicular to the plane of observation. 4.1. Background and introduction Grains are of FCC type and have the three effective slip systems; the orientation of each grain is specified by the acute angle ϕ of one of the slip planes with the direction of straining. Fig. 4. Unit cell of a free-stranding thin film with thickness h comprising columnar grains having width d. The film is subject to an overall uniaxial strain ε under plane strain conditions perpendicular to the plane of observation. Grains are of FCC type and have the three effective slip systems; the orientation of each grain is specified by the acute angle ϕ of one of the slip planes with the direction of straining. Fig. 5. Predicted size dependence of the uniaxial stress–strain response of polycrystalline thin films from 2D DD simulations at constant dislocation mobility M and with the effective mobility Meff according to Eq. (8). There are eight randomly oriented grains in the unit cell, and grain sizes vary from small (d = 0.5 μm), through medium (d = 1.5 μm) to large (d = 3 μm) size. In all cases, the film thickness is h = d/1.5 so that the grain aspect ratio is constant. The response is averaged over multiple realizations and the error bars denote the standard deviation from the mean Fig. 6. Predicted size dependence of the uniaxial stress–strain response of single crystals from 2D-DDP simulations at constant dislocation mobility M and with the effective mobility Meff according to Eq. (8). Grain sizes d vary as indicated, with the film thickness given by h = d/1.5. The response is averaged over multiple realizations and the error bars denote the standard deviation from the mean. Fig. 5. Predicted size dependence of the uniaxial stress–strain response of polycrystalline thin films from 2D DD simulations at constant dislocation mobility M and with the effective mobility Meff according to Eq. (8). There are eight randomly oriented grains in the unit cell, and grain sizes vary from small (d = 0.5 μm), through medium (d = 1.5 μm) to large (d = 3 μm) size. In all cases, the film thickness is h = d/1.5 so that the grain aspect ratio is constant. The response is averaged over multiple realizations and the error bars denote the standard deviation from the mean. Fig. 6. 4.1. Background and introduction Predicted size dependence of the uniaxial stress–strain response of single crystals from 2D-DDP simulations at constant dislocation mobility M and with the effective mobility Meff according to Eq. (8). Grain sizes d vary as indicated, with the film thickness given by h = d/1.5. The response is averaged over multiple realizations and the error bars denote the standard deviation from the mean. Fig. 6. Predicted size dependence of the uniaxial stress–strain response of single crystals from 2D-DDP simulations at constant dislocation mobility M and with the effective mobility Meff according to Eq. (8). Grain sizes d vary as indicated, with the film thickness given by h = d/1.5. The response is averaged over multiple realizations and the error bars denote the standard deviation from the mean. Fig. 5. Predicted size dependence of the uniaxial stress–strain response of polycrystalline thin films from 2D DD simulations at constant dislocation mobility M and with the effective mobility Meff according to Eq. (8). There are eight randomly oriented grains in the unit cell, and grain sizes vary from small (d = 0.5 μm), through medium (d = 1.5 μm) to large (d = 3 μm) size. In all cases, the film thickness is h = d/1.5 so that the grain aspect ratio is constant. The response is averaged over multiple realizations and the error bars denote the standard deviation from the mean. parameter values, [14] obtained very good agreement with the depen dence of the response on grain size and film thickness in thin Cu films found experimentally. governed by the same ratio (7), the loop diameter can be readily computed at every moment from the current dipole size s using Fig. 5 recapitulates some typical results from [14] for constant mobility, M = 104 (Pa s)−1, and confronts them with corresponding re sults obtained by assuming the effective mobility according to (8) instead of the constant mobility. Evidently, there is no significant in fluence of the mobility law on the predicted response. The mobility of individual dipoles will be different, but apparently the interplay be tween the collective motion of many dislocations in the various grains in the film and dislocation generation eventually yields an almost identical overall response. l D = D0 ̅̅̅̅̅̅̅̅̅ s/s0 √ . 4.2. Size effects in single and polycrystals Fig. 4 depicts the problem under consideration: a unit cell of length w out of a free-standing thin film of thickness h, which is subjected to a prescribed tensile strain ε. The unit cell contains eight grains of width d, each having three effective slip systems (see Fig. 1) but with random orientation ϕ relative to the tensile direction. Just like in [13,14], the boundaries of the grains that coincide with the traction-free boundaries along the top and the bottom of the film are transparent to dislocations, while the grain boundaries inside the film are assumed to be impene trable for dislocations. Furthermore, it is assumed that the grains contain dislocation sources at a density of ρnuc = 30 μm−2, which obey a log- normal distribution of strength as discussed in Section 2 with τ0 nuc = 47MPa, but are free of obstacles. In order to allow for a clean com parison, all parameter values are taken to be the same as in [14]: magnitude of Burgers vector b = 0.25 nm, shear modulus μ = 41GPa, Poisson ratio ν = 0.34; applied strain rate ˙ε = 2500 s−1. With these One might think that the influence of the mobility fades during the averaging over the grains in the film. To gain some insight in this, we also investigate the behavior of single crystals with the same aspect ratio d/h = 1.5 and the crystal orientation ϕ = 45◦(cf. Fig. 4). For each grain size, at least 10 realizations of randomly positioned sources have been analyzed (20 realizations for the middle grain size) and the mean stress response is shown in Fig. 6. We obtain a qualitatively similar grain size dependence as in the polycrystalline film, which again is not affected significantly by using the variable, effective mobility instead of the underlying constant mobility of a dislocation loop. The observed size dependence, with smaller grains being harder, is qualitatively consistent with the Hall–Petch effect recovered in the study 5 5 Computational Materials Science 187 (2021) 110129 Fig. 7. Model formulation for crack propagation under small-scale yielding conditions in an FCC crystal under mode-I conditions. (a) The crack coincides with a (010) plane, has its front in the [ 101 ] direction and grows in the [101] direction. 4.2. Size effects in single and polycrystals The slip systems (i) through (iii) are oriented in agreement with the effective slip systems proposed by Rice [20], and discrete dislocations are constrained to move inside a process window. Reproduced from [38]. (b) The traction–separation law, given in Eq. (9), used ahead of the crack tip to describe propagation of the crack. T. Katiyar and E. Van der Giessen Computational Materials Science 187 (2021) 110129 T. Katiyar and E. Van der Giessen yar and E. Van der Giessen Fig. 7. Model formulation for crack propagation under small-scale yielding conditions in an FCC crystal under mode-I conditions. (a) The crack coincides with a (010) plane, has its front in the [ 101 ] direction and grows in the [101] direction. The slip systems (i) through (iii) are oriented in agreement with the effective slip systems proposed by Rice [20], and discrete dislocations are constrained to move inside a process window. Reproduced from [38]. (b) The traction–separation law, given in Eq. (9), used ahead of the crack tip to describe propagation of the crack. Fig. 8. Normalized applied stress intensity as a function of crack extension Δa in a single FCC crystal with four different realizations of dislocation sources and obstacles, analyzed with a fixed dislocation mobility M as well as with effective mobility Meff according to Eq. (8). Fig. 8. Normalized applied stress intensity as a function of crack extension Δa in a single FCC crystal with four different realizations of dislocation sources and obstacles, analyzed with a fixed dislocation mobility M as well as with effective mobility Meff according to Eq. (8). 6 Computational Materials Science 187 (2021) 110129 T. Katiyar and E. Van der Giessen Computational Materials Science 187 (2021) 110129 Fig. 9. Variation of effective mobility Meff corresponding to dislocation loops of increasing aspect ratio α = p/q in α-iron. The representative average value 〈 Meff〉is indicated by the dashed line. of doubly periodic arrays of grains by Balint et al. [9]. It bears emphasis, however, that while the size effect in the latter study was attributed to the development of pile-ups against the impenetrable grain boundaries, here only the vertical grain boundaries are impenetrable, thereby diminishing the effect of pile-ups. At the same time, there is an addi tional inherent size dependence in the log-normal distribution of source strengths. 4.3. Crack growth in single crystals In the absence of dislocation activity, this cohesive law predicts unstable crack growth at an applied stress in tensity factor K0 = ̅̅̅̅̅̅̅̅̅̅̅̅̅̅ ϕn/CG √ with CG = 1 2 ( 1 −ν )/μ for isotropic elasticity (the expression for cubic elasticity can be found in [38]). M ( θ ) = ( M2 esin2θ + M2 scos2θ )1/2, (10) (10) with θ the orientation of the dislocation line direction with respect to the Burgers vector, as indicated in Fig. 3. Following Section 3, we impose that the plastic strain rates in 3D and 2D be the same, by equating Eqs. (4) and (5) to give the following intermediate form of the effective mobility: In the presence of dislocation activity that tends to shield the crack tip, part of the energy released by crack growth is dissipated, thus requiring higher values of KI for continued crack growth. This is illus trated in Fig. 8 for four realizations of source and obstacle distributions at a density of ρnuc = 32 μm−2 and ρobs = 32 μm−2, respectively. In these computations, since the crystal is infinitely large, the original size- independent source model of [6] is adopted, characterized by a Gaussian distribution of strength with mean value of 50 MPa and a standard deviation of 10 MPa. The obstacles have a strength of τobs = 150 MPa and the loading is applied at a constant rate of ˙KI = 100 GPa m1/2 s−1. Meff = 1/V 2L/V ∮( M2 esin2θ + M2 scos2θ )1/2 dl. (11) (11) With small re-arrangements, this expression can be written as Meff = ∮ dl/V 2L/V ∮( M2 esin2θ+M2 scos2θ )1/2dl ∮ dl =ρ3D ρ2D ∮( M2 esin2θ+M2 scos2θ )1/2dl ∮ dl , (12) (12) in terms of the dislocation densities for the loop and edge dislocation dipole, ρ3D = ∮ dl/V and ρ2D = 2L/V, respectively. When these densities are taken to have the same value, Eq. (12) can be further evaluated to yield We observe that for each realization, there is a noticeable effect of using the effective dislocation mobility instead of the constant mobility, M = 104 (Pa s)−1, as adopted in previous work [24,38]. 4.3. Crack growth in single crystals The overall stress–strain response of a single crystal still averages out over many discrete slip planes. By contrast, crack growth in a crystal with a small plastic zone is governed by the local stress field near the crack tip. This stress field is the superposition of the remote elastic field and the stress fields of the dislocations present in the crystal. A fraction of these dislocations have caused plastic relaxation and shield the crack tip, but also anti-shielding dislocations get generated. Indeed, Clever inga et al. [24] have concluded that discrete dislocation plasticity near crack tips has two opposite effects: energy dissipation associated with stress relaxation and the development of localized, high-stress regions by dislocation structures close to the crack tip. i Fig. 9. Variation of effective mobility Meff corresponding to dislocation loops of increasing aspect ratio α = p/q in α-iron. The representative average value 〈 Meff〉is indicated by the dashed line. The problem studied first in [24] and more recently within the framework of anisotropic elasticity in [38] is sketched in Fig. 7(a). Small-scale yielding conditions around the tip of an infinitely long crystal are assumed, with the remote mode-I loading prescribed in terms of an incrementally increasing KI elastic field. Crack growth is described in terms of a cohesive law that is based on the universal binding law and characterized by the following relation between the opening traction Tn and the separation Δn: mobility can be safely used in the case of FCC crystals. Building on a similar notion, in this section, we develop the effective 2D mobility in BCC crystals based on the same principle of conserving the plastic strain. One difference with FCC crystals is that the dislocation mobility in BCC is not constant along a loop: the screw mobility Ms is often signif icantly lower than the mobility of edge segments, Me. To account for the variability of dislocation mobility with the character of the dislocation segment, the glide mobility of the loop is taken to follow the widely used BCC0 construction [39] where the mobility of any dislocation segment is given by Tn ( Δn ) = σmax Δn δn exp ( −Δn δn + 1 ) . (9) (9) As a consequence, crack propagation is governed by two parameters: the cohesive strength σmax and the critical opening δn, or the work of sepa ration ϕn = exp(1)σmaxδn. 4.2. Size effects in single and polycrystals It is the latter feature that is responsible for the agreement of the entire stress–strain curve with experiments in [14], while the size independent nucleation model of [6] used in [13] only yields agreement in terms of the yield strength. 4.3. Crack growth in single crystals However, the magnitude of the effect is smaller than the variability in the crack extension curves among different source and obstacle distributions, so that all previous discrete dislocation results on crack growth paper are essentially unchanged when taking the effect of dislocation loop size on mobility into account. Meff = Ms E [ 1 −M2 e M2 sα2 ]/ E [ 1 −1 α2 ] , (13) (13) where α = p/q is the aspect ratio of the elliptical loop and E[⋅] denotes the complete elliptic integral, E[k] = ∫π/2 0 ̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ 1 −ksin2θ √ dθ, which emerges as a consequence of integration along an elliptical arc. 6. Methodology The sensitivity of the plastic flow stress to the applied loading rate has been characterized extensively in the literature [44–47]. In general, the flow stress is observed to rise slowly at low strain rates and to in crease more rapidly at high strain rates (typically for ˙ε > 103 s−1). The increase in flow stress is attributed to different mechanisms in the two strain rate regimes: plastic strain at low strain rates is governed by thermal activation while at high strain rates, the mobility of dislocations is the limiting mechanism [44,48]. In addition, the 2D-DDP simulations on FCC polycrystals by Agnihotri and Van der Giessen [49] suggested that the Frank-Read mechanism is rate-limiting in the lower end of this high strain-rate regime. Recently, Cui et al. [50] emphasized the motion of screw dislocations in controlling the strain-bursts in tungsten pillars at high strain rates. As the rate sensitivity of the flow stress at high strain rates is directly coupled to the mobility of dislocations [49], we will compare 2D-DDP predictions of the rate sensitivity of flow stress at high strain rates in iron (at room temperature) against experiments. The experimental data to be used is the flow stress of as-received iron over a wide range of strain rates reported by Weston [51]. Of the data repro duced in Fig. 10, only the three highest strain rates will be used, to ensure we are in the mobility-controlled regime. The three flow stress values at the highest strain rates reported in Fig. 10 are the peak stress values (of the yield point phenomenon) occurring at ̃10% total strain [51] which is too large to be reached in 2D-DDP. Instead of these peak stresses, we will use the yield stress (flow stress at 0.2% plastic strain) at the highest strain rates, as extracted from the stress–strain curves in Weston [51]. It should be noted that the rate sensitivity of flow stress is independent of the strain at which the flow stress is measured, as was also observed by Ostwaldt et al. [52]. Fig. 10. Rate dependent flow stress of as-received Iron [51], shown with ‘▿’ symbols. The yield strength at the three highest strain rates has been super posed with ‘o’ symbols. α-iron at room temperature. The same value of Me has been found in a molecular dynamics study of the 1/2〈111〉(110) edge dislocation in BCC Fe [40], but Queyreau et al. 6. Methodology [41] reports a slightly lower value. Following Tang and Marian [42], we take the screw mobility to be two orders of magnitude smaller than Me for α-iron at room temperature. For these values of Ms and Me, the dependence of the effective mobility according to Eq. (13) on α = p/q is shown in Fig. 9. Assuming that loops are initially circular (α = 1), Meff(α) begins at a value that is signifi cantly lower than the pure edge mobility Me, and decrease with increasing α. This trend is easily understood by noting that the loop becomes more screw-like with increasing aspect ratio, and therefore the low screw mobility plays an increasingly important role. The maximum aspect ratio αmax a dislocation loop can achieve is limited by Me/Ms (= 100 in this case), as also reported by Kang et al. [43] for faceted dislocation loops in BCC. This can be readily appreci ated by considering an arbitrary dislocation loop with initial (at time t = 0) semi-major and -minor axes, p0 and q0, respectively. Since the major and minor axes at any time t > 0 are governed by the mobilities of pure edge and screw segments, respectively, the time-dependence of the aspect ratio is given by To predict the rate sensitivity in iron, the polycrystalline 2D-DDP framework with elastic anisotropy of Shishvan et al. [53] is employed. Polycrystalline iron is modeled as a two-dimensional array of square grains sized d, as shown in Fig. 11(a). Each grain is endowed with three effective slip systems consistent with plain-strain deformation in BCC α = p0 + Meτbt q0 + Msτbt . (14) α = p0 + Meτbt q0 + Msτbt . (14) In the limit of t→∞, this expression reduces to α→Me/Ms, implying that in real dislocation microstructures, loops can have any aspect ratio Fig. 11. (a) BCC Polycrystal comprising of two-dimensional array of square and randomly oriented (illustrated with grey scale) grains of size d and subjected to uniaxial strain rate loading (b) the plane-strain state of a BCC crystal: three effective slip systems under plain-strain condition on [101]–[010] plane. These slip- systems are randomly oriented with respect to the global coordinate system x-y. Fig. 11. 5. Effective 2D mobility in BCC crystals We further explore the dependence of Meff on α for values Me = 10000 Pa−1s−1 and Ms = 100 Pa−1s−1, which are representative for The results from the previous section indicate that the effective 2D 7 Fig. 10. Rate dependent flow stress of as-received Iron [51], shown with ‘▿’ symbols. The yield strength at the three highest strain rates has been super posed with ‘o’ symbols. T. Katiyar and E. Van der Giessen Computational Materials Science 187 (2021) 110129 T. Katiyar and E. Van der Giessen between the minimum of 1 and the maximum of αmax. If one would want to incorporate the associated variation in effective mobility seen in Fig. 9, one would need to keep track of the aspect ratio of every loop in 3D that is represented in the 2D simulations. Not only would this require a major modification of existing codes, it would nearly defeat the simplification obtained in a 2D computation. As an alternative, and without a bias to any aspect ratio, we here propose to take the mean value of the effective mobility over the possible α-range as a represen tative value of the effective mobility; this evaluates to 〈Meff〉= 475 Pa−1s−1 for α-iron. In the subsequent section, we describe the methodology to validate the obtained effective mobility, by comparison with the rate sensitivity of α-iron in 2D-DDP. 6. Methodology The grain size in Weston [51]’s experiments is ̃45 μm, but a grain size of 10 μm is used in the computations; the latter value was found to be sufficiently large to prevent grain size effects, yet small enough for the calculations to be computationally convenient (see Appendix A). Material parameters for single crystalline iron are b = 0.25 nm and C11 = 229 GPa, C12 = 134 GPa and C44 = 115 GPa. Due to the limited number of grains in the computational cell, the predicted overall Young’s modulus may deviate slightly from the experimental value; we adopt the correction proposed by Shishvan et al. [53] to ensure that the overall Young’s modulus corresponds to the experi mental value E = 210 GPa [54]. procedure as the one used successfully by Shishvan and Van der Giessen [14] for obstacle-free copper thin films. i The approach is as follows. First recall from Section 2 that the dis tribution of source strengths is assumed to follow-up a log-normal dis tribution that is governed by the grain size and by the theoretical strength. For the chosen grain size of 10 μm, this leads to a log-normal distribution with a mean source strength τLN nuc = 207MPa. The off-set value τ0 nuc in (1) is initially set to an arbitrary value of 237 MPa, and will eventually be tuned to fit the calculated yield stress to the experi mental value at the highest strain rate. Next, the source density is set so that there are enough dislocations sources to properly sample the log- normal distribution. Since in this study there are more, and larger, grains, we do not need as many sources as used by Shishvan and Van der Giessen [14]: a density of ρnuc = 20 μm−2 ensures that the log-normal distribution of strengths is represented accurately (i.e., with a maximum error of 12.5% at a confidence level of 90%). This value can be used for all strain rates since Agnihotri and Van der Giessen [49] have shown that the rate sensitivity of an FCC polycrystal is independent of the source density. Thus, we are left with two unspecified material pa rameters, i.e. obstacle density ρobs and obstacle strength τobs, both of which were not considered in the thin-film studies of [14] as they con tained a very low initial dislocation density. 6. Methodology Whereas hardening in those copper thin films, with columnar grain sizes around 1 μm, arises from dislocation pile-ups against the grain boundaries, hardening in poly crystalline bulk iron with grains of > 10 μm in 2D will have to arise from obstacles such as forest dislocations and small precipitates. Leaving the details to the subsequent subsection, the obstacle parameters will be determined by fitting the experimental strain hardening rate. The grains are initially dislocation free and dislocation nucleation is incorporated through point sources randomly distributed over the slip planes with a source density ρnuc. These sources mimic the Frank-Read mechanism in two dimensions (see Section 2); that is, when the resolved shear stress at a source exceeds its strength τnuc, an edge- dislocation dipole is created, with its sign being determined by the di rection of the resolved shear stress. 6. Methodology (a) BCC Polycrystal comprising of two-dimensional array of square and randomly oriented (illustrated with grey scale) grains of size d and subjected to uniaxial strain rate loading (b) the plane-strain state of a BCC crystal: three effective slip systems under plain-strain condition on [101]–[010] plane. These slip- systems are randomly oriented with respect to the global coordinate system x-y. Computational Materials Science 187 (2021) 110129 Fig. 12. (a) Uniaxial tensile stress–strain curves at ˙ε = 5200 s−1 with ρobs/ρnuc = 3. Contrary to the experimental results, the predictions for this obstacle density exhibit strain softening even at significantly high τobs/τnuc ratios (b) Uniaxial tensile stress–strain curves at ˙ε = 5200s−1 and ρobs/ρnuc = 7. The stress response for the case τobs/τnuc = 7 (shown in orange solid color) satisfies the experimentally observed strain hardening rate. T. Katiyar and E. Van der Giessen T. Katiyar and E. Van der Giessen T. Katiyar and E. Van der Giessen Computational Materials Science 187 (2021) 110129 Computational Materials Science 1 Fig. 12. (a) Uniaxial tensile stress–strain curves at ˙ε = 5200 s−1 with ρobs/ρnuc = 3. Contrary to the experimental results, the predictions for this obstacle density exhibit strain softening even at significantly high τobs/τnuc ratios (b) Uniaxial tensile stress–strain curves at ˙ε = 5200s−1 and ρobs/ρnuc = 7. The stress response for the case τobs/τnuc = 7 (shown in orange solid color) satisfies the experimentally observed strain hardening rate. crystals [20], see Fig. 11(b). By geometry, the effective lengths of the Burgers vectors on the three slip systems (i), (ii) and (iii) are b/ ̅̅̅ 3 √ , b and b, respectively. The grains are randomly oriented such that there is no texture in the polycrystal. All grain boundaries are assumed to be impenetrable to dislocations and there is a distance of 200b between parallel slip planes in a slip system. The top and bottom surface of the bulk polycrystal are assumed to be traction free, while a tensile strain rate is imposed at the ends. The computational cell is taken to comprise a 4 × 8 array of rectangular grains of size d = 10 μm. The choice of four layers is motivated by previous DDP studies that have shown that free- surface effects fade from three layers onwards [36] while eight grains per layer were enough to give a statistically meaningful average tensile response of copper [14]. 7. Results Fig. 12(a) shows the predicted tensile stress–strain response of iron at ρobs = 3ρnuc for a wide range of obstacle strengths. Each of the predicted stress–strain curves presented here and in the sequel, is the mean response of the polycrystal averaged over ten realizations of crystal orientations and random distributions of sources and obstacles. The results in Fig. 12(a) reveal, in particular, significant strain softening after the yield point. This softening is to be attributed to the inefficiency of the obstacles to reduce the free-flight distance of dislocations; in the absence of any obstacles, the softening is even more pronounced (as demon strated in Appendix A). As expected, the softening reduces at higher obstacle strengths. However, even an obstacle strength as high as τobs = 21τnuc is not sufficient to produce realistic strain hardening as seen in The DDP model contains four parameters that are, as of yet, un known: ρnuc, τnuc, ρobs and τobs. Together with the dislocation mobility, these parameters control plasticity properties such as strain hardening and yield stress. Inspired by the power of the approach demonstrated in [14] for free-standing copper thin films, we will fit these source and obstacle parameters to the uniaxial response of iron for a particular strain rate, and then predict the flow stress at other strain rates. Since the highest strain rates can be analyzed most quickly, parameter identifi cation is carried out for the highest strain rate reported in Fig. 10: ˙ε = 5200 s−1. There is no unique way to fit the four source and obstacle parameters from a given stress–strain curve; we will adopt a similar 9 Fig. 13. Uniaxial tensile stress–strain curves corresponding to two different values of τnuc at ˙ε = 5200 s−1. Reducing the τnuc from 444 MPa to 237 MPa uniformly shifts down the plastic response to the desired yield stress level, indicated by the dotted black lines. T. Katiyar and E. Van der Giessen Computational Materials Science 187 (2021) 110129 T. Katiyar and E. Van der Giessen to bring down the mean source strength τnuc from 444 MPa to 237 MPa. The change in τnuc, while maintaining the ratio τobs/τnuc, shifts the plastic branch of the stress–strain curves to the desired yield stress level without affecting the shape, as demonstrated in Fig. 13. 8. Discussion To validate the effective-mobility approach, 2D-DDP simulations were employed to compare the prediction of the rate sensitivity of the yield stress at high strain rates in polycrystalline iron with experiments. Over the range of strain rates for which this comparison was possible, our predictions deviate from the experimental yield strength by less than 8%. Based on this comparison, we conclude that the proposed projection method provides a good estimate of the effective mobility Meff of α-iron dislocations in 2D-DDP. Fig. 13. Uniaxial tensile stress–strain curves corresponding to two different values of τnuc at ˙ε = 5200 s−1. Reducing the τnuc from 444 MPa to 237 MPa uniformly shifts down the plastic response to the desired yield stress level, indicated by the dotted black lines. experiments. This suggests there are insufficient obstacles in the simu lation cell. In previous 2D-DDP studies, the strain hardening slope has been observed to be much more sensitive to ρobs than to τobs. Therefore, we increased the obstacle density ρobs to 7ρnuc, closer to values used in [55,8]. A higher obstacle density not only leads to an increased proba bility of a dislocation getting pinned but also reduces the mean free path for dislocation glide, thus leading to a diminished plastic strain rate and increased strain hardening. Further, to assess the extent to which the effective mobility impacts the rate sensitivity prediction, we also calculated the rate sensitivity of iron by just using the screw mobility Ms as the mobility of the 2D dis locations and reported the result in Fig. 14(b) (see Appendix B for details of the fitting procedure). The rate sensitivity predicted for this mobility is observed to deviate substantially from the experimental data, with a deviation of at least 25% for the lowest strain rate considered here. The predicted σf vs ˙ε curve is steeper than that of the prediction by effective mobility, which is qualitatively consistent with the rate sensitivity of the yield stress in FCC crystals at high strain rates studied previously [49]. In fact, a slightly higher value of the effective mobility can even give a near-perfect agreement with the experimental rate-sensitivity, albeit at the cost of trial-and-error. For a similar range of obstacle strengths as considered before, Fig. 12 (b) indeed shows that ρobs = 7ρnuc remedies the strain softening. 7.2. Rate sensitivity prediction In the foregoing, the predicted yield stress at the highest strain rate has been fit to the experimental value. To test the rate sensitivity of the yield strength, the same material parameters are used for tensile simu lations at lower strain rates: ˙ε = 2000 and 200 s−1. The resulting stress–strain curves at all strain rates are plotted in Fig. 14(a), while the extracted strain rate dependence of the yield strength is reported in Fig. 14(b). The predictions of rate sensitivity of yield stress by the effective mobility gives good agreement with the experimental values, with the prediction being slightly over-sensitive. 8. Discussion Now, the value of τobs is chosen such that predicted strain hardening complies with the experimental strain hardening rate which at ˙ε = 5200s−1 is estimated to be ∼10000 MPa from [51]. This leads to selecting τobs = 7τnuc (cf. Fig. 12 (b)). Lastly, we observe that the yield stress (at 0.2% plastic strain) pre dicted at the determined values of ρobs and τobs differs significantly from the experimental yield stress of ∼330 MPa. To fit the predicted yield stress to the experimental value, the fit parameter τ0 nuc is adjusted, so as The input parameters to the effective mobility computed from Eq. (13) are the mobilities of pure screw and edge segments (Ms and Me). In this work, the edge and screw mobilities were assumed to be constant in Fig. 14. (a) Uniaxial stress–strain curves of polycrystalline iron at various strain rates, obtained using the effective 2D mobility according to Section 5. (b) Predicted strain rate sensitivity of the yield stress in comparison with the experimental data from Weston [51]. The error bars represent the standard deviation of the yield stresses obtained from the 10 realizations of grain orientations and distributions of dislocation sources and obstacles. rates, obtained using the effective 2D mobility according to Section 5. (b) Predicted ata from Weston [51]. The error bars represent the standard deviation of the yield Fig. 14. (a) Uniaxial stress–strain curves of polycrystalline iron at various strain rates, obtained using the effective 2D mobility according to Section 5. (b) Predicted strain rate sensitivity of the yield stress in comparison with the experimental data from Weston [51]. The error bars represent the standard deviation of the yield stresses obtained from the 10 realizations of grain orientations and distributions of dislocation sources and obstacles. 10 Computational Materials Science 187 (2021) 110129 T. Katiyar and E. Van der Giessen Fig. 15. The effective mobility Meff versus aspect ratio of dislocation loop α = p/q for Iron, Tantalum and Molybdenum at room temperature. Ms = 100 Pa−1s−1 and Me = 2000 Pa−1s−1 for Ta [60] and Ms = 25 Pa−1s−1 and Me = 1000 Pa−1s−1 for Mo [61]. For each metal, the maximum aspect ratio αmax is given by the mobility ratio Me/Ms. mobilities is directly incorporated into the effective mobility function through Ms and Me, as proposed by Po et al. [57]. 8. Discussion One of the characteristic features of BCC metals is the existence of an athermal transition temperature Ta above which thermal activation of the slow moving screw segments is no longer rate limiting. Above Ta (for example, Ta = 340 K and 465 K, respectively for α-Fe and Ta [58]), the mobilities of screw and edge segments become comparable. The dislo cation dynamics and the resulting dislocation microstructures then are akin to that of FCC crystals [59]. The proposed 3D–2D conversion of dislocation mobility is consistent with this. When Me = Ms and α = 1, as is the case in FCC, Eq. (13) reduces to Meff = Me by virtue of the fact that E[0] = π/2. The projection method estimates the effective mobility of 2D dislo cations in iron to be ̃475 Pa−1s−1 by employing the values of Ms and Me at room temperature. For other BCC metals, for example, Tantalum and Molybdenum, the method estimates the effective mobility values at room temperature to be 323 and 97 Pa−1s−1, respectively, as shown in Fig. 15. Unfortunately, we are not aware of experimental rate-sensitivity data against which these values can be validated. A Peierls barrier for screw segments is not included in the formula tion here: the analysis begins with the assumption that the shear stress levels are high enough to set the dislocation loop into motion. If the Peierls stress is relevant, for instance in case of fracture studies [62], one may consider to explicitly incorporate it in 2D-DDP in the form of an effective Peierls stress τeff p , defined in a similar way as we have defined the effective mobility Meff. This requires defining the Peierls stress along the elliptical dislocation loop by a smoothly varying function, such as, for example, τp(θ) = (τ2 esin2θ + τ2 s cos2θ) 1/2, with τe and τs the Peierls stress for pure edge and screw segments. A uniform effective τeff p along the loop is then defined through Fig. 15. The effective mobility Meff versus aspect ratio of dislocation loop α = p/q for Iron, Tantalum and Molybdenum at room temperature. Ms = 100 Pa−1s−1 and Me = 2000 Pa−1s−1 for Ta [60] and Ms = 25 Pa−1s−1 and Me = 1000 Pa−1s−1 for Mo [61]. For each metal, the maximum aspect ratio αmax is given by the mobility ratio Me/Ms. 8. Discussion stress such that the dislocation velocity follows a linear mobility law. While this typically holds true for edge segments, the screw segments move by thermally-assisted kink-pair nucleation and propagation at moderate to low temperatures. Their velocity follows an Arrhenius law which is nonlinear in stress. Several three-dimensional discrete dislo cation dynamics (3D-DDD) models, such as [56,57], have formulated phenomenological prescriptions of the nonlinear dependence of the screw velocity on stress and other variables, such as the length of the screw segments. While such details are important in 3D-DDD models, we have ignored them in the current analysis, primarily to demonstrate the viability of the notion of an effective mobility. Moreover, we note that the non-linearity decreases with increasing temperature and is essen tially absent at room temperature, i.e. the target temperature of this study. It also bears emphasis that the temperature dependence of the ∮ ellipse τp ( θ ) dl = ∮ ellipse τeff p dl. (15) (15) When τp(θ) is assumed to have a similar form as the mobility according to the BCC0 construction in (10), this procedure gives a similar expression for τeff p as for Meff (Eq. 13) with Ms and Me replaced with τe and τs. The τeff p at the initiation of the loop expansion can be estimated by substituting α = 1 in the equation. Just for illustration, using τe = 24.8 MPa and τs = 248 MPa for Tantalum at room temperature [60], we find τeff p ≈160 MPa. When τp(θ) is assumed to have a similar form as the mobility according to the BCC0 construction in (10), this procedure gives a similar expression for τeff p as for Meff (Eq. 13) with Ms and Me replaced with τe and τs. The τeff p at the initiation of the loop expansion can be estimated by substituting α = 1 in the equation. Just for illustration, using τe = 24.8 MPa and τs = 248 MPa for Tantalum at room temperature [60], we find τeff p ≈160 MPa. Fig. A.1. (a) Uniaxial stress–strain curves for three different grain sizes at ˙ε = 2800 s−1 (b) Corresponding dislocation microstructures at the end of simulations. ress–strain curves for three different grain sizes at ˙ε = 2800 s−1 (b) Corresponding dislocation microstructures at the end of Fig. A.1. 8. Discussion (a) Uniaxial stress–strain curves for three different grain sizes at ˙ε = 2800 s−1 (b) Corresponding dislocation micro 11 Computational Materials Science 187 (2021) 110129 T. Katiyar and E. Van der Giessen Fig. B.1. Uniaxial stress-strain curves of polycrystalline iron at various strain rates, obtained using the screw mobility Ms.. 2. We have proposed a 3D-to-2D procedure to combine the screw and edge mobilities of BCC loops into an effective mobility for use in 2D, based on the equality of the plastic slip rate produced by dislocation loops in 3D and that by dislocation dipoles in 2D. 2. We have proposed a 3D-to-2D procedure to combine the screw and edge mobilities of BCC loops into an effective mobility for use in 2D, based on the equality of the plastic slip rate produced by dislocation loops in 3D and that by dislocation dipoles in 2D. 3. The effective mobility of the BCC dislocations, strictly speaking, depends on the aspect ratio of the loop in 3D. For simplicity we propose to use the average value over all possible aspect ratios, so that the effective mobility can be directly computed from the edge and screw mobilities. 3. The effective mobility of the BCC dislocations, strictly speaking, depends on the aspect ratio of the loop in 3D. For simplicity we propose to use the average value over all possible aspect ratios, so that the effective mobility can be directly computed from the edge and screw mobilities. 4. The effective mobility thus obtained from the 3D-to-2D procedure is found to provide a good prediction of the rate sensitivity of yield of iron at room temperature. This gives some confidence that this method is useful for other BCC metals too. Acknowledgment The authors gratefully acknowledge the Netherlands Organisation for Scientific Research (www.nwo.nl) and the Materials Innovation Institute M2i (www.m2i.nl) for supporting this work. The work was performed within the framework of the Partnership Program of the aforementioned organisations and under the project number F22.2.13518a. We also thank the following industrial partners (via M2i) for their valuable support: Prorail, SKF, DAF, Tata Steel and Philips. The computations reported in Section 4 were carried out by Thiadrik Tiesma within the framework of his MSc research project. CRediT authorship contribution statement Fig. B.1. Uniaxial stress-strain curves of polycrystalline iron at various strain rates, obtained using the screw mobility Ms.. T. Katiyar: Conceptualization, Investigation, Methodology, Soft ware, Validation, Formal analysis, Data curation, Writing - original draft. E. Van der Giessen: Conceptualization, Methodology, Resources, Writing - review & editing, Supervision, Project administration, Funding acquisition. It bears emphasis that even after careful consideration of glide mobility and Peierls stress, 2D-DDP analyses are based on strong ide alizations and ignore a number of important 3D mechanisms, and hence, can only provide a qualitative description of plasticity. When carefully parameterized, however, it has been shown for FCC to possess quanti tative predictive abilities for the size-dependent yield strength in thin films [14]. The main goal of 2D-DDP is to capture the dislocation plas ticity rather than the dislocation microstructure. With this work, we have aimed to incorporate the slow-screw-dominated plastic deforma tion of BCC crystals in 2D in the form of an effective mobility. Other pertinent features of BCC crystal plasticity, such as non-Schmid behavior, need further attention and efforts to be addressed in 2D. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. 9. Concluding remarks In summary, the key observations and conclusions from this study are: 1. Dislocation dynamics methods for FCC crystals in the past have benefited from the fact that the mobility is uniform along the dislo cation loop. The screw segments of BCC dislocation loops, however, have a much lower mobility than edge segments. 2D discrete dislo cation plasticity for BCC crystals should therefore adopt an effective mobility that, somehow, incorporates this mobility difference. Data availability The raw/processed data required to reproduce these findings cannot be shared at this time due to technical or time limitations. References Shawki, Pressure-shear impact and the dynamic viscoplastic response of metals, Mech. Mater. 4 (1985) 375–385. l f h f [16] L.P. Kubin, Dislocations. Mesoscale Simulations and Plastic Flow, Oxford University Press, Oxford, 2013, https://doi.org/10.1093/acprof:oso/ 9780198525011.001.0001. [46] W. Tong, R.J. Clifton, S. Huang, Pressure-shear impact investigation of strain rate history effects in oxygen-free high-conductivity copper, J. Mech. Phys. Solids 40 (1992) 1251–1294. [17] R.J. Amodeo, N.M. Ghoniem, Dislocation dynamics. II. Applications to the formation of persistent slip bands, planar arrays, and dislocation cells, Phys. Rev. B 41 (1987) 6968–6976. [47] R.W. Armstrong, S.M. Walley, High strain rate properties of metals and alloys, Int. Mater. Rev. 53 (2008) 105–128. [48] G. Regazzoni, U.F. Kocks, P.S. Follansbee, Dislocation kinetics at high strain rates, Acta Metall. 35 (1987) 2865–2875. [18] L.P. Kubin, G. Canova, M. Condat, B. Devincre, V. Pontikis, Y. Br´echet, Dislocation microstructures and plastic flow: a 3D simulation, Solid State Phenom. 23–24 (1992) 455–472. [49] P.K. Agnihotri, E. Van der Giessen, On the rate sensitivity in discrete dislocation plasticity, Mech. Mater. 90 (2015) 37–46. [19] J. L´epinoux, L.P. Kubin, The dynamic organization of dislocation structures, Scr. Metall. 21 (1987) 833–838. i [50] Y. Cui, G. Po, P. Srivastava, K. Jiang, V. Gupta, N.M. Ghoniem, The role of slow screw dislocations in controlling fast strain avalanche dynamics in body-centered cubic metals, Int. J. Plast 124 (2020) 117–132. [20] J.R. Rice, Tensile crack tip fields in elastic-ideally plastic crystals, Mech. Mater. 6 (1987) 317–335. [51] G.M. Weston, Flow stress of shock-hardened Remco iron over strain rates from 0.001 to 9000 s-1, J. Mater. Sci. Lett. 11 (1992) 1361–1363. [21] C. Lemarchand, B. Devincre, L.P. Kubin, Homogenization method for a discrete- continuum simulation of dislocation dynamics, J. Mech. Phys. Solids 49 (2001) 1969–1982. [52] D. Ostwaldt, J.R. Klepaczko, P. Klimanek, Compression tests of polycrystalline alpha-iron up to high strains over a large range of strain, J. Phys. IV France 07 (1997) 385–390. [22] A.A. Benzerga, An analysis of exhaustion hardening in micron-scale plasticity, Int. J. Plast 24 (2008) 1128–1157. [53] S.S. Shishvan, S. Mohammadi, M. Rahimian, E. Van der Giessen, Plane-strain discrete dislocation plasticity incorporating anisotropic elasticity, Int. J. Solids Struct. 48 (2011) 374–387. [23] A. Widjaja, A. Needleman, E. Van der Giessen, The effect of indenter shape on sub- micron indentation according to discrete dislocation plasticity, Modell. Simul. Mater. Sci. Eng. 15 (2006) S121. [54] H.M. Ledbetter, R.P. References Vlassak, Bauschinger and size effects in thin-film plasticity, Acta Mater. 54 (2006) 5449–5460. [8] S. Papanikolaou, H. Song, E. Van der Giessen, Obstacles and sources in dislocation dynamics: Strengthening and statistics of abrupt plastic events in nanopillar compression, J. Mech. Phys. Solids 102 (2017) 17–29. [38] S.S. Shishvan, E. Van der Giessen, Mode I crack analysis in single crystals with anisotropic discrete dislocation plasticity: I. Formulation and crack growth, Modell. Simul. Mater. Sci. Eng. 21 (2013) 065006. [9] D.S. Balint, V.S. Deshpande, A. Needleman, E. Van der Giessen, Discrete dislocation plasticity analysis of the grain size dependence of the flow strength of polycrystals, Int. J. Plast 24 (2008) 2149–2172. [39] W. Cai, V.V. Bulatov, Mobility laws in dislocation dynamics simulations, Mater. Sci. Eng., A 387–389 (2004) 277–281. [40] S. Yu, C.Y. Wang, T. Yu, The kink-pair nucleation in edge dislocation motion, Solid State Sci. 11 (2009) 733–739. [10] H.H.M. Cleveringa, E. Van der Giessen, A. Needleman, Discrete dislocation analysis of bending, Int. J. Plasticity 15 (1999) 837–868. [41] S. Queyreau, J. Marian, M.R. Gilbert, B.D. Wirth, Edge dislocation mobilities in bcc Fe obtained by molecular dynamics, Phys. Rev. B – Condensed Matter Mater. Phys. 84 (2011). [11] V.S. Deshpande, A. Needleman, E. Van der Giessen, Discrete dislocation modeling of fatigue crack propagation, Acta Mater. 50 (2002) 831–846. [12] M.P. O’Day, W.A. Curtin, Bimaterial interface fracture: a discrete dislocation model, J. Mech. Phys. Solids 53 (2005) 359–382. [42] M. Tang, J. Marian, Temperature and high strain rate dependence of tensile deformation behavior in single-crystal iron from dislocation dynamics simulations, Acta Mater. 70 (2014) 123–129. [13] L. Nicola, Y. Xiang, J.J. Vlassak, E. Van der Giessen, A. Needleman, Plastic deformation of freestanding thin films: experiments and modeling, J. Mech. Phys. Solids 54 (2006) 2089–2110. [43] K. Kang, V.V. Bulatov, W. Cai, Singular orientations and faceted motion of dislocations in body-centered cubic crystals, Proc. Nat. Acad. Sci. 109 (2012) 15174–15178. [14] S.S. Shishvan, E. Van der Giessen, Distribution of dislocation source length and the size dependent yield strength in freestanding thin films, J. Mech. Phys. Solids 58 (2010) 678–695. [44] P. Follansbee, G. Regazzoni, U. Kocks, in: J. Hardling (Ed.), Mechanical Properties at high rates of strain, Institute of Physics, Bristol, 1984, pp. 71–80. [15] J.W. Christian, Some surprising features of the plastic deformation of body- centered cubic metals and alloys, Metall. Trans. A 14 (1983) 1237–1256. [45] R.W. Klopp, R.J. Clifton, T.G. References [1] G. Po, M.S. Mohamed, T. Crosby, C. Erel, A. El-Azab, N. Ghoniem, Recent progress in discrete dislocation dynamics and its applications to micro plasticity, JOM 66 (2014) 2108–2120. [29] A.J.E. Foreman, The bowing of a dislocation segment, Philos. Mag. 15 (1967) 1011–1021. [30] S.P. Fitzgerald, Frank-Read sources and the yield of anisotropic cubic crystals, Philos. Mag. Lett. 90 (2010) 209–218. [2] J.A. El-Awady, H. Fan, A.M. Hussein, Advances in discrete dislocation dynamics modeling of size-affected plasticity, in: Springer Series in Materials Science vol. 245, 2016, pp. 337–371, https://doi.org/10.1007/978-3-319-33480-6_11. [31] D. Caillard, Kinetics of dislocations in pure Fe. Part I. In situ straining experiments at room temperature, Acta Mater. 58 (2010) 3493–3503. [3] R. Sills, A. Aghaei, W. Cai, Advanced time integration algorithms for dislocation dynamics simulations of work hardening, Modell. Simul. Mater. Sci. Eng. 24 (2016) 45019. [32] D. Caillard, A TEM in situ study of alloying effects in iron. II - Solid solution hardening caused by high concentrations of Si and Cr, Acta Mater. 61 (2013) 2808–2827. [33] S.P. Fitzgerald, S.L. Dudarev, Interaction between dislocations in bcc iron at high temperature, J. Nucl. Mater. 386–388 (2009) 67–70. [34] S P Fi ld S A b S L D d W C i Di l i d i i l i [4] D. Weygand, L.H. Friedman, E. Van der Giessen, A. Needleman, Aspects of boundary-value problem solutions with three-dimensional dislocation dynamics, Modell. Simul. Mater. Sci. Eng. 10 (2002) 437–468. [34] S.P. Fitzgerald, S. Aubry, S.L. Dudarev, W. Cai, Dislocation dynamics simulation of Frank-Read sources in anisotropic α-Fe, Modell. Simul. Mater. Sci. Eng. 20 (2012). [5] H. Akhondzadeh, R. Sills, S. Papanikolaou, E. Van der Giessen, W. Cai, Geometrically projected discrete dislocation dynamics, Modell. Simul. Mater. Sci. Eng. (2018). [35] L. Nicola, M. Gurtin, E. Van der Giessen, Effect of defect energy on strain-gradient predictions of confined single-crystal plasticity, J. Mech. Phys. Solids 53 (2005) 1280–1294. g [6] E. Van der Giessen, A. Needleman, Discrete dislocation plasticity: a simple planar model, Modell. Simul. Mater. Sci. Eng. 3 (1995) 689–735. [36] R. Kumar, L. Nicola, E. Van der Giessen, Density of grain boundaries and plasticity size effects: a discrete dislocation dynamics study, Mater. Sci. Eng., A 527 (2009) 7–15. i [7] P.J. Guruprasad, A.A. Benzerga, Size effects under homogeneous deformation of single crystals: a discrete dislocation analysis, J. Mech. Phys. Solids 56 (2008) 132–156. [37] Y. Xiang, J.J. Appendix A To select an appropriate grain size to model polycrystalline iron, the stress–strain response of iron single crystals with different grain size d is calculated at an applied strain rate of 2800 s−1. Inspired by [14], the source density and the mean source strength are ρnuc = 30 μm−2 and τnuc = 180 MPa, respectively, and the crystals are obstacle-free. The Burgers vector and the cubic elastic constants are taken to have the same values as adopted in the main text. l As seen in Fig. A.1(a), the crystals with d = 10 μm and 25 μm exhibit a steady-state flow stress level preceded by an abrupt drop in the stress level. This dip is an artefact originating from the lack of initial dislocations and obstacles. The activation of a large number of weak sources at the beginning of plasticity leads to a burst in dislocation activity. In the case without obstacles considered in this Appendix, the nucleated dislocations freely traverse long distances across the slip planes before stopping at the boundary. Notably, the stress response of the crystal with d = 1 μm shows a strong size- effect, as also elucidated by the corresponding dislocation microstructure in Fig. A.1(b), revealing long pile-ups of dislocations against the crystal boundaries. The length of pile-ups is significantly shorter in the larger grain sizes. Crystals with the experimental grain size ̃45 μm [51] would also flow steadily at a constant stress level but would require excessive computational time due to huge number of dislocation sources. Therefore, the grain size of d = 10 μm is selected which is not only computationally less expensive but also prevent any significant size effects. The flow stress of the 25 μm sized grain is only slightly lower. In so far as this difference in flow stress between a grain of d = 10 μm and a bigger one is significant, it can be accounted for by varying the strength of sources τnuc as also discussed in the methodology section. 12 T. Katiyar and E. Van der Giessen Computational Materials Science 187 (2021) 110129 Appendix B In order to predict the rate sensitivity of iron using the screw mobility Ms, we follow the fitting procedure for the highest strain rate as described in the subSections 7.1 and 7.2. The mean source strength τnuc is initially set to an arbitrary value 444 MPa and a sufficiently high density of obstacles ρobs = 7ρnuc is chosen. The obstacle strength τobs is fit to the experimental strain hardening and is found to be 7τnuc. Lastly, to match the experimental yield stress, the source strength τnuc is adjusted to 208 MPa. With these parameters being fit to the highest strain rate, the stress response of the iron polycrystal at the lower strain rates is predicted and shown in Fig. B.1. The predicted values of 0.2% yield strength are included in Fig. 14(b). [28] I. Kovacs, The critical shear stress needed for the operation of a dislocation segment as a frank-read source in anisotropic crystals, Physica Status Solidi (B) 3 (1963) 140–144. [28] I. Kovacs, The critical shear stress needed for the operation of a dislocation segment as a frank-read source in anisotropic crystals, Physica Status Solidi (B) 3 (1963) 140–144. [62] Q. Chen, V.S. Deshpande, E. Van der Giessen, A. Needleman, Friction stress effects on mode i crack growth predictions, Scripta Mater. 48 (2003) 755–759. [61] H.M. Zbib, T. Díaz De La Rubia, M. Rhee, J.P. Hirth, 3D dislocation dynamics: stress-strain behavior and hardening mechanisms in fcc and bcc metals, J. Nucl. Mater. 276 (2000) 154–165. [59] W.A. Spitzig, A.S. Keh, The effect of orientation and temperature on the plastic flow properties of iron single crystals, Acta Metall. 18 (1970) 611–622. References Reed, Elastic properties of metals and alloys, I. Iron, nickel, and iron-nickel alloys, J. Phys. Chem. Ref. Data 2 (1973) 531–618. g [24] H.H.M. Cleveringa, E. Van der Giessen, A. Needleman, A discrete dislocation analysis of mode I crack growth, J. Mech. Phys. Solids 48 (2000) 1133–1157. [24] H.H.M. Cleveringa, E. Van der Giessen, A. Needleman, A discrete dislocation analysis of mode I crack growth, J. Mech. Phys. Solids 48 (2000) 1133–1157. [25] E. Tarleton, D. Balint, J. Gong, A. Wilkinson, A discrete dislocation plasticity study of the micro-cantilever size effect, Acta Mater. 88 (2015) 271–282. [55] S.S. Chakravarthy, W.A. Curtin, Effect of source and obstacle strengths on yield stress: a discrete dislocation study, J. Mech. Phys. Solids 58 (2010) 625–635. [ ] ilb i d d d f [25] E. Tarleton, D. Balint, J. Gong, A. Wilkinson, A discrete dislocation plasticity study of the micro-cantilever size effect, Acta Mater. 88 (2015) 271–282. [56] M.R. Gilbert, S. Queyreau, J. Marian, Stress and temperature dependence of screw dislocation mobility in α-Fe by molecular dynamics, Phys. Rev. B – Condensed Matter Mater. Phys. 84 (2011). [26] A. Benzerga, Y. Br´echet, A. Needleman, E. Van der Giessen, Incorporating three- dimensional mechanisms into two-dimensional dislocation dynamics, Modell. Simul. Mater. Sci. Eng. 12 (2003) 159–196. [57] G. Po, Y. Cui, D. Rivera, D. Cereceda, T.D. Swinburne, J. Marian, N. Ghoniem, A phenomenological dislocation mobility law for bcc metals, Acta Mater. 119 (2016) 123–135. [27] G. DeWit, J.S. Koehler, Interaction of dislocations with an applied stress in anisotropic crystals, Phys. Rev. 116 (1959) 1113–1120. 13 l [60] Z.Q. Wang, I.J. Beyerlein, An atomistically-informed dislocation dynamics model for the plastic anisotropy and tension-compression asymmetry of BCC metals, Int. J. Plast 27 (2011) 1471–1484. Computational Materials Science 187 (2021) 110129 T. Katiyar and E. Van der Giessen T. Katiyar and E. Van der Giessen Computational Materials Science 187 (2021) 110129 Computational Materials Science 187 (2021) 110129 [58] T. Suzuki, Y. Kamimura, H.O. Kirchner, Plastic homology of bcc metals, Philos. Mag. A: Phys. Condensed Matter, Struct., Defects Mech. Properties 79 (1999) 1629–1642. [59] W.A. Spitzig, A.S. Keh, The effect of orientation and temperature on the plastic flow properties of iron single crystals, Acta Metall. 18 (1970) 611–622. [62] Q. Chen, V.S. Deshpande, E. Van der Giessen, A. Needleman, Friction stress effects on mode i crack growth predictions, Scripta Mater. 48 (2003) 755–759. l [60] Z.Q. Wang, I.J. Beyerlein, An atomistically-informed dislocation dynamics model for the plastic anisotropy and tension-compression asymmetry of BCC metals, Int. J. Plast 27 (2011) 1471–1484. 14 14
https://openalex.org/W2777498278	https://europepmc.org/articles/pmc5739439?pdf=render	English	null	Users’ thoughts and opinions about a self-regulation-based eHealth intervention targeting physical activity and the intake of fruit and vegetables: A qualitative study	PloS one	2,017	cc-by	9,973	RESEARCH ARTICLE Editor: Jacobus P. van Wouwe, TNO, NETHERLANDS Received: January 18, 2017 Accepted: December 6, 2017 Published: December 21, 2017 Methods Twenty adults from the general population (mean age = 42.65, 11 women) and twenty adults with type 2 diabetes (mean age = 64.30, 12 women) performed ‘MyPlan 1.0’, which is a self- regulation-based eHealth intervention designed to increase physical activity and the intake of fruit and vegetables in the general population. The opinions and preferences of end-users were explored using a think aloud procedure and a questionnaire. During a home visit, par- ticipants were invited to think aloud while performing ‘MyPlan 1.0’. The thoughts were tran- scribed verbatim and inductive thematic analysis was applied. Copyright: © 2017 Poppe et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Abstract Citation: Poppe L, Van der Mispel C, De Bourdeaudhuij I, Verloigne M, Shadid S, Crombez G (2017) Users’ thoughts and opinions about a self-regulation-based eHealth intervention targeting physical activity and the intake of fruit and vegetables: A qualitative study. PLoS ONE 12(12): e0190020. https://doi.org/10.1371/journal. pone.0190020 Purpose EHealth interventions are effective in changing health behaviours, such as increasing physi- cal activity and altering dietary habits, but suffer from high attrition rates. In order to create interventions that are adapted to end-users, in-depth investigations about their opinions and preferences are required. As opinions and preferences may vary for different target groups, we explored these in two groups: the general population and a clinical sample. Users’ thoughts and opinions about a self- regulation-based eHealth intervention targeting physical activity and the intake of fruit and vegetables: A qualitative study Louise Poppe1,2☯, Celien Van der Mispel1,2☯, Ilse De Bourdeaudhuij1, Maïte´ Verloigne1, Samyah Shadid3, Geert Crombez2,4 Louise Poppe1,2☯, Celien Van der Mispel1,2☯, Ilse De Bourdeaudhuij1, Maïte´ Verloigne1, Samyah Shadid3, Geert Crombez2,4 Louise Poppe1,2☯, Celien Van der Mispel1,2☯, Ilse De Bourdeaudhuij1, Maïte´ Verloigne1, Samyah Shadid3, Geert Crombez2,4 1 Department of Movement and Sports Sciences, Faculty of Medicine and Health Sciences, Ghent University, Ghent, Belgium, 2 Department of Experimental-Clinical and Health Psychology, Faculty of Psychology and Educational Sciences, Ghent University, Ghent, Belgium, 3 Department of Endocrinology, Ghent University Hospital, Ghent, Belgium, 4 Centre for Pain Research, University of Bath, Bath, United Kingdom a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 ☯These authors contributed equally to this work. louise.poppe@ugent.be ☯These authors contributed equally to this work. * louise.poppe@ugent.be Introduction The prevalence of many chronic diseases, such as type 2 diabetes, cardiovascular diseases and cancer, is high and still rising [1]. A healthy lifestyle, including sufficient physical activity (PA) and fruit and vegetable intake, can prevent these diseases, or alter their negative consequences [2–6]. However, only 35% of the Western adult population follows the guidelines of 30 min- utes moderate- to vigorous PA on 5 days per week (and preferably every day), whereas 24% meets the norm of eating 5 or more servings of fruit and vegetables per day [6–8]. Even people for whom having a healthy lifestyle is key in the management of their disease, such as patients with type 2 diabetes, often fail to be physically active on a regular basis and fail to conform to a disease-specific dietary regime [9–11]. Consequently, there is a strong need for effective and easy-to-implement interventions that target a healthy lifestyle. Internet-based interventions have the potential to reach a large part of the population, while still being able to offer a personal approach. Indeed, via computer tailoring e- (elec- tronic) health features can be adapted to the user’s specific needs. Such interventions are promising in changing health behaviours [12], such as increasing PA levels and altering dietary habits [13–18], especially when interventions are informed by solid theory and use behaviour change techniques that are evidence-based [16, 19–21]. However, even when eHealth interven- tions are theory- and evidence-based, attrition is often high and affects the potential effect of these programmes [22, 23]. In order to address this challenge, it has been recommended to involve users in the design and testing of eHealth tools [24]. The experience and opinion of end-users about core elements of online interventions may further guide eHealth development and understand eHealth usage [25]. As eHealth interventions are often used in various target groups [26], it is important to investigate possible differences in perceptions, opinions and preferences between user groups. Only by doing so the needs of different target populations can be taken into account. In this paper, we investigate the perspectives of users on a web-based programme (the web- site ‘MyPlan 1.0’) that aims to increase PA and fruit and vegetable intake. Conclusions BOF.24J.2015.0007.01). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. BOF.24J.2015.0007.01). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Although originally created for the general population, ‘MyPlan 1.0’ can also be used in adults with type 2 diabetes. Nevertheless, more adaptations are needed to make the eHealth intervention more convenient and less time-consuming. Furthermore, users’ ideas regarding a healthy lifestyle should be taken into account when designing online interventions. Competing interests: The authors have declared that no competing interests exist. Competing interests: The authors have declared that no competing interests exist. Users’ thoughts and opinions about a self-regulation-based eHealth intervention Results Both groups had similar opinions regarding health behaviours and ‘MyPlan 1.0’. Participants generally liked the website, but often experienced it as time-consuming. Furthermore, they regularly mentioned that a mobile application would be useful to remind them about their goals on a daily basis. Finally, users’ ideas about how to pursue health behaviours often hin- dered them to correctly use the website. Funding: MV and LP are supported by a grant from the Fund for Scientific Research Flanders (FWO) (grant numbers: MV: FWO13/PDO/191; LP:FWO.3F0.2015.0045.01; www.fwo.be). The project is also funded by BOF (Ghent University Special Research Fund) (grant number: 1 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Participants We wanted to include twenty participants from the general population and twenty participants with type 2 diabetes. The intended sample size was based on a similar study using a think aloud interview [35]. Furthermore, the meta-analysis of Hwang and Salvendy (2010) indicated that 10±2 participants is sufficient for usability tests such as the think aloud method [36]. The participants from the general population were recruited via an available database, con- sisting of individuals who had expressed their interest to participate in studies of the Ghent Health Psychology Research Group via a website (http://www.healthpsychology.ugent.be/ vrijwilligers), and via the snowball sampling technique. Participants were purposively sampled to have an equal distribution of men versus women, participants with low versus high educa- tion level and younger versus older persons. Inclusion criteria were being 18 years old and Dutch speaking. Participants with type 2 diabetes were recruited via advertisements distrib- uted by the Diabetes Association Flanders and the Ghent University Hospital and via the snowball sampling technique. For this group, we aimed to create an equal distribution in men versus women. Patients had to be 18 years old, Dutch speaking and being 1 month post- diagnosis to be eligible for participation in the study. Six persons from the general population were not willing to take part in the study. One per- son with type 2 diabetes could not participate because she did not have a computer. Conse- quently, another person with type 2 diabetes was recruited. Some of the participants were acquaintances of the interviewers. The study was approved by the Committee of Medical Ethics of the Ghent university hospi- tal (B670201526613) and written informed consent was obtained. There was no reimburse- ment for participation in the study. Introduction ‘MyPlan 1.0’ is grounded in self-regulation theory [27, 28], and guides users through the processes of behav- iour change via different evidence-based techniques, such as setting specific goals (goal-set- ting), deciding how these goals can be reached (action planning), foreseeing barriers and possible solutions (coping planning) and monitoring the behaviour change process. These behavioural targets and self-regulation techniques are relevant for the general population as well as for individuals with chronic diseases [29]. The potential use of ‘MyPlan 1.0’ by patients with type 2 diabetes was also suggested by general practitioners who reflected upon the useful- ness of that intervention for their practice [30]. Indeed, disease management of patients with type 2 diabetes also includes alterations in PA and diet [31]. Developing an inclusive tool may then be non-stigmatizing and more cost-effective than creating an intervention specifically designed for adults with type 2 diabetes. 2 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Users’ thoughts and opinions about a self-regulation-based eHealth intervention ‘MyPlan 1.0’ is effective in increasing PA and the intake of fruit and vegetables (trial regis- tration on ClinicalTrials.gov: NCT02211040) [32, 33]. Adults of the general population per- ceive the intervention as feasible and acceptable [32, 34]. Notwithstanding, only 24% of the users completed the entire intervention [34]. Such high attrition rates may indicate that ‘MyPlan 1.0’ was designed too much from a top-down perspective (i.e. from a theoretical point of view) and highlights the need for more in-depth research regarding the users’ perspectives. Assessing users’ experiences while performing the online programme may guide further adap- tations to the programme, and may also inform other researchers about how possible end- users experience online programmes and how programmes can be adapted to better meet the users’ needs. The aim of this paper is to understand the experience of potential users, both adults from the general population and individuals with type 2 diabetes, via a think aloud procedure and a self-report questionnaire. Because self-report questionnaires only allow general inferences and are sensitive to recall bias, we also used a think aloud procedure to assess users’ thoughts while performing each step of the programme. In so doing, users provide more immediate reactions in comparison with the opinions expressed in retrospective focus groups and interviews. PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 The intervention ‘My Plan 1.0’ ‘MyPlan 1.0’ is an eHealth tool designed to increase PA and the intake of fruit and vegetables in the general population [37]. The fully-automated and freely accessible website (www. mijnactieplan.be) incorporates several self-regulation techniques [27] and consists of three modules: a start module, a first follow-up module (one week after the start module) and a sec- ond follow-up module (one month after the start module). 3 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Users’ thoughts and opinions about a self-regulation-based eHealth intervention Fig 1. Overview of the start module. https://doi.org/10.1371/journal.pone.0190020.g001 Fig 1. Overview of the start module. https://doi.org/10.1371/journal.pone.0190020.g001 https://doi.org/10.1371/journal.pone.0190020.g001 Fig 1 displays an overview of the components of the start module and the order in which they appear. First, users choose whether they would like to increase their PA, their intake of fruit, or their intake of vegetables, and answer questions assessing demographic information. Thereafter, they fill out a questionnaire to assess the baseline level of their selected health behaviour. PA is assessed by the International Physical Activity Questionnaire (IPAQ-L) [38]. Fruit and vegetable intake is assessed by the Flemish Fruit Test and Vegetable Test [39]. Users also answer questions about personal determinants for behaviour change, such as outcome- expectancies and self-efficacy. This is done for research purposes. Next, users receive tailored feedback based on their answers on the questionnaire assessing the chosen health behaviour, and are asked whether they would like to make a plan to change this behaviour. Users can decide to make a plan or leave the website. By going through the coping and action planning components, users respectively look for solutions to tackle possible barriers (e.g. “I will put my running shoes at the door so I don’t forget about my plan”) and create their own specific plan to be more physically active, to eat more fruit or to eat more vegetables (e.g. “Every Tuesday I will run for 30 minutes”). During this process, users have the opportunity to create implemen- tation intentions in the form of an if-then plan (e.g. “If I come home from work, I will run for 30 minutes in the neighbourhood) [40]. Their specific plans together with possible barriers and solutions are then shown in a printable format. Finally, the website invites users to moni- tor their behaviour change and to send their plan to friends and family in order to receive social support. PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Protocol and procedures The protocol of the think aloud procedure was based on a think aloud study by Yardley and colleagues (2010) [35]. Two female researchers (LP and CVDM) and two female master stu- dents in Clinical Psychology (ED and VVH) performed the think aloud procedure. LP has a Master’s degree in Experimental and Theoretical Psychology. CVDM has a Master’s degree in Clinical Psychology. The master students were trained by LP and CVDM. Participants were visited at home by one of the researchers. They first completed a questionnaire assess- ing demographic information. Participants with type 2 diabetes also reported the time since their diagnosis. Participants received the following instructions for the think aloud procedure: “‘MyPlan’ helps people to live a healthier life. Via this tool you can choose whether you like to eat more vegetables, consume more fruit or increase your physical activity. ‘MyPlan’ will give you advice about your current health behaviours and help you with making a plan to change these behaviours step by step. Currently, we try to improve the programme, and you can help us with this. We will require you to perform the programme and develop your own plan. When going through the programme, please state out aloud what comes to your mind. Please do not refrain from giving critical remarks, as we can learn a lot from these comments. Also positive experiences can be stated.” To let participants become acquainted with verbalizing their thoughts, a short exercise was provided. Participants were instructed to imagine their house and count the windows. They had to state out aloud how they imagined walking through each room and counting the win- dows. After the exercise, participants went through the website on their own computer at their own pace. Verbalizations were voice-recorded and the computer screen was filmed via a tablet. In contrast to the standard intervention, participants were instructed to start the first follow- up module immediately after completing the start module. Participants did not perform the second follow-up session, because it is similar to the first follow-up session. When participants forgot to verbalize their thoughts, prompts such as “please try to say out aloud what comes to your mind” or “what comes to your mind when you see this?” were given by the researcher. When participants completed the second module of the programme, the researcher asked them how they generally perceived the website and what they liked or disliked. The intervention ‘My Plan 1.0’ PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 4 / 19 Users’ thoughts and opinions about a self-regulation-based eHealth intervention During the follow-up modules (one week and one month after the start module), users complete questions to assess whether they have been able to change the health behaviour of their choice and to receive tailored feedback on their change. Thereafter users have the possi- bility to adapt their plan based upon the success or failure of their previous plan. PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Data analysis For the Website Evaluation Questionnaire, sum scores for each scale were calculated by adding the scores for each item of a subscale (minimum = 1; maximum = 5), resulting in a possible maximum score of 15 for each scale. Inductive thematic analysis was conducted in order to identify recurring patterns in par- ticipants’ perceptions about ‘MyPlan 1.0’. Thematic analysis has been defined as “a method for identifying, analysing and reporting patterns (themes) within data” ([42] page 6). Induc- tive or “bottom-up” thematic analysis codes the data without a pre-existing framework to do so [42]. This technique was chosen for two reasons. First, by using a data-driven way of cod- ing we maximally explore users’ perceptions instead of framing their ideas according to a spe- cific theory. Second, we wanted to explore whether different themes would emerge as a function of group, i.e. the participants from the general population and the type 2 diabetes group. In order to code the data we followed the analysis process described by Braun and Clarke (2006). In a first step, all the recordings were transcribed verbatim. None of the transcripts were returned to the participants for comments. Two researchers (CVDM and LP) read the tran- scripts to get acquainted with the data. While reading, some general findings were written down. In the second step, the data were read again, and initial codes were generated using qualitative data analysis software nVivo 11 (QSR International Pty. Ltd. Version 11, 2015). During this phase, no limit was set on the amount of generated codes. CVDM and LP each coded the data of 10 randomly selected participants from the general population and 10 partic- ipants with type 2 diabetes. In the third step, codes related to different opinions or experiences (e.g. codes expressing opinions regarding lay-out versus codes expressing experiences of awareness) identified in more than one participant’s transcript were brought together in differ- ent themes and a first differentiation between main and subthemes was established (e.g. opin- ions regarding lay-out and user-friendliness are both related to the design of the website, but not to the usefulness of the website). The codes were now classified under the themes accord- ing to the principles of internal homogeneity and external heterogeneity. Themes with a low number of codes were removed or integrated within other themes. Protocol and procedures Finally, participants filled out the Dutch version of the Website Evaluation Questionnaire [41]. This questionnaire consists of three subscales, each having three items. The first subscale assesses users’ perceptions regarding the personal relevance of the eHealth tool. The second subscale measures the extent to which users experience the tool enjoyable and attractive (i.e. their engagement with the tool). The third subscale assesses whether users feel like the tool helped them with self-assessment and goal-setting. All items are rated on a Likert-scale, rang- ing from 1 (strongly disagree) to 5 (strongly agree). The home visits were carried out from March 2016 until May 2016, and lasted approxi- mately 75 minutes. 5 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Users’ thoughts and opinions about a self-regulation-based eHealth intervention Participants Twenty adults from the general population and twenty adults with type 2 diabetes participated in the study. Demographic information for both groups is shown in Table 1. Data analysis This phase was carried out by CVDM for the data of the participants from the general population, whereas LP did the same for the data of the participants with type 2 diabetes. Next, all data was read again but with the identified themes in mind. This was done to check whether the data was well captured by the themes. Finally, CVDM and LP discussed and defined the final themes for both groups. If no consensus was reached, a third researcher (LD) was consulted. The participants did not provide feedback on the findings. The completed version of the consolidated criteria for reporting qualitative research (COREQ) checklist is added as S1 File. Website Evaluation Questionnaire Table 2 shows the results from the Website Evaluation Questionnaire for both the general sam- ple and the individuals with type 2 diabetes. A fairly good score was given for each of the sub- scales by both groups. 6 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Users’ thoughts and opinions about a self-regulation-based eHealth intervention Table 1. Demographic information. General Sample (n = 20) Diabetes Sample (n = 20) Mean age (SD, range) 42.65 years (14.47, 20– 60) 64.30 years (15.30, 18–83) Women 11 (55%) 12 (60%) Level of education Primary school 0 (0%) 1 (5%) Lower secondary education 1 (5%) 3 (15%) Higher secondary education 13 (65%) 3 (15%) College 3 (15%) 9 (45%) University 4 (20%) 3 (15%) Marital status Single 11 (55%) 3 (15%) Married 8 (40%) 15 (75%) Cohabiting 1 (5%) 0 (0%) Widowhood 0 (0%) 2 (10%) Mean reported time since diagnosis (SD, range) NA 145.80 months (95.21, 6–324) https://doi.org/10.1371/journal.pone.0190020.t001 Table 1. Demographic information. Think aloud procedure The content of the remarks from the general sample and the adults with type 2 diabetes was very similar. No group-specific themes were identified. Consequently, the results of both groups are discussed together. Table 3 gives an overview of the themes and shows the impor- tance of each subtheme in the dataset by displaying the number of participants who endorsed the subtheme. Knowledge. Many participants stated their opinion about a healthy lifestyle (i.e. the posi- tive effects of sufficient PA and the intake of fruit and vegetables). Some users agreed that hav- ing a healthy lifestyle has a positive effect on their physical and mental wellbeing, whereas others disagreed. Below some quotes of participants are provided. Underlined text indicates that the participant was reading the content from the website out aloud. Table 2. Results from the Website Evaluation Questionnaire. Questions according to subscales Scale Range General Sample Mean (SD, range) Diabetes Sample Mean (SD, range) Personal Relevance 3–15 10.45 (2.16, 6–14) 10.85 (1.93, 6–14) The information and advice provided by the website were of personal relevance to me 1–5 3.70 (0.73, 2–5) 3.65 (0.75, 2–5) The website addressed my specific problems 1–5 3.10 (0.91, 1–4) 3.45 (0.83, 2–5) The information and advice provided by the website were appropriate for me 1–5 3.65 (0.99, 1–5) 3.75 (0.72, 2–5) Engagement 3–15 10.90 (2.97, 5–14) 11.35 (2.03, 8–15) The website kept my attention 1–5 3.85 (0.99, 2–5) 4.05 (0.83, 2–5) The website was engaging 1–5 3.45 (1.15, 1–5) 3.80 (0.70, 3–5) I found the website enjoyable to use 1–5 3.60 (1.23, 1–5) 3.50 (1.00, 1–5) Goal-setting 3–15 11.10 (2.02, 7–13) 11.65 (1.93, 7–15) The website helped me to plan 1–5 3.60 (0.75, 2–4) 3.80 (0.83, 2–5) The website helped me to think about my own behaviour 1–5 4.00 (0.79, 2–5) 4.00 (0.79, 2–5) The website helped me to set goals regarding my PA / fruit intake/ vegetable intake 1–5 3.50 (1.15, 1–5) 3.85 (0.81, 2–5) htt //d i /10 1371/j l 0190020 t002 Table 2. Results from the Website Evaluation Questionnaire. 7 / 19 Users’ thoughts and opinions about a self-regulation-based eHealth intervention Table 3. Number of participants that addressed a subtheme. PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Think aloud procedure Theme General Sample (N = 20) Diabetes Sample (N = 20) Knowledge Perceptions regarding a healthy lifestyle 16 (80%) 13 (65%) Perceptions regarding behaviour change 18 (90%) 15 (75%) Design of the intervention General appreciation of the website 14 (70%) 14 (70%) User friendliness 19 (95%) 20 (100%) Time-efficiency 15 (75%) 18 (90%) Lay-out 8 (40%) 4 (20%) Usefulness of the website Opinion on the motivational value of the website 16 (80%) 10 (20%) Opinion on the informative value of the website 10 (50%) 9 (45%) Feelings of awareness elicited by the website 14 (70%) 10 (50%) Personal relevance 15 (75%) 19 (95%) Recommendations 10 (50%) 6 (30%) https://doi.org/10.1371/journal.pone.0190020.t003 Table 3. Number of participants that addressed a subtheme. “Then I will have a smaller chance on getting diseases (cf. when eating more vegetables). . . I hope so, but I don’t think it has much to do with it.” (Female, 57) “If I eat more fruit, then I will have a smaller chance on getting diseases. I do think that eating fruit might indeed be healthy.” (Male, 23) “My mental wellbeing will be better (cf. when being more physically active). . . yeah, since I will exercise and lose weight.” (Male, 49) “My mental wellbeing will be better. . . those things (cf. mental wellbeing and eating fruit) are poorly related, it hardly affects your mental state.” (Male, 78, type 2 diabetes) Users often expressed their uncertainty about which actions actually are health-related. “I make fresh soup. Does that also count?” (Female, 83, type 2 diabetes) “I once read that eating fruit after a meal is not healthy because of the fermentation process. . . I will practically never eat it after a meal.” (Female, 64, type 2 diabetes) “I often eat yogurt with fruit, but I guess that does not count?” (Female, 48) Some users shared their opinion about behaviour change and stated that creating action Some users shared their opinion about behaviour change and stated that creating action plans was unnecessary. In addition, users often believed that creating specific plans is difficult due to changes in their week schedules. PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 8 / 19 Users’ thoughts and opinions about a self-regulation-based eHealth intervention Users often stated that they did not need or want any social support when trying to have a healthy lifestyle. Think aloud procedure Consequently, many people were reluctant to send their plans to friends or family. “Do you think other people will support you? What has that to do with it? I think it has nothing to do with each other.” (M l 23) “Do you think other people will support you? What has that to do with it? I think it has nothing to do with each other.” (Male, 23) PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 (Male, 23) “I’m not going to send it, I’m doing this for myself” (Female, 73, type 2 diabetes) urthermore, many users stated that they did not want to monitor their behaviour change. “I’m not going to monitor this, I have my brain. I don’t want to monitor it. I’ll remember it, I don’t need a book or diary, no. I’m not going to do this, I’m not a child.” “In my case, it’s hard to define whether I have reached my goal and if I reached it, then I don’t have to write it down. I don’t really see the point to start counting how many times I took the stairs. I would rather not keep track of this every day.” (Male, 58, type 2 diabetes) The design of the intervention. Participants also verbalized their general appreciation of the website and often described ‘My Plan’ as a questionnaire rather than an interactive tool that may help users with behaviour change. “I don’t like to fill in questionnaires, I think it’s a waste of time and it’s not useful for me. I don’t think I would fill it in a second time.” (Female, 20) 9 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Users’ thoughts and opinions about a self-regulation-based eHealth intervention Almost all users had complaints regarding user-friendliness. Generally, there were four types of remarks. First, participants reported difficulties to fill out the questionnaires to assess PA, or intake of fruit or vegetables. “I don’t really know how much vegetables weigh. A tomato, how much does it weigh? The rec- ommendation is 300 grams, but I don’t know how I should envision 300 grams. And one vege- table weighs more than another one. . .” (Female, 57) Second, most participants found it hard to know how to answer questions, and doubted whether they were actually doing what the website requested them to do. “The advice gives you a recapitulation and then you think “I didn’t answer the questions well”, maybe because I didn’t understand them, I don’t know.” (Female, 69, type 2 diabetes) Third, participants often reported difficulties with making the ‘if-then plan’. “I find “if-then” difficult. . . It’s hard to put it into words, because well. . . PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 (Male, 47) “I think that making a plan helps you stick to the goal. If you really have a plan, I think you believe that you. . . otherwise you wouldn’t make it! So, I will do it.” Nevertheless, participants who did not like the idea of making specific plans, were not moti- vated by the action planning component of the intervention. “I don’t really think it’s necessary. I don’t think people will need the website to fill-in what they have done. I don’t need it. I also don’t think it will have a large influence on whether or not I will reach my goals, to eat fruit five times a week. That’s my opinion.” “If this would go together with a dietician or a doctor or do it in a hospital, then I can imagine that it will be followed, but people on their own, they fill it in and forget about it.” “If this would go together with a dietician or a doctor or do it in a hospital, then I can imagine that it will be followed, but people on their own, they fill it in and forget about it.” (Female, 76, type 2 diabetes) Participants often verbalized their opinion about the informative value of the website and stated that they had learned something. These statements were mostly evoked when partici- pants read the guidelines regarding PA and the intake of fruit and vegetables or when the web- site displayed their BMI. “Two pieces of fruit a day, interesting!” (Male, 56) “Eating olives also counts? It’s strange that it’s a fruit, it surprises me!” (Male, 60) “I find this (cf. the action plan) informative” (Male, 78) (Male, 23) I am an economist, but you can’t say three or five or seven, “if-then” just gives you too much freedom.” (Male, 47) Finally, the website offered users the possibility to send their plans to friends and family. However, users often did not know the email address of the persons they wanted to send it to. “Would you like to show your plan to someone. . . yeah, why not. Email address, I don’t know by heart. . .” (Male, 78, type 2 diabetes) “Would you like to show your plan to someone. . . yeah, why not. Email address, I don’t know by heart. . .” (Male, 78, type 2 diabetes) Many participants had remarks about the time-efficiency and stated that it took too much time to go through the website. Furthermore, filling out the questionnaires was often perceived as a waste of time. “But it does take some time and you need to read it carefully too. And now we are lucky, but if you fill it out during the evening and the phone rings or at the office and someone walks in, you could be distracted, maybe it should be formulated more concisely.” “I think it’s—well not for me, because I have time—I think it takes too long and beats aroun the bush. It’s too long, I would have forgotten about it!” “I think it’s—well not for me, because I have time—I think it takes too long and beats around the bush. It’s too long, I would have forgotten about it!” (Female, 76, type 2 diabetes) (Female, 76, type 2 diabetes) (Female, 76, type 2 diabetes) Participants who had remarks about the lay-out. They often disliked how the website looked, but appreciated the large font of the text. 10 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 attractive website. . . Look at it. ng. . . It’s for old people.” as a large font, it invites you to es) ite. Opinions about the moti ts found it stimulating to mak ntake of fruit or vegetables. cf. the action plan) is well-defin sers’ thoughts and opinions about Users’ thoughts and opinions about a self-regulation-based eHealth intervention “I really don’t find it an attractive website. . . Look at it. It is ugly as sin, it is not attractive, it is not interactive, it’s nothing. . . PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 (Male, 23) It’s for old people.” (Male, 58, type 2 diabetes) Usefulness of the website. Opinions about the motivational value of the website were mostly positive. Participants found it stimulating to make specific plans to be more physically active or to increase their intake of fruit or vegetables. “Two pieces of fruit a day, interesting!” “Two pieces of fruit a day, interesting!” (Male, 56) “Eating olives also counts? It’s strange that it’s a fruit, it surprises me!” (Male, 60) “I find this (cf. the action plan) informative” (Male, 78) PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 11 / 19 Users’ thoughts and opinions about a self-regulation-based eHealth intervention Participants often expressed feelings of awareness when reading the guidelines about PA and the intake of fruit and vegetables, or when filling out the questionnaire regarding these behaviours. “Damn, that’s not much. Horrible. . .” (Female, 56, type 2 diabetes) “I think it’s good that you are getting confronted with it and they highlight it and that they really inculcate it.” (Male, 78, type 2 diabetes) Perceptions regarding personal relevance were also stated. Although the website was cre- ated for adults in the general population, some participants felt like they did not belong to the target group and believed that the website was poorly tailored to their situation. “It really depends on your age. You will not work in the garden when your back is starting to hurt. And if you are young, you will do more household chores, you will paint and perform renovations, but if it is done then it’s done.” “Physical activity during work. . . Well, first you need to have work!” (Male, 57) Participants with type 2 diabetes often stated that the suggestions for fruit did not fit within their dietary scheme. “I shouldn’t eat apples, it’s all sugars they say. . . Mangos. . . I also don’t eat it.” (Male, 78, type 2 diabetes) While going through the website, some people shared their ideas about how ‘MyPlan 1.0’ could be adapted. Most recommendations involved creating a mobile application so people would be easily reminded to their goals. “If you really want to do it well, then you have to monitor it. But I am not good at doing this, so an app would be really ideal to do this.” so an app would be really ideal to do this. (Male, 47) “With an app you will be able to reach much more since you can send messages” (Male, 60) “It would be nice if the website showed graphs and if you could create your own tables.” (Male, 49) “I find it non-committal. I would rather send stuff on a daily basis. PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Discussion This study investigated the perspectives of users on a self-regulation-based eHealth tool labelled ‘MyPlan 1.0’ via a self-report questionnaire and a think aloud procedure. Noteworthy, the results of the Website Evaluation Questionnaire and the think aloud procedure indicate that participants from the general population and patients with type 2 diabetes had similar per- ceptions about health behaviours and ‘MyPlan 1.0’, except regarding the suggestions proposed by the website for fruit intake. This difference is appropriate. The norms for the amount of fruit during the day are different for people with type 2 diabetes than for the general popula- tion. Several participants with type 2 diabetes stated to avoid some types of fruit because of the potential impact upon their sugar level. Notwithstanding, research shows that the consump- tion of fresh fruit, when limiting the portions based on the choice of fruit and adjusting the insulin amounts to these choices, should be encouraged rather than discouraged in this popu- lation [10, 43, 44]. Hence, eHealth tools targeting adults with type 2 diabetes may also provide users with correct and up to date information about dietary guidelines. For example, when users with type 2 diabetes create a plan to replace unhealthy snacks by a piece of fruit the pro- gramme may explain why consuming snacks marketed specifically for people with type 2 dia- betes should be discouraged and why it is beneficial to consume fresh fruit [44]. The same holds for guidelines regarding PA. For example, when users with type 2 diabetes create a plan to go hiking, information regarding the importance of regular blood glucose checks could be shown [45]. Nevertheless, it is promising that similar comments and remarks were stated by the sample with type 2 diabetes and the sample from the general population. This may indicate that there is a large overlap between the needs and expectations of both groups regarding an eHealth tool targeting a healthy lifestyle. If specific adaptations (e.g. adding information about the beneficial effects of fruit intake or about the importance of foot care) are made, ‘MyPlan 1.0’ has the potential to be a suitable tool for this clinical group as well. These findings may encourage other researchers to adapt existing interventions to new target populations as it may be more cost-effective and less stigmatizing than creating new condition-specific interventions. According to the Website Evaluation Questionnaire users experienced ‘MyPlan 1.0’ as engaging. “Two pieces of fruit a day, interesting!” I would observe it more closely and react rapidly to people following a plan. You would get an automatic email saying “Have you eaten two pieces of fruit, yes or no?” and then they have to fill in which ones.” (Female, 41, type 2 diabetes) PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 12 / 19 Users’ thoughts and opinions about a self-regulation-based eHealth intervention PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Discussion This was better captured by the think aloud procedure: users described the tool rather as a long questionnaire than an inter- vention. They stated that it took too much time to go through the website and answer all the questions. Furthermore, they described that some questions were difficult and confusing. Although researchers are probably well aware that users prefer short and easy interventions [53], it remains an issue to specifically address. For example, Yardley and colleagues [35] also found that pages with extensive questions from their health-care website were often perceived as excessive by the users. In order to overcome this problem, researchers should check with users whether interven- tions are sufficiently short and do not contain too much text, and adapt when deemed neces- sary. It is clear from our findings that extensive questionnaires should be avoided in online interventions, and kept to a minimum. Evidently, research on the efficacy of interventions may require multiple questionnaires at baseline, during and after intervention, and at follow- up moments. When this is the case, we recommend that participants are explicitly informed that these questionnaires are only for research purposes and are not part of the intervention. Another possibility to reduce the length of the online intervention is to only include the neces- sary components of behavioural change. Therefore, more research is warranted to identify the optimal combination of self-regulation techniques. Apart from frustrations regarding time efficiency, some users stated that they disliked the lay-out of ‘MyPlan 1.0’. This problem can be tackled by involving members of the target population early on in the developing process. Sev- eral authors have provided guidelines to adapt the lay-out of websites to specific target popula- tions (for example see [54, 55]). Finally, many users suggested that a mobile application would be useful for monitoring their behaviour and helping them remind about their goal. Indeed, self-monitoring has proven to be one of the most effective techniques for behaviour change [56]. A mobile appli- cation targeting a healthy lifestyle (mHealth) could indeed be a more convenient tool for self- monitoring and lead to a more sustained behaviour change by providing daily reminders and support. Furthermore, mHealth has shown to be promising in altering health behaviours [57, 58]. This study has several strengths. First, the think aloud method allows us to grasp immedi- ate thoughts and reactions of users, not compromised by recall bias or researcher suggestions. Discussion This finding is further corroborated by the think aloud procedure. Users reported that the website provided new information and raised awareness about the selected beha- vioural target. The results of the questionnaire furthermore show that the participants per- ceived the website as being personally relevant. However, the think aloud procedure revealed that personal relevance can still be improved: some participants stated that the website was poorly tailored to their specific situation. The effectiveness of offering a personalised approach by using tailored feedback and showing information (such as for example success stories) based on the user’s characteristics is well-established in research [18, 46–48] and should there- fore be included as a standard element in all behaviour change interventions. The results of the Website Evaluation Questionnaire indicate that users valued the website as helpful for goal-setting and self-assessment. This finding was also corroborated by the think aloud results: most of the users found it useful to make a tangible and concrete plan. Literature has shown that action planning can indeed be an effective technique for behaviour change [28, 49]. Nevertheless, action planning strategies should be offered in a way that is understandable for users. For example, although implementation intentions have shown to be effective in facil- itating behaviour change [40, 50–52], users experienced great difficulties in creating their own implementation intentions. Consequently, implementation intentions should be accompanied by clear instructions and examples. Important to note however is that a fair amount of users reported that they did not see the point of making a specific plan for PA, fruit or vegetable PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 13 / 19 Users’ thoughts and opinions about a self-regulation-based eHealth intervention intake. Likewise, some users did not understand why they should monitor their behaviour change or why they would seek for social support. This can be due to a lack of knowledge about how behavioural change is achieved. Therefore, future interventions should include the rationale for these components and give information to users about behaviour change and the self-regulation techniques that can be of aid. Even more, one could give the users more auton- omy regarding which behaviour change techniques they want to apply in a specific situation. An important finding of our study is that the Website Evaluation Questionnaire did not reveal the frustrations of users about the design of the website. PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Discussion Second, the perceptions of two different groups were investigated and found to be similar. As such, ‘MyPlan 1.0’ seems to have potential in clinical samples, such as patients with type 2 diabetes. Finally, the sample was heterogeneous and there was an equal distribution regard- ing sex, age and education. This study has also some limitations. First, a shortcoming of the think aloud method is that performing the intervention website interferes with thinking out aloud. Some users reported difficulties to simultaneously read text and to think aloud. Fur- thermore, users were more likely to elaborate on the things they found difficult or superflu- ous than highlighting the aspects they appreciated. Second, there are no norms available for the Website Evaluation Questionnaire. Therefore, the interpretation of low or high scores is arbitrary. Moreover, one should be cautious when interpreting results of the Website Evalua- tion Questionnaire, because users tend to give higher scores on the questionnaire compared PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 14 / 19 Users’ thoughts and opinions about a self-regulation-based eHealth intervention to their actual experience [59]. Furthermore, although the calculated means of the Website Evaluation Questionnaire show that the intervention was perceived as engaging, relevant and helpful, a large range was found on all three subscales. This suggests that a certain group of users did not evaluate the intervention positively. Third, the fact that users went through the follow-up module immediately after filling out the start module compromises the ecological validity of the comments regarding the follow-up. Fourth, there is always the potential that results are biased. The–even unintended- influence of researchers (e.g. training, profession) may result in a confirmation bias [60]. However, we used a strict protocol for the think aloud procedure (e.g. the use of predefined prompts) and for the analyses of the data (e.g. the use of double coding). Furthermore, the interviewers were not involved in the development of ‘MyPlan 1.0’. Bias may also result from social desirability, especially because the researcher was sitting next to the participant [61]. In so doing, participants may have provided a more positive view about ‘MyPlan 1.0’ by neglecting the problems of the programme. However, we explicitly asked participants not to refrain from remarks, and stressed the constructive nature of these remarks for the further optimization of the programme. Discussion Finally, we did not assess specific characteristics of the clinical sample such as their treatment options or the presence of late-complications. This information might have given us more in-depth knowledge regarding our study population. To conclude, this study used a think aloud procedure and a questionnaire to gain insight in the perceptions and preferences of the users of a self-regulation-based eHealth interven- tion. The presented study showed that the intervention, providing small adaptations, can also be used in tertiary prevention of type 2 diabetes. We thus argue that ‘MyPlan 1.0’ might be able to help adults with type 2 diabetes to adopt a healthier way of living which in turn will have a positive impact on the further development of their disease (i.e. better glycemic con- trol [45] and a reduced risk of diabetic complications [62]). Furthermore, we found that there are still strong efforts needed to make eHealth interventions more convenient and less time-consuming. Finally, users’ ideas regarding health and behaviour change can form possi- ble hindrances and should be taken into account. This study could be a first step in the devel- opment of an engaging eHealth intervention, but more research is needed to investigate how behaviour change techniques can be more conveniently implemented in eHealth. The use of mHealth can contribute to this process. Adaptations made to ‘MyPlan 1.0’ based on this study are described in S2 File. In further developmental phases, perspectives of users should again be explored in order to make constant improvements regarding personal relevance and user-friendliness. Acknowledgments We thank Evy Declerck and Victoria Van Hecke for their help with the data collection and Laurent Degroote for his contribution to the data analysis process. Author Contributions Conceptualization: Louise Poppe, Celien Van der Mispel, Ilse De Bourdeaudhuij, Maïte´ Ver- loigne, Geert Crombez. Data curation: Louise Poppe, Celien Van der Mispel. Formal analysis: Louise Poppe, Celien Van der Mispel. Funding acquisition: Louise Poppe, Celien Van der Mispel, Ilse De Bourdeaudhuij, Geert Crombez. Investigation: Celien Van der Mispel. Methodology: Louise Poppe, Ilse De Bourdeaudhuij, Maïte´ Verloigne, Geert Crombez. Project administration: Ilse De Bourdeaudhuij, Maïte´ Verloigne, Geert Crombez. Resources: Samyah Shadid, Geert Crombez. Supervision: Ilse De Bourdeaudhuij, Maïte´ Verloigne, Samyah Shadid, Geert Crombez. Visualization: Louise Poppe, Celien Van der Mispel. Writing – original draft: Louise Poppe, Celien Van der Mispel, Ilse De Bourdeaudhuij, Maïte´ Verloigne, Samyah Shadid, Geert Crombez. Writing – review & editing: Louise Poppe, Celien Van der Mispel, Ilse De Bourdeaudhuij, Maïte´ Verloigne, Geert Crombez. S1 File. Completed COREQ checklist. (PDF) S2 File. Adaptations made in ‘MyPlan 2.0’. This file describes how ‘MyPlan 2.0’ was adapted based on users’ remarks and frustrations discussed in this study. (PDF) S3 File. Think aloud Interviews. This file contains the transcribed interviews. (ZIP) S3 File. Think aloud Interviews. This file contains the transcribed interviews. (ZIP) S4 File. Website Evaluation Questionnaire data. This file contains the data from the Website Evaluation Questionnaire. (SAV) S4 File. Website Evaluation Questionnaire data. This file contains the data from the Website Evaluation Questionnaire. (SAV) PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 15 / 19 Users’ thoughts and opinions about a self-regulation-based eHealth intervention References 1. Geneau R, Stuckler D, Stachenko S, McKee M, Ebrahim S, Basu S, et al. Raising the priority of prevent- ing chronic diseases: a political process. The Lancet. 2010; 376(9753):1689–98. http://dx.doi.org/10. 1016/S0140-6736(10)61414-6. 2. Boeing H, Bechthold A, Bub A, Ellinger S, Haller D, Kroke A, et al. Critical review: vegetables and fruit in the prevention of chronic diseases. Eur J Nutr. 2012; 51(6):637–63. https://doi.org/10.1007/s00394- 012-0380-y PMID: 22684631 3. Haskell WL, Lee IM, Pate RR, Powell KE, Blair SN, Franklin BA, et al. Physical activity and public health: Updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Med Sci Sports Exerc. 2007; 39(8):1423–34. https://doi.org/10.1249/mss. 0b013e3180616b27 PMID: 17762377 4. Lindstro¨m J, Ilanne-Parikka P, Peltonen M, Aunola S, Eriksson JG, Hemio¨ K, et al. Sustained reduction in the incidence of type 2 diabetes by lifestyle intervention: follow-up of the Finnish Diabetes Prevention Study. The Lancet. 2006; 368(9548):1673–9. 5. Lock K, Pomerleau J, Causer L, Altmann DR, McKee M. The global burden of disease attributable to low consumption of fruit and vegetables: implications for the global strategy on diet. Bulletin of the World Health Organization. 2005; 83(2):100–8. PMID: 15744402 6. WHO. Global recommendations on physical activity for health. Geneva (2010). 7. Kimmons J, Seymour J, Serdula M. Trends in fruit and vegetable consumption among US men and women, 1994–2005. Prev Chronic Dis. 2008. 8. Hallal PC, Andersen LB, Bull FC, Guthold R, Haskell W, Ekelund U, et al. Global physical activity levels: surveillance progress, pitfalls, and prospects. Lancet. 2012; 380(9838):247–57. https://doi.org/10.1016/ S0140-6736(12)60646-1 PMID: 22818937 PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 16 / 19 Users’ thoughts and opinions about a self-regulation-based eHealth intervention 9. Deshpande AD, Baker EA, Lovegreen SL, Brownson RC. Environmental correlates of physical activity among individuals with diabetes in the rural midwest. Diabetes Care. 2005; 28(5):1012–8. PMID: 15855559 10. Nelson KM, Reiber G, Boyko EJ. Diet and exercise among adults with type 2 diabetes findings from the third national health and nutrition examination survey (NHANES III). Diabetes Care. 2002; 25 (10):1722–8. PMID: 12351468 11. Thomas N, Alder E, Leese G. Barriers to physical activity in patients with diabetes. Postgrad Med J. 2004; 80(943):287–91. https://doi.org/10.1136/pgmj.2003.010553 PMID: 15138320 12. Wantland DJ, Portillo CJ, Holzemer WL, Slaughter R, McGhee E. The effectiveness of web-based vs. non-web-based interventions: a meta-analysis of behavioral change outcomes. J Med Internet Res. 2004; 6. https://doi.org/10.2196/jmir.6.4.e40 PMID: 15631964 13. References Alexander GL, McClure JB, Calvi JH, Divine GW, Stopponi MA, Rolnick SJ, et al. A Randomized Clinical Trial Evaluating Online Interventions to Improve Fruit and Vegetable Consumption. Am J Public Health. 2010; 100(2):319–26. https://doi.org/10.2105/AJPH.2008.154468 PMID: 20019315 14. Broekhuizen K, Kroeze W, van Poppel MN, Oenema A, Brug J. A Systematic Review of Randomized Controlled Trials on the Effectiveness of Computer-Tailored Physical Activity and Dietary Behavior Pro- motion Programs: an Update. Ann Behav Med. 2012; 44(2):259–86. https://doi.org/10.1007/s12160- 012-9384-3 PMID: 22767052 15. Davies CA, Spence JC, Vandelanotte C, Caperchione CM, Mummery WK. Meta-analysis of internet- delivered interventions to increase physical activity levels. Int J Behav Nutr Phys Act. 2012; 9(1):1–13. https://doi.org/10.1186/1479-5868-9-52 PMID: 22546283 16. Greaves CJ, Sheppard KE, Abraham C, Hardeman W, Roden M, Evans PH, et al. Systematic review of reviews of intervention components associated with increased effectiveness in dietary and physical activity interventions. BMC public health. 2011; 11(1):1. 17. Hurling R, Catt M, De Boni M, Fairley BW, Hurst T, Murray P, et al. Using internet and mobile phone technology to deliver an automated physical activity program: Randomized controlled trial. J Med Inter- net Res. 2007; 9. https://doi.org/10.2196/jmir.9.2.e7 PMID: 17478409 18. Krebs P, Prochaska JO, Rossi JS. A meta-analysis of computer-tailored interventions for health behav- ior change. Prev Med. 2010; 51(3):214–21. 19. Michie S, Johnston M, Francis J, Hardeman W, Eccles M. From theory to intervention: mapping theoret- ically derived behavioural determinants to behaviour change techniques. Applied psychology. 2008; 57(4):660–80. 20. Webb T, Joseph J, Yardley L, Michie S. Using the internet to promote health behavior change: A sys- tematic review and meta-analysis of the impact of theoretical basis, use of behvaior change techniques, and mode of delivery. J Med Internet Res. 2010; 12. https://doi.org/10.2196/jmir.1376 PMID: 20164043 21. Vandelanotte C, Mu¨ller AM, Short CE, Hingle M, Nathan N, Williams SL, et al. Past, Present, and Future of eHealth and mHealth Research to Improve Physical Activity and Dietary Behaviors. J Nutr Educ Behav. 2016; 48(3):219–28.e1. http://dx.doi.org/10.1016/j.jneb.2015.12.006. PMID: 26965100 22. Eysenbach G. The Law of Attrition. J Med Internet Res. 2005; 7(1). https://doi.org/10.2196/jmir.7.1.e11 PMID: 15829473 23. Murray E, White IR, Varagunam M, Godfrey C, Khadjesari Z, McCambridge J. Attrition Revisited: Adherence and Retention in a Web-Based Alcohol Trial. J Med Internet Res. 2013; 15(8):10. https://doi. org/10.2196/jmir.2336 PMID: 23996958 24. Pagliari C. Design and Evaluation in eHealth: Challenges and Implications for an Interdisciplinary Field. J Med Internet Res. 2007; 9(2):e15. Epub 27.5.2007. PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 References https://doi.org/10.2196/jmir.9.2.e15 PMID: 17537718. 25. Yardley L, Morrison L, Bradbury K, Muller I. The Person-Based Approach to Intervention Development: Application to Digital Health-Related Behavior Change Interventions. J Med Internet Res. 2015; 17 (1):15. https://doi.org/10.2196/jmir.4055 PMID: 25639757 26. Perez HR, Nick MW, Mateo KF, Squires A, Sherman SE, Kalet A, et al. Adapting a weight management tool for Latina women: a usability study of the Veteran Health Administration’s MOVE! 23 tool. BMC Med Inform Decis Mak. 2016; 16(1):128. https://doi.org/10.1186/s12911-016-0368-2 PMID: 27716279 27. Maes S, Karoly P. Self-Regulation Assessment and Intervention in Physical Health and Illness: A Review. Appl Psychol Health Well Being. 2005; 54(2):267–99. https://doi.org/10.1111/j.1464-0597. 2005.00210.x 28. Schwarzer R. Modeling health behavior change: How to predict and modify the adoption and mainte- nance of health behaviors. Applied Psychology. 2008; 57(1):1–29. 17 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 Users’ thoughts and opinions about a self-regulation-based eHealth intervention 29. Thoolen B, de Ridder D, Bensing J, Gorter K, Rutten G. Beyond Good Intentions: the development and evaluation of a proactive self-management course for patients recently diagnosed with Type 2 diabetes. Health Educ Res. 2008; 23(1):53–61. https://doi.org/10.1093/her/cyl160 PMID: 17289660 30. Plaete J, Crombez G, DeSmet A, Deveugele M, Verloigne M, De Bourdeaudhuij I. What do general practitioners think about an online self-regulation programme for health promotion? Focus group inter- views. BMC family practice. 2015; 16(1):3. 31. Inzucchi SE, Bergenstal RM, Buse JB, Diamant M, Ferrannini E, Nauck M, et al. Management of hyper- glycaemia in type 2 diabetes: a patient-centered approach. Position statement of the American Diabe- tes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetologia. 2012; 55(6):1577–96. https://doi.org/10.1007/s00125-012-2534-0 PMID: 22526604 32. Plaete J, Crombez G, Van der Mispel C, Verloigne M, Van Stappen V, De Bourdeaudhuij I. Effect of the Web-Based Intervention MyPlan 1.0 on Self-Reported Fruit and Vegetable Intake in Adults Who Visit General Practice: A Quasi-Experimental Trial. J Med Internet Res. 2016; 18(2):e47. Epub 29.02.2016. https://doi.org/10.2196/jmir.5252 PMID: 26929095. 33. Van Dyck D, Plaete J, Cardon G, Crombez G, De Bourdeaudhuij I. Effectiveness of the self-regulation eHealth intervention ’MyPlan1.0.’ on physical activity levels of recently retired Belgian adults: a random- ized controlled trial. Health Educ Res. 2016; 31(5):653–64. Epub 2016/07/17. https://doi.org/10.1093/ her/cyw036 PMID: 27422898. 34. Plaete J, De Bourdeaudhuij I, Verloigne M, Crombez G. Acceptability, feasibility and effectiveness of an eHealth behaviour intervention using self-regulation: ‘MyPlan’. Patient Educ Couns. 2015; 98(12):1617– 24. http://dx.doi.org/10.1016/j.pec.2015.07.014. 35. References PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 18 / 19 Users’ thoughts and opinions about a self-regulation-based eHealth intervention 49. Sniehotta FF, Scholz U, Schwarzer R. Bridging the intention–behaviour gap: Planning, self-efficacy, and action control in the adoption and maintenance of physical exercise. Psychol Health. 2005; 20(2):143–60. 50. Adriaanse MA, Vinkers CD, De Ridder DT, Hox JJ, De Wit JB. Do implementation intentions help to eat a healthy diet? A systematic review and meta-analysis of the empirical evidence. Appetite. 2011; 56(1):183–93. https://doi.org/10.1016/j.appet.2010.10.012 PMID: 21056605 51. Be´langer-Gravel A, Godin G, Amireault S. A meta-analytic review of the effect of implementation inten- tions on physical activity. Health Psychol Rev. 2013; 7(1):23–54. 52. Gollwitzer PM, Sheeran P. Implementation intentions and goal achievement: A meta-analysis of effects and processes. Adv Exp Soc Psychol. 2006; 38:69–119. 53. Brouwer W, Oenema A, Crutzen R, de Nooijer J, de Vries NK, Brug J. What makes people decide to visit and use an Internet-delivered behavior-change intervention? A qualitative study among adults. Health Education. 2009; 109(6):460–73. 54. Nielsen J, Loranger H. Prioritizing web usability: Pearson Education; 2006. 55. Johnson R, Kent S. Designing universal access: web-applications for the elderly and disabled. Cogni- tion, Technology & Work. 2007; 9(4):209–18. https://doi.org/10.1007/s10111-007-0063-2 56. Michie S, Abraham C, Whittington C, McAteer J, Gupta S. Effective techniques in healthy eating and physical activity interventions: a meta-regression. Health Psychology. 2009; 28(6):690. https://doi.org/ 10.1037/a0016136 PMID: 19916637 57. Arora S, Peters AL, Agy C, Menchine M. A Mobile Health Intervention for Inner City Patients with Poorly Controlled Diabetes: Proof-of-Concept of the TExT-MED Program. Diabetes Technol Ther. 2012; 14 (6):492–6. https://doi.org/10.1089/dia.2011.0252 PMID: 22524591 58. Glynn LG, Hayes PS, Casey M, Glynn F, Alvarez-Iglesias A, Newell J, et al. Effectiveness of a smart- phone application to promote physical activity in primary care: the SMART MOVE randomised con- trolled trial. Br J Gen Pract. 2014; 64(624):E384–E91. https://doi.org/10.3399/bjgp14X680461 PMID: 24982490 59. Elling S, Lentz L, De Jong M, editors. Website evaluation questionnaire: development of a research- based tool for evaluating informational websites. International Conference on Electronic Government; 2007: Springer. 60. Nickerson RS. Confirmation bias: A ubiquitous phenomenon in many guises. Review of general psy- chology. 1998; 2:175. 61. Richman WL, Kiesler S, Weisband S, Drasgow F. A meta-analytic study of social desirability distortion in computer-administered questionnaires, traditional questionnaires, and interviews. Journal of Applied Psychology. 1999; 84(5):754. 62. Boule´ NG, Haddad E, Kenny GP, Wells GA, Sigal RJ. References Yardley L, Morrison LG, Andreou P, Joseph J, Little P. Understanding reactions to an internet-delivered health-care intervention: accommodating user preferences for information provision. BMC Med Inform Decis Mak. 2010; 10:10. 36. Hwang W, Salvendy G. Number of people required for usability evaluation: the 10±2 rule. Communica- tions of the ACM. 2010; 53(5):130–3. 37. Plaete J, De Bourdeaudhuij I, Verloigne M, Oenema A, Crombez G. A Self-Regulation eHealth Interven- tion to Increase Healthy Behavior Through General Practice: Protocol and Systematic Development. JMIR Res Protoc. 2015; 4(4):e141. Epub 22.12.2015. https://doi.org/10.2196/resprot.4835 PMID: 26694101. 38. Craig CL, Marshall AL, Sjostrom M, Bauman AE, Booth ML, Ainsworth BE, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003; 35(8):1381–95. https://doi.org/10.1249/01.MSS.0000078924.61453.FB PMID: 12900694 39. Plaete J, De Bourdeaudhuij I, Crombez G, Steenhuyzen S, Dejaegere L, Vanhauwaert E, et al. The Reliability and Validity of Short Online Questionnaires to Measure Fruit and Vegetable Intake in Adults: The Fruit Test and Vegetable Test. PLoS One. 2016; 11(7):12. https://doi.org/10.1371/journal.pone. 0159834 PMID: 27441632 40. Gollwitzer PM. Implementation intentions—Strong effects of simple plans. Am Psychol. 1999; 54 (7):493–503. https://doi.org/10.1037//0003-066x.54.7.493 41. Morrison L, Moss-Morris R, Michie S, Yardley L. Optimizing engagement with Internet-based health behaviour change interventions: Comparison of self-assessment with and without tailored feedback using a mixed methods approach. Br J Health Psychol. 2014; 19(4):839–55. https://doi.org/10.1111/ bjhp.12083 PMID: 24308806 42. Braun V, Clarke V. Using thematic analysis in psychology. Qual Res Psychol. 2006; 3(2):77–101. https://doi.org/10.1191/1478088706qp063oa 43. Ley SH, Hamdy O, Mohan V, Hu FB. Prevention and management of type 2 diabetes: dietary compo- nents and nutritional strategies. The Lancet. 2014; 383(9933):1999–2007. https://doi.org/10.1016/ S0140-6736(14)60613-9. 44. Excellence NIfHaC. Type 2 diabetes in adults: management (NG28) London: NICE; 2017. 45. Sigal RJ, Kenny GP, Wasserman DH, Castaneda-Sceppa C, White RD. Physical activity/exercise and Type 2 diabetes A consensus statement from the American Diabetes Association. Diabetes Care. 2006; 29(6):1433–8. https://doi.org/10.2337/dc06-9910 PMID: 16732040 46. Noar S, Benac C, Harris M. Does tailoring matter? Meta-analytic review of tailored print health behavior change interventions. Psychol Bull. 2007; 133. https://doi.org/10.1037/0033-2909.133.4.673 PMID: 17592961 47. Smeets T, Kremers S, De Vries H, Brug J. Effects of tailored feedback on multiple health behaviors. Ann Behav Med. 2007; 33(2):117–23. https://doi.org/10.1080/08836610701307801 PMID: 17447863 48. Pruitt J, Adlin T. The persona lifecycle: keeping people in mind throughout product design: Morgan Kaufmann; 2010. PLOS ONE \| https://doi.org/10.1371/journal.pone.0190020 December 21, 2017 References Effects of exercise on glycemic control and body mass in type 2 diabetes mellitus: a meta-analysis of controlled clinical trials. JAMA. 2001; 286(10):1218– 27. PMID: 11559268 19 / 19
https://openalex.org/W3201606505	https://ieeexplore.ieee.org/ielx7/6287639/9312710/09474435.pdf	English	null	An Improved Dynamic Collaborative Filtering Algorithm Based on LDA	IEEE access	2,021	cc-by	6,959	Received May 21, 2021, accepted June 29, 2021, date of publication July 5, 2021, date of current version September 10, 2021. Digital Object Identifier 10.1109/ACCESS.2021.3094519 INDEX TERMS Collaborative ﬁltering, LDA, topic model, time tag. In recent years, an increasing number of studies have used topic models and review texts to generate recommendations. In the recommendation algorithm, to ﬁnd the nearest neigh- bor of the target user, we need to measure the similarity between users. However, the number of users and items in the recommendation system is very large, users often only score a small number of items, and the user rating data are extremely sparse, which makes the nearest neighbor set obtained by the traditional similarity measurement method not accurate enough, resulting in a reduction in the recommendation qual- ity of the algorithm. In addition, the time factor is very important context information in recommendation systems. Users’ interests vary greatly in different periods of time. The topics in a document evolve over time. We compare the inaugural addresses of two presidents in different periods and present them in the form of word clouds in Figure 1. The higher the word frequency, the larger the display size. It can be seen in the ﬁgure that although they were all inaugural addresses, they used different words, and the use of high-frequency words was also quite different. Therefore, when the time span of the corpus is large, the topic model obviously needs to consider the time factor and the dynamic change in language. However, most methods ignore the time factor. Therefore, we propose a hybrid dynamic collaborative ﬁltering algorithm (HDCF) that can capture the evolution of topics in a collaborative ﬁltering algorithm. An Improved Dynamic Collaborative Filtering Algorithm Based on LDA MENG DI-FEI 1, LIU NA 1, LI MING-XIA1,2,3, AND SU HAO-LONG1 1School of Information Science and Engineering, Dalian Polytechnic University, Dalian 116001, China 2School of Computer Science and Technology, University of Chinese Academy of Sciences, Beijing 100049, China 3Shenyang Institute of Computing Technology, Chinese Academy of Sciences, Shenyang 110168, China Corresponding author: Liu Na (liuna@dlpu.edu.cn) MENG DI-FEI 1, LIU NA 1, LI MING-XIA1,2,3, AND SU HAO-LONG1 1School of Information Science and Engineering, Dalian Polytechnic University, Dalian 116001, China 2School of Computer Science and Technology, University of Chinese Academy of Sciences, Beijing 100049, China 3Shenyang Institute of Computing Technology, Chinese Academy of Sciences, Shenyang 110168, China Corresponding author: Liu Na (liuna@dlpu.edu.cn) This work was supported in part by the National Natural Science Foundation of China under Grant 61402069, in part by the National Key Research and Development Plan under Grant 2017YFC0821003, and in part by the Natural Science Foundation of Liaoning Province under Grant 20180550395. ABSTRACT Currently, available collaborative ﬁltering (CF) algorithms often utilize user behavior data to generate recommendations. The similarity calculation between users is mostly based on the scores, without considering the explicit attributes of the users with proﬁles, as these are difﬁcult to generate, or their preferences over time evolve. This paper proposes a collaborative ﬁltering algorithm named hybrid dynamic collaborative ﬁltering (HDCF), which is based on the topic model. Considering that the user’s evaluation of an item will change over time, we add a time-decay function to the subject model and give its variational inference model. In the collaborative ﬁltering score, we generate a hybrid score for similarity calculation with the topic model. The experimental results show that this algorithm has better performance than currently available algorithms on the MovieLens dataset, Netﬂix dataset and la.fm dataset. under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ VOLU I. INTRODUCTION With the advent of the Internet, the availability of informa- tion continues to rapidly increase. This abundance makes it challenging for users to effectively ﬁnd the information they need in the high volume of data they can access. To address this problem, recommendation systems have been proposed to help users quickly obtain useful information depending on their past preferences or other sources. Most established recommendation algorithms start with ﬁnding a set of customers. This set of customers has pur- chased and rated products that overlap with the products that the current user has purchased and rated. The algo- rithm aggregates the products from these similar customers, excludes the products that the user has purchased or rated, and recommends the remaining products to the user. The purpose of collaborative ﬁltering (CF) is to suggest new items or to predict the utility of a certain item for a particular user based on the user’s previous preferences and the opinions of other like-minded users. Researchers have proposed many CF algorithms that can be divided into two main categories: user-based and item-based. In this research, we focus on user-based methods. The associate editor coordinating the review of this manuscript and approving it for publication was Ting Li . 122568 VOLUME 9, 2021 M. D.-Fei et al.: Improved Dynamic Collaborative Filtering Algorithm Based on LDA FIGURE 1. Example of topic evolution. FIGURE 1. Example of topic evolution. attention in the past decade, resulting in signiﬁcant progress and adoption by some successful commercial systems. Equation (1) the traditional CF algorithm assigns the same weight to all items of interest, ignoring the inﬂuence of user time on interest when calculating the similarity. In fact, the user’s attention to the item will change over time, and the attention will affect the user’s interest. Therefore, the algo- rithm should consider the time dimension. Aiming at the above problems, this paper introduces a time-decay func- tion in the LDA model that gives different weights to items according to the time users look at them. The corpus is divided into time slices according to the time attribute of the document. The distribution of current topic words is deter- mined by the distribution of previous time, the time-decay rate and the weight of words. We also give the approximate variational posterior of this model. I. INTRODUCTION p y y Along with the deep study of recommendation sys- tems, researchers introduced the context environment into recommendation systems, such as time, location, mood, active state, and network condition. Traditional CF tech- niques cannot track users’ preferences over a period of time [1]–[9]. Therefore, temporal dynamics emerged in recommendation systems. Considering the time interval between purchases, Wang et al. proposed an opportunity model. The model determines the items to be recommended and the best time to recommend a speciﬁc product [10]. Mustansar Ali Ghazanfar et al. presented a novel structure learning technique called the kernel-mapping recommender (KMR) to make reliable recommendations under sparse, cold-start, and long-tail scenarios. In 2011, they presented a fast incremental algorithm for building the model. In 2012, they constructed user-based and item-based KMR, combin- ing user-based and item-based KMR and kernels from fea- ture information [11], [12]. Cheng JiuJun proposed a user spatiotemporal behavior pattern method combining mobile personalized attributes and context information. Considering the popularity of users’ interests, the popularity of subjects and the impact of users, they built a model based on users’ interests [13]. Nour EI Islem Karabadji considered the rec- ommended range when selecting the appropriate group. This helped to make accurate and diverse recommendations simul- taneously [14]. Niu Z proposed a new knowledge-based topic model that combined the Dirichlet tree and integrated the must-links into topic modeling for object discovery. In partic- ular, to better deal with the polysemy phenomenon of visual vocabulary, a must-link was redeﬁned such that it only con- strained one or more speciﬁc topics rather than all topics, thus signiﬁcantly improving topic coherence [15]. Zhang Xiong proposed a method that further improved recommendation accuracy. Within a semantic environment in a document, entities of different topics were present. Another entity that Equation (2) in the processing of similarity computation, we calculate a hybrid user similarity score, that is, the inte- gration of the topic model and traditional similarity. In this way, our approach is different from the established recommendation algorithms to improve the recommendation quality. This paper is organized as follows. Section II describes related work, including CF algorithms and topic models in the literature. Section III describes the proposed algorithm, and Section IV presents the results of applying this algorithm to different datasets. In Section V, we conclude and discuss further research directions. II. RELATED WORK II. RELATED WORK d Recommender systems are often based on CF, which relies only on past user behavior. For example, their previous transactions or product ratings do not require the creation of explicit proﬁles. Notably, CF techniques do not require domain knowledge and avoid the need for extensive data collection. In addition, relying directly on user behavior allows the discovery of complex and unexpected patterns that would be difﬁcult or impossible to proﬁle using known data attributes. As a consequence, CF has attracted much 122569 122569 VOLUME 9, 2021 M. D.-Fei et al.: Improved Dynamic Collaborative Filtering Algorithm Based on LDA FIGURE 2. Graphical model representation of LDA. appeared in the same document at the same time was used to help disambiguate the referenced content to a certain extent [16]. Zhang and Liu presented a cross-domain recom- mender system based on kernel-induced knowledge transfer. This method effectively transfers knowledge through over- lapping entities and alleviates data sparsity issues [17], [18]. Das J proposed a CF method based on clustering, which used two kinds of hierarchical space to divide the data structure: K-d tree and quadtree. They clustered or partitioned the user space according to the user’s location and then used the generated clustering results to predict the target user’s score [19]. Zhang, Pengfei, et al. proposed decomposing the matrix into two nonnegative matrices and then integrat- ing the time weighting into the evaluation matrix in the collaborative ﬁltering algorithm [20]. Yu X et al. proposed two auxiliary domains, i.e., user-side domain and item-side domain, to solve the sparsity problem. In this method, both the user side and item side can not only share informa- tion but also infer domain-independent user and item fea- tures. They also proposed another cross-domain collaborative ﬁltering algorithm to alleviate the sparsity problem. This method ﬁrst formulates the recommendation problem as a classiﬁcation problem in the target domain. Then, Funk-SVD decomposition is used to extract extra user and item features. Finally, the C4.5 decision tree algorithm is used to predict missing ratings [21], [22]. Wang joined LDA and a listwise model to generate collaborative ﬁltering results [23]. Zhang used a multichannel feature vector to calculate the similarity between items [24]. FIGURE 2. Graphical model representation of LDA. large-scale corpus. This method assumes that each word is extracted from a hidden topic. The LDA model can be expressed as a probability graph model, as shown in Figure 2. III. PROPOSED APPROACH Zdn = multi (θd) (3) (3) The basic idea of CF algorithms is to provide item recom- mendations or predictions based on the opinions of other like-minded users. The opinions of users can be obtained explicitly from the users or by using implicit measures. In a typical CF scenario, there is a list of m users U={u1,u2...um} and a list of n items I={i1,i2...in}. Each user ui has a list of items rui, which the user has expressed their opinions about. CF algorithms represent the entire m×n user-item data as a ratings matrix. Each entry ai;j represents the preference score (ratings) of the ith user on the jth item. Each individual rating is within a numerical scale; the value zero indicates that the user has not yet rated the item. For this topic number, we obtain the probability distribution of word Wd,n as follows: For this topic number, we obtain the probability distribution of word Wd,n as follows: wd,n = multi ϕzdn (4) (4) A k-dimensional Dirichlet random variable θ can take values in the (k −1)-simplex. It can be expressed as the following probability density on this simplex: p(θ \| α) = 0 Pk i=1 αi Qk i=1 0 (αi) θα1−1 1 · · · θαk−1 k (5) (5) where the parameters α1, α2 . . . αk > 0 and 0(x) are the gamma functions. A key step in CF algorithms is to calculate the similarity between users. It is typically based on a cosine approach or a correlation in similarity computation. However, they cannot ﬁnd deeper relations between words. A topic model is another option. In this section, we ﬁrst introduce the basic idea of LDA, then introduce our proposed method, give its varia- tional inference model, then introduce the similarity com- putation of the HDCF model, and ﬁnally the collaborative ﬁltering prediction model. Given the parameters α and β, the joint distribution of a topic mixture is: p(θ, z, w \| α, β) = p(θ \| α) N Y n=1 p (zn \| θ) p (wn \| zn, β) (6) Integrating over θg and summing over z, the distribution of a document is: p(w \| α, β)= Z p(θ \| α) N Y n=1 X zn p (zn \| θ) p (wn \| zn, β) ! dθ (7) II. RELATED WORK d The shadow circle in the ﬁg- ure represents the observed variable, the nonshadow circle represents the potential variable, the arrow represents the conditional dependency between the two variables, the box represents the replicate sampling, and the number of repli- cates is in the lower right corner of the box. The LDA assumes that the prior distribution of the doc- ument topics is the Dirichlet distribution. That is, for each document d, its topic distribution θd is θd = Dirichlet(⃗α) (1) (1) where α is the proportion parameters. where α is the proportion parameters. where α is the proportion parameters. The LDA assumes that the prior word distribution in a topic is a Dirichlet distribution. That is, for each topic k, its word distribution ϕ is: ϕd = Dirichlet( ⃗β) (2) (2) where β is a vector of the topic parameters. For each word Wd,n in document d, we can obtain its topic number Zdn distribution from the topic distribution θd as follows: B. GENERATING THE HDCF ALGORITHM Most current topic models are based on the assumption that documents in a corpus are not sequential; in other words, doc- uments in a corpus are interchangeable. In fact, this simpliﬁed assumption is inappropriate and inconsistent with the actual situation. 1) A large number of corpora, such as scientiﬁc and technological literature databases and news databases, are temporal corpora, and the documents in them have the time attribute. Some speciﬁc text information can only appear in a certain time period. In addition, many corpora span hundreds of years; for these corpora, it is obviously inappropriate to ignore the time order attribute. 2) Language changes over time, and the topic is bound to evolve over time. FIGURE 3. The process of HDCF algorithm. in Figure 3. This method not only uses time-delay function, but also gives the derivation process of Variational Inference in phase 2. The nearest neighbor set is used instead of all data sets, so as to reduce the sparsity of data in phase 3. To verify the reliability and accuracy of the algorithm, the number of baselines is increased in the experiment, and the experimental results std. deviation (with error bars) are given. The traditional CF algorithm assigns the same weight to all items of interest, ignoring the inﬂuence of user time on interest when calculating the similarity. This is obviously unreasonable. Because the attention of users to items will change with the retention time of items, the attention affects the interest of users. Therefore, the algorithm should consider the time dimension to improve the accuracy of the recommen- dation prediction. Generally, the recently followed items with high attention of users show their recent interests and hobbies, so users have a high degree of interest in such items, whereas users have a low degree of interest in items. According to the time attribute of the document, the corpus is divided into T time slices. Time slice 1 is the oldest time slice, time slice T is the latest time slice, and a subset of documents on each time slice is D ={D1,D2...DT }. Intro- ducing the weight of time attribute r(r >0). Let the last browsing time of the topic be ti, the current time be t, and the distribution of current topic words is determined by the distribution of previous times, the time-decay rate and the weight of words. A. UNDERLYING EXISTING LDA LDA is an unsupervised machine learning technique that can be used to identify potentially hidden topic information in a (7) VOLUME 9, 2021 122570 VOLUME 9, 2021 M. D.-Fei et al.: Improved Dynamic Collaborative Filtering Algorithm Based on LDA FIGURE 3. The process of HDCF algorithm. Taking the product of the probabilities of single documents, the probability of a corpus is: Taking the product of the probabilities of single documents, the probability of a corpus is: Taking the product of the probabilities of single documents, h b bili f i the probability of a corpus is: p(D \| α, β) = M Y d=1 Z p (θd \| α) × Nd Y n=1 X zdn p (zdn \| θd) p (wdn \| zdn, β) ! dθd (8) (8) Each topic z corresponds to a distribution over an item, and each user has a distribution θ on a potential topic. The approach for rating items can be described as follows: the users choose a topic based on their interests and then choose an item according to the distribution over this topic. More speciﬁcally, the process is as follows: 1. For each document d, choose θ ∼Dir (α). 2. For each word w for document d (a) Choose a topic zn ∼multinomial (θ). (b) Choose a word wn from p (wn\|zn, β), a multinomial probability conditioned on the topic zn. D. SIMILARITY COMPUTATION OF HDCF MODEL An important step in CF algorithms is to calculate the similar- ity between users. This result is used to establish a proximity- based neighborhood between the target user and some similar users. The main goal of neighborhood formation is to ﬁnd an ordered list of l users N = {n1,n2...nl} for each user u. For example, sim(u,N1) is maximum, sim(u,N2) is the next maximum, and so on. The proximity between two users is usually measured by a cosine-based similarity. However, both the cosine and correlation approaches utilize a bag of words. They cannot mind the semantic relationships between words. Additionally, the distribution of topics is affected by the time-decay rate. θd t = θd · r(t) (10) (10) Thus, the dynamic topic model is established as shown in Figure 4. Let each document on a time slice correspond to a ϕt and θd t by the LDA model. For our HDCF method, the generation process of documents on time slice t is as follows: In our method, we ﬁrst deﬁne the matching degree δt between document d and topic over word wn on time slice t as Algorithm 1 Generate process of documents on time slice t Algorithm 1 Generate process of documents on time slice t Algorithm 1 Generate process of documents on time slice t Input: Document D, hyperparameters α, β, topic number K Output: topic z, parameters ϕ, θ for each topic k if time slice T==1 choose ϕt ∼Dir(βk,1) else calculate ϕt ∼Dir(βk,t) according to Equation (9) for each document d calculate θd t ∼Dir(α) for each word w in document d generate the topic from Zd n,t ∼multi(θd t ) generate a word from wn,t ∼multi(βZn,t,t) × TY t=1 Dt Y d=1 q θd t \| γ d t Nd Y n=1 q zd n,t \| φd t,n !! (12) Input: Document D, hyperparameters α, β, topic number K (12) The rest of the iterative and updating process of our model is very similar to that of the traditional LDA, but the two are obviously different. The distribution of topics in our method is related to the degree of time decay, which makes the iteration and updating process on each document associated with the previous slice process. B. GENERATING THE HDCF ALGORITHM Aiming at the above problems, this paper introduces the HDCF model on the basis of reference [25] and gives different weights to items according to the time users look at the items. Based on the weight of the time attribute, the changes in users’ interests can be reﬂected to realize the time-based simi- larity calculation of interests. The shorter the time for users to consume items, the higher their corresponding interests will be, which should have a higher weight, and vice versa. The difference between our model and reference [25] lies in: The process of HDCF algorithm is divided into 4 phases as shown ϕt = ϕt−1 · ω · r(t) ω = Wt Wt−1 r(t) = 1 1 + eα(ti−t) (9) (9) r(t) = 1 1 + eα(ti−t) (9) where W(t) is the number of words of time slice t and r(t) is the rate of time decay. VOLUME 9, 2021 122571 M. D.-Fei et al.: Improved Dynamic Collaborative Filtering Algorithm Based on LDA FIGURE 4. LDA with time decay. βt\|βt−1 ∼N(βt, σ 2I). With Gaussian ‘‘variational observa- tions’’, we denote ˆβk,1, · · · ˆβk,T as the variational distribution ofβk,1, · · · βk,T that retained the sequential structure of the topic. Thus, using the standard mean-ﬁeld approximation, the approximate variational posterior is q(β, z, θ \| λ, φ, γ ) = q βt, zd n,t, θd t \| ˆβt, φd t , γ d t = K Y k=1 q βt \| ˆβt Dt Y d=1 q θd t , zd t \| γ d t , φd t = K Y k=1 q βt \| ˆβt Dt Y d=1 q θd t \| γ d t Nd Y n=1 q zd n,t \| φd t,n ! = K Y k=1 q βk,1, · · · φk,T \| ˆβk,1, · · · ˆβk,T × TY t=1 Dt Y d=1 q θd t \| γ d t Nd Y n=1 q zd n,t \| φd t,n !! (12) FIGURE 4. LDA with time decay. B. EVALUATION METRICS Another important step in the CF algorithm is to generate recommendations based on predictions. After calculating the similarity between the items, the user can predict the unrated items based on their similarity. There are numerous measures available to evaluate the rec- ommendation quality. We use the mean absolute error (MAE) and the root-mean-square error (RMSE) to evaluate our algo- rithm. They are common measures in recommender systems. MAE is an average of the absolute errors between predictions and eventual outcomes, while RMSE is the square root of the average of the square of all of the errors. MAE and RMSE are given by There are numerous measures available to evaluate the rec- ommendation quality. We use the mean absolute error (MAE) and the root-mean-square error (RMSE) to evaluate our algo- rithm. They are common measures in recommender systems. To solve the problem of data sparsity, we use the nearest target users’ neighbors method for prediction [27]. Setting NBSu as the nearest neighbor for user u, the prediction rui of user u for item i are obtained by the ratings of user u for the nearest neighbors set NBSu. MAE is an average of the absolute errors between predictions and eventual outcomes, while RMSE is the square root of the average of the square of all of the errors. MAE and RMSE are given by rui = ru + P n∈NBSu simu,n × (rn,i −rn) P n∈NBSu ( simu,n ) (17) MAE (ui) = P u∈U ru,i −ˆru,i n (20) RMSE (ui) = sX u∈U ru,i −ˆru,i 2 /n (21) (17) (20) (21) where sim(u,n) is the similarity between user u and user n, rn,i is the rating of user n for item i, and rn is the average rating of user u and user n. Obviously, the lower the value of MAE or RMSE is, the higher the accuracy of the recommended results. This method calculates the prediction on user u for item i by computing the sum of the user’s ratings for items that are similar to i. For the inﬂuence of context time, a rating the user made long ago should be less inﬂuential than a rating made now. To accomplish this, a time delay function fu,i of active user u for tag ti is used in the prediction computation. divergence to evaluate the similarity of users. divergence to evaluate the similarity of users. In this method, the prediction on user u for item i is calculated by the average of the user’s ratings. The ratings are weighted by the similarity between items, which is similar to item i. However, the recently followed items with high user attention show their recent interests and hobbies, so users have a high degree of interest in such items, whereas users have a low degree of interest in items. Therefore, considering the time inﬂuence on the prediction results, the time-decay function is introduced into the prediction process in this paper. simLDA i,j = exp−KL(ui,uj) KL(ui, uj) = KL(ui\|\|uj) + KL(uj\|\|ui) = X k∈m ln θi θk · θi + X k∈m ln θj θk · θj (15) Considering the advantages of traditional similarity and topic similarity, we use a hybrid similarity for words in this paper. simc i,j = P u∈U ru,i · ru,j r P u∈U r2 u,i · P u∈U r2 u,j simac i,j = P u∈U (ru,i −ru)(ru,j −ru) r P u∈U (ru,i −ru)2 P u∈U (ru,j −ru)2 simp i,j = P u∈U (ru,i −ri)(ru,j −rj) r P u∈U (ru,i −ri)2 P u∈U (ru,j −rj)2 simc i,j = P u∈U ru,i · ru,j r P u∈U r2 u,i · P u∈U r2 u,j simac i,j = P u∈U (ru,i −ru)(ru,j −ru) r P u∈U (ru,i −ru)2 P u∈U (ru,j −ru)2 simp i,j = P u∈U (ru,i −ri)(ru,j −rj) r P u∈U (ru,i −ri)2 P u∈U (ru,j −rj)2 simi,j = λ(1 3(simc i,j + simp i,j + simac i,j)) + (1 −λ)simLDA i,j simc i,j = P u∈U ru,i · ru,j r P u∈U r2 u,i · P u∈U r2 u,j simac i,j = P u∈U (ru,i −ru)(ru,j −ru) r P u∈U (ru,i −ru)2 P u∈U (ru,j −ru)2 simp i,j = P u∈U (ru,i −ri)(ru,j −rj) r P u∈U (ru,i −ri)2 P u∈U (ru,j −rj)2 simi,j = λ(1 3(simc i,j + simp i,j + simac i,j)) + (1 −λ)simLDA i,j fu,i = X j∈Mi simi,j· exp(− tu,j −tu,i tu ) (18) (18) Recommendation prediction is modiﬁed as Recommendation prediction is modiﬁed as ∧ ru,i = fu,j · ru,j P j∈Mi fu,j (19) (19) A. DATASET We use the MovieLens dataset for our ﬁrst part of the exper- iments. This dataset consists of 100,000 ratings (1-5) for 943 users of 1,682 movies. We use 80% of the dataset as the training set and 20% as the test set. In addition to the MovieLens dataset, we also use the Netﬂix dataset and last.fm dataset in the comparison experiments. IV. EXPERIMENTS Our experiment is divided into two parts. The ﬁrst part determines all kinds of parameters required by the algorithm, including the number of topics, neighbor size, and hybrid similarity parameter. We compare performance with other traditional methods in the second part. (16) where the similarity simc i,j is calculated by the cosine mea- surement, the adjusted cosine similarity simp i,j improves the above limitation by subtracting the user’s average score, correlation-based similarity simac i,j measures similarity with the Pearson correlation coefﬁcient, ri is the average of ri, λ is adjusting parameter. C. VARIATIONAL INFERENCE δt = θd t · Z d n,t Z d n,t = wn P j=1 Zd j,t Wt (13) According to the Bayesian algorithm, the posterior probabil- ity of document D at time slice t is p(θ, z, w\|α, β) = p(θd t , zd n,t\|wn,t, α, βt) = p(θd t , zd n,t, wn,t\|, α, βt) p(wn,t\|, α, βt) (11) (13) (11) where Z d n,t is the distribution of words on topic and Wt is the number of words on time slice t. Because of the coupling between θ, z, β, the above probability cannot be directly calculated, so we intro- duce variational inference. We assume that latent variable θd t is formed by an independent distribution γt and that latent variable Zd n,t is formed by an independent dis- tribution φd t . In reference [25], the evolution of topic- term probability with Gaussian noise can be expressed as Then, the rij of document di for word wj in the CF algorithm is modiﬁed by ri,j = δt · ϕt (14) (14) Then, we regard the topic distribution as the rating of the document-word matrix and use the KL (Kullback-Leibler) VOLUME 9, 2021 VOLUME 9, 2021 122572 VOLUME 9, 2021 VOLUME 9, 2021 M. D.-Fei et al.: Improved Dynamic Collaborative Filtering Algorithm Based on LDA C. PARAMETER OPTIMIZATION EXPERIMENTS We started our experiments by dividing the dataset into train- ing and test sections (80% for training and 20% for testing). Then, we measured the sensitivity of some parameters before VOLUME 9, 2021 122573 M. D.-Fei et al.: Improved Dynamic Collaborative Filtering Algorithm Based on LDA TABLE 1. Description of first 5 distribution. FIGURE 5. Number of topic result. FIGURE 6. Effect of neighborhood size and cluster number. FIGURE 7. Result of hybrid user similarity parameter. FIGURE 8. Std. deviation (with error bars) on movielens dataset. TABLE 1. Description of first 5 distribution. TABLE 1. Description of first 5 distribution. FIGURE 7. Result of hybrid user similarity parameter. FIGURE 5. Number of topic result. FIGURE 7. Result of hybrid user similarity parameter. FIGURE 5. Number of topic result. FIGURE 8. Std. deviation (with error bars) on movielens dataset. FIGURE 8. Std. deviation (with error bars) on movielens dataset. FIGURE 6. Effect of neighborhood size and cluster number. FIGURE 8. Std. deviation (with error bars) on movielens dataset. FIGURE 6. Effect of neighborhood size and cluster number. FIGURE 9. Std. deviation (with error bars) on netlix dataset. running the main experiment. These parameters include the number of topics K, neighborhood size, number of clusters, and the hybrid user similarity parameter. Based on the results, we ﬁxed the optimum values of these parameters and used them for the rest of the experiments. There are three main parameters in the LDA: α, β, and K. Parameter α denotes the distribution of topics for each document, β denotes the distribution of words for each topic, and K denotes the number of topics. We set α = 0.1 and β = 50/K in our experiment. We set the number of topics K to range from 10 to 100, incrementing by 10 each time. The MAE results are shown in Figure 5. FIGURE 9. Std. deviation (with error bars) on netlix dataset. of MAE occurs when K=50. As a result, we set the number of topics K=50 in the following experiment. We also give the ﬁrst ﬁve distributions of topics and words over the topics in TABLE 1 when the number of topics It is clear in Figure 5 that MAE performance ﬂuctuates slightly with different numbers of topics. The minimum value 122574 122574 VOLUME 9, 2021 M. D.-Fei et al.: Improved Dynamic Collaborative Filtering Algorithm Based on LDA TABLE 2. VOLUME 9, 2021 C. PARAMETER OPTIMIZATION EXPERIMENTS Performance results with MAE on movielen dataset. TABLE 3. Performance results with MAE on netflix dataset. TABLE 4. Performance results with MAE on fa.lm dataset. TABLE 5. Performance results with RMSE on movielen dataset. k 20 I i i i h h h h i i l h h h i hb h d i TABLE 2. Performance results with MAE on movielen dataset. TABLE 3. Performance results with MAE on netflix dataset. TABLE 4. Performance results with MAE on fa.lm dataset. TABLE 5. Performance results with RMSE on movielen dataset. TABLE 5. Performance results with RMSE on movielen dataset. k = 20. It is interesting that even though the topic is latent and there is no direct real-world explanation, the top item corresponding to each topic allows us to have some form of interpretation of the item found by the model. these parameters, we change the neighborhood size from 10 to 160 and the cluster number from 5 to 50. Figure 4 depicts our experimental results. The neighborhood size and the number of clusters affect the prediction quality. As shown in Figure 6, the MAE is lower when the number of clusters is larger. Therefore, we choose 50 clusters to compare our proposed algorithm with the other algorithms. The ﬁrst line is the ﬁrst 5 topics with their probability. The remaining line is the word distribution over the corresponding topic. The topic name and the word is in bold. Our proposed algorithm uses the topic model and some traditional similarity approach to calculate a hybrid user The neighborhood size and the number of clusters have a signiﬁcant inﬂuence on the prediction quality. To determine 122575 VOLUME 9, 2021 VOLUME 9, 2021 M. D.-Fei et al.: Improved Dynamic Collaborative Filtering Algorithm Based on LDA TABLE 6. Performance results with RMSE on netflix dataset. TABLE 7. Performance results with RMSE on la.fm dataset. TABLE 8. T-test with MAE. TABLE 7. Performance results with RMSE on la.fm dataset. TABLE 8. T-test with MAE. TABLE 7. Performance results with RMSE on la.fm dataset. TABLE 8. T-test with MAE. TABLE 7. Performance results with RMSE on la.fm dataset. TABLE 7. Performance results with RMSE on la.fm dataset. TABLE 8. T-test with MAE. TABLE 8. T-test with MAE. FIGURE 10. Std. deviation (with error bars) on la.fm dataset. similarity. It is calculated as Equation (16). C. PARAMETER OPTIMIZATION EXPERIMENTS We set the param- eter λ =0.1, 0.2, 0.3, 0.4, 0.5. Figure 5 shows the results. similarity. It is calculated as Equation (16). We set the param- eter λ =0.1, 0.2, 0.3, 0.4, 0.5. Figure 5 shows the results. similarity. It is calculated as Equation (16). We set the param- eter λ =0.1, 0.2, 0.3, 0.4, 0.5. Figure 5 shows the results. , , , , g From Figure 7, we determine that the hybrid parameter λ is 0.5. We present our experimental results on the MovieLens 10 M dataset, Netﬂix dataset and last.fm dataset after we obtain the optimal values of the parameters. The la.fm dataset contains social networking, tagging, and music artist listening information from a set of 2K users from Last.fm online music system. The Netﬂx dataset contains over 100 million ratings. The ratings were collected between October 1998 and December 2005. Each rating has a customer id, a movie id, the date of the rating, and the value of the rating. In these two dataset, we generate topic model for ‘‘title’’ and ‘‘tag’’ while the number of topic is K = 5. We use these three baselines to demonstrate the ability to semantically interpret LDA. CF-cos represents CF with cosine as the similarity measure. CF-pcc represents CF with the Pearson correlation coefﬁcient as the similarity measure. CF-ac represents CF with adjusted cosine as the similarity measure. We also compare three methods [23], [24], [28], CF-LDA, CF-MCFV and ROST to FIGURE 10. Std. deviation (with error bars) on la.fm dataset. verify the time attribute and hybrid attribute. We set time slide T = 10. The results of MAE on different datasets are shown in TABLE 2, TABLE 3 and TABLE 4. The RMSE results on different datasets are shown in TABLE 5, TABLE 6 and TABLE 7. In addition, we use the t-test to test the signiﬁcance and give one-sample test results in Table 8. We also give std. devi- ation (with error bars) with MAE in Figure 8, Figure 9 and Figure 10. We conclude from these results. The MAE and 122576 122576 VOLUME 9, 2021 VOLUME 9, 2021 M. D.-Fei et al.: Improved Dynamic Collaborative Filtering Algorithm Based on LDA RMSE of the proposed method are signiﬁcantly lower than those of the other methods. Experimental results show that the proposed method has high recommendation accuracy. [12] M. A. Ghazanfar, A. Prügel-Bennett, and S. REFERENCES REFERENCES [1] T.-M. Chang and W.-F. Hsiao, ‘‘LDA-based personalized document rec- ommendation,’’ in Proc. PACIS, 2013, p. 13. [19] J. Das, S. Majumder, P. Gupta, and K. Mali, ‘‘Collaborative recommenda- tions using hierarchical clustering based on K-d trees and quadtrees,’’ Int. J. Uncertainty, Fuzziness Knowl.-Based Syst., vol. 27, no. 4, pp. 637–668, Aug. 2019. [2] Q. Liu, E. Chen, H. Xiong, C. H. Q. Ding, and J. Chen, ‘‘Enhancing collaborative ﬁltering by user interest expansion via personalized ranking,’’ IEEE Trans. Syst., Man, Cybern. B, Cybern., vol. 42, no. 1, pp. 218–233, Feb. 2012. [20] P. Zhang, Z. Zhang, T. Tian, and Y. Wang, ‘‘Collaborative ﬁltering rec- ommendation algorithm integrating time windows and rating predictions,’’ Appl. Intell., vol. 49, pp. 3146–3157, Aug. 2019. [3] F. Ortega, J. Bobadilla, A. Hernando, and A. Gutiérrez, ‘‘Incorporating group recommendations to recommender systems: Alternatives and per- formance,’’ Inf. Process. Manage., vol. 49, no. 4, pp. 895–901, Jul. 2013. [21] X. Yu, Y. Chu, F. Jiang, Y. Guo, and D. Gong, ‘‘SVMs classiﬁcation based two-side cross domain collaborative ﬁltering by inferring intrin- sic user and item features,’’ Knowl.-Based Syst., vol. 141, pp. 80–91, Feb. 2018. [4] J. Wang and Y. Zhang, ‘‘Opportunity model for e-commerce recommenda- tion: Right product; right time,’’ in Proc. 36th Int. ACM SIGIR Conf. Res. Develop. Inf. Retr., Jul. 2013, pp. 303–312. [5] Ç. Bakir and S. Albayrak, ‘‘User based and item based collaborative ﬁltering with temporal dynamics,’’ in Proc. 22nd Signal Process. Commun. Appl. Conf. (Siu), Apr. 2014, pp. 252–255. [22] X. Yu, F. Jiang, J. Du, and D. Gong, ‘‘A cross-domain collaborative ﬁltering algorithm with expanding user and item features via the latent factor space of auxiliary domains,’’ Pattern Recognit., vol. 94, pp. 96–109, Oct. 2019. [6] Z. Wang, J. Liao, Q. Cao, H. Qi, and Z. Wang, ‘‘Friendbook: A semantic- based friend recommendation system for social networks,’’ IEEE Trans. Mobile Comput., vol. 14, no. 3, pp. 538–551, Mar. 2015. [23] W. Han and X. Hong-Bin, ‘‘Collaborative ﬁltering recommendation algo- rithm mixing LDA model and list-wise model,’’ Comput. Sci., vol. 46, no. 9, pp. 216–222, 2019. [7] H. P. Xuan, J. J. Jung, B. K. H. Nam, and T. T. Nguyen, ‘‘User timeline and interest-based collaborative ﬁltering on social network,’’ in Proc. Int. Conf. Context-Aware Syst. Appl. Springer, 2015, pp. 132–140. pp [24] H.-R. Zhang, F. Min, Z.-H. Zhang, and S. V. CONCLUSION Collaborative ﬁltering (CF) is a popular recommendation algorithm that makes predictions and recommendations based on the ratings or behaviors of other users in the system. The traditional CF algorithm assigns the same weight to all items of interest, ignoring the inﬂuence of user time on interest when calculating the similarity. This paper introduces the HDCF model and gives different weights to items according to the time users look at the items. In the future, we will continue to study how to improve the dynamic topic model to improve the quality of recommendations. [14] N. E. I. Karabadji, S. Beldjoudi, H. Seridi, S. Aridhi, and W. Dhiﬂi, ‘‘Improving memory-based user collaborative ﬁltering with evolutionary multi-objective optimization,’’ Expert Syst. Appl., vol. 98, pp. 153–165, May 2018. [15] Z. Niu, G. Hua, L. Wang, and X. Gao, ‘‘Knowledge-based topic model for unsupervised object discovery and localization,’’ IEEE Trans. Image Process., vol. 27, no. 1, pp. 50–63, Jan. 2018. [16] Z. Xiong, C. Fu-Cai, and H. Rui-Yang, ‘‘Semi-supervised entity disam- biguation method research based on biterm topic model,’’ Chin. J. Elec- tron., vol. 46, no. 3, pp. 607–613, 2018. [17] J. Lu, J. Xuan, G. Zhang, and X. Luo, ‘‘Structural property-aware mul- tilayer network embedding for latent factor analysis,’’ Pattern Recognit., vol. 76, pp. 228–241, Apr. 2018. [18] Q. Zhang, J. Lu, D. Wu, and G. Zhang, ‘‘A cross-domain recommender system with kernel-induced knowledge transfer for overlapping entities,’’ IEEE Trans. Neural Netw. Learn. Syst., vol. 30, no. 7, pp. 1998–2012, Jul. 2019. C. PARAMETER OPTIMIZATION EXPERIMENTS Szedmak, ‘‘Kernel-mapping recommender system algorithms,’’ Inf. Sci., vol. 208, pp. 81–104, Nov. 2012. [13] J.-J. Cheng, L.-F. Hu, S. Gao, J.-J. Liu, and Y. Yang, ‘‘An optimized col- laborative ﬁltering method to construct spatial-temporal behavior pattern- based user interest model,’’ IEEJ Trans. Electr. Electron. Eng., vol. 12, no. 2, pp. 221–227, Mar. 2017. REFERENCES Wang, ‘‘Efﬁcient collaborative ﬁltering recommendations with multi-channel feature vectors,’’ Int. J. Mach. Learn. Cybern., vol. 10, no. 5, pp. 1165–1172, May 2019. [8] L. Tang, ‘‘Thresholding for top-k recommendation with temporal dynam- ics,’’ 2015, arXiv:1506.02190. [Online]. Available: https://arxiv.org/abs/ 1506.02190 [25] L. Na et al., ‘‘A hybrid user-based collaborative ﬁltering algorithm with topic model,’’ Appl. Intell., Mar. 2021, doi: 10.1007/s10489-021- 02207-7. [9] F. Yang, J. Mou, J. Liu, C. Ma, and H. Yan, ‘‘Characteristic analysis of the fractional-order hyperchaotic complex system and its image encryption application,’’ Signal Process., vol. 169, Apr. 2020, Art. no. 107373. [26] D. M. Blei and J. D. Lafferty, ‘‘Dynamic topic models,’’ in Proc. 23rd Int. Conf. Mach. Learn. (ICML), 2006, pp. 113–120. [10] J. Wang, B. Sarwar, and N. Sundaresan, ‘‘Utilizing related products for post-purchase recommendation in e-commerce,’’ in Proc. 5th ACM Conf. Recommender Syst. (RecSys), 2011, pp. 329–332. [27] A.-L. Deng, Y.-Y. Zhu, and B.-L. Shi, ‘‘A collaborative ﬁltering recom- mendation algorithm based on item rating prediction,’’ J. Softw., vol. 14, no. 9, pp. 54–65, 2003. [11] M. A. Ghazanfar, S. Szedmak, and A. Prugel-Bennett, ‘‘Incremental ker- nel mapping algorithms for scalable recommender systems,’’ in Proc. IEEE 23rd Int. Conf. Tools Artif. Intell., Nov. 2011, pp. 1077–1084, doi: 10.1109/ICTAI.2011.183. [28] B. Li, X. Zhu, R. Li, and C. Zhang, ‘‘Rating knowledge sharing in cross- domain collaborative ﬁltering,’’ IEEE Trans. Cybern., vol. 45, no. 5, pp. 1068–1082, May 2015. 122577 VOLUME 9, 2021
https://openalex.org/W2406627495	http://journals.iucr.org/e/issues/2015/10/00/tk5380/tk5380.pdf	English	null	Crystal structure of 2-cyano-1-methylpyridinium tetrafluoroborate	Acta crystallographica. Section E, Crystallographic communications	2,015	cc-by	3,627	data reports data reports ISSN 2056-9890 Crystal structure of 2-cyano-1-methyl- pyridinium tetrafluoroborate Francesca A. Vaccaro,a Lynn V. Koplitza and Joel T. Magueb* aDepartment of Chemistry, Loyola University, New Orleans, LA 70118, USA, and bDepartment of Chemistry, Tulane University, New Orleans, LA 70118, USA. Correspondence e-mail: joelt@tulane.edu aDepartment of Chemistry, Loyola University, New Orleans, LA 70118, USA, and bDepartment of Chemistry, Tulane University, New Orleans, LA 70118, USA. Correspondence e-mail: joelt@tulane.edu Received 26 August 2015; accepted 26 August 2015 Edited by E. R. T. Tiekink, University of Malaya, Malaysia The asymmetric unit of the title salt, C7H7N2 +BF4 , comprises two independent but nearly identical formula units. The solid- state structure comprises corrugated layers of cations and anions, formed by C—H F hydrogen bonding, that are approximately parallel to (010). Further C—H F hydrogen bonding consolidates the three-dimensional architecture. The sample was reﬁned as a two-component non-merohedral twin. Table 1 Hydrogen-bond geometry (A˚ , ). D—H A D—H H A D A D—H A C1—H1A F7i 0.98 2.50 3.407 (6) 154 C1—H1B F8ii 0.98 2.54 3.498 (6) 166 C1—H1C F3iii 0.98 2.47 3.214 (5) 132 C2—H2 F7i 0.95 2.29 3.190 (5) 157 C3—H3 F1iv 0.95 2.46 3.294 (6) 147 C5—H5 F1v 0.95 2.45 3.306 (5) 149 C8—H8A F2i 0.98 2.48 3.159 (6) 126 C8—H8C F3ii 0.98 2.55 3.437 (6) 151 C9—H9 F3ii 0.95 2.52 3.392 (6) 152 C9—H9 F4ii 0.95 2.59 3.476 (6) 156 C10—H10 F6ii 0.95 2.54 3.167 (6) 123 C12—H12 F5i 0.95 2.49 3.277 (6) 141 Symmetry codes: (i) x þ 2; y 1 2; z þ 1; (ii) x þ 1; y 1 2; z þ 1; (iii) x; y 1; z; (iv) x þ 2; y 1 2; z þ 2; (v) x þ 1; y 1 2; z þ 2. Keywords: crystal structure; salt; C—H F interactions. CCDC reference: 1420782 CCDC reference: 1420782 CCDC reference: 1420782 Koplitz, L. V., Mague, J. T., Kammer, M. N., McCormick, C. A., Renfro, H. E. & Vumbaco, D. J. (2012). Acta Cryst. E68, o1653. Sheldrick, G. M. (2008). Acta Cryst. A64, 112–122. Sheldrick, G. M. (2015a). Acta Cryst. A71, 3–8. Sheldrick, G. M. (2015b). Acta Cryst. C71, 3–8. Spek, A. L. (2009). Acta Cryst. D65, 148–155. Zhang, Y. (2009). Acta Cryst. E65, o2373. 1. Related literature For structures of other salts of the 2-cyano-1-methyl- pyridinium cation, see: Koplitz et al. (2012); Kammer et al. (2013). For structures of salts of the isomeric 2-cyanoanilinium cation, see: Zhang (2009); Cui & Chen (2010). Data collection: APEX2 (Bruker, 2014); cell reﬁnement: SAINT (Bruker, 2014); data reduction: SAINT; program(s) used to solve structure: SHELXT (Sheldrick, 2015a); program(s) used to reﬁne structure: SHELXL2014 (Sheldrick, 2015b); molecular graphics: DIAMOND (Brandenburg & Putz, 2012); software used to prepare material for publication: SHELXTL (Sheldrick, 2008) and PLATON (Spek, 2009). Acknowledgements JTM thanks Tulane University for support of the Tulane Crystallography Laboratory. Supporting information for this paper is available from the IUCr electronic archives (Reference: TK5380). Data collection: APEX2 (Bruker, 2014); cell reﬁnement: SAINT (Bruker, 2014); data reduction: SAINT; program(s) used to solve structure: SHELXT (Sheldrick, 2015a); program(s) used to reﬁne structure: SHELXL2014 (Sheldrick, 2015b); molecular graphics: DIAMOND (Brandenburg & Putz, 2012); software used to prepare material for publication: SHELXTL (Sheldrick, 2008) and PLATON (Spek, 2009). Acknowledgements JTM thanks Tulane University for support of the Tulane Crystallography Laboratory. Supporting information for this paper is available from the IUCr electronic archives (Reference: TK5380). /S2056989015016011 Vaccaro et al. o697 Vaccaro et al. o697 Vaccaro et al. o697 Vaccaro et al. o697 Acta Cryst. (2015). E71, o697–o698 doi:10.1107/S2056989015016011 References Brandenburg, K. & Putz, H. (2012). DIAMOND. Crystal Impact GbR, Bonn, Germany. Bruker (2014). APEX2, SADABS and SAINT. Bruker AXS Inc., Madison, Wisconsin, USA. Cui, L.-J. & Chen, X.-Y. (2010). Acta Cryst. E66, o467. Kammer, M. N., Koplitz, L. V. & Mague, J. T. (2013). Acta Cryst. E69, o1281. p ( ) y Zhang, Y. (2009). Acta Cryst. E65, o2373. ( ) y Spek, A. L. (2009). Acta Cryst. D65, 148–155. S1. Comment The asymmetric unit consists of two independent formula units. A portion of the C—H···F hydrogen bonding network which aids the packing of the several ions is shown in Fig. 1 with fuller depictions appearing in Figs 2 and 3. The solid state structure comprises corrugated layers of cations and anions formed by C—H···F hydrogen bonding between them and approximately parallel to (010). These layers are held to one another by additional C—H···F interactions. S2. Experimental To 0.64 g (0.5 mmol) of 2-cyano-1-methylpyridinium iodide dissolved in 8.5 ml of 95% ethanol was added 1.08 g (0.55 mmol) of solid silver tetrafluoroborate with stirring. The reaction mixture was filtered to remove the precipitated AgI and the filtrate allowed to evaporate to dryness. From the resulting mass, crystals suitable for X-ray diffraction were selected. S3 Refinement data reports References Brandenburg, K. & Putz, H. (2012). DIAMOND. Crystal Impact GbR, Bonn, Germany. Bruker (2014). APEX2, SADABS and SAINT. Bruker AXS Inc., Madison, Wisconsin, USA. Cui, L.-J. & Chen, X.-Y. (2010). Acta Cryst. E66, o467. Kammer, M. N., Koplitz, L. V. & Mague, J. T. (2013). Acta Cryst. E69, o1281. Acta Cryst. (2015). E71, o697–o698 o698 Vaccaro et al. C7H7N2 +BF4 supporting information supporting information Acta Cryst. (2015). E71, o697–o698 [doi:10.1107/S2056989015016011] Acta Cryst. (2015). E71, o697–o698 [doi:10.1107/S2056989015016011] Acta Cryst. (2015). E71, o697–o698 [doi:10.1107/S2056989015016011] Acta Cryst. (2015). E71, o697–o698 Figure 1 Figure 1 Perspective view of the asymmetric unit with 50% probability ellipsoids. The C—H···F interaction is shown by a dotted Figure 1 Perspective view of the asymmetric unit with 50% probability ellipsoids. The C—H···F interaction is shown by a dotted line. Figure 1 Perspective view of the asymmetric unit with 50% probability ellipsoids. The C—H···F interaction is shown by a dotted line. g Perspective view of the asymmetric unit with 50% probability ellipsoids. The C—H···F interaction is shown by a dotted line. Figure 2 Packing viewed down the a axis showing an edge view of two corrugated layers and the C—H···F interactions (dotted lines) holding them together. S3. Refinement The H-atoms attached to carbon were placed in calculated positions (C—H = 0.95 - 0.98 Å). All were included as riding contributions with isotropic displacement parameters 1.2 - 1.5 times those of the attached atoms. In the late stages of the refinement a consistent pattern of Fo2 >> Fc2 suggested twinning not yet accounted for. Use of the TwinRotMat routine in PLATON (Spek, 2009) generated the twin law -1 0 0 0 - 1 0 0 0 1, inclusion of which enabled satisfactory refinement as a 2-component twin. sup-1 Acta Cryst. (2015). E71, o697–o698 supporting information Figure 1 Perspective view of the asymmetric unit with 50% probability ellipsoids. The C—H···F interaction is shown by a dotted li Acta Cryst. (2015). E71, o697–o698 Figure 3 g Packing viewed down the b axis providing a plan view of the corrugated sheets with C—H···F interactions shown as dotted lines g Packing viewed down the b axis providing a plan view of the corrugated sheets with C—H···F interactions shown as dotted lines. Figure 2 Figure 2 Packing viewed down the a axis showing an edge view of two corrugated layers and the C—H···F interactions (dotted lines) holding them together. g Packing viewed down the a axis showing an edge view of two corrugated layers and the C—H···F interactions (dotted lines) holding them together. sup-2 Acta Cryst. (2015). E71, o697–o698 supporting information Figure 3 Packing viewed down the b axis providing a plan view of the corrugated sheets with C—H···F interactions shown as dotted lines. 2-Cyano-1-methylpyridinium tetrafluoroborate y y py Crystal data C7H7N2+·BF4− Mr = 205.96 Monoclinic, P21 a = 7.9704 (16) Å b = 7.5527 (15) Å c = 14.570 (3) Å β = 90.312 (3)° V = 877.1 (3) Å3 Z = 4 F(000) = 416 Dx = 1.560 Mg m−3 Mo Kα radiation, λ = 0.71073 Å Cell parameters from 8457 reflections θ = 2.6–29.0° µ = 0.15 mm−1 T = 150 K Block, colourless 0.14 × 0.13 × 0.08 mm Data collection Bruker SMART APEX CCD diffractometer Radiation source: fine-focus sealed tube Graphite monochromator Detector resolution: 8.3660 pixels mm-1 φ and ω scans Absorption correction: multi-scan (SADABS; Bruker, 2014) Tmin = 0.70, Tmax = 0.99 16120 measured reflections 4566 independent reflections 3779 reflections with I > 2σ(I) Rint = 0.057 θmax = 29.3°, θmin = 2.6° h = −10→10 k = −10→10 l = −19→19 Refinement Refinement on F2 Least-squares matrix: full R[F2 > 2σ(F2)] = 0.050 wR(F2) = 0.122 S = 1.08 4566 reflections 256 parameters 1 restraint Primary atom site location: structure-invariant direct methods Secondary atom site location: difference Fourier map Hydrogen site location: inferred from neighbouring sites F(000) = 416 Dx = 1.560 Mg m−3 Mo Kα radiation, λ = 0.71073 Å Cell parameters from 8457 reflections θ = 2.6–29.0° µ = 0.15 mm−1 T = 150 K Block, colourless 0.14 × 0.13 × 0.08 mm 16120 measured reflections 4566 independent reflections 3779 reflections with I > 2σ(I) Rint = 0.057 θmax = 29.3°, θmin = 2.6° h = −10→10 k = −10→10 l = −19→19 Special details Experimental. The diffraction data were obtained from 3 sets of 400 frames, each of width 0.5° in ω, collected at φ = 0.00, 90.00 and 180.00° and 2 sets of 800 frames, each of width 0.45° in φ, collected at ω = -30.00 and 210.00°. The scan time was 10 sec/frame. Geometry. All e.s.d.'s (except the e.s.d. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell e.s.d.'s are taken into account individually in the estimation of e.s.d.'s in distances, angles and torsion angles; correlations between e.s.d.'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell e.s.d.'s is used for estimating e.s.d.'s involving l.s. planes. pp ( p ) g g p Refinement. Refinement of F2 against ALL reflections. The weighted R-factor wR and goodness of fit S are based on F2, conventional R-factors R are based on F, with F set to zero for negative F2. The threshold expression of F2 > σ(F2) is used only for calculating R-factors(gt) etc. and is not relevant to the choice of reflections for refinement. R-factors based on F2 are statistically about twice as large as those based on F, and R- factors based on ALL data will be even larger. H-atoms attached to carbon were placed in calculated positions (C—H = 0.95 - 0.98 Å). All were included as riding contributions with isotropic displacement parameters 1.2 - 1.5 times those of the attached atoms. In the late stages of the refinement a consistent pattern of Fo2 >> Fc2 suggested twinning not yet accounted for. Use of the TwinRotMat routine in PLATON (Spek, 2009) generated the twin law -1 0 0 0 - 1 0 0 0 1, inclusion of which enabled satisfactory refinement as a 2- component twin. Refinement 1 restraint Primary atom site location: structure-invariant direct methods Secondary atom site location: difference Fourier map Hydrogen site location: inferred from neighbouring sites sup-3 Acta Cryst. (2015). E71, o697–o698 H-atom parameters constrained p w = 1/[σ2(Fo2) + (0.0572P)2 + 0.091P] where P = (Fo2 + 2Fc2)/3 (Δ/σ)max < 0.001 Δρmax = 0.33 e Å−3 p w = 1/[σ2(Fo2) + (0.0572P)2 + 0.091P] where P = (Fo2 + 2Fc2)/3 (Δ/σ)max < 0.001 Δρmax = 0.33 e Å−3 supporting information Δρmin = −0.27 e Å−3 Absolute structure: The absolute structure could not be determined with certainty in this light- atom structure Δρmin = −0.27 e Å−3 Absolute structure: The absolute structure could not be determined with certainty in this light- atom structure Special details Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å2) x y z Uiso/Ueq N1 0.7369 (5) 0.1294 (4) 0.8819 (3) 0.0220 (7) N2 0.3067 (5) 0.0851 (7) 0.8963 (3) 0.0416 (11) C1 0.6868 (6) 0.1642 (6) 0.7854 (3) 0.0273 (10) H1A 0.7826 0.2126 0.7518 0.041 H1B 0.5943 0.2497 0.7841 0.041* H1C 0.6504 0.0534 0.7565 0.041* C2 0.8990 (6) 0.1331 (6) 0.9067 (3) 0.0264 (9) H2 0.9807 0.1629 0.8620 0.032* C3 0.9511 (6) 0.0952 (7) 0.9947 (3) 0.0299 (10) H3 1.0671 0.0973 1.0100 0.036* C4 0.8341 (6) 0.0543 (6) 1.0601 (3) 0.0283 (10) H4 0.8678 0.0281 1.1213 0.034* C5 0.6652 (6) 0.0520 (6) 1.0352 (3) 0.0274 (9) H5 0.5819 0.0254 1.0795 0.033* C6 0.6204 (5) 0.0884 (6) 0.9464 (3) 0.0232 (8) C7 0.4472 (6) 0.0856 (7) 0.9165 (3) 0.0295 (10) B1 0.7236 (7) 0.6589 (6) 0.8140 (3) 0.0242 (10) F1 0.7151 (4) 0.5195 (3) 0.8770 (2) 0.0328 (6) F2 0.8600 (4) 0.7654 (4) 0.8350 (2) 0.0361 (7) F3 0.5771 (3) 0.7603 (4) 0.8200 (2) 0.0389 (7) F4 0.7353 (4) 0.5925 (4) 0.72573 (19) 0.0446 (8) N3 0.7497 (5) 0.3325 (5) 0.3758 (2) 0.0246 (8) N4 1.1617 (6) 0.4550 (8) 0.3990 (3) 0.0454 (12) C8 0.8099 (6) 0.3192 (7) 0.2807 (3) 0.0316 (10) onal atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å2) Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å2) x y z Uiso/Ueq sup-4 Acta Cryst. (2015). Special details E71, o697–o698 supporting information supporting information H8A 0.9110 0.2455 0.2790 0.047 H8B 0.8361 0.4378 0.2575 0.047* H8C 0.7225 0.2655 0.2421 0.047* C9 0.5886 (6) 0.2988 (6) 0.3939 (3) 0.0292 (10) H9 0.5148 0.2630 0.3460 0.035* C10 0.5294 (6) 0.3159 (6) 0.4825 (3) 0.0302 (10) H10 0.4148 0.2926 0.4954 0.036* C11 0.6365 (6) 0.3668 (7) 0.5519 (3) 0.0310 (10) H11 0.5962 0.3775 0.6129 0.037* C12 0.8039 (6) 0.4028 (6) 0.5327 (3) 0.0304 (10) H12 0.8794 0.4387 0.5798 0.036* C13 0.8566 (6) 0.3847 (6) 0.4439 (3) 0.0255 (9) C14 1.0267 (7) 0.4226 (7) 0.4190 (3) 0.0327 (10) B2 0.7732 (7) 0.8421 (7) 0.3142 (3) 0.0273 (10) F5 0.8153 (4) 0.9543 (4) 0.3856 (2) 0.0435 (8) F6 0.6889 (5) 0.6960 (4) 0.3476 (2) 0.0500 (9) F7 0.9153 (4) 0.7890 (6) 0.2691 (2) 0.0569 (10) F8 0.6664 (4) 0.9294 (4) 0.2535 (2) 0.0464 (8) Atomic displacement parameters (Å2) U11 U22 U33 U12 U13 U23 N1 0.0256 (18) 0.0142 (16) 0.0263 (17) 0.0013 (14) 0.0025 (15) −0.0004 (14) N2 0.029 (2) 0.059 (3) 0.038 (2) 0.004 (2) 0.0041 (18) −0.002 (2) C1 0.037 (3) 0.022 (2) 0.023 (2) −0.0010 (18) −0.0026 (18) 0.0022 (17) C2 0.023 (2) 0.023 (2) 0.033 (2) −0.0011 (16) 0.0046 (19) −0.0008 (19) C3 0.025 (2) 0.033 (2) 0.032 (2) 0.0022 (19) −0.0025 (19) −0.002 (2) C4 0.031 (2) 0.029 (2) 0.025 (2) 0.0000 (18) −0.0006 (18) −0.0012 (18) C5 0.031 (2) 0.024 (2) 0.027 (2) 0.0002 (18) 0.0039 (19) −0.0005 (18) C6 0.023 (2) 0.0174 (19) 0.029 (2) 0.0007 (16) 0.0041 (17) −0.0020 (17) C7 0.029 (2) 0.028 (2) 0.032 (2) 0.0002 (18) 0.0046 (19) −0.001 (2) B1 0.029 (2) 0.018 (2) 0.026 (2) −0.0007 (19) 0.001 (2) 0.0018 (18) F1 0.0440 (15) 0.0189 (12) 0.0355 (14) −0.0006 (12) −0.0028 (13) 0.0047 (10) F2 0.0336 (15) 0.0254 (14) 0.0492 (17) −0.0069 (12) −0.0028 (13) 0.0005 (12) F3 0.0311 (15) 0.0235 (13) 0.062 (2) 0.0040 (12) 0.0021 (14) 0.0036 (13) F4 0.069 (2) 0.0343 (16) 0.0304 (14) −0.0020 (16) 0.0062 (15) −0.0067 (13) N3 0.0326 (19) 0.0141 (15) 0.0270 (18) 0.0007 (14) 0.0007 (15) −0.0004 (14) N4 0.034 (2) 0.060 (3) 0.042 (2) −0.008 (2) 0.000 (2) 0.005 (2) C8 0.041 (3) 0.025 (2) 0.028 (2) −0.004 (2) 0.004 (2) 0.0001 (19) C9 0.031 (2) 0.023 (2) 0.034 (2) −0.0023 (17) −0.005 (2) 0.0003 (19) C10 0.029 (2) 0.023 (2) 0.039 (3) −0.0003 (18) 0.0034 (19) 0.005 (2) C11 0.037 (2) 0.029 (2) 0.027 (2) 0.0017 (19) 0.001 (2) 0.0019 (19) C12 0.034 (2) 0.026 (2) 0.032 (2) 0.000 (2) −0.003 (2) 0.0016 (19) C13 0.029 (2) 0.0147 (18) 0.033 (2) 0.0023 (16) −0.0026 (19) 0.0026 (17) C14 0.035 (3) 0.031 (2) 0.032 (2) −0.001 (2) −0.002 (2) 0.002 (2) B2 0.033 (3) 0.023 (2) 0.026 (2) −0.002 (2) 0.002 (2) 0.004 (2) F5 0.061 (2) 0.0328 (16) 0.0363 (16) 0.0103 (14) −0.0076 (16) −0.0078 (13) F6 0.072 (2) 0.0196 (14) 0.059 (2) 0.0019 (15) 0.0198 (17) 0.0098 (14) Atomic displacement parameters (Å2) sup-5 Acta Cryst. supporting information F7 0.0376 (17) 0.072 (3) 0.061 (2) 0.0042 (17 F8 0.057 (2) 0.0294 (16) 0.0522 (18) −0.0021 (1 Geometric parameters (Å, º) N1—C2 1.340 (6) N3—C9 N1—C6 1.360 (6) N3—C13 N1—C1 1.484 (6) N3—C8 N2—C7 1.157 (6) N4—C14 C1—H1A 0.9800 C8—H8A C1—H1B 0.9800 C8—H8B C1—H1C 0.9800 C8—H8C C2—C3 1.376 (7) C9—C10 C2—H2 0.9500 C9—H9 C3—C4 1.372 (7) C10—C11 C3—H3 0.9500 C10—H10 C4—C5 1.392 (6) C11—C12 C4—H4 0.9500 C11—H11 C5—C6 1.369 (6) C12—C13 C5—H5 0.9500 C12—H12 C6—C7 1.446 (6) C13—C14 B1—F4 1.384 (6) B2—F7 B1—F2 1.385 (6) B2—F6 B1—F1 1.398 (5) B2—F5 B1—F3 1.400 (6) B2—F8 C2—N1—C6 118.7 (4) C9—N3—C1 C2—N1—C1 120.4 (4) C9—N3—C8 C6—N1—C1 120.9 (4) C13—N3—C N1—C1—H1A 109.5 N3—C8—H8 N1—C1—H1B 109.5 N3—C8—H8 H1A—C1—H1B 109.5 H8A—C8—H N1—C1—H1C 109.5 N3—C8—H8 H1A—C1—H1C 109.5 H8A—C8—H H1B—C1—H1C 109.5 H8B—C8—H N1—C2—C3 122.1 (4) N3—C9—C1 N1—C2—H2 118.9 N3—C9—H9 C3—C2—H2 118.9 C10—C9—H C4—C3—C2 119.5 (4) C11—C10—C C4—C3—H3 120.3 C11—C10—H C2—C3—H3 120.3 C9—C10—H C3—C4—C5 118.8 (4) C10—C11—C C3—C4—H4 120.6 C10—C11—H C5—C4—H4 120.6 C12—C11—H C6—C5—C4 119.4 (4) C13—C12—C C6—C5—H5 120.3 C13—C12—H C4—C5—H5 120.3 C11—C12—H N1—C6—C5 121.5 (4) N3—C13—C F7 0.0376 (17) 0.072 (3) 0.061 (2) 0.0042 (17) 0.0157 (15) −0.0251 (19) F8 0.057 (2) 0.0294 (16) 0.0522 (18) −0.0021 (14) −0.0197 (17) 0.0099 (14) Geometric parameters (Å, º) N1—C2 1.340 (6) N3—C9 1.336 (6) N1—C6 1.360 (6) N3—C13 1.362 (6) N1—C1 1.484 (6) N3—C8 1.473 (6) N2—C7 1.157 (6) N4—C14 1.143 (7) C1—H1A 0.9800 C8—H8A 0.9800 C1—H1B 0.9800 C8—H8B 0.9800 C1—H1C 0.9800 C8—H8C 0.9800 C2—C3 1.376 (7) C9—C10 1.382 (7) C2—H2 0.9500 C9—H9 0.9500 C3—C4 1.372 (7) C10—C11 1.375 (7) C3—H3 0.9500 C10—H10 0.9500 C4—C5 1.392 (6) C11—C12 1.392 (7) C4—H4 0.9500 C11—H11 0.9500 C5—C6 1.369 (6) C12—C13 1.370 (6) C5—H5 0.9500 C12—H12 0.9500 C6—C7 1.446 (6) C13—C14 1.434 (7) B1—F4 1.384 (6) B2—F7 1.372 (6) B1—F2 1.385 (6) B2—F6 1.382 (6) B1—F1 1.398 (5) B2—F5 1.382 (6) B1—F3 1.400 (6) B2—F8 1.391 (6) C2—N1—C6 118.7 (4) C9—N3—C13 120.6 (4) C2—N1—C1 120.4 (4) C9—N3—C8 119.4 (4) C6—N1—C1 120.9 (4) C13—N3—C8 119.9 (4) N1—C1—H1A 109.5 N3—C8—H8A 109.5 N1—C1—H1B 109.5 N3—C8—H8B 109.5 H1A—C1—H1B 109.5 H8A—C8—H8B 109.5 N1—C1—H1C 109.5 N3—C8—H8C 109.5 H1A—C1—H1C 109.5 H8A—C8—H8C 109.5 H1B—C1—H1C 109.5 H8B—C8—H8C 109.5 N1—C2—C3 122.1 (4) N3—C9—C10 119.9 (5) N1—C2—H2 118.9 N3—C9—H9 120.0 C3—C2—H2 118.9 C10—C9—H9 120.0 C4—C3—C2 119.5 (4) C11—C10—C9 120.0 (4) C4—C3—H3 120.3 C11—C10—H10 120.0 C2—C3—H3 120.3 C9—C10—H10 120.0 C3—C4—C5 118.8 (4) C10—C11—C12 119.9 (4) C3—C4—H4 120.6 C10—C11—H11 120.0 C5—C4—H4 120.6 C12—C11—H11 120.0 C6—C5—C4 119.4 (4) C13—C12—C11 118.0 (5) C6—C5—H5 120.3 C13—C12—H12 121.0 C4—C5—H5 120.3 C11—C12—H12 121.0 N1—C6—C5 121.5 (4) N3—C13—C12 121.5 (4) F7 0.0376 (17) 0.072 (3) 0.061 (2) 0.0042 (17) 0.0157 (15) −0.0251 (19) F8 0.057 (2) 0.0294 (16) 0.0522 (18) −0.0021 (14) −0.0197 (17) 0.0099 (14) sup-6 Acta Cryst. Special details (2015). E71, o697–o698 supporting information (2015). supporting information E71, o697–o698 supporting information N1—C6—C7 116.7 (4) N3—C13—C14 117.6 (4) C5—C6—C7 121.7 (4) C12—C13—C14 120.9 (5) N2—C7—C6 177.1 (5) N4—C14—C13 179.1 (6) F4—B1—F2 111.1 (4) F7—B2—F6 109.9 (4) F4—B1—F1 109.9 (4) F7—B2—F5 110.0 (4) F2—B1—F1 109.5 (4) F6—B2—F5 110.0 (4) F4—B1—F3 108.5 (4) F7—B2—F8 109.7 (4) F2—B1—F3 108.8 (4) F6—B2—F8 107.8 (4) F1—B1—F3 109.1 (4) F5—B2—F8 109.5 (4) C6—N1—C2—C3 0.7 (7) C13—N3—C9—C10 −0.4 (7) C1—N1—C2—C3 −177.6 (4) C8—N3—C9—C10 −178.0 (4) N1—C2—C3—C4 −0.9 (8) N3—C9—C10—C11 −0.3 (7) C2—C3—C4—C5 0.2 (8) C9—C10—C11—C12 0.7 (8) C3—C4—C5—C6 0.7 (7) C10—C11—C12—C13 −0.3 (7) C2—N1—C6—C5 0.2 (6) C9—N3—C13—C12 0.7 (7) C1—N1—C6—C5 178.5 (4) C8—N3—C13—C12 178.4 (4) C2—N1—C6—C7 −179.9 (4) C9—N3—C13—C14 −178.6 (4) C1—N1—C6—C7 −1.6 (6) C8—N3—C13—C14 −0.9 (7) C4—C5—C6—N1 −0.9 (7) C11—C12—C13—N3 −0.3 (7) C4—C5—C6—C7 179.2 (5) C11—C12—C13—C14 178.9 (5) Hydrogen-bond geometry (Å, º) D—H···A D—H H···A D···A D—H···A C1—H1A···F7i 0.98 2.50 3.407 (6) 154 C1—H1B···F8ii 0.98 2.54 3.498 (6) 166 C1—H1C···F3iii 0.98 2.47 3.214 (5) 132 C2—H2···F7i 0.95 2.29 3.190 (5) 157 C3—H3···F1iv 0.95 2.46 3.294 (6) 147 C5—H5···F1v 0.95 2.45 3.306 (5) 149 C8—H8A···F2i 0.98 2.48 3.159 (6) 126 C8—H8C···F3ii 0.98 2.55 3.437 (6) 151 C9—H9···F3ii 0.95 2.52 3.392 (6) 152 C9—H9···F4ii 0.95 2.59 3.476 (6) 156 C10—H10···F6ii 0.95 2.54 3.167 (6) 123 C12—H12···F5i 0.95 2.49 3.277 (6) 141 S mmetr codes: (i) +2 1/2 +1; (ii) +1 1/2 +1; (iii) 1 ; (i ) +2 1/2 +2; ( ) +1 1/2 +2 N1—C6—C7 116.7 (4) N3—C13—C14 117.6 (4) C5—C6—C7 121.7 (4) C12—C13—C14 120.9 (5) N2—C7—C6 177.1 (5) N4—C14—C13 179.1 (6) F4—B1—F2 111.1 (4) F7—B2—F6 109.9 (4) F4—B1—F1 109.9 (4) F7—B2—F5 110.0 (4) F2—B1—F1 109.5 (4) F6—B2—F5 110.0 (4) F4—B1—F3 108.5 (4) F7—B2—F8 109.7 (4) F2—B1—F3 108.8 (4) F6—B2—F8 107.8 (4) F1—B1—F3 109.1 (4) F5—B2—F8 109.5 (4) C6—N1—C2—C3 0.7 (7) C13—N3—C9—C10 −0.4 (7) C1—N1—C2—C3 −177.6 (4) C8—N3—C9—C10 −178.0 (4) N1—C2—C3—C4 −0.9 (8) N3—C9—C10—C11 −0.3 (7) C2—C3—C4—C5 0.2 (8) C9—C10—C11—C12 0.7 (8) C3—C4—C5—C6 0.7 (7) C10—C11—C12—C13 −0.3 (7) C2—N1—C6—C5 0.2 (6) C9—N3—C13—C12 0.7 (7) C1—N1—C6—C5 178.5 (4) C8—N3—C13—C12 178.4 (4) C2—N1—C6—C7 −179.9 (4) C9—N3—C13—C14 −178.6 (4) C1—N1—C6—C7 −1.6 (6) C8—N3—C13—C14 −0.9 (7) C4—C5—C6—N1 −0.9 (7) C11—C12—C13—N3 −0.3 (7) C4—C5—C6—C7 179.2 (5) C11—C12—C13—C14 178.9 (5) Hydrogen-bond geometry (Å, º) Hydrogen-bond geometry (Å, º) D—H···A D—H H···A D···A D—H···A C1—H1A···F7i 0.98 2.50 3.407 (6) 154 C1—H1B···F8ii 0.98 2.54 3.498 (6) 166 C1—H1C···F3iii 0.98 2.47 3.214 (5) 132 C2—H2···F7i 0.95 2.29 3.190 (5) 157 C3—H3···F1iv 0.95 2.46 3.294 (6) 147 C5—H5···F1v 0.95 2.45 3.306 (5) 149 C8—H8A···F2i 0.98 2.48 3.159 (6) 126 C8—H8C···F3ii 0.98 2.55 3.437 (6) 151 C9—H9···F3ii 0.95 2.52 3.392 (6) 152 C9—H9···F4ii 0.95 2.59 3.476 (6) 156 C10—H10···F6ii 0.95 2.54 3.167 (6) 123 C12—H12···F5i 0.95 2.49 3.277 (6) 141 sup-7 Acta Cryst. Symmetry codes: (i) −x+2, y−1/2, −z+1; (ii) −x+1, y−1/2, −z+1; (iii) x, y−1, z; (iv) −x+2, y−1/2, −z+2; (v) −x+1, y−1/2, −z+2. Acta Cryst. (2015). E71, o697–o698 Symmetry codes: (i) −x+2, y−1/2, −z+1; (ii) −x+1, y−1/2, −z+1; (iii) x, y−1, z; (iv) −x+2, y−1/2, −z+2; (v) −x+1, y−1/2, −z+2. supporting information (2015). E71, o697–o698
https://openalex.org/W2807845858	https://digital.csic.es/bitstream/10261/245151/1/Prognostic_Gonzalez_PV_Art2018.pdf	English	null	Prognostic stratification of adult primary glioblastoma multiforme patients based on their tumor gene amplification profiles	Oncotarget	2,018	cc-by	11,412	Prognostic stratification of adult primary glioblastoma multiforme patients based on their tumor gene amplification profiles María González-Tablas1, Inês Crespo2,3, Ana Luísa Vital2,3, Álvaro Otero4, Ana Belén Nieto5, Pablo Sousa4, María Carmen Patino-Alonso5, Luis Antonio Corchete6, Hermínio Tão7, Olinda Rebelo8, Marcos Barbosa7,9, Maria Rosário Almeida2, Ana Filipa Guedes2, María Celeste Lopes2,3, Pim J. French10, Alberto Orfao1,11,* and María Dolores Tabernero1,11,* 1Centre for Cancer Research (CIC IBMCC-CSIC/USAL), Department of Medicine, CIBERONC, University of Salamanca, Salamanca, Spain for Neuroscience and Cell Biology, University of Coimbra, Coimbra, Portugal 2Centre for Neuroscience and Cell Biology, University of Coimbra, Coimbra, Portugal 3Faculty of Pharmacy, University of Coimbra, Coimbra, Portugal 6Departamento de Hematología, Hospital Universitario, IBSAL, IBMCC (USAL-CSIC), Salamanca, Spain Neurosurgery Service, University Hospital of Coimbra, Coimbra, Portugal Neuropathology Laboratory, Neurology Service, University Hospital of Coimbra, Coimbra, Portugal 9Faculty of Medicine, University of Coimbra, Coimbra, Portugal 10 0Department of Neurology, Erasmus MC, Rotterdam, The Netherlands 11Instituto Biosanitario de Salamanca (IBSAL), Salamanca, Spain Both authors have equally contributed to this work and should be considered as last authors 11Instituto Biosanitario de Salamanca (IBSAL), Salamanca, Spain 1Instituto Biosanitario de Salamanca (IBSAL), Salamanca, Spain Both authors have equally contributed to this work and should be considered as last authors Correspondence to: María Dolores Tabernero, email: taberner@usal.es Keywords: glioblastoma; classification; subtypes; gene amplification; survival Received: February 28, 2018 Accepted: May 14, 2018 Published: June 15, 2018 Correspondence to: María Dolores Tabernero, email: taberner@usal.es Correspondence to: María Dolores Tabernero, email: taberner@usal.es Keywords: glioblastoma; classification; subtypes; gene amplification; survival Keywords: glioblastoma; classification; subtypes; gene amplification; survival Received: February 28, 2018 Accepted: May 14, 2018 Published: June 15, 2018 Copyright: González-Tablas et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Copyright: González-Tablas et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Oncotarget, 2018, Vol. 9, (No. 46), pp: 28083-28102 Oncotarget, 2018, Vol. 9, (No. 46), pp: 28083-28102 www.oncotarget.com www.oncotarget.com Prognostic stratification of adult primary glioblastoma multiforme patients based on their tumor gene amplification profiles p INTRODUCTION into classical and non-classical GBM, based on gene expression profiling (GEP) data, was first proposed [27]. Later on, Phillips et al [13] defined three molecular subtypes of GBM according to a combination of GEP data and numerical alterations of chromosomes 7 and 10: proneural, proliferative and mesenchymal GBM. This classification was subsequently redefined by the Cancer Genome Atlas Consortium (TCGA) [14] into four subtypes – proneural, proliferative, mesenchymal and neural GBM, using a combination of GEP data and CNA, together with the pattern of somatic mutations. Despite overlapping data is used in the later two classifications, both approaches are not equivalent, at the same time they are rather complex to be reproducibly applied in routine laboratory diagnostics. In addition, the prognostic impact of both approaches remains controversial because e.g. the proneural subgroup of Philips [13] has a longer survival, while proneural tumors defined according to the TCGA classification [14], have a poorer outcome. In turn, despite EGFR amplification is a defining event for the classical subtype of GBM, in the TCGA classification it appears in >95% of the neural and mesenchymal tumor subtypes, but also in 54% of the proneural subtype; similarly, the same EGFR mutations (e.g. the EGFRvIII variant) can also be detected across all above subtypes of GBM; in addition, neither the criteria used to define the proneural class based on focal amplification of the 4q12 locus harboring the PDGFRA gene (with or without EGFR amplification), nor the evaluation of the expression of the NEFL, GABRA1, SYT1 and SLC12A5 genes, are currently applied for routine diagnostic classification of GBM. Altogether this hampers fast and reproducible risk-stratification of GBM patients at diagnosis, based on this classification. Primary glioblastoma multiforme (GBM) is the most common and malignant subtype of glial tumors [1]. From the clinical and biological point of view, GBM includes a rather heterogeneous group of tumors that vary by site of origin, histophatological features, tumor microenvironment [2] and genetics [3]. They are usually resistant to radio/chemotherapy and show overall survival (OS) rates of a few months to years, making them unvariably lethal [4, 5]. Criteria used for the histological classification and grading of GBM have been recently revised in the 4th World Health Organization (WHO) classification of GBM [1]; however, this classification still fails in distinguishing subgroups (or variants) of primary GBM that display clearly distinct clinical and biological behaviours, despite sharing similar histopathological features [6–9]. ABSTRACT Several classification systems have been proposed to address genomic heterogeneity of glioblastoma multiforme, but they either showed limited prognostic value and/or are difficult to implement in routine diagnostics. Here we propose a prognostic stratification model for these primary tumors based on tumor gene amplification profiles, that might be easily implemented in routine diagnostics, and potentially improve the patients management. Gene amplification profiles were prospectively evaluated in 80 primary glioblastoma multiforme tumors using single- nucleotide polymorphism arrays and the results obtained validated in publicly available data from 267/347 cases. Gene amplification was detected in 45% of patients, and chromosome 7p11.2 including the EGFR gene, was the most frequently amplified chromosomal region – either alone (18%) or in combination with amplification of DNA sequences in other chromosomal regions (10% of cases). Other frequently amplified DNA sequences included regions in chromosomes 12q(10%), 4q12(7%) and 1q32.1(4%). Based on their gene amplification profiles, glioblastomas were subdivided into: i) tumors with no gene amplification (55%); ii) tumors with chromosome 7p/EGFR gene amplification (with or without amplification of other chromosomal www.oncotarget.com Oncotarget 28083 regions) (38%); and iii) glioblastoma multiforme with a single (11%) or multiple (6%) amplified DNA sequences in chromosomal regions other than chromosome 7p. From the prognostic point of view, these amplification profiles showed a significant impact on overall survival of glioblastoma multiforme patients (p>0.001). Based on these gene amplification profiles, a risk-stratification scoring system was built for prognostic stratification of glioblastoma which might be easily implemented in routine diagnostics, and potentially contribute to improved patient management. regions) (38%); and iii) glioblastoma multiforme with a single (11%) or multiple (6%) amplified DNA sequences in chromosomal regions other than chromosome 7p. From the prognostic point of view, these amplification profiles showed a significant impact on overall survival of glioblastoma multiforme patients (p>0.001). Based on these gene amplification profiles, a risk-stratification scoring system was built for prognostic stratification of glioblastoma which might be easily implemented in routine diagnostics, and potentially contribute to improved patient management. Gene amplification profiles and chromosomal regions involved in GBM SNP-arrays showed CNA containing > 7 DNA copies of the same DNA sequence for ≥1 chromosome/ chromosomal region in all (80/80) cases analyzed from series 1. In the majority of patients (45/80; 56%) such CNA involved genetic amplification of DNA sequences from ≥1 chromosomal region (Table 1 and Figure 1). As expected, DNA sequences at chromosome 7p11.2 containing the EGFR gene, were the most frequently amplified sequences (30/80 cases; 38%), followed by DNA sequences at the 12q (14/80; (18%), 4q (8/80; 10%) and 1q (5/80; 6%) chromosomal regions. Amplification of DNA sequences at chromosome 7p11.2 was found either alone (17/80; 21%) or in combination with amplification of DNA sequences at other regions in chromosomes 1q, 4q, 5q, 6q, 11p, 11q, 12q, 17p and 17q (13/80; 16% of cases) (Table 1). The most frequent combination of simultaneously amplified DNA sequences at chromosomal regions involving 7p11.2 were those of 7p and 12q (7/80 cases; 9%) and of 7p and 11p (2/80 cases; 3%) (Table 1). Isolated amplification of DNA sequences at chromosomal regions other than 7p11.2 was found in 10/80 cases (13%); this included isolated amplification of 4q12 (where the PDGFRA gene is coded) in 4/80 cases (5%), 12q14.1 (where the CDK4 gene is coded) in 3/80 cases (4%), 1q32.1 (this chromosomal region contains the MDM4 gene) in another 2/80 tumors (3%) and 16q in one tumor (1%) without any annotated gene being coded in the amplified DNA sequences from this later chromosomal region (Table 1). RESULTS were also cases with polyploidies in the absence of gene amplification and vice versa. p Tumors with isolated amplification of DNA sequences at a single chromosomal region had a median of 9 genes involved (range: 1 to 40 genes). Gene amplification at 7p11.2 included the smallest number of affected genes (median: 5 vs 16 genes for gene amplification involving DNA sequences at chromosomes 12q, 1q and 4q). In turn, cases showing amplified DNA sequences at several chromosomal regions had a median of 19 genes involved (range: 1 to 45 genes). The most commonly (>5% of cases) amplified genes per chromosomal region included: i) 9 genes at chromosome 1q32.1 (SOX13, ETNK2, REN, KISS1, GOLT1A, PLEKHA6, PIK3C2B, MDM4 and LRRN2); ii) 4 genes at chromosome 4q12 (SCFD2, FIP1L1, PDGFRA and KIT); iii) 5 genes in chromosome 7p11.2 (EGFR, LANCL2, VSTM2A, VOPP1 and SEC61G), and; iv) multiple genes at chromosome 12 which had 3 cytobands involved, i.e. 12q14.1 (CDK4, METTL1, CYP27B1, AVIL, CTDSP2, METT21B, AGAP2, OS9, TSFM), 12q13.3 (B4GALNT1, KIF5A, PIP4K2C, DTX3, SLC26A10, MARS, DCTN2, ARHGEF25) and 12q15 (ATP23, MDM2, CPM) (Table 3). When individually considered, the EGFR gene (38%) together with the LANCL2 gene (23%), coded also at chromosome 7p11.2, were the two most frequently amplified genes (Table 3). Of note, amplification of none of the genes detected here had been previously described in healthy individuals [33] and/or publicly available GBM databases. INTRODUCTION Besides histopathology, molecular genetics data has also confirmed the high heterogeneity of GBM, both at the intertumoral and intratumoral levels [3, 10– 12]; in addition, molecular genetics data also proved useful for improving the diagnosis, classification, and prognostic stratification of GBM [13, 14], with therapeutic implications [14–16]. However, despite all advances achieved via the study of the methylation status, gene mutations (e.g. IDH1 or PT53 genes) [17] and affected oncogenic pathways [11, 18–20], the precise mechanisms involved in the pathogenesis of GBM still remain far from being fully understood. In turn, routine implementation of molecular genetics into the diagnostic classification of GBM still remains limited, due to the complexity of the genetic findings involved. Among other genetic abnormalities, the copy number aberration (CNA) profile of tumor cells, over the copy number variation (CNV) of individual patients, has been recognized [21–24] as a useful prognostic tool in GBM [14]. Thus, some copy number aberrations involving one or multiple genes that affect a significant fraction of the tumors, (e.g. CNA associated with either gains of chromosome 7 and/or amplification of the epidermal growth factor receptor (EGFR) gene [18], and losses of DNA sequences in chromosome 10 [19], together with other less frequent alterations involving DNA sequences in chromosomes 12q13-15 [25], 4q and 1q [26], have all been associated with the outcome of GBM patients. Despite this, risk stratification of GBM Here we investigated the gene amplification profiles (GAP) across the whole tumor genome of a series of 80 GBM tumors, as detected by high-density single-nucleotide polymorphism (SNP)-arrays, and evaluated their impact on overall survival (OS) of GBM patients. Based on the results obtained, a GAP-based risk- stratification model was built and validated in a series of 267 GBM tumors from a total of 7 GBM series publicly available at the GEO and ArrayExpress databases and/or whose data was kindly provided by the authors [15, 24, 28–32], in addition to our own cases. www.oncotarget.com Oncotarget 28084 Classification of GBM tumors based on their gene amplification profiles (GAP) Based on the presence of gene amplification, their specific subtype and the number of chromosomal regions affected by gene amplification (Figure 1), GBM tumors from group 1 (series 1) were divided into five different subgroups: i) tumors which had no gene amplification (n=35; 44%); ii) tumors with isolated amplification of DNA sequences at chromosome 7p including the EGFR gene (n=17; 21%); iii) GBM with isolated amplification of DNA sequences at a chromosomal region different from chromosome 7p (n=10; 13%); iv) tumors with amplifications of DNA sequences from multiple (≥2) chromosomal regions including that of the EGFR gene (n=13; 16%), and; v) tumors with amplification of DNA sequences at ≥2 chromosomal regions which did not include amplification of the EGFR gene (n=5; 6%) (Table 4 and Figure 2). In turn, genetic amplification at 12q was detected in 14/80 cases (18%) including isolated amplification in 3 GBM and combined 12q gene amplification with gene amplification at other chromosome regions (1p, 1q, 4q, 11p, 11q, 7q, or 17q) in 11 tumors (Table 1), including amplification at chromosome 7p in 7 cases (9%). Gene amplification at 4q was found in 8/80 cases (10%) either alone (4/80; 5% or in association with amplification of DNA sequences at the 12q and/or 7p (4/80; 5%) chromosomal regions (Table 1). Overall, coexistence of genetically amplified DNA sequences at ≥ 2 chromosomal regions in the same tumor (median of 2 altered chromosomal regions; range: 2-4 regions) was found in 18 cases (23%). In 13 of these 18 tumors, the EGFR gene was involved (Table 1). Of note, gene amplification involving regions at chromosomes 7q (4%), 11p (3%), 17q (3%), 5q (1%), 6q (1%), 11q (1%) and 17p (1%) included variable chromosomal patterns and numbers of genes involved (Table 1 and Table 2). Although gene amplification was frequently associated with chromosomal gains (polyploidy), there The above GAP were further investigated in the other 7 series of GBM used for validation purposes, for a total of 267 primary GBM patients (Table 4). Classification of GBM tumors based on their gene amplification profiles (GAP) Around half of these cases ˗156/267 (58%)˗ did not show genetic amplification at any of the chromosomal regions investigated (Table 4); in contrast, 69 (26%) tumors had gene amplification at www.oncotarget.com Oncotarget 28085 Table 1: Major subsets of GBM that carried different gene amplification profiles, as identified in our series (series 1) of GBM patients (n=80) grouped according to the location, type and number of chromosomal regions involved (n=45/80 tumors) Tumor group Tumor ID Amplified chromosome(s) N. of amplified chromosomal band(s)/genes Amplified Chromosomal bands Gene amplification at a single chromosomal region (n=27) EGFR gene involved (n=17;21%) G94 7p 2/4 7p11.2/7p12.1# G55 2/2 G91 1/2 7p11.2 G80 1/3 G72 1/2 G68 1/2 G67 1/1 G56 1/2 G44ϑ 1/3 G40ϑ 1/3 G37ϑ 1/2 G30ϑ 1/1 GBM3ϑ 1/17 GBM7ϑ 1/3 GBM11ϑ 1/7 GBM12ϑ 1/25 GBM17ϑ 1/2 EGFR gene not involved (n=10;13%) G73 4q 3/35 4q11/4q12/4q13.3 G12 1/29 4q11#/4q12 GBM1ϑ 1/16 4q12 GBM14ϑ 1/7 4q12 G51ϑ 12q 2/14 12q13.3/12q14.1 G46ϑ 2/5 12q14.1-q14.3/12q15 G25ϑ 1/8 12q14.1 G79 1q 1/12 1q32.1 G54 1/18 G10 16q# 1/0 16q12.1-q12.2 Gene amplification at multiple chromosomal regions (n=18) EGFR gene & other amplicons involved (n=13;16%) G39ϑ 7p, 12q 1/2, 2/9 7p11.2/12q13.3/12q14.1 G41 1/4, 2/11 7p11.2/12q14/12q15 G53ϑ 3/6, 1/2 7p11.2/7p21.3/7p22.1/12q15 G70 1/1, 2/20 7p11.2/12q13.3/12q14.1 G83 1q, 7p 1/16, 1/1 1q32.1/7p11.2 G65 1/26, 1/12 GBM19ϑ 7p, 7q 4/5 7p21/7p12.3/7p11.2/7q22.3 G23ϑ 7p, 11p 1/2, 1/17 7p11.2/11p13 GBM4ϑ 1q,7p,12q 1/18, 1/6, 2/11 1q32.1/7p11.2/12q13.3/12q14.1 G82 4q, 7p, 12q 2/6, 1/1, 3/23 4q12/4q13.3/7p11.2/12q13.12- q13.13/12q13.3/12q14.1/12q15 G90 5q#, 6q#, 7p 1/1 5q34/6q25/7p11.2 G81 7p, 17p, 17q 1/2, 2/11 7p11.2/17p13.1/17q25.1 G71 7p, 11p, 11q, 12q 2/5, 3/16, 2/19 7p11.2/7p12.1/11p15.3/11p11.2#/ 11q13.3/11q25/12q13.3/12q14.1 Table 1: Major subsets of GBM that carried different gene amplification profiles, as identified in our series (series 1) of GBM patients (n=80) grouped according to the location, type and number of chromosomal regions involved (n=45/80 tumors) Table 1: Major subsets of GBM that carried different gene amplification profiles, as identified in our series (series 1) of GBM patients (n=80) grouped according to the location, type and number of chromosomal regions involved (n=45/80 tumors) (Continued ) www.oncotarget.com www.oncotarget.com Oncotarget 28086 Tumor group Tumor ID Amplified chromosome(s) N. of amplified chromosomal band(s)/genes Amplified Chromosomal bands EGFR gene not involved (n=5;6%) GBM13ϑ 4q, 12q 1/7, 1/11 4q12/12q14.1 GBM22ϑ 1/19, 2/20 4q12/12q13.3/12q14.1 G08ϑ 1p, 7q 3/30, 1/7 1p12/1p13/1p21/7q21.2-q21.3 G88 4q, 7q, 12q 2/13, 1/3, 4/30 4q12/4q13.3/7q31.2/12q13.3/12q14.1/12q1 5/12q21.1 G89 1p, 12q, 17q 1/3, 2/21, 3/6 1p36.21/12q13.3/12q14.1/17q11/17q12#/17q2 1/17q22#/17q24 ID: Case identification code. Classification of GBM tumors based on their gene amplification profiles (GAP) Most frequently involved genes per chromosomal band included: 1q32.1, MDM4; 4q12, PDGFRA; 12q14.1, CDK4; 12q15, MDM2; 7q31.2,MET. #: Amplified chromosomal region without any annotated gene in it; ϑ: EGFR gene mutation studied in this sample. ID: Case identification code. Most frequently involved genes per chromosomal band included: 1q32.1, MDM4; 4q12, PDGFRA; 12q14.1, CDK4; 12q15, MDM2; 7q31.2,MET. #: Amplified chromosomal region without any annotated gene in it; ϑ: EGFR gene mutation studied in this sample. GBM patients, both in the test series (series 1) alone (p<0.001; Figure 3A), and in the whole cohort (series 1-8; p<0.001) (Figure 3C). In detail, cases that did not show gene amplification or that displayed amplification of the EGFR gene (alone or in combination with amplification of genes in other chromosomal regions) (groups 1, 2, and 4, respectively) showed a significantly better outcome than patients with isolated amplification of genes other than the EGFR gene (group 3) and cases with amplification of multiple chromosomal regions which did not involve the EGFR region (group 5): median OS of 14, 18 and 14 months vs 6 (p=0.001, p<0.001 and p=0.007, respectively) and 8 months (p=0.03, p<0.001 and p=0.003, respectively) in the test series alone, and of 14, 18 and 14 months vs 8 (p=0.03, p=0.003 and p=0.16, respectively) and 8 (p<0.001, p<0.001 and p=0.001, respectively) months in the whole cohort, respectively (Figure 3, panels A and C, respectively). When cases were re-grouped according to i) the absence of gene amplification or presence of EGFR gene amplification vs ii) occurrence of other GAP, the prognostic impact of the re-grouped GAP was enhanced, both when the test series alone and the whole cohort of patients analyzed, were considered: median OS rates of 15 vs 6 months (p< 0.001) and of 15 vs 8 months (p<0.001), respectively (Figure 3 panels B and D, respectively). Subsequent multivariate analysis of prognostic factors showed that the tumor GAP (p<0.001), together with the administration of chemotherapy (p<0.001) were the best combination of independent prognostic factors to predict patient OS (Table 5). chromosome 7p11.2 amplification –46/267 cases (17%) with isolated EGFR amplification and 23/267 cases (9%) showed amplification of DNA sequences at multiple chromosomal regions including amplification of the EGFR gene– (Table 2 and Table 4). DISCUSSION Despite histopathology remains the gold standard for the diagnosis of GBM, it provides limited information about patient outcome. Consequently, new classifications have been proposed in recent years for GBM in which molecular genetics data derived from chromosomal alteration profiles, DNA mutational status and GEP data, is used for the definition of tumor subgroups of Classification of GBM tumors based on their gene amplification profiles (GAP) Similarly to what was observed in the test series, gene amplification at 7p, in association with gene amplification at chromosome 12q, was found in 6 tumors (2%) and gene amplification at chromosome 7p together with amplification of DNA sequences at chromosomal regions other than 12q was found in another 17 cases (6%), –e.g. gene amplification at the 1q or 4q chromosomal regions in 3/267 (1.1%) and 4/267 (1.5%) cases, respectively, and at 7q or 3q in 3/267 (1.1%) and 2/267 (0.8%) tumors, respectively – (Table 2). DNA sequences from other chromosomal regions which were also found to be amplified at lower frequencies than those amplified at 7p, included DNA sequences at chromosomes 12q (10/267; 4%) and 4q (6/267; 2%) (Table 2). Genetic amplification at other chromosomal regions such as 7q, 12p and 13q was found in two tumors each, and gene amplification at the 6p, 8q, 15q and 17q chromosomal regions was found in a single tumor each (Table 2). Upon merging the test and validation series (series 1 to 8), the following distribution into the 5 subgroups of GBM defined by their distinct GAP (Table 4), was found: 191 cases (55%) had no gene amplification (group 1), 63/347 cases (18%) showed isolated amplification of EGFR) (group 2); 37/347 (11%) displayed isolated amplification of genes other than EGFR (group 3); 36/347 (10%) had genetic amplification at multiple chromosomal regions, including amplification of the EGFR gene (group 4); and 20/347 (6%) showed multiple amplified DNA sequences at ≥ 2 distinct chromosomal regions, which did not include amplification of the EGFR gene (group 5) (Table 4). Prognostic impact of gene amplification profiles in GBM From the prognostic point of view, the above defined GAP showed a significant impact on OS of www.oncotarget.com Oncotarget 28087 trating examples of the cytogenetic profiles representative of the distinct patterns of gene amplifi rom series 1, used for the definition of the 5 different subgroups of GBM based on their gene am Figure 1: Illustrating examples of the cytogenetic profiles representative of the distinct patterns of gene amplification identified in GBM tumors from series 1, used for the definition of the 5 different subgroups of GBM based on their gene amplification profiles. www.oncotarget.com Oncotarget 28088 t t Table 2: Frequency of genetic amplification involving different chromosomal regions in the test (series 1) vs the validation series (series 2-8) of GBM patients analyzed Number of amplified regions Chromosomal regions involved Amplified GBM cases Series 1 n=45/80 (56%) Series 2-8 n=111/267 (42%) One chromosomal region amplified (n=100; 29%) 7p 12q 4q 1q 12p 7q 13q 6p 8q 15q 16q 17q 17 (21%) 3 (4%) 4 (5%) 2 (3%) (0) (0) (0) (0) (0) (0) 1(1.2%) (0) 46(17%) 10 (4%) 6 (2%) 1 (0.4%) 2 (0.8%) 2 (0.8%) 2 (0.8%) 1 (0.4%) 1 (0.4%) 1 (0.4%) (0) 1 (0.4%) Subtotal 27 (34%) 73 (27%) Two chromosomal regions amplified (n=39; 11%) 7p,12q 1q,7p 1q,4q 7p,7q 3q,7p 4q,7p 4q,12q 1q,5p 1p7q 2q,12q 4q,9p 5p,7p 5q,7p 7p,11p 7p,9p 7p,14q 7p, Xp 17q,20q 20p,20q 4 (5%) 2 (3%) (0) 1 (1.2%) (0) (0) 2 (3%) (0) 1 (1.2%) (0) (0) (0) (0) 1 (1.2%) (0) (0) (0) (0) (0) 4 (1%) 3 (1%) 3 (1%) 2 (0.8%) 2 (0.8%) 2 (0.8%) 2 (0.8%) 1 (0.4%) (0) 1 (0.4%) 1 (0.4%) 1 (0.4%) 1 (0.4%) (0) 1 (0.4%) 1 (0.4%) 1 (0.4%) 1 (0.4%) 1 (0.4%) Subtotal 11 (14%) 28 (10%) ≥Three chromosomal regions amplified (n=17; 5%) 1p,1q,10q 1p,7p,12q 1p,12q,17q 1q,7p,12q 4p,4q,12p 4q,7p,12q 4q,7q,12p 4q7q12q 4q,7p,18q 5q,6q,7p 6p,7q,12q 7p,17p,17q 2p,7p,12p,12q 4q,7p,7q,15q 7p11p11q12q 7q,12q,17q,20q 1p,4p,7p,11q,14q (0) (0) 1 (1.2%) 1 (1.2%) (0) 1 (1.2%) (0) 1 (1.2%) (0) 1 (1.2%) (0) 1 (1.2%) (0) (0) 1 (1.2%) (0) (0) 1 (0.4%) 1 (0.4%) (0) (0) 1 (0.4%) (0) 1 (0.4%) (0) 1 (0.4%) (0) 1 (0.4%) (0) 1 (0.4%) 1 (0.4%) (0) 1 (0.4%) 1 (0.4%) Subtotal 7 (9%) 10 (4%) Results expressed as number (percentage) of cases showing gene amplification/chromosome arm. cytobands was calculated using only the 250k-SNP_Nsp and/or STY arrays (n=50) due to the absence of probes for these specific genes in the 50k array. he SNP6 array; The incidence of amplified genes localized in the 12q14.1, 12q13.3 and 12q15 #genes included in the SNP6 array; The incidence of amplified genes localized in the 12q14.1, 12q13.3 and 12q15 cytobands was calculated using only the 250k-SNP_Nsp and/or STY arrays (n=50) due to the absence of probes for these specific genes in the 50k array. sing only the 250k-SNP_Nsp and/or STY arrays (n=50) due to the absence of probes for these ray Prognostic impact of gene amplification profiles in GBM Table 2: Frequency of genetic amplification involving different chromosomal regions in the test (series 1) vs the validation series (series 2-8) of GBM patients analyzed Results expressed as number (percentage) of cases showing gene amplification/chromosome arm. www.oncotarget.com Oncotarget 28089 Table 3: Frequency and chromosomal localization of recurrently amplified genes (> 5% of cases) in GBM from series 1 (n=45/80 tumors), and other previously reported series of GBM (n=111/267 tumors) as detected by SNP- arrays Amplified genes Amplified chromosomal regions Symbol Frequency of gene amplification Chr Amplified/Total cases (%) Cytoband Series 1 (n=45) Series 2-8 (n=111) Chr 7 119/347 (34%) 7p11.2 EGFR LANCL2 VSTM2A VOPP1 SEC61G 67% 40% 31% 27% 11% 57% 19% 29% 7% 19% Chr 12 34/347 (10%) 12q14.1 CDK4 CYP27B1 METTL1 AVIL CTDSP2 METT21B AGAP2 OS9 TSFM 29% 29% 27% 27% 27% 22% 22% 16% 16% 14% 2%* 14%* 18%* 2%* 16%* 16%* 2%* - 12q13.3 B4GALNT1 KIF5A PIP4K2C SLC26A10# DTX3 MARS ARHGEF25 DCTN2# 16% 11% 11% 11% 9% 9% 9% 7% 14%* 10%* - - 12%* 4%* 4%* - 12q15 ATP23 MDM2 CPM 16% 9% 9% 14%* 12% 11% Chr 4 26/347 (7%) 4q12 SCFD2 FIP1L1 PDGFRA KIT 18% 18% 18% 7% 7% 14% 10% 10% Chr 1 14/347 (4%) 1q32.1 SOX13 ETNK2 REN KISS1 GOLT1A PLEKHA6 PIK3C2B MDM4 LRRN2 11% 11% 11% 11% 11% 11% 11% 11% 9% 5% 3% 2% 3% 3% 5% 5% 3% 2% #genes included in the SNP6 array; The incidence of amplified genes localized in the 12q14.1, 12q13.3 and 12q15 cytobands was calculated using only the 250k-SNP Nsp and/or STY arrays (n=50) due to the absence of probes for these Table 3: Frequency and chromosomal localization of recurrently amplified genes (> 5% of cases) in GBM from series 1 (n=45/80 tumors), and other previously reported series of GBM (n=111/267 tumors) as detected by SNP- arrays Chr 12 #genes included in the SNP6 array; The incidence of amplified genes localized in the 12q14.1, 12q13.3 and 12q15 cytobands was calculated using only the 250k-SNP_Nsp and/or STY arrays (n=50) due to the absence of probes for these specific genes in the 50k array. #genes included in the SNP6 array; The incidence of amplified genes localized in the 12q14.1, 12q13.3 and 12q15 cytobands was calculated using only the 250k-SNP_Nsp and/or STY arrays (n=50) due to the absence of probes for these specific genes in the 50k array. Prognostic impact of gene amplification profiles in GBM www.oncotarget.com www.oncotarget.com Oncotarget 28090 Table 4: Overall survival and genetic features of GBM patients from our series (series 1; n=80) and the seven series of GBM patients previously reported in the literature (series 2-8; n=267) and included in this study for a total of 347 GBM investigated Variables GBM patient series Series 1 Crespo et al. (GSE 42631) Series 2 Chen et al. (GSE 19612) Series 3 Beroukhim et al. (GSE19399/ GSE9635) Series 4 Bralten et al. NA Series 5 Hodgson et al. (GSE 14804) Series 6 Yin et al. (EMEXP-1330) Series 7 Kuga et al. (GSE 10922) Series 8 Solomon et al. (GSE 13021) Total N. of cases (n=347) 80 24 120 15 12 53 13 30 N. of cases with annotated OS (n=273) 80 24 104 15 NA 50 NA NA Median OS months (range) 15 (0-83) 15 (1-31) 18 (1-67) 10 (4-28) NA 17 (0-90) NA NA N. of SNP probes investigated 5-18x105 5x105 1-2.5x105 2.5x105 0.5x105 0.5-2.5x105 0.5x105 2.5x105 New genetical subsets NO gene AMP (n=191; 55%) 35 (44%) 11 (46%) 75 (62%) 9 (60%) 7 (58%) 31 (58%) 7 (54%) 16 (53%) Isolated EGFR AMP (n=63; 18%) 17 (21%) 7 (29%) 20 (17%) 1 (7%) 2 (17%) 10 (19%) 2 (15%) 4 (13%) Isolated non- EGFR AMP (n=37; 11%) 10 (13%) 1 (4%) 11 (9%) 3 (20%) 2 (17%) 7 (13%) 1 (8%) 2 (7%) Multiple AMP including EGFR (n=36; 10%) 13 (16%) 3 (13%) 8 (7%) 1 (7%) 0 3 (6%) 2 (15%) 6 (20%) Multiple AMP without EGFR (n=20; 6%) 5 (6%) 2 (8%) 6 (5%) 1 (7%) 1 (8%) 2 (4%) 1 (8%) 2 (7%) * 23 additional GBM tumors not available in the GEO data repository were hybridized with the Cytoscan750K (n=11 samples) and Cytoscan HD (n=12 samples) SNP-arrays; AMP: genetic amplification; GAP: genetic amplification profile; GSE: genomic repository series code; OS: overall survival; N: number; NA: not available. Prognostic impact of gene amplification profiles in GBM Table 4: Overall survival and genetic features of GBM patients from our series (series 1; n=80) and the seven series of GBM patients previously reported in the literature (series 2-8; n=267) and included in this study for a total of 347 GBM investigated * 23 additional GBM tumors not available in the GEO data repository were hybridized with the Cytoscan750K (n=11 samples) and Cytoscan HD (n=12 samples) SNP-arrays; AMP: genetic amplification; GAP: genetic amplification profile; GSE: genomic repository series code; OS: overall survival; N: number; NA: not available. * 23 additional GBM tumors not available in the GEO data repository were hybridized with the Cytoscan750K (n=11 samples) and Cytoscan HD (n=12 samples) SNP-arrays; AMP: genetic amplification; GAP: genetic amplification profile; GSE: genomic repository series code; OS: overall survival; N: number; NA: not available. distinct embryonic tissue origins [13, 14]. However, these classification models are difficult to implement in routine diagnostics and/or remain of relatively limited prognostic value [34–36]. MDM2, MDM4, PDGFRA and CDK4 genes encoded in chromosomes 12q15, 1q32.1, 4q12 and 12q14.1 [11, 14, 38], respectively. Of note, gene amplification is a common genetic alteration across different malignancies and it usually involves (wild-type or mutated) genes that show oncogenic potential (i.e are capable of reproducing the tumor); thereby, it might confer a malignant phenotype associated with a variable outcome, depending on the specific genes amplified and/or overexpressed. Despite this, no study has been reported so far in which the impact of different GAP on OS has been investigated in a large series of GBM patients. Among other cytogenetic/molecular alterations, gene amplification, particularly gene amplification involving the EGFR gene, represents one of the most common genetic changes in GBM [11, 20-22, 37]. Thus, EGFR amplification at chromosome 7p11.2 can be found in between one third and half of all GBM patients as the only chromosomal region amplified – including a variable number of amplified genes – or it can be find in the same tumor in combination with amplification of genes located at other chromosomal regions, such as the Here we investigated the GAP across the whole GBM tumor genome and analyzed their impact on www.oncotarget.com www.oncotarget.com Oncotarget 28091 patient OS, based on one of the largest series of GBM patients reported so far in the literature. Prognostic impact of gene amplification profiles in GBM SNP-arrays were used for both sensitive identification of CNA involving specific DNA sequences across the whole tumor genome and detailed delineation of the amplified genes; in order to avoid CNV due to germinal single nucleotide polymorphisms, insertions and deletions potentially associated with an increased predisposition to GBM fic DNA sequences across the whole tumor genome associated with an increased predisposition to GB re 2: Distribution of the most frequently altered chromosomes (A) and chromosomal regions (B) showing isolated or multi fications both in cases from series 1 (n=45/80) and in patients (111/267) from the seven distinct series previously reported in ure by others. Results are shown as percentage values from all cases analyzed. Figure 2: Distribution of the most frequently altered chromosomes (A) and chromosomal regions (B) showing isolated or multiple amplifications both in cases from series 1 (n=45/80) and in patients (111/267) from the seven distinct series previously reported in the literature by others. Results are shown as percentage values from all cases analyzed. Figure 2: Distribution of the most frequently altered chromosomes (A) and chromosomal regions (B) showing isolated or multiple amplifications both in cases from series 1 (n=45/80) and in patients (111/267) from the seven distinct series previously reported in the literature by others. Results are shown as percentage values from all cases analyzed. www.oncotarget.com Oncotarget Oncotarget 28092 Figure 3: Prognostic impact on overall patient survival of distinct gene amplification profiles of GBM patients from the test cohort – series Figure 3: Prognostic impact on overall patient survival of distinct gene amplification profiles of GBM patients from the test cohort – series 1 alone (n=76); panels (A) and (B) – and after pooling our cases with 181 additional cases from seven series of GBM patients previously reported in the literature by other groups (panels (C) and (D); n= 257 GBM). In panels (A) and (C), GBM patients were grouped as having i) no gene amplification; ii) isolated EGFR amplification; iii) gene amplification at multiple chromosomal regions including EGFR gene amplification; iv) isolated amplification of DNA sequences from a single chromosomal region other than 7p11.2 (i.e not including amplification of the EGFR gene), and; v) gene amplification at multiple chromosomal regions in the absence of EGFR amplification. Prognostic impact of gene amplification profiles in GBM In panels (B) and (D), patients were grouped as: i) cases showing either no gene amplification or having EGFR gene amplification and ii) patients showing genetic amplification at one or more chromosomal regions which did not involve the EGFR gene. Median overall survival is expressed in months and was calculated for 257 patients from series 1-8. Cases who were alive but had a follow-up of less than 18 months and/or died within the 1st month after surgery (n= 16) were excluded from OS analyses. Figure 3: Prognostic impact on overall patient survival of distinct gene amplification profiles of GBM patients from the test cohort – series 1 alone (n=76); panels (A) and (B) – and after pooling our cases with 181 additional cases from seven series of GBM patients previously reported in the literature by other groups (panels (C) and (D); n= 257 GBM). In panels (A) and (C), GBM patients were grouped as having i) no gene amplification; ii) isolated EGFR amplification; iii) gene amplification at multiple chromosomal regions including EGFR gene amplification; iv) isolated amplification of DNA sequences from a single chromosomal region other than 7p11.2 (i.e not including amplification of the EGFR gene), and; v) gene amplification at multiple chromosomal regions in the absence of EGFR amplification. In panels (B) and (D), patients were grouped as: i) cases showing either no gene amplification or having EGFR gene amplification and ii) patients showing genetic amplification at one or more chromosomal regions which did not involve the EGFR gene. Median overall survival is expressed in months and was calculated for 257 patients from series 1-8. Cases who were alive but had a follow-up of less than 18 months and/or died within the 1st month after surgery (n= 16) were excluded from OS analyses. Figure 3: Prognostic impact on overall patient survival of distinct gene amplification profiles of GBM patients from the test cohort – series 1 alone (n=76); panels (A) and (B) – and after pooling our cases with 181 additional cases from seven series of GBM patients previously reported in the literature by other groups (panels (C) and (D); n= 257 GBM). Prognostic impact of gene amplification profiles in GBM In panels (A) and (C), GBM patients were grouped as having i) no gene amplification; ii) isolated EGFR amplification; iii) gene amplification at multiple chromosomal regions including EGFR gene amplification; iv) isolated amplification of DNA sequences from a single chromosomal region other than 7p11.2 (i.e not including amplification of the EGFR gene), and; v) gene amplification at multiple chromosomal regions in the absence of EGFR amplification. In panels (B) and (D), patients were grouped as: i) cases showing either no gene amplification or having EGFR gene amplification and ii) patients showing genetic amplification at one or more chromosomal regions which did not involve the EGFR gene. Median overall survival is expressed in months and was calculated for 257 patients from series 1-8. Cases who were alive but had a follow-up of less than 18 months and/or died within the 1st month after surgery (n= 16) were excluded from OS analyses. www.oncotarget.com Oncotarget 28093 acteristics of the GBM patients included in the series 1 (n=76) and their association to disease Table 5: Clinical characteristics of the GBM patients included in the series 1 (n=76) and their association to disease outcome outcome Clinical/genetic characteristics Patient distribution Overall survival Univariate analysis Multivariate analysis Median (range) p-value Hazard Ratio (95% CI) p-value Age >30 2 (2%) 67 (2.7-67) <.001 31-45 9 (12%) 15 (8-21) 46-60 21 (28%) 14 (2-63) 61-75 34 (45%) 13 (2-83) >75 10 (13%) 6 (2-11) Karnofsky Index # >70 48 (64%) 15 (2-83) .004 ≤70 27 (36%) 10 (2-30) Type of Surgery Complete resection 24 (32%) 15 (2-83) .012 Partial resection 41 (54%) 13 (2-67) No resection 11 (14%) 6 (2-21) Adjuvant chemotherapy# Stupp 43 (67%) 18 (2-83) <.001 3 (2-4) <.001 BCNU 10 (16%) 13 (5-67) No chemotherapy 11 (17%) 6 (2-11) Gene amplification profile No gene AMP 34 (45%) 13 (2-67) <.001 Isolated EGFR AMP 16 (21%) 17 (9-83) Multiple AMP including EGFR 12 (16%) 13 (2-29) Isolated non-EGFR AMP 9 (12%) 6 (2-15) Multiple AMP without EGFR 5 (6%) 8 (2-9) No gene AMP+EGFR AMP 62 (82%) 14 (2-83) <.001 6 (2-12) <.001 AMP without EGFR 14 (18%) 6 (2-15) CI: confidence interval; #: Data of Karnofsky index and adjuvant chemotherapy were only available in only 75 and 65 GBM patients respectively; Stupp: radiotherapy plus temozolamide; BCNU: carmustine; AMP: genetic amplification. Prognostic impact of gene amplification profiles in GBM CI: confidence interval; #: Data of Karnofsky index and adjuvant chemotherapy were only available in only 75 and 65 GBM patients respectively; Stupp: radiotherapy plus temozolamide; BCNU: carmustine; AMP: genetic amplification. CI: confidence interval; #: Data of Karnofsky index and adjuvant chemotherapy were only available patients respectively; Stupp: radiotherapy plus temozolamide; BCNU: carmustine; AMP: genetic am (e.g. the rs1801320 SNP in the RAD51 DNA repair gene [39]), paired tumor and peripheral blood (PB) samples were analyzed per patient. by genetic amplification at the 12q, 4q12 and 1q32.1 chromosomal regions, and to a lower extent also, at 1q, 5q, 6q, 7q and at regions in both arms of chromosomes 11 or 17. These results confirm and extend on previous observations by our [20–22] and other groups [40, 41] which indicate that EGFR is the most frequently amplified oncogene in GBM, where it is detected in up to 40% of primary GBM tumors in association Overall, our results showed the presence of gene amplification in the majority (>50%) of tumors investigated. As expected, genetic amplification involving DNA sequences at the 7p11.2 chromosomal region was by far the most frequent alteration, followed www.oncotarget.com Oncotarget 28094 with a better outcome, compared to cases that show either no gene amplification or amplification of genes other than EGFR [20, 37]. At present, it is well- established that activation of the EGFR gene via gene amplification and/or mutations, up-regulates the RAS/ RAF/MAPK and PI3K signaling pathways, translating into a tumor phenotype consisting of: i) abnormally high cell proliferation and ii) survival of tumor cells, and iii) an increased angiogenesis [11, 42]. Of note, here we confirm via mapping of the amplified region in chromosome 7p11.2, that this alteration frequently includes also the LANCL2 gene, in addition to other genes adjacent to the EGFR and LANCL2 genes [20, 43]. Although the LANCL2 gene codes for a protein involved in up-regulation of AKT and cell survival, and an increased cell sensitivity to adriamycin [44, 45], its precise mechanism of action remains elusive. tirosine-protein kinase cell surface receptor of the PDGF growth factor, which promotes cell proliferation and migration. Finally, amplification of DNA sequences at chromosome 1q32.1 involved the MDM4 gene together with the SOX13, ETNK2, KISS1, GOLT1A, PLEKHA6, REN and PIK3C2B genes. Prognostic impact of gene amplification profiles in GBM MDM4 inhibits TP53 and TP73, mediating cell cycle arrest via binding to their transcriptional activation domain, at the same time it inhibits degradation of MDM2 [48], whereas PIK3C2B belongs to the PI3K gene family and activates signaling for cell proliferation, survival and migration [49]. In turn, the DTX3 ubiquitin ligase gene probably acts both as a positive and negative regulator of Notch, depending on the developmental stage and cell context [50]. Altogether, these findings indicate that the distinct GAP here reported for GBM tumors might confer a distinct biological and (also) clinical behavior to these tumors. Thus, based on the presence vs absence of gene amplification and its subtypes, five different patterns were defined among our GBM patients, which showed an association with OS. From the prognostic point of view, these five GAP could be further re-grouped into two major risk-groups including: i) patients with either no gene amplification or EGFR gene amplification associated with or without amplification of genes coded at chromosomal regions other than chromosome 7p11.2, with a significantly longer OS; and, ii) cases presenting with amplification of one or more chromosomal regions that did not include EGFR gene amplification, and that were associated with a significantly poorer outcome. The prognostic impact of this later classification was further confirmed in a larger cohort of GBM patients previously reported in the literature and, together with the type of treatment administered, emerged as the most powerful combination of independent prognostic factor for GBM patients. However, the precise mechanisms involved in determining the distinct survival rates of these two molecular groups of patients, still remain to be elucidated and deserve further investigations. In line with previous observations [11, 38, 46], EGFR gene amplification at the 7p11.2 chromosomal region, was found either as the only amplified DNA sequence, or in association with amplification of DNA sequences at other chromosomal regions and genes such as the PDGFRA, MDM2, MDM4 and CDK4 genes. Such combined pattern of amplification of multiple genes at distinct chromosomal regions might lead to unique malignant transformation profiles for which the underlying mechanisms are still poorly understood; however, in our series, it did not prove to confer a more adverse outcome vs isolated amp 7p11.2. In contrast, isolated amplification of DNA sequences at chromosomal regions other than that containing the EGFR gene, were associated with a significantly shorter OS of GBM patients. Prognostic impact of gene amplification profiles in GBM Of note, these later alterations most frequently affected genes encoded in the 12q, 4q and 1q chromosomal regions, and they typically involved multiple genes [15]. Thus, amplification of DNA sequences at the 12q13-14 chromosomal region usually included the CDK4 gene together with the METTL1, CYP27B1, AVIL, CTDSP2, METT21B, AGAP2 and OS9 genes, while genetic amplification at 12q15 affected the MDM2 oncogene in a significant fraction of all (primary) GBM tumors analyzed. CDK4 is a member of the Ser/Thr protein kinase family, required for the cell cycle transition from the G1 to S-phase [47]; thus, CDK4 phosphorylates the Rb gene product leading to its inactivation and the release of proteins required for cell cycle progression, at the same time it also down-regulates TP53 [47]. In turn, AVIL binds actin and promotes the development of neuronal cells, while the MDM2 gene codes for a nuclear-localized E3- ubiquitin ligase whose transcription is recognized as the main p53 negative regulator [48]. In our series, genetic amplification at the 4q11-12 chromosomal region, systematically affected the PDGFRA gene together with the SCFD2 and FIP1L1 genes. PDFGRA [11] codes for a Despite several classification models have been previously proposed which address the genomic heterogeneity of GBM [13, 14, 27] and identify tumors with different cellular origins, so far they have proven to be of limited prognostic value [34-36, 51] and/or difficult to be used in routine diagnostics due to the complexity of the information they require to classify GBM patients at diagnosis. In contrast, here we propose a relatively simple prognostic stratification model for GBM tumors based on their gene amplification profiles that might be easily implemented in routine laboratory diagnostics, and that will potentially contribute to a better management of the patients. www.oncotarget.com www.oncotarget.com Patients and samples Overall, 347 GBM tumors were studied. These included two groups of adult patients: the first group consisted of 80 caucasian GBM patients (group 1, series 1) with histological diagnosis of primary GBM based on the WHO criteria (38 males and 42 females; mean age of 62±13 years, ranging from 24 to 84 years) (Table 6). Fifty-seven of these 80 patients (71%) were admitted to the University Hospital of Coimbra (Coimbra, Portugal) and their genomic data has been deposited in the genomic repository GEO (series code number: GSE42631) and 23 (29%) were from the University Hospital of Salamanca (Spain). Each patient from series 1 gave his/ her informed consent prior to entering the study, and the study was approved by the local Ethics Committees of both institutions, according to the Declaration of Helsinki. For each patient within this first group, tumor samples containing representative areas of (fresh) tumor tissues were obtained by surgical resection, immediately (<30 min) snap-frozen in liquid nitrogen, and stored at -80ºC for further SNP-array studies; in parallel, a PB sample was also collected from each patient. Prior to the SNP-array studies, a section was cut from the stored tissue blocks and assessed by conventional histopathological procedures for its tumoral cell contents. Specimens with ≥75% tumoral Prognostic impact of gene amplification profiles in GBM In line with this, we have recently patented an array containing this combination of probes [52] to assess the above referred www.oncotarget.com Oncotarget 28095 Table 6: Clinical and biological characteristics of GBM patients from series 1 who were analyzed by single- nucleotide polymorphism arrays in this study (n=80) Oncotarg 28096 www.oncotarget.com Table 6: Clinical and biological characteristics of GBM patients from series 1 who were analyzed by single- nucleotide polymorphism arrays in this study (n=80) Case ID Age Gender Karnofsky Index (%) Location Brain hemisphere Overall survival or follow-up Exitus Treatment Type of surgery Chemotherap GBM1 80 M ND Temporal L 6 Yes T ND GBM2 75 M 80 Fronto- temporal R 22 Yes T Stupp GBM3 61 F 90 Parietal R 83 Yes T Stupp GBM4 73 M 100 Temporal L 19 Yes P Stupp GBM5 38 M 90 Frontal R 15 Yes T Stupp GBM6 49 F 60 Frontal R 30* No P Stupp GBM7 41 M 40 Temporal R 11 Yes P Stupp GBM8 57 M 90 Tempo- parietal R 6 Yes T Stupp GBM9 72 M 90 Temporal R 28 No P Stupp GBM10 62 M 100 Parietal R 28 No T Stupp GBM11 71 M 80 Temporal R 27 Yes P Stupp GBM12 50 F 100 Temporal R 13 Yes T Stupp GBM13 72 F 70 Temporal R 9 Yes P Stupp GBM14 78 F 100 Frontal R 6 Yes T - GBM15 61 F 100 Frontal R 25 No T Stupp GBM16 54 M 100 Fronto- parietal R 2 Yes T Stupp GBM17 52 M 80 Frontal R 63 Yes P Stupp GBM18 57 F 90 Temporal R 10 Yes P Stupp GBM19 68 M 90 Occipital L 10 Yes T Stupp GBM20 82 F 80 Frontal L 7 Yes T - GBM21 77 M 70 Temporal R 6 Yes P - GBM22 69 F 100 Frontal L 8 Yes T Stupp GBM23 24 F 80 Frontal L 21 Yes P Stupp G97 53 M 80 Temporal R 21 Yes T Stupp G94 79 F 80 Temporal R 9 Yes P - G93 63 M 80 Occipital R 29 Yes T Stupp G92 54 F 80 Parietal R 15 Yes T Stupp G91 73 F 60 Occipital R 13 Yes P Stupp G90 57 F 60 Parietal L 5 Yes B - G89 51 M 80 Temporal R 2 Yes P - G88 71 M 80 Parietal R 8 Yes P Stupp G87 45 M 80 Temporal L 16 Yes P Stupp/ Sequential (Continued www.oncotarget.com Oncotarget 28096 Case ID Age Gender Karnofsky Index (%) Location Brain hemisphere Overall survival or follow-up Exitus Treatment Type of surgery Chemotherapy G83 75 M 70 Temporal R 10 Yes P - G82 78 M 70 Frontal R 2 Yes B - G81 62 F 70 Frontal R 13 Yes P Stupp G80 43 M 80 Frontal R 18 Yes T Stupp G79 71 F 60 Occipital R 6 Yes B - G73 78 F 60 Parietal L 4 Yes B - G72 77 F 70 Temporal L 1 Yes P - G71 66 F 60 Parietal R 10 Yes P Sequential G70 56 F 80 Occipital L 21 Yes P Stupp G68 72 M 70 Insular L 26 Yes T Stupp G67 68 F 80 Parietal R 35 Yes P Stupp G66 60 M 80 Occipital R 14 Yes T Stupp G65 69 F 60 Parietal L 1 Yes P - G64 57 M 60 Occipital L 8 Yes P Sequential G63 61 F 60 Insular R 13 Yes P Sequential G62 57 F 90 Occipital R 18 Yes T Stupp G57 34 M 90 Frontal R 8 Yes T Stupp G56 65 M 80 Frontal L 13 Yes P Stupp G55 54 F 80 Frontal R 17 Yes P Stupp G54 65 F 60 Parietal L 6 Yes P - G53 74 M 60 Frontal L 29 Yes T Stupp G52 56 M 90 Frontal L 21 Yes B Stupp G51 60 M 60 Temporal R 2 Yes B - G50 84 M 70 Temporal R 11 Yes P - G46 62 M 60 Frontal L 3 Yes P - G45 76 F 60 Temporal R 10 Yes P - G44 48 M 80 Frontal L 22 Yes P PCV G43 67 F 70 Temporal R 7 Yes P - G42 67 M 80 Temporal R 2 Yes P - G41 44 F 60 Frontal R 14 Yes B Sequential G40 45 F 80 Frontal R 15 Yes P BCNU+TMZ G39 70 F 50 Frontal R 18 Yes P Stupp G37 70 M 80 Temporal R 32 Yes T Stupp G35 50 F 80 Frontal L 2 Yes P - G34 69 M 60 Temporal R 5 Yes B - G31 71 F 90 Frontal R 7 Yes P - (Continued ) www.oncotarget.com Oncotarget 28097 Case ID Age Gender Karnofsky Index (%) Location Brain hemisphere Overall survival or follow-up Exitus Treatment Type of surgery Chemotherapy G30 71 F 70 Temporal R 9 Yes B - G29 49 F 80 Parietal L 12 Yes B Sequential G25 68 M 80 Frontal L 6 Yes P Stupp G23 50 F 70 Frontal R 14 Yes B Stupp G17 30 F 90 Temporal R 67 Yes P Sequential G15 79 M 80 Parietal L 5 Yes T Sequential G14 69 F 70 Frontal R 0 Yes B - G13 39 F 90 Frontal R 20 Yes P Sequential G12 74 M 70 Temporal R 1 Yes B - G10 35 F 80 Temporal L 15 Yes P Stupp G8 67 F 90 Deep NA 9 Yes P Stupp G6 70 F 80 Temporal R 19 Yes P Stupp GAP, and that we hope can be commercially available for routine diagnostics soon. MATERIALS AND METHODS The second group of GBM patients included 267 unselected cases from 7 different series of GBM (series 2 to 8; group 2) previously reported in the literature [15, 24, 28–32]. Data from cases included in one of these series (series 4; n=15 cases) were kindly provided by Bralten et al [32], while data about the patients and tumor samples from the other six series was accessed from publicly available data bases – GSE19612 (series 2; 24 cases); GSE19399 and GSE9635 (series 3; 104 and 16 cases, respectively); GSE14804 (series 5; 12 cases); E-MEXP1330 (series 6, 53 cases); GSE10922 (series 7; 13 cases) and GSE13021 (series 8; 30 cases)–. Prognostic impact of gene amplification profiles in GBM cells, in the absence of significant contamination by normal brain parenchyma and tumoral necrosis, were selected for further DNA and RNA extraction. Statistical analysis The statistical significance of differences observed between groups was assessed by the Student T and the Mann-Whitney U tests, for parametric and non- parametric (continuous) variables, respectively; for categorical variables, the X2 test was used. Overall, 257 GBM who survived for >1 month after surgery and had a minimum follow-up of 18 months (for patients remaining alive) were included in OS analyses. Survival curves were plotted according to the method of Kaplan and Meier, and the (two-sided) log-rank test was used to assess the statistical significance of differences in OS among distinct groups of patients. Multivariate analysis of prognostic factors for OS was performed using the Cox stepwise regression model. In this part of the study, only those variables showing a significant association with RFS in the univariate analysis were included (Table 5). P-values <0.05 were considered to be associated with statistical significance. For all statistical analysis, the SPSS software (SPSS 17.0, IBM SPSS, Armonk, NY, USA), was used. p In the test series (series 1), 4 different types of SNP-arrays were used. These included: i) the GeneChip Human Mapping 500K Array Set (n=35 tumors), which provides information about >500,000 SNPs according to the NCBI/hg19 assembly (262,264 SNPs in the Nsp array and 238,304 SNPs in the Sty array); ii) the Genome-Wide Human SNP Array 6.0 (n=22 cases), which contains probes for 906,600 SNPs and 945,826 non-polymorphic probes featuring a total of >1.8 million probes (Affymetrix); and, iii) the CytoScan 750K and Cytoscan HD arrays (n=23 tumors) which contain probes for 200,436 SNPs and 743,304 non-polymorphic probes (Affymetrix). Data about DNA probes was analyzed with the Console Genotyping software (version 3.0.2; Affymetrix). In addition, the dChip 2010 software (http// www.dchip.org; Dana Farber Cancer Institute, Boston, MA, USA) was used to calculate CNA values. To plot CNAs according to their chromosomal location, the Chromosome Analysis Suite (CHAS) was used. The Hg 19 human genome sequence was used as reference to name the amplified genes, as defined by CNAs values > 4.8 (arbitrary units) typically corresponding to > 7 DNA copies (in order to exclude polyploidies) [29]. Mutational analyses following the standard operating procedures (SOPs) of the Spanish DNA Bank Carlos III (University of Salamanca, Salamanca, Spain) (http://www.bancoadn.org). Analysis of IDH1 and IDH2 gene mutations was based on DNA extracted from formalin-fixed and paraffin- embebed tissues (n=54) using the QIAamp DNA Mini kit (Qiagen, Germany) according to the instructions of the manufacture. Exon 4 DNA of both the IDH1 and the IDH2 genes was amplified by PCR and sequenced on a capillary automated sequencer (CEQ 8000; Beckman-Coulter, Hialeah, FL, USA); mutational analysis of the sequence data was performed using the Sequencher, (version 4.7) software (Genes Codes, Ann Arbor, MI, USA). None of the 54 primary GBM cases analyzed showed IDH1 or IDH2 mutations. For the investigation of CNAs by SNP-arrays, DNA from frozen tumor tissues and their paired PB samples was used in order to exclude individual CNV due to germline SNPs, small insertions and deletions. Briefly, extracted DNA (250ng per array) was digested with restriction enzymes and ligated to the corresponding adaptors, following conventional Affymetrix procedures (Affymetrix Inc, Thermo-Fisher Scientific, Waltham, MA, USA). A generic primer that recognizes the adaptor sequence was used in triplicate, to amplify adaptor- ligated DNA fragments via polymerase chain reaction (PCR). The amplified DNA was then fragmented, labeled, and hybridized to the corresponding SNP- array (please see below). After hybridization, chips were washed in an Affymetrix Fluidics Station 450 (Affymetrix) and the hybridized sequences were labeled using streptavidin-phycoerythrin, and assayed by fluorescence detection using a GeneChip Scanner 3000 (Affymetrix). The allelotype at a locus was then determined based on probe-associated fluorescence intensity data for oligonucleotides complementary to the reference sequences that covered the corresponding SNP position. DNA extraction and identification of copy number alterations by SNP-arrays DNA from frozen tumor samples (n=80 tumors from series 1) was purified using the QIAamp DNA Mini Kit (Qiagen, Valencia, CA, USA), according to the instructions of the manufacturer. The yield and purity of the extracted DNA were determined using a NanoDrop-1000 spectrophotometer (Nano-Drop Technologies Inc, Wilmington, DE, USA), and they systematically showed absorbance (A) values >1.5 at 260/230nm and ≥1.8 at 260/280nm wavelengths, respectively. DNA integrity was evaluated by conventional electrophoretic procedures in a 1% agarose gel. RNA extraction and cDNA synthesis were performed www.oncotarget.com www.oncotarget.com Oncotarget 28098 REFERENCES 1. Louis DN, Perry A, Reifenberger G, von Deimling A, Figarella-Branger D, Cavenee WK, Ohgaki H, Wiestler OD, Kleihues P, Ellison DW. The 2016 World Health Organization Classification of Tumors of the Central Nervous System: a summary. Acta Neuropathol. 2016; 131:803-20. 14. Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, Miller CR, Ding L, Golub T, Mesirov JP, Alexe G, Lawrence M, O'Kelly M, et al. Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell. 2010; 17:98-110. 2. Charles NA, Holland EC, Gilbertson R, Glass R, Kettenmann H. The brain tumor microenvironment. Glia. 2012; 60:502-14. 3. Sottoriva A, Spiteri I, Piccirillo SG, Touloumis A, Collins VP, Marioni JC, Curtis C, Watts C, Tavare S. Intratumor heterogeneity in human glioblastoma reflects cancer evolutionary dynamics. Proc Natl Acad Sci U S A. 2013; 110:4009-14. 15. Yin D, Ogawa S, Kawamata N, Tunici P, Finocchiaro G, Eoli M, Ruckert C, Huynh T, Liu G, Kato M, Sanada M, Jauch A, Dugas M, et al. High-resolution genomic copy number profiling of glioblastoma multiforme by single nucleotide polymorphism DNA microarray. Mol Cancer Res. 2009; 7:665-77. 4. Lee CY. Strategies of temozolomide in future glioblastoma treatment. Onco Targets Ther. 2017; 10:265-70. 16. Brennan CW, Verhaak RG, McKenna A, Campos B, Noushmehr H, Salama SR, Zheng S, Chakravarty D, Sanborn JZ, Berman SH, Beroukhim R, Bernard B, Wu CJ, et al. The somatic genomic landscape of glioblastoma. Cell. 2013; 155:462-77. 5. Van Meir EG, Hadjipanayis CG, Norden AD, Shu HK, Wen PY, Olson JJ. Exciting new advances in neuro-oncology: the avenue to a cure for malignant glioma. CA Cancer J Clin. 2010; 60:166-93. 17. Ichimura K, Pearson DM, Kocialkowski S, Backlund LM, Chan R, Jones DT, Collins VP. IDH1 mutations are present in the majority of common adult gliomas but rare in primary glioblastomas. Neuro Oncol. 2009; 11:341-7. 6. Krakstad C, Chekenya M. Survival signalling and apoptosis resistance in glioblastomas: opportunities for targeted therapeutics. Mol Cancer. 2010; 9:135. 7. Gravendeel LA, Kouwenhoven MC, Gevaert O, de Rooi JJ, Stubbs AP, Duijm JE, Daemen A, Bleeker FE, Bralten LB, Kloosterhof NK, De Moor B, Eilers PH, van der Spek PJ, et al. Intrinsic gene expression profiles of gliomas are a better predictor of survival than histology. Cancer Res. 2009; 69:9065-72. 18. Hatanpaa KJ, Burma S, Zhao D, Habib AA. FUNDING 11. Szerlip NJ, Pedraza A, Chakravarty D, Azim M, McGuire J, Fang Y, Ozawa T, Holland EC, Huse JT, Jhanwar S, Leversha MA, Mikkelsen T, Brennan CW. Intratumoral heterogeneity of receptor tyrosine kinases EGFR and PDGFRA amplification in glioblastoma defines subpopulations with distinct growth factor response. Proc Natl Acad Sci U S A. 2012; 109:3041-6 This work was supported by RETICC RD06/0020/0035, RD06/0020/0059 and RD12/0036/0048 grants from Red Temática de Investigación Cooperativa en Cáncer (RTICC), Instituto de Salud Carlos III, Ministerio de Economía y Competitividad, (Madrid, Spain and FONDOS FEDER), AES PI16/000476 (Instituto de Salud Carlos III, Madrid, Spain and FONDOS FEDER), GRS909A14 (JCYL) and CB16/12/00400 grant (CIBER- ONC, Instituto de Salud Carlos III, Ministerio de Economía y Competitividad, Madrid, Spain and FONDOS FEDER). 12. The Cancer Genome Atlas (TCGA) Research Network. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature. 2008; 455:1061-8. 13. Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD, Misra A, Nigro JM, Colman H, Soroceanu L, Williams PM, Modrusan Z, Feuerstein BG, et al. Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell. 2006; 9:157-73. Author contributions Specific contributions of each author are: Álvaro Otero, Pablo Sousa, Hermínio Tão, Olinda Rebelo, Marcos Barbosa and Pim J French have provided samples and clinical or histopathological data; María González-Tablas, Inês Crespo, Ana Luísa Vital and María Dolores Tabernero have processed the samples; Luis Antonio Corchete, Ana Belen Nieto and María Carmen Patino-Alonso have analyzed SNP-data; María Dolores Tabernero, Alberto Orfao, Maria Celeste Lopes and María González-Tablas have design the study, drafted and writing the manuscript. Maria Rosario Almeida and Ana Filipa Guedes analyzed the IDH1 and IDH2 mutational status. All authors have contributed significantly to this article and all of them have read and approved the manuscript. For the validation series, data on 7 different SNP- array chips (Affymetrix 50K, 100K, 250K and/or 500K SNP-arrays) previously reported by others [15, 23, 24, 28–32] were used for the analysis of CNA. The number of common SNPs for the 100K and 500k arrays, and for the 500K and cytoscan HD arrays, was of 21,144 and 65,535 SNPs, respectively, for a total of around 11,000 SNPs in common to all SNPs-arrays used. Oncotarget www.oncotarget.com 28099 CONFLICTS OF INTEREST mechanisms and markers. Acta Neuropathol. 2015; 129:829-48. https://doi.org/10.1007/s00401-015-1432-1. The authors declared that they have no conflicts of interest. 10. Virk SM, Gibson RM, Quinones-Mateu ME, Barnholtz- Sloan JS. Identification of variants in primary and recurrent glioblastoma using a cancer-specific gene panel and whole exome sequencing. PLoS One. 2015; 10:e0124178. REFERENCES Epidermal growth factor receptor in glioma: signal transduction, neuropathology, imaging, and radioresistance. Neoplasia. 2010; 12:675-84. 19. Rong Y, Belozerov VE, Tucker-Burden C, Chen G, Durden DL, Olson JJ, Van Meir EG, Mackman N, Brat DJ. Epidermal growth factor receptor and PTEN modulate tissue factor expression in glioblastoma through JunD/ activator protein-1 transcriptional activity. Cancer Res. 2009; 69:2540-9. 8. Cloughesy TF, Mischel PS. New strategies in the molecular targeting of glioblastoma: how do you hit a moving target? Clin Cancer Res. 2011; 17:6-11. 9. Aldape K, Zadeh G, Mansouri S, Reifenberger G, von Deimling A. Glioblastoma: pathology, molecular www.oncotarget.com Oncotarget 28100 glioblastomas revealed by genomewide analysis of laser- microdissected tissues. Neuro Oncol. 2008; 10:995-1003. 20. Crespo I, Tao H, Nieto AB, Rebelo O, Domingues P, Vital AL, Patino Mdel C, Barbosa M, Lopes MC, Oliveira CR, Orfao A, Tabernero MD. Amplified and homozygously deleted genes in glioblastoma: impact on gene expression levels. PLoS One. 2012; 7:e46088. 31. Solomon DA, Kim JS, Cronin JC, Sibenaller Z, Ryken T, Rosenberg SA, Ressom H, Jean W, Bigner D, Yan H, Samuels Y, Waldman T. Mutational inactivation of PTPRD in glioblastoma multiforme and malignant melanoma. Cancer Res. 2008; 68:10300-6. 21. Crespo I, Vital AL, Nieto AB, Rebelo O, Tao H, Lopes MC, Oliveira CR, French PJ, Orfao A, Tabernero MD. Detailed characterization of alterations of chromosomes 7, 9, and 10 in glioblastomas as assessed by single-nucleotide polymorphism arrays. J Mol Diagn. 2011; 13:634-47. 32. Bralten LB, Kloosterhof NK, Gravendeel LA, Sacchetti A, Duijm EJ, Kros JM, van den Bent MJ, Hoogenraad CC, Sillevis Smitt PA, French PJ. Integrated genomic profiling identifies candidate genes implicated in glioma-genesis and a novel LEO1-SLC12A1 fusion gene. Genes Chromosomes Cancer. 2010; 49:509-17. 22. Vital AL, Tabernero MD, Crespo I, Rebelo O, Tao H, Gomes F, Lopes MC, Orfao A. Intratumoral patterns of clonal evolution in gliomas. Neurogenetics. 2010; 11:227-39. 33. Zarrei M, MacDonald JR, Merico D, Scherer SW. A copy number variation map of the human genome. Nat Rev Genet. 2015; 16:172-83. 23. Beroukhim R, Mermel CH, Porter D, Wei G, Raychaudhuri S, Donovan J, Barretina J, Boehm JS, Dobson J, Urashima M, Mc Henry KT, Pinchback RM, Ligon AH, et al. The landscape of somatic copy-number alteration across human cancers. Nature. 2010; 463:899-905. 34. Lin N, Yan W, Gao K, Wang Y, Zhang J, You Y. REFERENCES Prevalence and clinicopathologic characteristics of the molecular subtypes inmalignant glioma: a multi-institutional analysis of 941 cases. PLoS One. 2014; 9:e94871. 24. Hodgson JG, Yeh RF, Ray A, Wang NJ, Smirnov I, Yu M, Hariono S, Silber J, Feiler HS, Gray JW, Spellman PT, Vandenberg SR, Berger MS, et al. Comparative analyses of gene copy number and mRNA expression in glioblastoma multiforme tumors and xenografts. Neuro Oncol. 2009; 11:477-87. 35. Yan W, Zhang W, You G, Zhang J, Han L, Bao Z, Wang Y, Liu Y, Jiang C, Kang C, You Y, Jiang T. Molecular classification of gliomas based on whole genome gene expression: a systematic report of 225 samples from the Chinese Glioma Cooperative Group. Neuro Oncol. 2012; 14:1432-40. 25. Reifenberger G, Ichimura K, Reifenberger J, Elkahloun AG, Meltzer PS, Collins VP. Refined mapping of 12q13-q15 amplicons in human malignant gliomas suggests CDK4/ SAS and MDM2 as independent amplification targets. Cancer Res. 1996; 56:5141-5. 36. Crisman TJ, Zelaya I, Laks DR, Zhao Y, Kawaguchi R, Gao F, Kornblum HI, Coppola G. Identification of an Efficient Gene Expression Panel for Glioblastoma Classification. PLoS One. 2016; 11:e0164649. 26. Riemenschneider MJ, Knobbe CB, Reifenberger G. Refined mapping of 1q32 amplicons in malignant gliomas confirms MDM4 as the main amplification target. Int J Cancer. 2003; 104:752-7. 37. Hobbs J, Nikiforova MN, Fardo DW, Bortoluzzi S, Cieply K, Hamilton RL, Horbinski C. Paradoxical relationship between the degree of EGFR amplification and outcome in glioblastomas. Am J Surg Pathol. 2012; 36:1186-93. 27. Nutt CL, Mani DR, Betensky RA, Tamayo P, Cairncross JG, Ladd C, Pohl U, Hartmann C, McLaughlin ME, Batchelor TT, Black PM, von Deimling A, Pomeroy SL, et al. Gene expression-based classification of malignant gliomas correlates better with survival than histological classification. Cancer Res. 2003; 63:1602-7. 38. Furgason JM, Koncar RF, Michelhaugh SK, Sarkar FH, Mittal S, Sloan AE, Barnholtz-Sloan JS, Bahassi el M. Whole genome sequence analysis links chromothripsis to EGFR, MDM2, MDM4, and CDK4 amplification in glioblastoma. Oncoscience. 2015; 2:618-28. https://doi. org/10.18632/oncoscience.178. 28. Chen R, Nishimura MC, Bumbaca SM, Kharbanda S, Forrest WF, Kasman IM, Greve JM, Soriano RH, Gilmour LL, Rivers CS, Modrusan Z, Nacu S, Guerrero S, et al. A hierarchy of self-renewing tumor-initiating cell types in glioblastoma. Cancer Cell. 2010; 17:362-75. 39. Franceschi S, Tomei S, Mazzanti CM, Lessi F, Aretini P, La Ferla M, De Gregorio V, Pasqualetti F, Zavaglia K, Bevilacqua G, Naccarato AG. REFERENCES Association between RAD 51 rs1801320 and susceptibility to glioblastoma. J Neurooncol. 2016; 126:265-70. 29. Beroukhim R, Getz G, Nghiemphu L, Barretina J, Hsueh T, Linhart D, Vivanco I, Lee JC, Huang JH, Alexander S, Du J, Kau T, Thomas RK, et al. Assessing the significance of chromosomal aberrations in cancer: methodology and application to glioma. Proc Natl Acad Sci U S A. 2007; 104:20007-12. 40. Jeuken J, Sijben A, Alenda C, Rijntjes J, Dekkers M, Boots- Sprenger S, McLendon R, Wesseling P. Robust detection of EGFR copy number changes and EGFR variant III: technical aspects and relevance for glioma diagnostics. Brain Pathol. 2009; 19:661-71. 41. Srividya MR, Thota B, Arivazhagan A, Thennarasu K, Balasubramaniam A, Chandramouli BA, Hegde AS, Santosh V. Age-dependent prognostic effects of EGFR/p53 alterations in glioblastoma: study on a prospective cohort 30. Kuga D, Mizoguchi M, Guan Y, Hata N, Yoshimoto K, Shono T, Suzuki SO, Kukita Y, Tahira T, Nagata S, Sasaki T, Hayashi K. Prevalence of copy-number neutral LOH in www.oncotarget.com Oncotarget 28101 of 140 uniformly treated adult patients. J Clin Pathol. 2010; 63:687-91. in a conditional model of cellular senescence. Cell Cycle. 2015; 14:1164-73. 48. Pei D, Zhang Y, Zheng J. Regulation of p53: a collaboration between Mdm2 and Mdmx. Oncotarget. 2012; 3:228-35. https://doi.org/10.18632/oncotarget.443. 42. Turner KM, Sun Y, Ji P, Granberg KJ, Bernard B, Hu L, Cogdell DE, Zhou X, Yli-Harja O, Nykter M, Shmulevich I, Yung WK, Fuller GN, et al. Genomically amplified Akt3 activates DNA repair pathway and promotes glioma progression. Proc Natl Acad Sci U S A 2015; 112:3421-6. 49. Katso RM, Pardo OE, Palamidessi A, Franz CM, Marinov M, De Laurentiis A, Downward J, Scita G, Ridley AJ, Waterfield MD, Arcaro A. Phosphoinositide 3-Kinase C2beta regulates cytoskeletal organization and cell migration via Rac-dependent mechanisms. Mol Biol Cell. 2006; 17:3729-44. 43. Eley GD, Reiter JL, Pandita A, Park S, Jenkins RB, Maihle NJ, James CD. A chromosomal region 7p11.2 transcript map: its development and application to the study of EGFR amplicons in glioblastoma. Neuro Oncol. 2002; 4:86-94. 50. Kanu OO, Hughes B, Di C, Lin N, Fu J, Bigner DD, Yan H, Adamson C. Glioblastoma Multiforme Oncogenomics and Signaling Pathways. Clin Med Oncol. 2009; 3:39-52. 44. Zeng M, van der Donk WA, Chen JR. Lanthionine synthetase C-like protein 2 (LanCL2) is a novel regulator of Akt. Mol Biol Cell. 2014; 25:3954-61. 51. REFERENCES Le Mercier M, Hastir D, Moles Lopez X, De Nève N, Maris C, Trepant AL, Rorive S, Decaestecker C, Salmon I. A simplified approach for the molecular classification of glioblastomas. PLoS One. 2012; 7:e45475. 45. Landlinger C, Salzer U, Prohaska R. Myristoylation of human LanC-like protein 2 (LANCL2) is essential for the interaction with the plasma membrane and the increase in cellular sensitivity to adriamycin. Biochim Biophys Acta. 2006; 1758:1759-67. 52. Tabernero MD. A method for the global detection of chromosomal gains and losses, and an array for assessing genetic abnormalities. Spanish patent nº 2012318929 and European Trade Mark Register “FullChromArray” nº 11057627 and United States TM number 85826459. 2012 Nov 27. 46. Chakravarty D, Pedraza AM, Cotari J, Liu AH, Punko D, Kokroo A, Huse JT, Altan-Bonnet G, Brennan CW. EGFR and PDGFRA co-expression and heterodimerization in glioblastoma tumor sphere lines. Sci Rep. 2017; 7:9043. 47. Brookes S, Gagrica S, Sanij E, Rowe J, Gregory FJ, Hara E, Peters G. Evidence for a CDK4-dependent checkpoint www.oncotarget.com www.oncotarget.com Oncotarget 28102
https://openalex.org/W1524673471	https://ccforum.biomedcentral.com/track/pdf/10.1186/ccf-1999-121	English	null	Comparison of five sedation scoring systems by means of auditory evoked potentials	Critical care	1,999	cc-by	928	PublisherInfo PublisherName : BioMed Central PublisherLocation : London PublisherImprintName : BioMed Central PublisherInfo PublisherName : BioMed Central PublisherLocation : London PublisherImprintName : BioMed Central PublisherInfo PublisherName : BioMed Central PublisherLocation : London PublisherImprintName : BioMed Central Comments The drawbacks of AERs are that they require training and experience, and the effects of repeated auditory stimuli on a critically ill patient are unknown. The newer bispectral index monitor may be more useful in the ICU setting since it is simple to use by all staff and requires no auditory input. Alternatively in the absence of these expensive monitoring systems, oversedation may be avoided by switching off sedation at regular periods and not restarting infusions until the patient reaches a lesser sedation score. Aims The authors have compared an objective measure (AER) with various subjective scoring systems. The authors have compared an objective measure (AER) with various subjective s Evoked potentials, intensive care therapy, monitoring of sedation, sedation, sedation score Evoked potentials, intensive care therapy, monitoring of sedation, sedation, sedation score Introduction Auditory evoked responses (AEP) have been used as a means of monitoring awareness during anaesthesia. Attention has focused on the latency of the early cortical response Nb which indicates light anaesthesia and potential awareness when decreased to less than 44.5 ms. On the intensive care unit (ICU) sedation and analgesia is necessary to prevent ventilator dysynchrony and allow unpleasant procedures to take place. Sedation scoring systems are usually used to assess the level of sedation but unfortunately they are subjective and poor indicators of oversedation. Comparison of five sedation scoring systems by means of auditory evoked potentials Comparison of five sedation scoring systems by means of auditory evoked potentials ArticleInfo ArticleID : 4078 ArticleDOI : 10.1186/ccf-1999-121 ArticleCitationID : 121 ArticleSequenceNumber : 15 ArticleCategory : Paper Report ArticleFirstPage : 1 ArticleLastPage : 4 ArticleHistory : RegistrationDate : 1999–5–7 OnlineDate : 1999–5–7 ArticleCopyright : Current Science Ltd1999 ArticleGrants : ArticleContext : 130541111 ArticleInfo ArticleID : 4078 ArticleDOI : 10.1186/ccf-1999-121 ArticleCitationID : 121 ArticleSequenceNumber : 15 ArticleCategory : Paper Report ArticleFirstPage : 1 ArticleLastPage : 4 ArticleHistory : RegistrationDate : 1999–5–7 OnlineDate : 1999–5–7 ArticleCopyright : Current Science Ltd1999 ArticleGrants : ArticleContext : 130541111 Richard Venn Results The authors carried out 190 measurements in the 95 patients. Patients with higher APACHE II scores were more likely to be deeply sedated since, for example, severe acute respiratory distress syndrome (ARDS) patients required greater sedation to prevent ventilator dysynchrony and impaired gas exchange.The Ramsay score revealed the best correlation with changes in Nb latency (see table) Table Scoring system Kendall's coefficient Coefficient of determination(r2) P Ramsay 0.71 0.68 <0.05 Cohen -0.62 0.56 <0.05 Cambridge 0.68 0.61 <0.05 Bloomsbury -0.62 0.57 <0.05 Newcastle -0.64 0.59 <0.05 However scatterplots showed a large variation of Nb latencies with deeper levels of sedation; ie these scoring systems are poor discriminators at deeper sedation levels. The authors carried out 190 measurements in the 95 patients. Patients with higher APACHE II scores were more likely to be deeply sedated since, for example, severe acute respiratory distress syndrome (ARDS) patients required greater sedation to prevent ventilator dysynchrony and impaired gas exchange.The Ramsay score revealed the best correlation with changes in Nb latency (see table) Table Scoring system Kendall's coefficient Coefficient of determination(r2) P Ramsay 0.71 0.68 <0.05 Cohen -0.62 0.56 <0.05 Cambridge 0.68 0.61 <0.05 Bloomsbury -0.62 0.57 <0.05 Newcastle -0.64 0.59 <0.05 Table Scoring system Kendall's coefficient Coefficient of determination(r2) P Ramsay 0.71 0.68 <0.05 Cohen -0.62 0.56 <0.05 Cambridge 0.68 0.61 <0.05 Bloomsbury -0.62 0.57 <0.05 Newcastle -0.64 0.59 <0.05 However scatterplots showed a large variation of Nb latencies with deeper levels of sedation; ie these scoring systems are poor discriminators at deeper sedation levels. However scatterplots showed a large variation of Nb latencies with deeper levels of sedation; ie these scoring systems are poor discriminators at deeper sedation levels. Methods dy investigated 95 consecutive patients requiring sedation for ventilation on the IC The study investigated 95 consecutive patients requiring sedation for ventilation on the ICU. Midazolam and fentanyl were given for sedation and analgesia respectively, but additional agents could Midazolam and fentanyl were given for sedation and analgesia respectively, but additional agents could be used if sedation proved inadequate (propofol, methohexitone, gamma-hydroxybutyrate, ketamine or clonidine). The AER (midlatency Nbpeak) and corresponding level of sedation was recorded using five scoring systems [Ramsay, Cohen, Cambridge, Bloomsbury and Newcastle (Cook) Sedation Scores]. Kendall's correlation coefficients were calculated and regression analysis performed. be used if sedation proved inadequate (propofol, methohexitone, gamma-hydroxybutyrate, ketamine or clonidine). The AER (midlatency Nbpeak) and corresponding level of sedation was recorded using five scoring systems [Ramsay, Cohen, Cambridge, Bloomsbury and Newcastle (Cook) Sedation Scores]. Kendall's correlation coefficients were calculated and regression analysis performed. Discussion Derived electrophysiological measurements and evoked potentials have all been used to measure depth of sedation/anaesthesia but require training and experience to develop competency. However, since clinical scoring systems are poor discriminators at deeper sedation levels, the authors suggest that AERs may have a role where deeper levels of sedation are required for prolonged periods. It is postulated that the Ramsay score showed the best correlation since it includes a primitive reflex (glabellar tap) to discriminate asleep states. 1. Schulte-Tamburen AM, Scheier J, Briegel J, Schwender D, Peter K: Comparison of five sedation scoring systems by means of auditory evoked potentials. Intensive Care Med. 1999, 25: 377-382. References 1. Schulte-Tamburen AM, Scheier J, Briegel J, Schwender D, Peter K: Comparison of five sedation scoring systems by means of auditory evoked potentials. Intensive Care Med. 1999, 25: 377-382. This PDF file was created after publication.
https://openalex.org/W1998753324	https://europepmc.org/articles/pmc4116196?pdf=render	English	null	The Transaldolase, a Novel Allergen of Fusarium proliferatum, Demonstrates IgE Cross-Reactivity with Its Human Analogue	PloS one	2,014	cc-by	6,505	Abstract . These authors contributed equally to this work. Hong Chou1., Keh-Gong Wu2., Chang-Ching Yeh3, Hsiao-Yun Tai1, Ming F. Tam4, Yu-Sen Chen1, Horng- Der Shen1* Hong Chou1., Keh-Gong Wu2., Chang-Ching Yeh3, Hsiao-Yun Tai1, Ming F. Tam4, Der Shen1* 1 Department of Medical Research, Taipei Veterans General Hospital, Taipei, Taiwan, R.O.C., 2 Department of Pediatrics, Taipei Veterans General Hospital and National Yang-Ming University, Taipei, Taiwan, R.O.C., 3 Department of Obstetrics and Gynecology, Taipei Veterans General Hospital, Taipei, Taiwan, R.O.C., 4 Department of Biological Sciences, Carnegie Mellon University, Pittsburgh, Pennsylvania, United States of America Abstract Fusarium species are among airborne fungi and recognized as causative agents of human atopic disorders. However, Fusarium allergens have not been well characterized and the lack of information limits clinical diagnosis and treatment of fungal allergy. The purpose of this study is to identify and characterize important allergens of F. proliferatum. IgE-reacting F. proliferatum components were identified by immunoblot using serum samples from patients of respiratory atopic diseases. Characterization of allergens and determination of IgE cross-reactivity were performed by cDNA cloning, then homologous expression and immunoblot inhibition studies. We identified nine different F. proliferatum components that can be recognized by IgE antibodies in 17 (28%) of the 60 atopic sera tested. Components with molecular masses of about 43, 37.5 and 36.5 kDa with IgE-binding frequencies of about 88, 47 and 53%, respectively, were considered as important allergens of F. proliferatum. The 37.5 kDa IgE-binding component was putatively considered as a transaldolase protein of F. proliferatum. The full-length cDNA of F. proliferatum transaldolase was subsequently cloned. It encodes an open reading frame of 312 amino acids and has sequence identifies of 73 and 61%, respectively, with Cladosporium and human transaldolases. The purified recombinant F. proliferatum transaldolase can inhibit the IgE-binding against the 37.5 kDa component of F. proliferatum and the transaldolase allergen from Cladosporium cladosporioides. More importantly, the recombinant F. proliferatum transaldolase can inhibit IgE-binding against human transaldolase in a concentration-dependent manner. Thus, a novel and important F. proliferatum transaldolase allergen was identified. In addition to IgE cross-reactivity between the Fusarium and the Cladosporium transaldolase allergens, IgE cross-reactivity between the Fusarium and the human transaldolases also exists and might contribute to atopic manifestations in the absence of exogenous allergen exposure. Editor: Thomas H. Thatcher, University of Rochester Medical Center, United States of America Copyright: 2014 Chou et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was supported by grants from the National Science Council (Grant NSC 101-2320-B-075-005-MY2) and the Taipei Veterans General Hospital (V102C-066), Taipei, Taiwan, Republic of China. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * Email: hdshen@vghtpe.gov.tw . These authors contributed equally to this work. Editor: Thomas H. Thatcher, University of Rochester Medical Center, United States of America Citation: Chou H, Wu K-G, Yeh C-C, Tai H-Y, Tam MF, et al. (2014) The Transaldolase, a Novel Allergen of Fusarium proliferatum, Demonstrates IgE Cross-Reactivity with Its Human Analogue. PLoS ONE 9(7): e103488. doi:10.1371/journal.pone.0103488 Immunoblot inhibition For immunoblot inhibition studies, anti-transaldolase IgE- containing serum samples were firstly reacted with purified recombinant F. proliferatum transaldolase before incubating with PVDF blots containing F. proliferatum extracts, purified rCla c 14.0101, or purified recombinant human transaldolase at 4uC for 16 h. As controls, the blots were incubated with similar serum samples that have been pre-incubated with equivalent amounts of bovine serum albumin (BSA, Pierce, Rockford, IL, USA) or purified recombinant house dust mite allergen Der p 7 [21]. The blots were then washed and incubated with alkaline phosphatase- conjugated monoclonal anti-human IgE antibodies (Pharmingen) and developed with enzyme substrates as described [19,20]. In addition, two house dust mite-sensitized serum samples were firstly Transaldolase Allergen of Fusarium proliferatum in Europe. In India, skin prick tests with 60 allergens were performed on 48 patients with naso-bronchial allergy and results indicated that Aspergillus fumigatus, A. flavus, Alternaria teneis and F. solani were common fungal allergens [14]. In Greece, Gonianakis et al. [15] found that among 571 patients, 42% showed dermal positivity to allergens derived from Alternaria, Cladosporium, Fusarium, Aspergillus, and Mucor. Thus, there is a worldwide indication that Fusarium fungus may play a role in clinical allergy. However, our knowledge about allergens of this airborne Fusarium fungus is still quite limited [13,16] and standardized Fusarium extracts for clinical diagnostics are lacking. skimmed milk and incubated with serum samples for 16 h at 4uC. The membranes were washed, incubated with alkaline phosphatase-conjugated monoclonal anti-human IgE antibodies (Pharmingen, San Diego, CA, USA) then developed with enzyme substrates essentially as described [19,20]. Serum samples from a non-atopic healthy individual and two house dust mite (Derma- tophagoides pteronyssinus)-sensitized atopic individuals were used as controls. Crude extracts of F. proliferatum F. proliferatum strain BCRC 30972 was used in this study. It was isolated from the air of Taiwan and provided by the Food Industry Research and Development Institute, Hsinchu, Taiwan. It was cultured in a CYB medium without agitation at 26uC for five days. The CYB medium contains yeast carbon base (Difco Laboratories, Detroit MI, USA; 11.7 g/L), glucose (Mallinckrodt Baker, Inc., Phillipsburg, NJ, USA; 10 g/L) and casein enzymatic hydrolysate (Sigma Chemical Co., St. Louis, MO, USA; 10 g/L). Crude extracts of F. proliferatum were prepared essentially as described [19,20]. The protein content of crude fungal extracts was determined with a dye-binding assay according to the manufacturer’s instructions (Bio-Rad, Richmond, CA, USA). Preparation of recombinant fungal transaldolases The sixty serum samples used in this study were obtained from the Biobank at the Taipei Veterans General Hospital. All these serum samples were obtained from respiratory atopic patients (allergic rhinitis and/or atopic asthma) who attended the allergy clinics of the Taipei Veterans General Hospital and were stored in aliquots at 280uC. This study has been approved by the Institutional Review Board of the Taipei Veterans General Hospital. p g The F. proliferatum transaldolase was expressed with an N- terminal His6-tag. Appropriate primers (Fu-TAase-f, 5’- cgggatcc44tcttcctctctcgaacagctc64-3’ and Fu-TAase-r, 5’-aactg- cag1012ttaggcgagcttctccttgaggatgc987-3’) were used in the PCR amplification with the full-length cDNA encoding the F. proliferatum transaldolase prepared above as template. The PCR products were restricted then ligated into the pQE-80 vector (Qiagen Inc., Valencia, CA, USA) for protein expression in E. coli JM109 cells. The recombinant proteins were affinity-purified with Ni-NTA resin columns (Qiagen Inc.) according to the manufac- turer’s instructions. The His6-tagged recombinant C. cladospor- ioides transaldolase (Cla c 14.0101) used in this study was prepared as described previously [19]. Immuno-reactivity of the recombi- nant fungal and the recombinant human transaldolases (Novus Biologicals, Littleton, CO, USA; 0.5 mg/strip) against IgE antibodies was analyzed by SDS-PAGE-immunoblot. cDNA Cloning The cDNA encoding the F. proliferatum transaldolase was isolated with polymerase chain reactions (PCR) using an Affinity- Script Multiple Temperature cDNA Synthesis kit (Stratagene, La Jolla, Calif., USA) as previously described [19,20]. Primers TAase (5’-119aag/tac/aag/cc(a/c)/ca(a/g)/ga(t/c)/gc138-3’) and AP (5’- ggccacgcgtcgactagtact-(dt)16-3’) were used in the first set of PCR. The product obtained was used as a template in a subsequent PCR with primers TAase and AUAP (5’-ggccacgcgtcgactagtac-3’). The product from the nest PCR reaction was gel purified and inserted into the pGEM-T vector (Promega, Madison, WI, USA) for sequencing analysis. g g IgE cross-reactivity is an important component of fungal sensitization and could contribute significantly to allergy manifes- tation [17]. Thus, in addition to the identification and character- ization of fungal allergens, it is important to delineate IgE cross- reactivity between allergens from different fungal species and even more importantly, between fungal allergens and their human analogues. Previously, we have identified important IgE cross- reactive pan-serine protease fungal allergens from prevalent Penicillium and Aspergillus species [18]. Recently, in addition to serine proteases, the transaldolase has also been identified as a significant and IgE cross-reactive allergen family of Cladosporium and Penicillium species [19]. The purpose of this study is to identify and characterize allergens of Fusarium species. Our results show that the 37.5 kDa transaldolase is a novel and important allergen of F. proliferatum (Fus p 4.0101). In addition, Fus p 4.0101 demonstrated IgE cross-reactivity with the transaldolase allergen from C. cladosporioides (Cla c 14.0101) and, interestingly, with the human transaldolase. The full-length cDNA of the F. proliferatum transaldolase was obtained by 5’ rapid amplification of cDNA end (RACE) reaction. The template cDNA for the reaction was synthesized with reverse transcriptase (RT, Stratagene) and primer GSP-r1 (5’-529aagaga- gaacatgagggtgaggtt506-3’). An oligo-(dC) was added to the end of the purified cDNA with terminal deoxynucleotidyl transferase (Promega). Primers GSP-r2 (5’-336tcgacctcagttgagacctt317-3’) and 5R AAP (5’-ggccacgcgtcgactagtacgggiigggiigggiig-3’) were then used in the 5’-RACE reaction. The product was purified, subcloned, transformed and subsequently sequenced. The full-length cDNA of the F. proliferatum transaldolase was obtained by 5’ rapid amplification of cDNA end (RACE) reaction. The template cDNA for the reaction was synthesized with reverse transcriptase (RT, Stratagene) and primer GSP-r1 (5’-529aagaga- gaacatgagggtgaggtt506-3’). An oligo-(dC) was added to the end of the purified cDNA with terminal deoxynucleotidyl transferase (Promega). Primers GSP-r2 (5’-336tcgacctcagttgagacctt317-3’) and 5R AAP (5’-ggccacgcgtcgactagtacgggiigggiigggiig-3’) were then used in the 5’-RACE reaction. The product was purified, subcloned, transformed and subsequently sequenced. Introduction including Taipei [6–8]. Cladosporium and Alternaria are clinically important causative allergenic agents for patients sensitive to fungi [2]. Fusarium fungus can emit large amount of spores in rainy or humid enrivonment [9]. Thus, it is of interest to study the relevance of Fusarium fungus to allergic sensitization. Chang et al. [10] tested a list of 54 air-borne allergens in 66 bronchial asthma patients in the Taipei area, and 20 (30%) of the patients showed positive skin reaction to Fusarium extracts. O’Neil et al. [11] found that among 69 atopic individuals tested in United States, 17 (24%) of the patients had positive skin reactions to an extract of F. solani. Stroud et al. [12] reported that reactivity to fungi was found in 65% of chronic rhinitis patients and reactions to Fusarium (58%), Alternaria (39%) and Pullularia (38%) were particularly common. Using in-house extracts for EAST and immunoblot experiments, Hoff et al. [13] detected F. culmorum specific IgE antibodies in 23 (44%) of 52 subjects with suspected mould allergy The prevalence of human atopic diseases including allergic rhinitis and asthma is increasing during recent decades [1]. Sensitization to molds which are ubiquitous in our environment has been reported to be close to 80% of asthmatic patients [2]. Hence, it is important to characterize fungal allergens and subsequently provide a basis for better diagnosis and treatment of fungal allergy [2–4]. However, fungal allergens are difficult to be defined since multitudinous factors contribute to the results obtained [3,5]. Due to these inherent difficulties, the manufactur- ing of standardized and high quality fungal allergen extracts is not available in the United States [4]. It highlights the necessity of characterizing important fungal allergens [2–4]. Cladosporium species are the dominant airborne spores worldwide [3]. In addition, Alternaria, Aspergillus, Penicillium and Fusarium species are also airborne fungi in many areas July 2014 \| Volume 9 \| Issue 7 \| e103488 1 PLOS ONE \| www.plosone.org Transaldolase Allergen of Fusarium proliferatum July 2014 \| Volume 9 \| Issue 7 \| e103488 Immunoreactivity of recombinant transaldolase proteins Immunoreactivity of recombinant transaldolase proteins In this study, rFus p 4.0101 was expressed as N-terminal His6- tag proteins in E.coli and purified (Fig. 3, panel A). It has an apparent molecular mass of about 38 kDa upon SDS-PAGE analysis (data not shown). The Coomassie blue-stained protein profiles of rFus p 4.0101, rCla c 14.0101 and a recombinant human transaldolase obtained commercially are shown in Fig. 3, panel A. Serum sample nos. 1–3 from Fig. 1, panel B showed positive IgE-binding against rFus p 4.0101. These three serum samples have negative IgE-binding to rCla c 14.0101 and serum no. 2 showed positive IgE-binding against human transaldolase (Fig. 3, panel B). In addition, nine serum samples from asthmatic patients with IgE-binding against rCla c 14.0101 or recombinant human transaldolase [19] were included in this study. The IgE immunoblot reactivities of these sera (serum nos. 21–29) against these three different recombinant transaldolases are shown in Fig. 3, panel B. Among these nine serum samples, eight (serum nos. 21–26, 28 and 29) showed positive IgE-binding against rFus p 4.0101. Eight (serum nos. 21–28) of these nine serum samples showed IgE-binding against rCla c 14.0101. Furthermore, serum nos. 22, 24, 26, 27 and 29 have also IgE reactivity against recombinant human transaldolase. Serum from a non-atopic individual (serum no. 18) and from two house dust mite-sensitized atopic individuals (serum nos. 19 and 20) were used as controls and showed negative IgE immunoblot reactivity against all three recombinant transaldolases (Fig. 3, panel B). Since a 36.5 kDa IgE-binding component was identified as a transaldolase allergen of C. cladosporioides (Cla c 14.0101) [19]. We putatively concluded that the IgE-reacting 37.5 kDa compo- nent from F. proliferatum was possibly a transaldolase. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting Proteins in the crude fungal extractions or purified recombinant proteins were separated by SDS-PAGE [19,20] then transferred electrophoretically onto polyvinylidene difluoride (PVDF) mem- branes (0.45 mm, Millipore, Bedford, MS, USA). Protein compo- nents reacting against human IgE antibodies were determined as described [19,20]. The membranes were blocked with 1% July 2014 \| Volume 9 \| Issue 7 \| e103488 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org 2 Transaldolase Allergen of Fusarium proliferatum reacted with purified recombinant Der p 7, F. proliferatum transaldolase or BSA before incubating with PVDF blots containing Der p 7 at 4uC for 16 h. The blots were then incubated with alkaline phosphatase-conjugated monoclonal anti- human IgE antibodies (Pharmingen) and developed with enzyme substrates as above. GQ906475), P. chrysogenum (accession no. GQ925430), S. cerevisiae (accession no. AAB67752), and Homo sapiens (accession no. AAF40478) transaldolases, respectively (data not shown). The transaldolase from F. proliferatum has only one cysteine residue. However, three cysteine residues are conserved among A. fumigates, C. cladoporioides and P. chrysogenum transaldolases. In addition, one and two cysteine residues can be found in S. cerevisiae and Homo sapiens transaldolases, respectively (data not shown). The one potential N-glycosylation site of F. proliferatum transaldolase (156NLT158) is boxed in Fig. 2 and it is conserved among all the transaldolases mentioned above. cDNA cloning of the F. proliferatum transaldolase cDNA cloning of the F. proliferatum transaldolase The full-length cDNA encoding the F. proliferatum transaldo- lase was obtained through RT-PCR coupled with the 5’-end RACE reaction. The nucleotide (GenBank accession no. KF151224) and the deduced amino acid sequences of the open reading frame are presented in Fig. 2. A potential polyadenylation signal (AATCGA) for mRNAs of higher eukaryotes was found 15– 20 bases upstream from the poly-A tail. The mature F. proliferatum transaldolase protein, excluding the initiator methi- onine [22] has 322 residues and a calculated molecular mass of 35404 daltons, without considering the presence of further post- translational modifications. It has one cysteine (Cys240) and one putative N-glycosylation site (156NLT158) (Fig. 2). Furthermore, amino acid residues conserved among transaldolases (Asp18, Asn36, Glu97, Lys133, Asn156, Thr158, Ser178 and Arg183) can also be found (Fig. 2). These conserved residues are crucial to enzyme catalysis or substrate binding [23]. This transaldolase allergen has been designated as Fus p 4.0101 by the I.U.I.S. Allergen Nomenclature Sub-committee. Immunoblot inhibition The relationship between rFus p 4.0101 and the IgE-binding 37.5 kDa component of F. proliferatum was further delineated with immunoblot inhibition. Fig. 4 showed that serum sample no. 1 from panel B of Fig. 1 has IgE immunoblot reactivity against the 37.5 kDa component of F. proliferatum (Fig. 4, strip 1 of panel B). This immunoblot reactivity was inhibited when the same serum sample was pre-absorbed with 10 mg of rFus p 4.0101 (Fig. 4, strip 2 of panel B). Pre-absorption of the serum sample with 10 mg of BSA did not inhibit its IgE binding against the 37.5 kDa component of F. proliferatum (Fig. 4, strip 3 of panel B). Results obtained correlate the 37.5 kDa component with the F. prolifer- atum transaldolase and much of the IgE determinants on the native transaldolase were conserved in the recombinant Fus p 4.0101. Immunoblot reactivity against components of F. proliferatum p F. proliferatum crude extracts were separated by SDS-PAGE. The Coomassie blue-stained protein profile of the fungal extracts was shown in panel A of Fig. 1. The separated proteins were blotted onto PVDF membrane then reacted with patient sera. Among the 60 serum samples from respiratory atopic patients examined, 17 (28%) demonstrated IgE-binding against compo- nents of F. proliferatum (Fig. 1, strip nos. 1–17 of panel B). Human IgE antibodies reacted with at least nine different F. proliferatum components ranging in molecular mass from 92 to 30 kDa as shown in Fig. 1 and Table 1. Components of 43 and 36.5 kDa with IgE-binding frequencies of 88% (15/17) and 53% (9/17), respectively, may be considered as major allergens of F. proliferatum. In Fig. 1, the 36.5 kDa allergen has a relatively higher intensity of IgE-immunoblot reactivity than others. The 37.5 kDa component (indicated with an arrow, Fig. 1B) with IgE- binding frequency of 47% (8/17) was considered an important allergen of F. proliferatum. The 92, 83 and 48 kDa components with IgE-binding frequencies of about 35–41% were considered as significant allergens of F. proliferatum. The 40, 32 and 30 kDa components with IgE-binding frequencies of less than 20% were minor allergens of F. proliferatum. A serum sample from a non- atopic individual (serum no. 18) and two serum samples from house dust mite (D. pteronyssinus)-sensitized atopic individuals (serum nos. 19 and 20) were included as negative controls and shown in Fig. 1, panel B. Immunoglobulin E cross-reactivity equiseti, a 65 kDa protein was also found as a major allergen of this common Fusarium species [25]. The nature of the 65 kDa major allergen needs further study. The 45 kDa F. solani allergen has an N-terminal sequence of Lys-Gly-Arg-Thr-Glu-Phe-Ala, which does not show homology to any known fungal proteins [26]. Through cDNA cloning, Hoff et al. in Europe identified three F. culmorum allergens (Fus c 1, Fus c 2 and Fus c 3) [13]. The 11 kDa Fus c 1 (60S acidic ribosomal protein P2), 13 kDa Fus c 2 (thioredoxin-like protein) and 49 kDa Fus c 3 (not related to known proteins) have IgE-binding frequencies of 35, 50, and 15%, respectively, with sera from 26 individuals sensitized to F. culmorum [13]. Recently, Khosravi et al. in Iran showed with immunoblotting that F. solani has six major allergens with molecular masses of 24, 58.5, 64.5, 69, 72 and 97 kDa [27]. In this study, we did not detect a major IgE-binding F. proliferatum protein with molecular mass of about 65 kDa. Whether our 37.5 and 43 kDa IgE-reacting F. proliferatum components correspond to the 38 and 45 kDa F. solani allergens reported need further clarification. Furthermore, we did not detect low molecular mass IgE-binding F. proliferatum proteins correlate to the 11 and 13 kDa F. culmorum allergens. Whether our 48 kDa IgE-reacting F. proliferatum protein resembles the 49 kDa F. culmorum allergen also needs further elucidation. The utilization of different fungal strains, the variations in the culturing conditions, the variance of the methods used in preparing the fungal extracts, and the divergences in exposure and genetic background of individuals examined all contribute to discrepancies in results obtained from various studies and research groups [4]. Thus, studies of fungal allergens at molecular level to provide a basis for standardized fungal extracts is of major importance in clinical allergy. Serum sample no. 26 has also IgE-binding activity against recombinant human transaldolase (Fig. 3, panel B and Fig. 5, panel B). This activity can be inhibited dose dependently by pre- absorbing the serum with 2 or 20 mg of rFus p 4.0101. Inhibition of IgE binding was not detected when the same serum was pre- absorbed with 20 mg of BSA (Fig. 5, panel B). In addition, the IgE- binding activity against recombinant human transaldolase of serum no. Immunoglobulin E cross-reactivity The panel A of Fig. 5 showed that serum nos. 26 and 28 from Fig. 3, panel B has IgE-binding activity against rCla c 14.0101. rFus p 4.0101 can inhibit, dose dependently, this reactivity (Fig. 5, panel A). Pre-absorption of serum sample no. 26 with 50 mg of BSA or pre-absorption of serum sample no. 28 with 20 mg of rDer p 7 did not inhibit its IgE binding against rCla c 14.0101 (Fig. 5, panel A). The results suggest IgE cross-reactivity between transaldolases from Fusarium and Cladosporium fungi. Results from sequence alignment revealed that the F. prolifer- atum transaldolase has 255 (79%), 236 (73%), 247 (77%), 207 (64%) and 195 (61%) amino acids identical to that of A. fumigatus (accession no. XM748623), C. cladoporioides (accession no. PLOS ONE \| www.plosone.org July 2014 \| Volume 9 \| Issue 7 \| e103488 3 Transaldolase Allergen of Fusarium proliferatum Figure 1. Immunoblot reactivity against components in F. proliferatum extracts. (A) Coomassie blue-stained protein profile of F. proliferatum extracts. (B) IgE immunoblot profiles obtained with serum samples from respiratory atopic patients (serum nos. 1–17), a non-atopic individual (serum no. 18) and two house dust mite (D. pteronyssinus)-sensitized atopic individuals (serum nos. 19, 20). The arrow indicates the position of the 37.5 kDa component of F. proliferatum. doi:10.1371/journal.pone.0103488.g001 gure 1 Immunoblot reactivity against components in F proliferatum extracts (A) Coomassie blue-stained protein profile of F proliferatum Figure 1. Immunoblot reactivity against components in F. proliferatum extracts. (A) Coomassie blue-stained protein profile of F. proliferatum extracts. (B) IgE immunoblot profiles obtained with serum samples from respiratory atopic patients (serum nos. 1–17), a non-atopic individual (serum no. 18) and two house dust mite (D. pteronyssinus)-sensitized atopic individuals (serum nos. 19, 20). The arrow indicates the position of the 37.5 kDa component of F. proliferatum. doi:10.1371/journal.pone.0103488.g001 Allergens of the Fusarium species have been reported by three different groups. Verma et al. [16] reported from India that the culture filtrates from F. solani contained 18 allergenic proteins as determined by immunoblotting. A 65-kDa protein component reacted with IgE antibodies in all 15 patient sera tested and was considered as a major allergen [16]. The 45 and 14 kDa components reacted against IgE in 12 patients’ sera. The 41, 38, 35 and 30 kDa F. solani components reacted with IgE antibodies in 9 of the 15 patients’ sera tested [16]. In studies of allergens of F. Immunoglobulin E cross-reactivity 29 can also be inhibited dose dependently by pre- absorbing the serum with 5 or 20 mg of rFus p 4.0101 (Fig. 5, panel B). Inhibition of IgE binding was not detected when the same serum was pre-absorbed with 20 mg of Der p 7 (Fig. 5, panel B). Results in Fig. 5, panel C demonstrated that IgE-binding against recombinant Der p 7 in two house dust mite-sensitized atopic serum samples (serum nos. 19 and 20 in Fig. 1 and Fig. 3) can be inhibited by pre-absorbing both serum samples with 5 or 20 mg of rDer p 7. But similar amounts of Fus p 4.0101 or BSA cannot inhibit this binding activity. July 2014 \| Volume 9 \| Issue 7 \| e103488 Transaldolase Allergen of Fusarium proliferatum Table 1. Reactivity of IgE antibodies against components of F. proliferatum analyzed by SDS-PAGE-immunoblot. F. proliferatum component Strips of IgE-immunoblot No. kDa Frequency of IgE-binding (%) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1 92 41 (7/17) +* + + + + + + 2 83 35 (6/17) + + + + + + 3 48 35 (6/17) + + + + + + 4 43 88 (15/17) + + + + + + + + + + + + + + + 5 40 18 (3/17) + + + 6 37.5 47 (8/17) + + + + + + + + 7 36.5 53 (9/17) + + + + + + + + + 8 32 12 (2/17) + + 9 30 12 (2/17) + + Figure 2. The nucleotide and deduced amino acid sequences of Fus p 4.0101 (GenBank accession no. KF151224). Numbers to the right are the positions of the nucleotides and the deduced residues of the sequences. The stop codon TAA is denoted with an asterisk (). The potential N-glycosylation site is indicated in bold letters and boxed. Eight conserved amino acid residues that are involved in the catalysis and the substrate binding of transaldolases are shaded and in bold letters. Nucleotide sequences in grey correspond to primer sequences synthesized for PCR experiments in the cDNA cloning of Fus p 4.0101 as described in the Materials and Methods. The sequences corresponding to primers Fu-TAase-f and Fu-TAase-r used in the preparation of rFus p 4.0101 are boxed. doi:10.1371/journal.pone.0103488.g002 doi:10.1371/journal.pone.0103488.g002 indicates positive IgE-binding as shown in Fig. 1, panel B. doi:10.1371/journal.pone.0103488.t001 Fungal-atopic subjects often demonstrated a parallel and independent multiple sensitization to many different fungal species [13,17]. Patients may be sensitized by allergens from individual fungal species. Their multiple sensitizations could be caused by IgE cross-reactivity of a particular protein common among various fungal species. We have observed IgE cross-reactivity among the vacuolar serine protease major pan-fungal allergens [18,20]. Our results here indicate that the 37.5 kDa transaldolase is an important allergen of F. proliferatum. Transaldolases have been identified as important allergens of prevalent airborne Cladospo- rium (Cla c 14.0101) and Penicillium (Pen c 35) species [19]. IgE cross-reactivity between Cla c 14.0101 and Pen c 35 has been demonstrated [19]. Transaldolase allergens from Cladosporium, Penicillium and Fusarium species share 67% (217/323 residues) amino acid sequence identity. Discussion Fungi are prominent sources of allergens. However, they are still largely neglected in clinical practice and basic research [24]. Allergens of the frequent mold genera including Cladosporium, Alternaria, Aspergillus and Penicillium have been characterized and reported [2,3,18–20,24]. Fusarium is among airborne fungi that contribute to human respiratory atopic disorders worldwide. It is important to well characterize IgE-recognizing Fusarium components. In this study, among 60 respiratory atopic sera tested, 17 (28%) demonstrated IgE-binding against nine different components from F. proliferatum. The nature of the 43 and 36.5 kDa major F. proliferatum allergens is currently under investigation and will be published separately (Shen et al., manuscript in preparation). The 37.5 kDa component of F. proliferatum has an IgE-binding frequency of 47% (8/17). IgE- binding against this 37.5 kDa component can be inhibited by a recombinant F. proliferatum transaldolase (Fig. 4). Our results indicate that this important 37.5 kDa IgE-binding component is a F. proliferatum transaldolase. July 2014 \| Volume 9 \| Issue 7 \| e103488 July 2014 \| Volume 9 \| Issue 7 \| e103488 PLOS ONE \| www.plosone.org 4 Transaldolase Allergen of Fusarium proliferatum July 2014 \| Volume 9 \| Issue 7 \| e103488 Figure 3. Antigenicity of recombinant F. proliferatum, C. cladosporioides and human transaldolases. (A) Coomassie blue-stained protein profile of rFus p 4.0101, rCla c 14.0101 and recombinant human transaldolase on PVDF membranes. (B) IgE immunoblot reactivities of these three recombinant proteins analyzed by using serum samples nos. 1–3 from Fig. 1 (serum nos. 1–3) and nine serum samples from asthmatic patients (serum nos. 21–29) who showed previously IgE-binding reactvities against rCla c 14.0101 or recombinant human transaldolase. Sera from a non- atopic healthy individual (serum no. 18) and two house dust mite-sensitized atopic individuals (serum nos. 19 and 20) were included as controls. doi:10.1371/journal.pone.0103488.g003 Fusarium and Cladosporium transaldolase allergens. Results from IgE-immunoblot inhibition (Fig. 5, panel A) with serum nos. 26 and 28 in Fig. 3 confirmed the presence of IgE cross-reactivity between Fus p 4.0101 and Cla c 14.0101. Combining with previous results [19], it is suggestive that IgE cross-reactivity prevails among the transaldolase allergens from Fusarium, Penicillium and Cladosporium species. Similarly, Verma et al. detected allergenic cross-reactivity among the 14 kDa protein component of three different Fusarium species [28]. In addition, the 45 kDa F. solani major allergen has allergenic cross-reactivity with fungal extracts prepared from Alternaria, Cladosporium, Curvularia and Epicocum species [26]. Furthermore, Hoff et al. also demonstrated IgE cross-reactivity between F. culmorum and A. alternata allergens [13]. All these results provide important information in clinical fungal allergy. Fusarium and Cladosporium transaldolase allergens. Results from IgE-immunoblot inhibition (Fig. 5, panel A) with serum nos. 26 and 28 in Fig. 3 confirmed the presence of IgE cross-reactivity between Fus p 4.0101 and Cla c 14.0101. Combining with previous results [19], it is suggestive that IgE cross-reactivity prevails among the transaldolase allergens from Fusarium, Figure 4. Immunoblot inhibition of IgE-binding against nFus p 4.0101 in crude F. proliferatum extracts with purified rFus p 4.0101 and BSA as inhibitors. (A) Coomassie blue-stained protein profile of F. proliferatum extracts and protein molecular weight markers. (B) IgE binding against the 37.5 kDa component using serum no. 1 from figure 1B (lane 1); this binding activity was inhibited with 10 mg of rFus p 4.0101 (lane 2) but not BSA (lane 3). doi:10.1371/journal.pone.0103488.g004 Transaldolase from F. proliferatum, C. cladosporioides and homo sapiens share 54% (173/322 residues) amino acid sequence identity. Interestingly, five of the eleven rFus p 4.0101 IgE-positive sera (serum nos. Transaldolase Allergen of Fusarium proliferatum Our results in Fig. 3 showed that seven sera have IgE reactivity against both Fus p 4.0101 and Cla c 14.0101 and suggested IgE cross-reactivity exists between the PLOS ONE \| www.plosone.org July 2014 \| Volume 9 \| Issue 7 \| e103488 5 Transaldolase Allergen of Fusarium proliferatum Figure 3. Antigenicity of recombinant F. proliferatum, C. cladosporioides and human transaldolases. (A) Coomassie blue-stained protein profile of rFus p 4.0101, rCla c 14.0101 and recombinant human transaldolase on PVDF membranes. (B) IgE immunoblot reactivities of these three recombinant proteins analyzed by using serum samples nos. 1–3 from Fig. 1 (serum nos. 1–3) and nine serum samples from asthmatic patients (serum nos. 21–29) who showed previously IgE-binding reactvities against rCla c 14.0101 or recombinant human transaldolase. Sera from a non- atopic healthy individual (serum no. 18) and two house dust mite-sensitized atopic individuals (serum nos. 19 and 20) were included as controls. doi:10.1371/journal.pone.0103488.g003 July 2014 \| Volume 9 \| Issue 7 \| e103488 2, 22, 24, 26 and 29) and one of the rCla c 14.0101 IgE-positive serum (serum no. 27) showed IgE-binding to recombinant human transaldolase (Fig. 3, panel B). Similarly, among another eight fungal (Cla c 14.0101) transaldolase-positive sera tested previously, three of them showed IgE-binding against the recombinant human transaldolase [19]. In this study, IgE cross-reactivity between Fusarium and human transaldolase was further demonstrated (Fig. 5, panel B). In literature, evolutionarily conserved IgE-reactive human antigens corresponding to fungal and pollen allergens such as profilin, ribosomal P2 protein, and manganese superoxide dismutase proteins have been reported [29–31]. Recognition of these human counterparts by IgE antibodies might contribute to the stimulation of type I hypersensitive reactions in the absence of exogenous allergen exposure [30]. It might also play a role in certain chronic and severe allergic disorders [29,31]. Results obtained from this study provide further evidence in IgE cross-reactivity between an environmental fungal allergen and its human analogue which might contribute to disease manifestations. Figure 4. Immunoblot inhibition of IgE-binding against nFus p 4.0101 in crude F. proliferatum extracts with purified rFus p 4.0101 and BSA as inhibitors. (A) Coomassie blue-stained protein profile of F. proliferatum extracts and protein molecular weight markers. (B) IgE binding against the 37.5 kDa component using serum no. 1 from figure 1B (lane 1); this binding activity was inhibited with 10 mg of rFus p 4.0101 (lane 2) but not BSA (lane 3). doi:10.1371/journal.pone.0103488.g004 In conclusion, we identified an important novel transaldolase allergen of F. proliferatum. In addition, IgE cross-reactivities between Fusarium and Cladosporium transaldolase allergens as well as between Fusarium and human transaldolases were July 2014 \| Volume 9 \| Issue 7 \| e103488 July 2014 \| Volume 9 \| Issue 7 \| e103488 PLOS ONE \| www.plosone.org 6 Transaldolase Allergen of Fusarium proliferatum Figure 5. Inhibition of IgE-immunoblot reactivity against (A) rCla c 14.0101, (B) recombinant human transaldolase and (C) Der p 7. IgE-immunoblot experiments were carried out with serum nos. 19, 20, 26, 28 and 29 from Fig. 3. Serum was pre-absorbed with the amount of rFus p 4.0101, Der p 7 or BSA as indicated. * indicates not done. doi:10.1371/journal.pone.0103488.g005 Figure 5. Inhibition of IgE-immunoblot reactivity against (A) rCla c 14.0101, (B) recombinant human transaldolase and (C) Der p 7. IgE-immunoblot experiments were carried out with serum nos. 19, 20, 26, 28 and 29 from Fig. 3. Author Contributions Conceived and designed the experiments: HC KGW CCY HDS. Performed the experiments: HC HYT YSC. Analyzed the data: HC KGW HYT CCY HDS. Contributed reagents/materials/analysis tools: HC KGW. Wrote the paper: HC KGW CCY MFT HDS. Conceived and designed the experiments: HC KGW CCY HDS. Performed the experiments: HC HYT YSC. Analyzed the data: HC KGW HYT CCY HDS. Contributed reagents/materials/analysis tools: HC KGW. Wrote the paper: HC KGW CCY MFT HDS. 10. Chang HN, Lin FM, Chang YF (1984) Correlation of skin test to RAST, and PRIST in asthma patients. Chin Med J 34: 376–383. Serum was pre-absorbed with the amount of rFus p 4.0101, Der p 7 or BSA as indicated. * indicates not done. doi:10.1371/journal.pone.0103488.g005 demonstrated. Our results provide important information in clinical fungal allergy. demonstrated. Our results provide important information in clinical fungal allergy. 9. von Wahl PG, Kersten W (1991) Fusarium and Didymella–neglected spores in the air. Aerobiologia 7: 111–117. 6. Amend AS, Seifert KA, Samson R, Bruns TD (2010) Indoor fungal composition is geographically patterned and more diverse in temperate zones than in the tropics. Proc Natl Acad Sci USA 107: 13748–13753. 8. MH Abu-Dieyeh, R Barham, K Abu-Elteen, R Al-Rashidi (2010) Seasonal variation of fungal spore populations in the atmosphere of Zarqa area, Jordan. Aerobiologia 26: 263–276. p 7. Li CS, Hsu LY, Chou CC, Hsieh KH (1995) Fungus allergens inside and outside the residences of atopic and control children. Arch Environ Health 50: 38–43. Transaldolase Allergen of Fusarium proliferatum 11. O’Neil CE, McCants ML, Salvaggio JE, Lehrer SB (1986) Fusarium solani: prevalence of skin reactivity and antigenic allergenic analysis. J Allergy Clin Immunol 77: 842–849. 21. Tai HY, Zhou JK, Chou H, Tam MF, Chen YS, et al. (2013) Epitope mapping and in silico characterization of interactions between Der p 7 allergen and MoAb WH9. PLOS ONE 8: e71269. 12. Stroud R, Calhoun K, Wright S, Kennedy K (2001) Prevalence of hypersensitivity to specific fungal allergens as determined by intradermal dilutional testing. Otolaryngol Head Neck Surg 125: 491–494. 22. Sprenger GA, Scho¨rken U, Sprenger G, Sahm H (1995) Transaldolase B of Escherichia coli K-12: cloning of its gene, talB, and characterization of the enzyme from recombinant strains. J Bacteriol 177: 5930–5936. 23. Thorell S, Gergely P Jr, Banki K, Perl A, Schneider G (2000) The three- dimensional structure of human transaldolase. FEBS Lett 475: 205–208. 13. Hoff M, Ballmer-Weber BK, Niggemann B, Cistero-Bahima A, San Miguel- Moncı´n M, et al. (2003) Molecular cloning and immunological characterization of potential allergens from the mould Fusarium culmorum. Mol Immunol 39: 965–975. 24. Crameri R, Garbani M, Rhyner C, Huitema C (2014) Fungi: the neglected allergenic sources. Allergy 69: 176–185. 14. Prasad R, Verma SK, Dua R, Kant S, Kushwaha RA, et al. (2009) A study of skin sensitivity to various allergens by skin prick test in patients of naso-bronchial allergy. Lung India 26: 70–73. 25. Verma J, Sridhara S, Rai D, Gangal SV (1998) Isolation and immunobiochem- ical characterization of a major allergen (65 kDa) from Fusarium equiseti. Allergy 53: 311–315. 26. Verma J, Singh BP, Sridhara S, Gaur SN, Arora N (2003) Purification and characterization of a cross-reactive 45-kD major allergen of Fusarium solani. Int Arch Allergy Immunol 130: 193–199. 15. Gonianakis MI, Neonakis IK, Gonianakis IM, Baritaki MA, Bouros D, et al. (2006) Mold allergy in the Mediterranean Island of Crete, Greece: a 10-year volumetric, aerobiological study with dermal sensitization correlations. Allergy Asthma Proc 27: 354–362. 27. Khosravi AR, Fatahinia M, Shokri H, Yadegari MH (2012) Allergens from Fusarium solani identified by immunoblotting in asthma patients In Iran. Arh Hig Rada Toksikol 63: 1–6. 16. Verma J, Gangal SV (1994) Fusarium solani: Immunochemical characterization of allergens. Int Arch Allergy Appl Immunol 104: 175–183. g 28. Verma J, Gangal SV (1994) Studies on Fusarium solani–Cross-reactivity among Fusarium species. Allergy 49: 330–336. g gy pp 17. References 1. Centers for Disease Control and Prevention (CDC) (2011) Vital signs: asthma prevalence, disease characteristics, and self-management education–United States, 2001–2009. MMWR Morb Mortal Wkly Rep 60: 547–552. 6. Amend AS, Seifert KA, Samson R, Bruns TD (2010) Indoor fungal composition is geographically patterned and more diverse in temperate zones than in the tropics. Proc Natl Acad Sci USA 107: 13748–13753. p 7. Li CS, Hsu LY, Chou CC, Hsieh KH (1995) Fungus allergens inside and outside the residences of atopic and control children. Arch Environ Health 50: 38–43. y p 2. Simon-Nobbe B, Denk U, Po¨ll V, Rid R, Breitenbach M (2008) The spectrum of fungal allergy. Int Arch Allergy Immunol 145: 58–86. 3. Horner WE, Helbling A, Salvaggio JE, Lehrer SB (1995) Fungal allergens. Clin Microbiol Rev 8: 161–179. 8. MH Abu-Dieyeh, R Barham, K Abu-Elteen, R Al-Rashidi (2010) Seasonal variation of fungal spore populations in the atmosphere of Zarqa area, Jordan. Aerobiologia 26: 263–276. 4. Esch RE (2004) Manufacturing and standardizing fungal allergen products. J Allergy Clin Immunol 113: 210–215. 4. Esch RE (2004) Manufacturing and standardizing fungal allergen products. J Allergy Clin Immunol 113: 210–215. 5. Bush RK, Yunginger JW (1987) Standardization of fungal allergens. Clin Rev Allergy 5: 3–21. 5. Bush RK, Yunginger JW (1987) Standardization of fungal allergens. Clin Rev Allergy 5: 3–21. 7 PLOS ONE \| www.plosone.org July 2014 \| Volume 9 \| Issue 7 \| e103488 Transaldolase Allergen of Fusarium proliferatum Transaldolase Allergen of Fusarium proliferatum Crameri R, Zeller S, Glaser AG, Vilhelmsson M, Rhyner C (2009) Cross- reactivity among fungal allergens: a clinically relevant phenomenon? Mycoses 52: 99–106. 29. Appenzeller U, Meyer C, Menz G, Blaser K, Crameri R (1999) IgE-mediated reactions to autoantigens in allergic diseases. Int Arch Allergy Immunol 118: 193–196. 18. Shen HD, Tam MF, Chou H, Han SH (1999) The importance of serine proteinases as aeroallergens associated with asthma. Int Arch Allergy Immunol 119: 259–264. 30. Valenta R, Mittermann I, Werfel T, Garn H, Renz H (2009) Linking allergy to autoimmune disease. Trends Immunol 30: 109–116. 19. Chou H, Tam MF, Chiang CH, Chou CT, Tai HY, et al. (2011) Transaldolases are novel and IgE cross-reacting fungal allergens. Clin Exp Allergy 41: 739–749. 31. Mayer C, Appenzeller U, Seelbach H, Achatz G, Oberkofler H, et al. (1999) Humoral and cell-mediated autoimmune reactions to human acidic ribosomal P2 protein in individuals sensitized to Aspergillus fumigatus P2 protein. J Exp Med 189: 1507–1512. 20. Chou H, Tam MF, Lee LH, Chiang CH, Tai HY, et al. (2008) Vacuolar serine protease is a major allergen of Cladosporium cladosporioides. Int Arch Allergy Immunol 146: 277–286. PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org July 2014 \| Volume 9 \| Issue 7 \| e103488 8
https://openalex.org/W1820383962	https://ojs.unud.ac.id/index.php/mtk/article/download/7821/5899	Indonesian	null	PENERAPAN METODE LEAST MEDIAN SQUARE-MINIMUM COVARIANCE DETERMINANT (LMS-MCD) DALAM REGRESI KOMPONEN UTAMA	E-Jurnal Matematika	2,013	cc-by	1,585	ISSN: 2303-1751 ISSN: 2303-1751 E-Jurnal Matematika Vol. 2, No.4, Nopember 2013, 6- 10 I PUTU EKA IRAWAN1, I KOMANG GDE SUKARSA2, NI MADE ASIH3 1,2,3Jurusan Matematika FMIPA Universitas Udayana, Bukit Jimbaran-Bali e-mail: 1ekaone17@yahoo.co.id, 2sukarsakomang@yahoo.co.id, 3asihmath77@gmail.com Keywords: Multicollinearity, Outlier, Principal Component Regression, LMS, MCD Keywords: Multicollinearity, Outlier, Principal Component Regression, LMS, MCD PENERAPAN METODE LEAST MEDIAN SQUARE-MINIMUM COVARIANCE DETERMINANT (LMS-MCD) DALAM REGRESI KOMPONEN UTAMA I PUTU EKA IRAWAN1, I KOMANG GDE SUKARSA2, NI MADE ASIH3 1 Mahasiswa Jurusan Matematika FMIPA Universitas Udayana 2,3 Staf Pengajar Jurusan Matematika FMIPA Universitas Udayana Abstract Principal Component Regression is a method to overcome multicollinearity techniques by combining principal component analysis with regression analysis. The calculation of classical principal component analysis is based on the regular covariance matrix. The covariance matrix is optimal if the data originated from a multivariate normal distribution, but is very sensitive to the presence of outliers. Alternatives are used to overcome this problem the method of Least Median Square-Minimum Covariance Determinant (LMS-MCD). The purpose of this research is to conduct a comparison between Principal Component Regression (RKU) and Method of Least Median Square - Minimum Covariance Determinant (LMS-MCD) in dealing with outliers. In this study, Method of Least Median Square - Minimum Covariance Determinant (LMS- MCD) has a bias and mean square error (MSE) is smaller than the parameter RKU. Based on the difference of parameter estimators, still have a test that has a difference of parameter estimators method LMS-MCD greater than RKU method. 2,3 Staf Pengajar Jurusan Matematika FMIPA Universitas Udayana 1 Mahasiswa Jurusan Matematika FMIPA Universitas Udayana Metode Minimum Covariance Determinant (MCD) Pada Metode ini dicari matriks varians-kovarians dari setiap himpunan bagian data yang banyak elemennya ℎ′, dengan ℎ′ = 𝑛+𝑝+1 2 . Selanjutnya diantara matriks varian-kovarians tersebut dipilih yang memiliki determinan terkecil, dengan 𝑝 menyatakan banyak variabel dan 𝑛 menyatakan banyak pengamatan [1]. Metode Least Median Square (LMS) Salah satu metode regresi robust yang juga sering digunakan dalam analisis regresi adalah metode LMS (Least Median Square). Metode ini mempunyai keuntungan untuk mengurangi pengaruh dari sisaan (residual) terhadap keakuratan koefisien regresi. Penduga LMS diperoleh dengan mencari model regresi yang meminimumkan median kuadrat sisaan (𝑒𝑖 2) dari himpunan- himpunan bagian pada data yang masing-masing sebanyak ℎ, dengan 𝑒𝑖 2 = 𝑦𝑖−𝑥𝑖 𝑇𝑏 2 [3]. Nilai ℎ ditentukan dengan menggunakan rumus ℎ= 𝑛 2 + (𝑝+1) 2 , dengan 𝑛= ukuran sampel (banyaknya data), 𝑝= banyaknya parameter. 1. Pendahuluan Multikolinearitas (multicollinearity) merupakan suatu kondisi pada regresi linear berganda dengan variable-variabel bebas saling berkorelasi. Regresi Komponen Utama merupakan metode yang mengatasi multikolinearitas dengan cara mengombinasikan teknik analisis komponen utama dengan analisis regresi. Pada analisis komponen utama klasik, perhitungannya didasarkan pada matriks kovarian biasa. Matriks kovarian ini akan optimal apabila data berasal dari suatu distribusi normal multivariat, tetapi sangat sensitif terhadap adanya pencilan (oulier) [2]. Alternatif yang digunakan untuk mengatasi masalah ini adalah Metode Least Median Square-Minimum Covariance Determinant (LMS-MCD). Adapun tujuan dalam penelitian ini adalah melihat nilai bias dan MSE parameter dari metode RKU setelah matriks kovarian biasa diganti dengan matriks kovarian 6 e-Jurnal Matematika Vol. 2, No. 4, Nopember 2013, 6-10 robust (LMS-MCD). Pada penelitian ini, setelah matriks kovarian biasa yang digunakan pada RKU diganti dengan matriks kovarian robust (LMS-MCD) memiliki nilai bias dan mean square error (MSE) parameter lebih kecil dibandingkan dengan RKU. Berdasarkan selisih penduga parameter dari 100 ulangan, masih ada ulangan yang memiliki selisih penduga parameter dari metode LMS-MCD yang lebih besar dari metode RKU. 2. Metode Penelitian Data yang digunakan adalah data simulasi yang diperoleh dengan membangkitkan data yang berdistribusi normal dengan menggunakan bantuan program R i386 2.15.2. Penyelesaian Analisis Komponen Utama dan Least Median Square – Minimum Covariance Determinant (LMS-MCD) pada data yang dibangkitkan menggunakan software MINITAB 15 dan software R i386 2.15.2. Pada simulasi data ini, banyaknya sampel yang digunakan adalah 100 amatan (dinotasikan 𝑛= 100) . Sedangkan presentase pencilan yang digunakan adalah 10%, 15%, 𝑑𝑎𝑛 20%. Adapun tahapan yang dilakukan dalam penelitian ini adalah sebagai berikut: 1. Membangkitkan nilai sisaan 𝑒 , 𝑋1~ 50,5 , 𝑋2~ 100,10 , 𝑋3~(30,5), 𝑋4 = 𝑋1 + 𝑋3 + 𝑒, Menentukan nilai 𝑌 dengan cara 𝑌= 𝛽0 + 𝛽1𝑋1 + 𝛽2𝑋2 + 𝛽3𝑋3 + 𝛽4𝑋4 + 𝑒. Dimana 𝛽0 = 2 dan 𝛽1 = 𝛽2 = 𝛽3 = 𝛽4 = 1 2. Melakukan regresi komponen utama pada 𝑋1, 𝑋2, 𝑋3, 𝑋4 terhadap Y, sehingga diperoleh nilai penduga yang fit (𝛽 𝑓𝑖𝑡) dengan menggunakan R i386 2.15.2. 2. Melakukan regresi komponen utama pada 𝑋1, 𝑋2, 𝑋3, 𝑋4 terhadap Y, sehingga diperoleh nilai penduga yang fit (𝛽 𝑓𝑖𝑡) dengan menggunakan R i386 2.15.2. 3. Membangkitkan pencilan yang berdistribusi 𝑁(40,0.05) sebanyak 10% dari banyak amatan 𝑛= 100 dengan menggunakan program R i386 2.15.2. dan 3. Membangkitkan pencilan yang berdistribusi 𝑁(40,0.05) sebanyak 10% dari banyak amatan 𝑛= 100 dengan menggunakan program R i386 2.15.2. dan 7 Penerapan Metode LMS – MCD dalam Regresi Komponen Utama I Putu Eka Irawan, Komang Gde Sukarsa, Ni Made Asih memasukan pencilan pada data simulasi dengan cara menjumlahkan pencilan pada masing-masing variabel bebas yaitu 𝑋1, 𝑋2, 𝑋3 memasukan pencilan pada data simulasi dengan cara menjumlahkan pencilan pada masing-masing variabel bebas yaitu 𝑋1, 𝑋2, 𝑋3 4. Menghitung nilai 𝑌𝑝 dari peubah bebas yang sudah terkontaminasi pencilan. Dengan 𝑌𝑝= 𝛽0 + 𝛽1𝑋1 + 𝛽2𝑋2 + 𝛽3𝑋3 + 𝛽4𝑋4 + 𝑒, 𝛽0 = 2 dan 𝛽1 = 𝛽2 = 𝛽3 = 𝛽4 = 1. 5. Melakukan analisis Regresi Komponen Utama pada data yang sudah terkontaminasi pencilan dengan menggunakan software R i386 2.15.2 sehingga diperoleh 𝛽 (𝑅𝐾𝑈)𝑝 6. Melakukan analisis data menggunakan R i386 2.15.2 dengan menggunakan metode Least Median Square – Minimum Covariance Determinant (LMS- MCD) sehingga diperoleh 𝛽 (𝐿𝑀𝑆−𝑀𝐶𝐷)𝑝 . 7. Ulangi langkah 3 sampai 6 untuk presentase 15%, 20%. 8. Menganalisa dan membandingkan hasil perhitungan Regresi Komponen Utama dengan Least Median Square-Minimum Covariance Determinant (LMS-MCD) berdasarkan bias parameter dan Mean Square Error (MSE). 3. Hasil dan Pembahasan Teknik perhitungan analisis komponen utama menggunakan matriks kovarian biasa. Matriks kovarian ini akan optimal apabila data berasal dari suatu distribusi normal multivariat. Akan tetapi, matriks kovarian ini sensitif terhadap adanya pencilan. Apabila pada data terdapat pencilan, maka matriks kovarian akan kehilangan efisiensinya, sehingga diperlukan suatu matriks kovarian yang robust terhadap pencilan. Penaksir robust yang digunakan untuk kovarian adalah Minimum Covariance Determinant (MCD). Setelah matriks kovarian biasa pada metode RKU diganti dengan matriks kovarian robust, maka diperoleh komponen- komponen utama yang robust. Selanjutnya komponen-komponen utama yang robust ini diregresikan terhadap peubah tidak bebas dengan menggunakan metode Least Median Square. Berikut hasil perhitungan bias parameter dari metode RKU sebelum matriks kovarian biasa diganti dengan matriks kovarian robust dan sesudah matriks kovarian biasa diganti dengan matriks kovarian robust (LMS-MCD) yang disajikan dalam tabel 1. 8 8 e-Jurnal Matematika Vol. 2, No. 4, Nopember 2013, 6-10 Tabel 1. Hasil Perhitungan Bias Parameter untuk 100 kali Ulangan Presentase Pencilan Metode Intercept PC1 PC2 PC3 PC4 10% RKU 20,0055 1,2313 0,5453 0,0308 0,4356 LMS-MCD 1,7792 0,0119 0,1443 0,0199 0,3892 15% RKU 30,0025 1,5189 0,6074 0,3046 0,5959 LMS-MCD 6,7617 0,0883 0,0272 0,0308 0,4804 20% RKU 40,0035 1,0302 0,2414 1,4092 1,1498 LMS-MCD 13,1999 0,0324 0,0826 0,0673 0,685 Tabel 1. Hasil Perhitungan Bias Parameter untuk 100 kali Ulangan Hasil perhitungan MSE parameter dapat dilihat pada tabel 2. Tabel 2. Hasil Perhitungan Mean Square Error Parameter untuk 100 kali Ulangan Presentase Pencilan Metode Intercept PC1 PC2 PC3 PC4 10% RKU 400,222 3,7153 3,2863 0,1961 2,4759 LMS-MCD 6,3224 1,1183 2,7129 0,0523 1,7137 15% RKU 900,152 3,6949 3,15 0,4917 3,1602 LMS-MCD 48,8361 1,761 2,8019 0,0324 2,0185 20% RKU 1600,28 2,9369 2,916 2,0797 2,7384 LMS-MCD 176,411 1,8399 2,5413 0,0437 2,1958 Tabel 2. Hasil Perhitungan Mean Square Error Parameter untuk 100 kali Ulangan Berdasarkan Tabel 1 dan 2, terlihat bahwa pada presentase pencilan 10%, 15%, dan 20% nilai bias parameter dan mean square error (MSE) parameter dari metode RKU sebelum matriks kovarian biasa diganti dengan matriks kovarian robust selalu lebih besar dibandingkan dengan nilai bias parameter dan mean square error (MSE) parameter dari metode RKU sesudah matriks kovarian biasa diganti dengan matriks kovarian robust (LMS-MCD). Selain itu, perbandingan untuk kedua metode juga dilakukan dengan menghitung selisih dari penduga parameter yang dihasilkan oleh kedua metode dengan penduga parameter dari metode RKU pada data tanpa pencilan. 3. Hasil dan Pembahasan Setelah dilakukan pengulangan 100 kali dan dihitung selisih penduga parameter dari masing-masing ulangan pada kedua metode, diperoleh bahwa secara umum metode LMS-MCD lebih bagus daripada metode RKU. 4. Kesimpulan Berdasarkan hasil yang diperoleh dapat disimpulkan, untuk presentase pencilan 10%, 15%, dan 20% pada data simulasi yang dibuat peubah-peubah bebasnya bermultikolinearitas menunjukkan bahwa metode RKU sesudah matriks kovarian biasa diganti dengan matriks kovarian robust (LMS-MCD) menunjukkan hasil yang lebih baik dibandingkan dengan metode RKU sebelum matriks kovarian biasa diganti dengan matriks kovarian robust. Untuk penelitian 9 I Putu Eka Irawan, Komang Gde Sukarsa, Ni Made Asih Penerapan Metode LMS – MCD dalam Regresi Komponen Utama selanjutnya disarankan menambah jumlah ulangan dalam simulasi dan membandingkan peubah bebas yang dibangkitkan dari sebaran normal yang mengalami multikolinearitas dengan peubah bebas yang dibangkitkan dari sebaran data normal multivariat. [3] Rousseeuw, P.J. 1984. Least Median of Square Regresson. Journal of the American Statistical Association. Daftar Pustaka [1] Aderlina, D. 2011. Metode Minimum Covariance Determinant pada Analisis Regresi Linear Berganda dengan Kasus Pencilan. Skripsi. Jurusan Matematika F MIPA Universitas Padjadjaran. [2] Notiragayu. 2008. Pembandingan Beberapa Metode Analisis Regresi Komponen Utama Robust. Prosiding Seminar Hasil Penelitian dan Pengabdian kepada Masyarakat, Universitas Lampung. [2] Notiragayu. 2008. Pembandingan Beberapa Metode Analisis Regresi Komponen Utama Robust. Prosiding Seminar Hasil Penelitian dan Pengabdian kepada Masyarakat, Universitas Lampung. [3] Rousseeuw, P.J. 1984. Least Median of Square Regresson. Journal of the American Statistical Association. 10 10
https://openalex.org/W3049492208	https://researchonline.jcu.edu.au/64111/1/64111.pdf	English	null	A Game-Based Approach to Exploring Gender Differences in Smallholder Decisions to Change Farming Practices: White Rice Production in Laos	Sustainability	2,020	cc-by	16,415	Sustainability 2020, 12, 6594; doi:10.3390/su12166594 A Game-Based Approach to Exploring Gender Diﬀerences in Smallholder Decisions to Change Farming Practices: White Rice Production in Laos Silva Larson 1,* , Anne (Giger)-Dray 2,3,* , Tina Cornioley 2,3, Manithaythip Thephavanh 4,5, Phomma Thammavong 6, Sisavan Vorlasan 7, John G. Connell 8, Magnus Moglia 9 , Peter Case 8,10 , Kim S. Alexander 8 and Pascal Perez 11 1 School of Science and Engineering, University of Sunshine Coast, Sippy Downs, QLD 4556, Australia 2 Forest Management and Development Group (ForDev), Institute of Terrestrial Ecosystems, ETH Zürich, CH-8092 Zurich, Switzerland; t.cornioley@gmail.com 1 School of Science and Engineering, University of Sunshine Coast, Sippy Downs, QLD 4556, Australia 1 School of Science and Engineering, University of Sunshine Coast, Sippy Downs, QLD 4556, Australia 2 Forest Management and Development Group (ForDev), Institute of Terrestrial Ecosystems, ETH Zürich, CH-8092 Zurich, Switzerland; t.cornioley@gmail.com g g y ppy 2 Forest Management and Development Group (ForDev), Institute of Terrestrial Ecosystems, ETH Zü CH-8092 Zurich, Switzerland; t.cornioley@gmail.com 3 Institute of Terrestrial Ecosystems (ITES), ETH Zürich, CH-8092 Zurich, Switzerland 4 School of Agriculture, Food and Wine, University of Adelaide, Adelaide, SA 5064, Australia; thephavanh.mntt@gmail.com 3 Institute of Terrestrial Ecosystems (ITES), ETH Zürich, CH-8092 Zurich, Switzerland 4 School of Agriculture, Food and Wine, University of Adelaide, Adelaide, SA 5064, Australia; thephavanh.mntt@gmail.com 5 National Agriculture and Forestry Research Institute (NAFRI), Nongviengkham Village, Xaythany Distric Vientiane Capital P.O. Box 7170, Laos 5 National Agriculture and Forestry Research Institute (NAFRI), Nongviengkham Village, Xaythany District, Vientiane Capital P.O. Box 7170, Laos 6 Faculty of Agriculture, National University of Laos, Nabong Campus, Pakxap-Mai Villag Xaythany District, Vientiane Capital P.O. Box 7322, Laos; phommath@gmail.com Faculty of Agriculture, National University of Laos, Nabong Campus, Pakxap-Mai Vi Xaythany District, Vientiane Capital P.O. Box 7322, Laos; phommath@gmail.com y g y g p p g Xaythany District, Vientiane Capital P.O. sustainability sustainability sustainability sustainability 1. Introduction What inﬂuences farmers’ decisions to adopt introduced agricultural technologies is a central question for international agricultural research projects. Smallholder farmers in Lao People’s Democratic Republic (Lao PDR), like many others in developing countries, are facing transformative changes in their traditional agriculture-based livelihoods. With approximately 62% of labour contributing to the agriculture sector, Lao PDR is still highly dependent on agriculture [1]. However, processes of intensiﬁcation of agriculture, market integration, industrialization and urbanization are underway [2–6]. Smallholder family farming, understood as a household which combines family, farm and commercial activity, still represents the backbone of the world’s agriculture [7]. Traditionally, subsistence farming was predominant in Laos, with smallholders dependent on cultivatable land for rice and livestock husbandry, whilst other nontimber forest and river products are used as supplementary food sources and marketable goods [8–10]. Some 72% of the total cultivated land area is dedicated to rice production with the traditional glutinous (“sticky”) rice varieties, preferred for local consumption and providing almost 70% of calorie and protein intake [11]. To improve rural livelihoods, however, the Lao government is encouraging farmers to move to commercial agricultural production, with plans to increase both rice yield per hectare and production of nonglutinous (“white”) rice varieties; those demanded by major export markets, e.g., China. The key decision for farmers is whether to continue traditional production of the “sticky” rice and ensure family sustenance or adopt the production of “white” rice for the market; the latter option potentially bringing substantial ﬁnancial beneﬁts. The ﬂip side of the switch to white rice is the threat it carries to preferred subsistence food for the family. Decisions on household agricultural practices are being made in the larger context of industrialization and urbanization, with nontraditional oﬀ-farm and nonfarming activities becoming integral to the way that households generate income [12–15]. Women are a vital part of the agriculture sector in Lao PDR, contributing to every aspect of agricultural production [16]. Similar to other emerging economies in Southeast Asia [17] and elsewhere [18–21], the division of labour in smallholder households is gendered, and men and women tend to have diﬀerent roles and responsibilities. In the rural agricultural context of Lao PDR, data are available that describe the diﬀerences between male- and female-headed households, but little is known about the role and contribution of rural women within male-headed households or the decision-making process in such households [22]. A Game-Based Approach to Exploring Gender Diﬀerences in Smallholder Decisions to Change Farming Practices: White Rice Production in Laos Box 7322, Laos; phommath@gmail.com 7 Provincial Agriculture and Forestry Oﬃce, Thameung Village 13000, Kaysone Phomvihane District Savannakhet Province, Laos; tee.vorlasan@yahoo.com avannakhet Province, Laos; tee.vorlasan@yahoo.com 8 College of Business, Law and Governance, James Cook University, Douglas, QLD 4811, Australia; john.connell@jcu.edu.au (J.G.C.); peter.case@jcu.edu.au (P.C.); kim.alexander56@gmail.com (K.S.A.) 9 Center for Urban Transitions, Faculty of Arts, Social Sciences and Humanities, Swinburne University of Technology, Hawthorn, VIC 3122, Australia; mmoglia@swin.edu.au nter for Urban Transitions, Faculty of Arts, Social Sciences and Humanities, Swinburne University of h l H h VIC 1 A l l @ d 10 Bristol Business School, University of the West of England (UWE), Bristol BS16 1QY, UK SMART Infrastructure Facility, Faculty of Engineering and Information Sciences, University of Wollongong Northﬁelds Av, Wollongong, NSW 2522, Australia; pascal@uow.edu.au * Correspondence: silva.larson@gmail.com (S.L.); anne.dray@usys.ethz.ch (A.D.) Received: 7 July 2020; Accepted: 8 August 2020; Published: 14 August 2020 Abstract: What inﬂuences farmers’ decisions to adopt agricultural technologies is an important question for international agricultural research projects. There are often interpersonal diﬀerences between women and men that inﬂuence the adoption of decisions and behaviours, but few studies in the literature focus on these factors. We describe a game-based approach to explore decision-making processes underpinning the adoption of new farming technologies and practices in Lao People’s Democratic Republic. Sowing a diﬀerent rice variety is the tailored technology. The game explored adoption behaviours inﬂuencing decisions on transitioning from growing glutinous rice, a traditional variety preferred for consumption, to “white” rice for commercial export to international markets. We conducted separate game-workshops with 36 women and 36 men in 4 villages of southern Laos that were transitioning from subsistence to commercial smallholder production. The gaming exposed various possible behaviours and decisions that women and men considered. Access to resources, both assets and information, was equal for all players, yet women were found to adopt new rice varieties more readily than men and to engage in cooperative behaviours in the game situation. The study highlighted the need for further gender-sensitive research into cooperation among women in the agricultural context—an understanding beneﬁcial for countries and regions undergoing agricultural transition. Sustainability 2020, 12, 6594; doi:10.3390/su12166594 www.mdpi.com/journal/sustainability www.mdpi.com/journal/sustainability 2 of 22 Sustainability 2020, 12, 6594 Keywords: gender; rural development; role-playing game; Laos; adoption; agrarian transition: Companion Modelling Keywords: gender; rural development; role-playing game; Laos; adoption; agrarian transition: Companion Modelling 1. Introduction Games also serve as models of strategic situations, allowing participants to anticipate the consequences of actions or decisions [35]. By posing challenges and confronting players with complex situations, games can act as a reality check, whereby participants revisit their pre-existing assumptions, preconceived perceptions, values and knowledge regarding how the “system” works currently as well as how it should work or might work in the future [38,39]. Games, when carefully facilitated, operate as “safe spaces” where participants can voice their motivations behind the plurality of their livelihood strategies, and the drivers and thought-processes that inﬂuence how they react or adapt to changes, without fear of direct consequences or repercussions in the real world [44,45]. Role-playing games have been used in the past to explore agrarian transition and facilitate understanding of farmers’ decision making [46–49]. However, this approach has been underutilized to study the impact of gender in agriculture and landscapes changes [50]. Research techniques commonly used for gender analyses include focus groups, wealth ranking and semistructured or structured interviews [51]. Villamor et al. [52] have expanded the methodological repertoire and broken new ground by using role-playing games to analyse gender diﬀerences in forestry research. In this paper, we transpose a similar game-based approach to analyse the inﬂuence of gender on the adoption of new technologies and farming practices. We report on our investigation of adoption behaviours as complex decision-making processes, using a participatory role-playing game following the Companion Modelling approach [53]. Central to this approach is the coconstruction of the models or games with local experts and stakeholders through an iterative process. In this paper, we aimed to understand and compare the decision-making process and behavioural choices of men and women in a context of agricultural innovations. To do so, we explored the diﬀerent adoption behaviours when men and women are confronted with several new farming practices and technologies. The speciﬁc technological intervention we used to explore adoption behaviours is the transition from growing glutinous (“sticky”) rice for home-consumption to growing “white” rice (Hom Savanh variety) for sale to international markets. To explore the viability, possible implications and likelihood of this transition, we developed and undertook tailored game workshops to elicit speciﬁc behaviours that prevent or facilitate adoption by smallholder farmers of “white” rice for export to China. Gender-speciﬁc workshops were conducted with 72 smallholder farmers (36 females and 36 males) in four villages of southern Laos. 1. Introduction Rapid and uneven economic growth occurring in Southeast Asian countries over recent decades has resulted in new and challenging inequities between social groups and between men and women [23]. There are many examples of young men and women migrating for wages and remittances [24–26], with young migrant women becoming the industrial working class of Southeast Asian cities [23]. Understanding speciﬁc farmers’ behaviours that would prevent or facilitate the adoption of introduced agricultural technologies is important, particularly in terms of understanding possible future farm management scenarios. Ambitious targets for increasing agricultural production and international exports set by the Lao government are based on the premise that agricultural productivity can be boosted by changed farming practices. Modiﬁcation of farming practices through the adoption of technologies often occurs at a much slower rate than hoped [27–32]. The current low adoption rates of new technologies [33], including white rice production, suggest that more needs to be understood about how farmers decide on their production goals and how the oﬀered technologies might ﬁt with such goals [12]. Recognizing the central role of gender in the process of agrarian and also a wider societal transition in Lao PDR, we argue for the need to explore the diﬀerences in behaviours between rural women and men. Sustainability 2020, 12, 6594 3 of 22 The adoption of new practices results from a complex decision process involving empirical knowledge and individual beliefs, perceptions and interests, combined with personal or collective assets in relation to the institutional context. Nonetheless, understanding decision-making processes, and the necessary intermediate step of eliciting mental models, is a challenging task. Game-based approaches have become a popular means in the ﬁeld of social–ecological research to study decision-making processes [34,35] and to understand the heterogeneity of mental model constructs [36]. Garcia et al. [35] argue that role-playing games can help to elicit mental models and illustrate decision-making through concrete representations of the needs, constraints and aspirations of stakeholders. Games allow participants to grasp information more easily [37], empower local communities, elicit local knowledge and decision-making processes [38–40] and illuminate complex socio-ecological systems [41,42]. In addition, games allow researchers to better understand how people devise strategic decisions when confronted with complex, sometimes conﬂicting and uncertain situations. Sensitive issues, such as illegal activities or corruption, that would rarely be communicated in traditional interviews, can emerge implicitly and explicitly during a game session [36,43]. 2.1. Study Context and Game Participants This study was part of a wider research project [55] in which qualitative and quantitative data were collected to better understand the issues implicated in the adoption of new technologies in two provinces of southern Laos: Savannakhet and Champasak. The southern provinces are crucial to the Lao PDR government’s agricultural strategy and planning as they are the site of seven large plains that have been designated as “suitable for intensive crop production”. The research study undertaken in these provinces involved over 950 participants, including 900 farmers and 50 stakeholders involved in data collection exercises [12,14,56–58]. Livelihoods of farmers in selected villages participating in the study were based on rice production and livestock as sources of sustenance and income, with additional income from oﬀ-farm work. Using a Bayesian network analysis, Moglia et al. [14] found that the key opportunities important for improving rice farming practices and increasing “white” rice production in this area were: (1) The use of direct seeding machines that allow farmers to signiﬁcantly reduce the time required to plant their ﬁelds, and hence opens up options for alternative uses of labour. There were several technical concerns about the suitability of direct seeding approaches in terms of its applicability in some geographical areas and resultant weed intrusion. (1) The use of direct seeding machines that allow farmers to signiﬁcantly reduce the time required to plant their ﬁelds, and hence opens up options for alternative uses of labour. There were several technical concerns about the suitability of direct seeding approaches in terms of its applicability in some geographical areas and resultant weed intrusion. 2) Access to the market and cooperative agreements between traders and farmer groups, to m order requirements. (3) The proper application of fertilizer necessary to increase yields. The soils in the area are generally of poor quality and fertilizer is often inadequately applied. The above-mentioned factors served as a basis to create the game elements, as described in the next section. In addition to data collected in the wider project, secondary data [59–63] were used to create the context of the game. Realistic game scenarios were constructed in consultation with a range of experts and in reference to the literature. A total of eight game workshops were undertaken in four villages—Thouat, Phonthong, Alan Wattana and Nonokkhien—located in Savannakhet Province. Overall, 72 smallholder farmers (36 men and 36 women) participated. 2.1. Study Context and Game Participants Provincial and district agricultural extension oﬃcers made all local arrangements, particularly appointments with the village head and stakeholder contacts. In each village, two game sessions took place, one with male participants in the morning and the other with female participants in the afternoon. Each game session lasted 3 h on average. The selection of the participants was facilitated by the Lao Ministry of Agricultural and Forestry oﬃcers and the national research team. Most female participants were aged between 41–60 years with the second-highest participant rate in the 25–40 years age group. Male participants were mostly aged 41–60 years and 25–40 years, but more men over 60 years participated than in the women’s groups. Overwhelmingly, men identiﬁed themselves as the head of the family with only a few sons, a son-in-law and a grandfather involved. Women mainly identiﬁed as the housewife, but 20% of women identiﬁed as the head of the family. 1. Introduction The gaming workshop allowed participants to examine possible consequences of selecting a new technology (white rice variety) in a no-risk, virtual manner [54]. We speciﬁcally explored: (1) potential diﬀerences in incomes achieved; (2) technical production choice (patterns of adoption); (3) the role cooperation plays in reaching targets or maximizing outcomes. In the next section, we introduce the context of rice production in Laos, the participants’ proﬁles, the game elements and the statistical methods employed to analyse the data. We then present the results across three indicators (income, production choices and cooperation) by comparing the male 4 of 22 Sustainability 2020, 12, 6594 and female game sessions. Finally, we discuss the implications for gender-sensitive studies of adoption of new farming practices and the role women play in regions undergoing agricultural transition. 2.2. Game Development A game in the form of “table top” role-playing was developed to mimic an archetypal village in the rice belt of Laos. The purpose of the study was to better understand the gender-speciﬁc interactions between farmers, extension oﬃcers and traders when being introduced to new technologies and farming practices. The game was designed to conﬁrm key issues identiﬁed during research activities that seem to prevent many farmers from adopting new technologies proposed by extension services. 5 of 22 Sustainability 2020, 12, 6594 During the workshop (Figure 1), the participants, in teams of 2 players, made decisions to manage a virtual farm in the face of future seasonal scenarios, with access to a suite of rice crops and farm production options. Players were provided with information about the cost of each option and expected returns or beneﬁts. The objective of the game was to demonstrate conditions under which new technologies could be adopted by a large number of farmers and how farmers, traders and extension oﬃcers could work together to ensure the uptake of new technologies. The games were centred on switching traditional “sticky rice” production to “white rice” for commercial sale. Players were also presented with the option of growing an “improved sticky rice” variety to create a portfolio of choices and options (and move away from a purely binary system of sticky rice versus white rice). As the game unfolded, players could also access a set of accompanying appropriate technologies such as farm machinery (direct seeding machine) and fertilizers. During the game, details were provided about the new technologies, weather forecasts and changing prices of commodities. At the end of each decision-making period, and again, at the end of each game year, players reviewed and discussed the consequences of their decisions on farm proﬁtability and improvement (or not) to their livelihoods. All game sessions were directly followed by a debrieﬁng phase when players were given the opportunity to reﬂect on the gaming experience [45]. The debrieﬁng was typically structured around the following questions: “how do you feel?”, “what happened during the game?”, “why did it happen?”, “what would you do diﬀerently?”, “how is the game reﬂecting (or not) your reality?” Game sessions were run by trained facilitators and presented in Lao language. Hereafter, we detail the players’ proﬁles and their objectives, the gamescape (i.e., the landscape of the game board) and the sequence of actions taken by players. 2.2.1. Player’s Proﬁles and Objectives The game included three diﬀerent players’ proﬁles: farmers, agricultural extension services and traders. The game was tailored to be played by 7 teams as follows: The game included three diﬀerent players’ proﬁles: farmers, agricultural extension services and traders. The game was tailored to be played by 7 teams as follows: • Five teams of farmers (with 2 participants per team when available): players were practicing male and female farmers from the participating villages. They all had experience with growing sticky rice, and white rice was grown in the villages by some but not all of them. The objective of each farmer’s team was to feed her/his family and pay “taxes” to cover school, health and electricity expenses. To be able to do so, farmers had to produce rice (either sticky, improved variety or white rice) and earn money tokens. They could also sell their livestock or send their labour forces (i.e., labour tokens) to Thailand. • A team of extensionists: players on this team were the local District Agriculture and Forestry Oﬃcers; hence, they were familiar with the participating village’s production and socio-cultural systems. They were also familiar with the new technologies being proposed in the game, as these technologies were selected based on their “real life” relevance. The extension oﬃcers’ team was given “the mission” to convince as many farmers as possible to adopt one (or more) of the following new technologies: (1) switching to “improved sticky rice”, (2) use of direct seeding machine and (3) use of high quality fertilizer. To encourage the uptake of new technologies, extension oﬃcers could distribute free “technical packages” (such as free fertilizer or use of the direct seeding machines); however, the distribution of “goods” for free per round was limited due to the limited (imaginary) budget of the District Agriculture and Forestry Oﬃce. No rules were set on how to distribute these “goods”; it was entirely up to the extension oﬃcers to decide on the process. A trader: this was the only role in the game that was not played by the “actual” actor. Rather, the role of the trader was scripted and performed by a member of the research team. Her/his job was to sell inputs and new technologies, such as improved sticky rice seeds, white rice seeds, fertilizer, direct seeding machine and cattle and fodder, to the farmer, enabling the farmer to make management and adoption choices. 2.2.2. The Gameboard 2 2 2 The Gameboard The game consisted of ﬁve individual gameboards—one for each participating farmer team of two players—that in their entirety embodied a typical village in the area under study and its surrounding farmland during the rainy season. A single board represented one farm and consisted of the following seven zones (from bottom to top, Figure 1b): The game consisted of five individual gameboards—one for each participating farmer team of two players—that in their entirety embodied a typical village in the area under study and its surrounding farmland during the rainy season. A single board represented one farm and consisted of the following seven zones (from bottom to top, Figure 1(b).): 1. An irrigated rice paddy ﬁeld located near the river (lowest area on the board). It had plenty of water, but in years with too much rain, it could be ﬂooded. 1. An irrigated rice paddy field located near the river (lowest area on the board). It had plenty of water, but in years with too much rain, it could be flooded. 2. Another irrigated paddy ﬁeld, but in years with a low amount of rain, the irrigation system could not provide enough water to this area. water, but in years with too much rain, it could be flooded. 2. Another irrigated paddy field, but in years with a low amount of rain, the irrigation system could not provide enough water to this area. 3. A rainfed paddy ﬁeld, where rice production depended on the quantity of rainfall during the rainy season. The soil was good. p g 3. A rainfed paddy field, where rice production depended on the quantity of rainfall during the rainy season. The soil was good. 4. Another rainfed paddy ﬁeld, where rice production depended on the quantity of rainfall during the rainy season. This area had bad soil. y g 4. Another rainfed paddy field, where rice production depended on the quantity of rainfall during the rainy season. This area had bad soil. 5. Grazing land near the farm. y 5. Grazing land near the farm. 6. The farmhouse. 6. The farmhouse. 7. This is a special area; it represented a factory in Thailand where players could send their family members (in the form of labour pawns) for oﬀ-farm income. 7. 2.2.1. Player’s Proﬁles and Objectives The trader appeared at the beginning of each round when he/she would pay a visit to the village. The trader re-appeared at the end of each turn and could A trader: this was the only role in the game that was not played by the “actual” actor. Rather, the role of the trader was scripted and performed by a member of the research team. Her/his job was to sell inputs and new technologies, such as improved sticky rice seeds, white rice seeds, fertilizer, direct seeding machine and cattle and fodder, to the farmer, enabling the farmer to make management and adoption choices. The trader appeared at the beginning of each round when he/she would pay a visit to the village. The trader re-appeared at the end of each turn and could 6 of 22 Sustainability 2020, 12, 6594 purchase the outputs from the farmers (rice, cattle). The trader ﬁxed the prices and expected quantities for the white rice at the start of the game. tainability 2020, 12, x FOR PEER REVIEW 6 of 23 purchase the outputs from the farmers (rice, cattle). The trader ﬁxed the prices and expected quantities for the white rice at the start of the game. stainability 2020, 12, x FOR PEER REVIEW 6 of 23 2.2.3. Rounds and Sequence of Events 2.2.3. Rounds and Sequence of Events Each round of the game represented one year, with only the rainy season being actively played, reflecting an annual single wet season crop from paddy production. Each gaming session consisted of five rounds (corresponding to 5 years). A description of the game elements and the sequences of events can be found in Appendix A. The game rules were as follows: Each round of the game represented one year, with only the rainy season being actively played, reﬂecting an annual single wet season crop from paddy production. Each gaming session consisted of ﬁve rounds (corresponding to 5 years). A description of the game elements and the sequences of events can be found in Appendix A. The game rules were as follows: • Rainfall: The timing of weather events was scripted and was identical across all workshops. The weather conditions were either ”good” or ”bad”, leading to different yields on the rainfed fields. ”Good” meant positive yields (corresponding to an average wet season), and ”bad” meant reduced yields (corresponding to a year with little rain). Additionally, floods could occur, influencing the lower irrigated rice paddy. Rounds 1, 3 and 5 were good rainy years, round 2 was a year with little rain, while flooding occurred in round 4. • Rainfall: The timing of weather events was scripted and was identical across all workshops. The weather conditions were either “good” or “bad”, leading to diﬀerent yields on the rainfed ﬁelds. “Good” meant positive yields (corresponding to an average wet season), and “bad” meant reduced yields (corresponding to a year with little rain). Additionally, ﬂoods could occur, inﬂuencing the lower irrigated rice paddy. Rounds 1, 3 and 5 were good rainy years, round 2 was a year with little rain, while ﬂooding occurred in round 4. y , g • Players’ tokens and actions: Farmers started with two types of tokens: labour pawns and money. Each farmer received eight labour pawns, with one pawn representing approximately 10 days work and eight money tokens of 500,000 Lao Kip each (~USD 55). Players could allocate labour entirely to one activity or a combination of: (1) rice cultivation on one of the four fields; (2) cattle herding; (3) off-farm work in Thailand; or (4) staying at home. • Players’ tokens and actions: Farmers started with two types of tokens: labour pawns and money. 2.2.2. The Gameboard 2 2 2 The Gameboard This is a special area; it represented a factory in Thailand where players could send their family members (in the form of labour pawns) for off-farm income. (a) Figure 1. Cont. (a) Figure 1. Cont. 7 of 22 7 of 23 Sustainability 2020, 12, 6594 Sustainability 2020, 12, x FOR (b) Figure 1. (a) Alan Wattana session with women paired in teams of 2; (b) single board for ”blue” player with the 7 areas (labelled from 1 to 7) and recording decisions grid filled by the observer. 1 2 3 4 5 6 7 Figure 1. (a) Alan Wattana session with women paired in teams of 2; (b) single board for “blue” player with the 7 areas (labelled from 1 to 7) and recording decisions grid ﬁlled by the observer. (b) Figure 1. (a) Alan Wattana session with women paired in teams of 2; (b) single board for ”blue” player with the 7 areas (labelled from 1 to 7) and recording decisions grid filled by the observer. Figure 1. (a) Alan Wattana session with women paired in teams of 2; (b) single board for “blue” player with the 7 areas (labelled from 1 to 7) and recording decisions grid ﬁlled by the observer. 2.2.4. Scenarios and the Special Focus on White Rice In each round, farmers had to allocate their labour/resources across their land plots and on the Thailand plot in case they opted for oﬀ-farm income. As detailed in 2.2 (above), the technical innovations that could occur in the game were based on the previous project ﬁndings, literature and consultations with the stakeholders and included the use of fertilizers, improved rice varieties and direct seeding machine. Technical innovations were introduced by the extension team along with the provision of sample inputs they could give to one of the ﬁve farmers’ teams of their choice. These were also available for sale by the trader. From rounds 2 to 4, a diﬀerent climatic or economic scenario that would aﬀect outcomes was introduced: drought in round 2 (lower yields on rainfed soils), global ﬁnancial crisis in round 3 (no oﬀ-farm income), ﬂood in round 4 (no yield on lower paddy). Additionally, from round 3 onwards, the economic scenario of interest was the traders’ pressure and opportunity to sell “white rice” to the market. The trader oﬀered the player a premium price if they could collectively produce a minimum of 29 bags of “white rice”. This threshold was established during the game calibration and crash-tests. Note that this scenario directly conﬂicted with the extensionist objective of promoting an improved rice variety for food security. 2.2.3. Rounds and Sequence of Events 2.2.3. Rounds and Sequence of Events Each farmer received eight labour pawns, with one pawn representing approximately 10 days work and eight money tokens of 500,000 Lao Kip each (~USD 55). Players could allocate labour entirely to one activity or a combination of: (1) rice cultivation on one of the four ﬁelds; (2) cattle herding; (3) oﬀ-farm work in Thailand; or (4) staying at home. 8 of 22 Sustainability 2020, 12, 6594 • Rice cultivation: Players chose between the traditional sticky rice, improved sticky rice and white rice. The game started with traditional sticky rice production only. In year two, improved sticky rice, which produced 1.5 more bags of rice than its traditional counterpart, was also available. White rice for the sale (commercial) production option was introduced into the game from year 3 onwards. Farmers could also choose to use external inputs, including a direct seeding machine (reducing workload) and fertilizers. • Cattle fattening: Each farmer initially received one cow token. Cattle could have 4 stages of fattening, represented by the ﬁelds coloured on the cow-card. The farmers could choose to fatten their cattle. To do so, they could either place the cow token on a grazing area or feed the cow token with fodder bought from the trader. If fattening the cattle, they had to allocate one labour pawn to remain with the cow token. The price received at sale depended on the stage of fattening of the cow token. • Players’ income: Players could generate income by selling rice, cattle or working oﬀ-farm. Sticky rice (whether from a traditional or improved seed) was sold for the ﬁxed price of 1 unit of money per bag of rice. White rice was sold for the same price, except when all the players succeeded in cumulatively producing at least 29 bags of white rice. In this case, they could sell white rice for a price premium of 1.5 unit of money per bag of rice. 2.2.4. Scenarios and the Special Focus on White Rice 2.3. Data Analysis Player’s decisions were documented throughout the game (Figure 1). Each pair of players’ income from sticky rice, white rice, oﬀ-farm work and selling cattle, as well as their total income, were recorded at each round. We also recorded at the end of each round: the rice type cultivated on each plot; fertilizer use; the cumulative number of bags of white rice produced; the cumulative number of direct seeding machines owned (DS machine owned); the fraction of those that were shared (DS machine sharing). Placement of these resources and the outcome of each season were carefully documented using a decision grid (Figure 1). The grid provided a numerical dataset that was used to analyse behavioural patterns. 9 of 22 Sustainability 2020, 12, 6594 We ﬁtted six linear mixed-eﬀects (lme) models to explore potential diﬀerences in strategies and outcomes between men and women over 5 rounds to account for the random eﬀects of the village and player ID (i.e., the team colour). One model was ﬁtted for each response variable of interest, namely total income, white rice income, sticky rice income, production of white rice, DS machine owned and DS machine sharing. The analyses were performed in R [64] with the lme4 package [65]. When the response variable was measured per round, per village and per player, we accounted for the random eﬀects of the village and player ID nested within the village. When the response variable was measured per round, per village and not per player, then we accounted for only one random eﬀect: village. To explore diﬀerences in income, we considered the ﬁxed eﬀects of the round of the game (rounds 1–5), gender and their interaction. For the models “white rice income” and “sticky rice income”, we further accommodated a quadratic eﬀect of the round of the game and its interaction with gender. For the models “white rice production”, “DS machine owned” and “DS machine sharing”, we accounted for a ﬁxed eﬀect for round and gender but not their interaction, as including the interaction led to model overﬁtting. We ﬁtted a lme model with a Poisson distribution for “DS machine owned” and binomial logistic regression for “DS machine sharing”. For the other lme models, we used a Gaussian distribution. We additionally ﬁtted two ﬁxed-eﬀect-only binomial logistic regressions instead of mixed eﬀect models, as the same models with random eﬀects did not converge [66]. 3. Results We reported on the following variables of interest from the “table top” role-playing game: income, production choices (adoption in terms of technical decision-making patterns—plot selection and use of fertilizer) and the role cooperation plays in maximizing outcomes. A number of other game elements, as presented in Appendix A, is not reported here, as no diﬀerences between male and female reactions were noted. We do, however, encourage further explorations of such themes. Statistics of the results presented here, and additional ﬁgures can be found in Appendix B. 2.3. Data Analysis To investigate potential diﬀerences between men and women in white rice cultivation, we ﬁtted a model on white rice cultivation explained by gender, round, plot properties and all possible two-way and three-way interactions. We further ﬁtted a regression on fertilizer use to assess the probability of using fertilizer on a plot depending on gender, round, plot properties and rice type. Interactions were not considered because of overﬁtting. 3.1. Income The income of both men and women increased over time and women achieved a higher income than men as the game progressed (Figure 2: intercept (men) = 3.760, se = 1.366, intercept (women vs. men) = −0.821, se = 1.520, slope (round) = 2.700, se = 0.323, slope (round women vs men) = 0.906, se = 0.460). Figure 2 shows the output of the mixed-eﬀect linear model with a Poisson distribution including Gender and Round as ﬁxed eﬀects, and village and players nested within the village as random eﬀects. 10 of 22 10 of 23 Sustainability 2020, 12, 6594 Sustainability 2020, 12, x FOR Figure 2. Income per player over rounds for men (20 teams) and women (20 teams). The dots represent not one observation but a group of observations sharing the same values. The size of the dot is proportional to the number of observations (n) with the same x and y values. Some of the dots for males are behind the dots for females (see transparency). The lines are predictions from a linear mixed effect model while the grey ribbons represent the corresponding 95% confidence intervals from the fixed effects. Figure 2. Income per player over rounds for men (20 teams) and women (20 teams). The dots represent not one observation but a group of observations sharing the same values. The size of the dot is proportional to the number of observations (n) with the same x and y values. Some of the dots for males are behind the dots for females (see transparency). The lines are predictions from a linear mixed eﬀect model while the grey ribbons represent the corresponding 95% conﬁdence intervals from the ﬁxed eﬀects. Figure 2. Income per player over rounds for men (20 teams) and women (20 teams). The dots represent not one observation but a group of observations sharing the same values. The size of the dot is proportional to the number of observations (n) with the same x and y values. Some of the dots for males are behind the dots for females (see transparency). The lines are predictions from a linear mixed effect model while the grey ribbons represent the corresponding 95% confidence intervals from the fixed effects. Figure 2. Income per player over rounds for men (20 teams) and women (20 teams). The dots represent not one observation but a group of observations sharing the same values. 3.1. Income The size of the dot is proportional to the number of observations (n) with the same x and y values. Some of the dots for males are behind the dots for females (see transparency). The lines are predictions from a linear mixed eﬀect model while the grey ribbons represent the corresponding 95% conﬁdence intervals from the ﬁxed eﬀects. Exploring the income outcomes more closely, we checked the money units generated from four sources of income (Figure 3): white rice (mean = 5.04 money units, sd = 5.84); sticky rice (mean = 4.80 money units, sd = 3.20); selling cattle (mean = 2.00 money units, sd = 2.73); off-farm workers (mean = 0.87 money units, sd = 1.09). We observed that before white rice was introduced in round 3, traditional ”sticky” rice (Figure 3c) was the main source of income for both men and women. White rice was introduced in round 3, and from then on, the majority of players’ income stemmed from white rice (Figure 3d). The income generated by selling cattle (Figure 3a) and from off-farm workers (Figure 3b) was marginal compared to income from selling white rice. Therefore, we only further investigated the trends in income from sticky rice (traditional and improved) and white rice. Exploring the income outcomes more closely, we checked the money units generated from four sources of income (Figure 3): white rice (mean = 5.04 money units, sd = 5.84); sticky rice (mean = 4.80 money units, sd = 3.20); selling cattle (mean = 2.00 money units, sd = 2.73); oﬀ-farm workers (mean = 0.87 money units, sd = 1.09). We observed that before white rice was introduced in round 3, traditional “sticky” rice (Figure 3c) was the main source of income for both men and women. White rice was introduced in round 3, and from then on, the majority of players’ income stemmed from white rice (Figure 3d). The income generated by selling cattle (Figure 3a) and from oﬀ-farm workers (Figure 3b) was marginal compared to income from selling white rice. Therefore, we only further investigated the trends in income from sticky rice (traditional and improved) and white rice. 11 of 22 11 of 23 Sustainability 2020, 12, 6594 Sustainability 2020 12 x FOR Figure 3. 3.1. Income The income generated from selling cattle in round 1 and the income from white rice in rounds 1 and 2, was zero because those options were not yet available to players. Figure 3. Boxplots of the observed income of players generated by four different sources of income, per round and gender, all villages cofounded, (a) from selling cattle, (b) from off-farm workers, (c) from sticky rice, whether standard or improved, (d) from white rice. The income from off-farm workers was zero in round 3, because in this round, workers sent off-farm failed to secure employment. The income generated from selling cattle in round 1 and the income from white rice in rounds 1 and 2, was zero because those options were not yet available to players. Figure 3. Boxplots of the observed income of players generated by four diﬀerent sources of income, per round and gender, all villages cofounded, (a) from selling cattle, (b) from oﬀ-farm workers, (c) from sticky rice, whether standard or improved, (d) from white rice. The income from oﬀ-farm workers was zero in round 3, because in this round, workers sent oﬀ-farm failed to secure employment. The income generated from selling cattle in round 1 and the income from white rice in rounds 1 and 2, was zero because those options were not yet available to players. There was no discernible difference in the increasing trend of income generated by sticky rice over time between men and women (Figure 3c). Women achieved a higher income from white rice than did men once white rice became available in the rounds (Figure 3d). There was no discernible diﬀerence in the increasing trend of income generated by sticky rice over time between men and women (Figure 3c). Women achieved a higher income from white rice than did men once white rice became available in the rounds (Figure 3d). There was no discernible difference in the increasing trend of income generated by sticky rice over time between men and women (Figure 3c). Women achieved a higher income from white rice than did men once white rice became available in the rounds (Figure 3d). There was no discernible diﬀerence in the increasing trend of income generated by sticky rice over time between men and women (Figure 3c). Women achieved a higher income from white rice than did men once white rice became available in the rounds (Figure 3d). 3.1. Income Boxplots of the observed income of players generated by four different sources of income, per round and gender, all villages cofounded, (a) from selling cattle, (b) from off-farm workers, (c) from sticky rice, whether standard or improved, (d) from white rice. The income from off-farm workers was zero in round 3, because in this round, workers sent off-farm failed to secure employment. The income generated from selling cattle in round 1 and the income from white rice in rounds 1 and 2, was zero because those options were not yet available to players. Figure 3. Boxplots of the observed income of players generated by four diﬀerent sources of income, per round and gender, all villages cofounded, (a) from selling cattle, (b) from oﬀ-farm workers, (c) from sticky rice, whether standard or improved, (d) from white rice. The income from oﬀ-farm workers was zero in round 3, because in this round, workers sent oﬀ-farm failed to secure employment. The income generated from selling cattle in round 1 and the income from white rice in rounds 1 and 2, was zero because those options were not yet available to players. Figure 3. Boxplots of the observed income of players generated by four different sources of income, Figure 3. Boxplots of the observed income of players generated by four diﬀerent sources of income, Figure 3. Boxplots of the observed income of players generated by four different sources of income, per round and gender, all villages cofounded, (a) from selling cattle, (b) from off-farm workers, (c) from sticky rice, whether standard or improved, (d) from white rice. The income from off-farm workers was zero in round 3, because in this round, workers sent off-farm failed to secure employment. The income generated from selling cattle in round 1 and the income from white rice in rounds 1 and 2, was zero because those options were not yet available to players. Figure 3. Boxplots of the observed income of players generated by four diﬀerent sources of income, per round and gender, all villages cofounded, (a) from selling cattle, (b) from oﬀ-farm workers, (c) from sticky rice, whether standard or improved, (d) from white rice. The income from oﬀ-farm workers was zero in round 3, because in this round, workers sent oﬀ-farm failed to secure employment. 3 2 Production Choices 3.2. Production Choices White rice production at the player level was determined not only by the amount of white rice cultivated but also by management decisions including plot choice and use of fertilizers White rice production at the player level was determined not only by the amount of white rice cultivated but also by management decisions, including plot choice and use of fertilizers. cultivated but also by management decisions, including plot choice and use of fertilizers. Women and men were equally likely to allocate most of their plots to white rice once it was available, from round 3 onward. The important distinction in terms of income, however, was that women were more likely to allocate their best plot—the paddy protected from flooding and with a high yield (paddy highland, PH, in Figure 4 —to white rice. This was not because men were less interested in maximising income, but because, as male players in village Alan Wattana explained during debriefing, “We need to plant more sticky rice because our family is eating sticky rice, and so we can’t produce enough white rice for trader” Women and men were equally likely to allocate most of their plots to white rice once it was available, from round 3 onward. The important distinction in terms of income, however, was that women were more likely to allocate their best plot—the paddy protected from ﬂooding and with a high yield (paddy highland, PH, in Figure 4—to white rice. This was not because men were less interested in maximising income, but because, as male players in village Alan Wattana explained during debrieﬁng, “We need to plant more sticky rice because our family is eating sticky rice, and so we can’t produce enough white rice for trader”. 12 of 22 Sustainability 2020, 12, 6594 Figure 4. The probability of allocating a plot to white rice cultivation depending on plot properties: rainfed with bad soil (RB), rainfed with good soil (RG), paddy highland (PH), paddy lowland (PL). Each circle represents the observed fraction of the plot of a given plot property allocated to white rice cultivation per game session. The size of the circle is proportional to the number of observations of the same values. The lines are predictions from a logistic regression, while the grey ribbons represent the corresponding 95% confidence intervals. Figure 4. 3 2 Production Choices 3.2. Production Choices The probability of allocating a plot to white rice cultivation depending on plot properties: rainfed with bad soil (RB), rainfed with good soil (RG), paddy highland (PH), paddy lowland (PL). Each circle represents the observed fraction of the plot of a given plot property allocated to white rice cultivation per game session. The size of the circle is proportional to the number of observations of the same values. The lines are predictions from a logistic regression, while the grey ribbons represent the corresponding 95% conﬁdence intervals. Figure 4. The probability of allocating a plot to white rice cultivation depending on plot properties: rainfed with bad soil (RB), rainfed with good soil (RG), paddy highland (PH), paddy lowland (PL). Each circle represents the observed fraction of the plot of a given plot property allocated to white rice cultivation per game session. The size of the circle is proportional to the number of observations of the same values. The lines are predictions from a logistic regression, while the grey ribbons represent the corresponding 95% confidence intervals. Figure 4. The probability of allocating a plot to white rice cultivation depending on plot properties: rainfed with bad soil (RB), rainfed with good soil (RG), paddy highland (PH), paddy lowland (PL). Each circle represents the observed fraction of the plot of a given plot property allocated to white rice cultivation per game session. The size of the circle is proportional to the number of observations of the same values. The lines are predictions from a logistic regression, while the grey ribbons represent the corresponding 95% conﬁdence intervals. Fertilizers were increasingly applied over time, with no noticeable difference between men and women (Appendix B). Players were more likely to apply fertilizers to white rice compared to sticky rice, while no significant difference was detected between improved sticky rice and white rice. From the game components (Appendix A), production increases the most by applying fertilizer to improved rice compared to either white rice or sticky rice. Fertilizer on improved sticky rice increased production by 1.5 bags of rice (average during a good year over the four plot types), while it increased production by only 1 bag of rice when applied to areas of white rice or sticky rice production. However, in optimum conditions, white rice could be sold for 1.5 units of money per bag compared to only 1 unit of money per bag for sticky rice. 3 2 Production Choices 3.2. Production Choices Thus, applying fertilizer on a plot of improved rice or white rice was equally profitable. Fertilizers were increasingly applied over time, with no noticeable diﬀerence between men and women (Appendix B). Players were more likely to apply fertilizers to white rice compared to sticky rice, while no signiﬁcant diﬀerence was detected between improved sticky rice and white rice. From the game components (Appendix A), production increases the most by applying fertilizer to improved rice compared to either white rice or sticky rice. Fertilizer on improved sticky rice increased production by 1.5 bags of rice (average during a good year over the four plot types), while it increased production by only 1 bag of rice when applied to areas of white rice or sticky rice production. However, in optimum conditions, white rice could be sold for 1.5 units of money per bag compared to only 1 unit of money per bag for sticky rice. Thus, applying fertilizer on a plot of improved rice or white rice was equally proﬁtable. increase farme 4. Discussion 4. Discussion In the game, the income of both men and women increased over time, but women achieved a higher income than men as the game progressed. Women achieved a higher income due to their higher levels of collaboration, tending more quickly to achieve the production of 29 bags of white rice required for a premium price. Women’s higher propensity to collaborate and cooperate is recorded in the literature, specifically when they are in a female-only environment [67,68], and this has also been noted in other experimental game settings [69,70]. Some studies, however, suggest that, overall, women and men can be equally cooperative, with sex differences in cooperation entirely context- dependent [71,72], whereas other research suggests that the effect of gender on cooperation remains inconclusive [73]. Nonetheless, it needs to be noted that most of the literature reported are on social dilemma studies that are conceptually concerned with situations involving a conflict between self- interest and the collective interest [71]. Gendered exploration of cooperation in a specific context of agriculture with the potential neutral-win and win-win outcomes, warrants further study as it In the game, the income of both men and women increased over time, but women achieved a higher income than men as the game progressed. Women achieved a higher income due to their higher levels of collaboration, tending more quickly to achieve the production of 29 bags of white rice required for a premium price. Women’s higher propensity to collaborate and cooperate is recorded in the literature, speciﬁcally when they are in a female-only environment [67,68], and this has also been noted in other experimental game settings [69,70]. Some studies, however, suggest that, overall, women and men can be equally cooperative, with sex diﬀerences in cooperation entirely context-dependent [71,72], whereas other research suggests that the eﬀect of gender on cooperation remains inconclusive [73]. Nonetheless, it needs to be noted that most of the literature reported are on social dilemma studies that are conceptually concerned with situations involving a conﬂict between self-interest and the collective interest [71]. Gendered exploration of cooperation in a speciﬁc context of agriculture with the potential neutral-win and win-win outcomes, warrants further study as it potentially carries more relevant policy implications than can be generated exclusively through social dilemma studies. 3.3. Cooperation 3.3. Cooperation The straight lines in colour are predictions from a linear-mixed effect model, while the grey ribbons represent the corresponding 95% confidence intervals from the fixed effects. (b) The regression coefficient estimates (dots) with their corresponding 95% confidence intervals (whiskers) of the fixed effects of the model used for prediction in (a). Coefficient estimates whose confidence intervals exclude zero can reasonably be considered to affect the response variable. Figure 5. Trends and corresponding coeﬃcients of white rice production at the session-level by round. (a) The production over the rounds for men in pink and women in blue. The dots represent observed data, and the size of the dot is proportional to the number of observations of the same values. The black horizontal line is drawn at 29, the threshold above which players receive a price premium for white rice. The straight lines in colour are predictions from a linear-mixed eﬀect model, while the grey ribbons represent the corresponding 95% conﬁdence intervals from the ﬁxed eﬀects. (b) The regression coeﬃcient estimates (dots) with their corresponding 95% conﬁdence intervals (whiskers) of the ﬁxed eﬀects of the model used for prediction in (a). Coeﬃcient estimates whose conﬁdence intervals exclude zero can reasonably be considered to aﬀect the response variable. In addition to cooperating to achieve the target of 29 bags of white rice to obtain the price premium, cooperation was also important in terms of needing to share the use of the available direct seeding machine. Although men and women were equally likely to own a direct seeding machine, women were more likely to share the available machines with other farmers. In addition, when machines were shared, players—even those unable to afford the cost of 6 money units to purchase— were able to use the machine and hence halve the number of labour tokens required to farm a plot. The labour tokens that became available could then be used to perform different actions to further increase farmers’ income In addition to cooperating to achieve the target of 29 bags of white rice to obtain the price premium, cooperation was also important in terms of needing to share the use of the available direct seeding machine. Although men and women were equally likely to own a direct seeding machine, women were more likely to share the available machines with other farmers. 3.3. Cooperation 3.3. Cooperation In addition, when machines were shared, players—even those unable to aﬀord the cost of 6 money units to purchase—were able to use the machine and hence halve the number of labour tokens required to farm a plot. The labour tokens that became available could then be used to perform diﬀerent actions to further increase farmers’ income. 3.3. Cooperation 3.3. Cooperation Trends and corresponding coefficients of white rice production at the session-level by round. Figure 5. Trends and corresponding coeﬃcients of white rice production at the session-level by round. Figure 5. Trends and corresponding coefficients of white rice production at the session-level by round. (a) The production over the rounds for men in pink and women in blue. The dots represent observed data, and the size of the dot is proportional to the number of observations of the same values. The black horizontal line is drawn at 29, the threshold above which players receive a price premium for white rice. The straight lines in colour are predictions from a linear-mixed effect model, while the grey ribbons represent the corresponding 95% confidence intervals from the fixed effects. (b) The regression coefficient estimates (dots) with their corresponding 95% confidence intervals (whiskers) of the fixed effects of the model used for prediction in (a). Coefficient estimates whose confidence intervals exclude zero can reasonably be considered to affect the response variable. Figure 5. Trends and corresponding coeﬃcients of white rice production at the session-level by round. (a) The production over the rounds for men in pink and women in blue. The dots represent observed data, and the size of the dot is proportional to the number of observations of the same values. The black horizontal line is drawn at 29, the threshold above which players receive a price premium for white rice. The straight lines in colour are predictions from a linear-mixed eﬀect model, while the grey ribbons represent the corresponding 95% conﬁdence intervals from the ﬁxed eﬀects. (b) The regression coeﬃcient estimates (dots) with their corresponding 95% conﬁdence intervals (whiskers) of the ﬁxed eﬀects of the model used for prediction in (a). Coeﬃcient estimates whose conﬁdence intervals exclude zero can reasonably be considered to aﬀect the response variable. Figure 5. Trends and corresponding coefficients of white rice production at the session-level by round. (a) The production over the rounds for men in pink and women in blue. The dots represent observed data, and the size of the dot is proportional to the number of observations of the same values. The black horizontal line is drawn at 29, the threshold above which players receive a price premium for white rice. 3.3. Cooperation 3.3. Cooperation More white rice was produced at the village level when women were playing, and in those sessions, the threshold of 29 bags of rice was reached earlier (Figure 5) to obtain the premium price. Women produced more white rice and hence achieved higher profits sooner than did men when playing the game. More white rice was produced at the village level when women were playing, and in those sessions, the threshold of 29 bags of rice was reached earlier (Figure 5) to obtain the premium price. Women produced more white rice and hence achieved higher proﬁts sooner than did men when playing the game. 13 of 22 Sustainability 2020, 12, 6594 Figure 5. Trends and corresponding coefficients of white rice production at the session-level by round. (a) The production over the rounds for men in pink and women in blue. The dots represent observed data, and the size of the dot is proportional to the number of observations of the same values. The black horizontal line is drawn at 29, the threshold above which players receive a price premium for white rice. The straight lines in colour are predictions from a linear-mixed effect model, while the grey ribbons represent the corresponding 95% confidence intervals from the fixed effects. (b) The regression coefficient estimates (dots) with their corresponding 95% confidence intervals (whiskers) of the fixed effects of the model used for prediction in (a). Coefficient estimates whose confidence intervals exclude zero can reasonably be considered to affect the response variable. Figure 5. Trends and corresponding coeﬃcients of white rice production at the session-level by round. (a) The production over the rounds for men in pink and women in blue. The dots represent observed data, and the size of the dot is proportional to the number of observations of the same values. The black horizontal line is drawn at 29, the threshold above which players receive a price premium for white rice. The straight lines in colour are predictions from a linear-mixed eﬀect model, while the grey ribbons represent the corresponding 95% conﬁdence intervals from the ﬁxed eﬀects. (b) The regression coeﬃcient estimates (dots) with their corresponding 95% conﬁdence intervals (whiskers) of the ﬁxed eﬀects of the model used for prediction in (a). Coeﬃcient estimates whose conﬁdence intervals exclude zero can reasonably be considered to aﬀect the response variable. Figure 5. increase farme 4. Discussion Using role-playing games (RPGs) in a female-only and male-only group setting, they found that women who played the RPGs approached land-use change in a more dynamic way than men from the same villages, reacting more positively to external investors and possible changes. The game context of our study might have aﬀected certain outcomes. Unlike in reality, women had access to production inputs and information about new practices equally to men. This, however, is not always the case in the “real” world, and an increasing number of studies acknowledge that it is access to inputs and information, not their behavioural propensity, that prevents women from the adoption of new and improved agricultural practices [76,77]. Peterman et al. [78] also found that across diﬀerent types of inputs, men generally have higher input measures than women and that this input gap was responsible for productivity diﬀerences between men and women observed in their study. On the whole, these results from various studies suggest that technology adoption decisions depend primarily on access to resources, including land, labour and information, rather than on gender per se [78–80]. If in a particular context, men tend to have better access to these resources than women, Doss and Morris [73] argue that the technologies will not beneﬁt men and women equally. Policy changes thus may be needed to increase women’s access to key resources rather than necessarily to focus on the motivational factors that might purportedly encourage women to adopt innovations. Of relevance here is also a study by Ogunlana [81], who argued that female farmers easily adopt innovations that enhance their economic status. In other words, a low participation of women in agricultural innovation adoption is caused by lack of information about the advantages of the innovations. Such ﬁndings have important policy implications because they suggest that ensuring the more widespread and equitable adoption of improved technologies may not require changes in the research system but rather the introduction of measures that ensure better access for women to complementary inputs, especially land, labour and agricultural extension services. Therefore, in order to predict whether the new technology will be adopted successfully by women as well as men, it is important to examine both the technology itself as well as the physical and institutional contexts in which the technology is implemented [76]. Improved communication of innovation beneﬁts [81] may also be a signiﬁcant factor. increase farme 4. Discussion agriculture with the potential neutral win and win win outcomes, warrants further study as it potentially carries more relevant policy implications than can be generated exclusively through social dilemma studies. While there was no difference between men and women in terms of fertilizer use, an important management decision-making difference was that women were more likely to allocate their best plot to white rice. Men were more reluctant to do so as they placed higher value on rice for subsistence. This supports findings from other studies that indicate that subsistence farming is the preferred While there was no diﬀerence between men and women in terms of fertilizer use, an important management decision-making diﬀerence was that women were more likely to allocate their best plot to white rice. Men were more reluctant to do so as they placed higher value on rice for subsistence. This supports ﬁndings from other studies that indicate that subsistence farming is the preferred livelihood strategy for fewer than 20% of men and 10% of women [74]. This ﬁnding indicates that not only is a transition to a modern agricultural economy well on its way in Lao PDR, but women 14 of 22 Sustainability 2020, 12, 6594 also appear more willing to exit the traditional farming context. Moglia et al. [74] also reported that almost two-thirds of women in their study opted for maximizing family income through oﬀ-farm employment, compared to nearly two-thirds of men who focused on either growing rice for subsistence or income-generation. In this study, we also found indications of income-maximization strategies: in addition to cooperating to reach the target number of rice bags and hence premium price, women were also cooperating by sharing machinery. These actions released additional labour to engage in alternative on-farm and oﬀ-farm income-generating activities. This ﬁnding is very much in line with other studies that show that nontraditional oﬀ-farm and nonfarming activities have become integral to the way households generate income in Lao PDR [12–15], as well as in other emerging economies [75]. y g [ ] g g [ ] Our ﬁndings that women changed farming practices earlier in the game than men might appear contrary to some literature claiming that women are slower innovation adopters. Similar to our study—and contrary to expectations and gender stereotypes—Villamor et al. [52] found that females were more active and dynamic than men in responding to external opportunities. increase farme 4. Discussion It also engaged local communities and had a capacity-building eﬀect [38–40,83]. Moreover, games provided the farmers with an opportunity for observing the implications of their decisions on their livelihoods. Villamor and Badmos [34] found that the game was instrumental for eliciting the players’ subjective perceptions, goals and expectations within the speciﬁc context described but were limited in facilitating social learning in terms of changes in values and norms. The gaming in our study was somewhat confronting to research teams and stakeholders present at the gaming sessions (extensionist services/local agricultural services) as it provided instant, strong and visual challenges to deeply rooted assumptions and myths. For stakeholders who played their own roles as extensionists, it generated new insights by enabling them to explore the potential and gendered consequences of strategic decisions in agriculture. We suggest, in line with Ogunlana [81], that promoters of new technologies, whether governments or donor organizations, need to better understand the economic advantages of the proposed innovations in the speciﬁc context. Our ﬁndings indicate that communication of economic advantages of innovation, and not the mere availability and access to innovations per se, is of policy importance in terms of encouraging greater levels of adoption by women. We advocate that game-based approaches can be considered when trialling new technologies/practices in the complex context of village and household decision-making. It enables potential beneﬁciaries to see how scenarios play out and can help counteract risks that are sometimes revealed during initial research activity phases of a project intervention. To avoid potential shortcomings of the gaming approach, games need to be representative so that the virtual village situations developed by a core team of scientists and targeted stakeholders correspond to the local social–ecological context [38,39]. Game authenticity, that is, the level at which it is similar to the actual context of the phenomena under study, is thus an important consideration when developing the game. This, in turn, requires comprehensive preparation, several rounds of crash-testing and ﬁne-tuning and team capacity building. Gaming sessions generate comprehensive sets of data and insights; hence, they require a lengthy interpretation process. In line with Fisher et al. [54], we found the game-workshop capable of enabling the exploration of complex issues within a limited timeframe. The approach enabled the identiﬁcation of information needs for farmers in a speciﬁc region. increase farme 4. Discussion Furthermore, the aspect of gendered power dynamics on the farm, noted in the previous paragraph, is of relevance here. Importantly, enhancing the economic status of the farm is by no means the same as enhancing the economic status of women on the farm. As Doss [82] argues for African women farmers, who are typically reported as less likely than men to adopt improved crop varieties and management systems, studies too often concentrate on how gender aﬀects technology adoption and not on how the introduction of new technologies aﬀect women’s well-being. Our results suggest that once women have access to assets and information equally with men and understand the potential economic beneﬁts of proposed actions, they might be faster adopters of improved technologies and practices than men. On the other hand, such a result might be due to diﬀerences in the farming experiences, risk aversion and other characteristics of males and females. In the case of Laos, as Moglia et al. [69] report, women have less emotional attachment to existing agricultural lifestyles, which might result in their greater Sustainability 2020, 12, 6594 15 of 22 propensity for adoption of new technologies. That said, however, if considerable risks are associated with adoption, women are more likely to move away from agriculture altogether [69]. As this ﬁnding is of signiﬁcant policy relevance, we strongly recommend further studies. The gaming approach methodology that was used was well received by both local experts and farmers. The approach allowed stakeholders to explore complex interplays and elicit speciﬁc behaviours. Gaming is a research methodology that combines community engagement and data collection and is particularly useful for exploring the potential consequences of strategic decisions in agriculture [54]. Overall, the game was perceived as engaging and educational. During the debrieﬁng discussions, farmers indicated that, for them, the game tool had an educational eﬀect because it demonstrated the concepts of strategic resource allocation of farming systems and the potential beneﬁts of longer-term planning horizons. It also resulted in a mind shift for local agricultural oﬃcers, who started playing the game with the focus on “technology”. Over time, however, they started to appreciate and tried to better understand the actual potential users of the technology—farmers—and their needs and motivations. As reported elsewhere in the literature, gaming aﬀorded farmers the opportunity to conceptualize their livelihood options and gave them insight into decision-making factors. increase farme 4. Discussion However, this speciﬁcity makes games more diﬃcult to extrapolate to diﬀerent geographic and temporal settings, as diﬀerent market conditions, prices, availability of inputs, etc., might change outcomes. It is, therefore, not so much the speciﬁc decisions per se, but the decision-making process and behavioural choices, that are the key learning factors of the game approach. Important to the understanding of gendered decision-making, games highlight the decisions that can be made on-farm, even in a highly simpliﬁed set-up. Sustainability 2020, 12, 6594 16 of 22 16 of 22 5. Conclusions Conﬂicts of Interest: The authors declare no conﬂict of interest. Data Availability Statement: Data collection was approved by the James Cook University Human Ethics Research Committee, approval number H6109. 5. Conclusions Gendered studies of adoption decision-making are limited. We used a game-based approach to explore decision-making processes behind the adoption of white rice farming, in the context of a gradual transition from subsistence to commercial smallholder production supported by Lao government. The approach was found to be well accepted and of beneﬁt to the players. Game-workshops allowed for the elicitation of behaviours and decisions that could be made by women and men in a farm context and in the absence of constraints of time, resources and logistics, based on data from a virtual scenario. Our results suggest that once women have access to assets and information equally with men, and understand the reported economic beneﬁts of proposed actions, they are faster adopters of improved technologies and practices than men. This is a key ﬁnding that carries wider policy implications. There is scope for developing the promising ﬁndings of this paper, and we recommend further research in the following areas: (a) gendered explorations of cooperation among women in the agricultural context; (b) exploration of beneﬁts of agriculture and agricultural innovation to women on the farm; (c) impacts of out-migration on availability of farm labour. Such studies, we suggest, would be beneﬁcial and widely applicable in the context of countries and regions undergoing agricultural transition from subsistence to commercial agricultural commodity production. Author Contributions: Conceptualization, P.P., A.D. and K.S.A.; methodology, A.D., P.P., M.T., P.T. and M.M.; validation, A.D. and P.P.; formal analysis, T.C.; investigation, A.D., P.P., M.M., P.T., M.T., P.S., J.G.C. and S.V.; resources, A.D.; data curation, A.D.; writing—original draft preparation, A.D. and S.L.; writing—review and editing, A.D., M.M., K.S.A., S.L., P.C.; visualization, A.D. and T.C.; supervision, K.S.A.; project administration, K.S.A., P.C.; funding acquisition, K.S.A., P.C. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Australian Centre for International Agricultural Research (ACIAR) (Project no. ASEM/2014/052: “Smallholder farmer decision-making and technology adoption in southern Lao PDR: opportunities and constraints”). Acknowledgments: We would like to thank staﬀbased in our Lao partner institutions for their support and assistance, namely, colleagues at: the National University of Laos, the National Agriculture and Forestry Institute and the Department of Technical Extension and Agro-Processing. Our thanks go to the workshop participants, the local organizers and our colleagues for their comments and play-testing the games prior to ﬁeldwork. Conﬂicts of Interest: The authors declare no conﬂict of interest. Appendix B. Additional Statistical Analysis and Figures Appendix B. Additional Statistical Analysis and Figures Fertilizer use, Figure A1: Players increasingly apply fertilizers to their plots over time, with no noticeable diﬀerence between men and women. Players were more likely to apply fertilizers to white rice compared to sticky rice, while no signiﬁcant diﬀerence was detected between improved rice and white rice. From the game mechanism, production increases the most by applying fertilizer to improved rice compared to either white rice or sticky rice. Fertilizers on improved rice increase production by 1.5 bag of rice (average during a good year over the four plot types), while it increases production by only 1 bag of rice when applied on white rice or sticky rice. However, in optimum conditions, white rice can be sold for 1.5 units of money per bag compared to only 1 unit of money per bag for sticky rice. Thus, applying fertilizer to a plot of improved rice or white rice is equally proﬁtable. Players were more likely to apply fertilizer on PH plots compared to PL and RB plots, with no marked diﬀerences between PH plots and RG plots. From the game mechanism, during good years (rounds 2, 3–5), adding fertilizer increases the production of rice by 1.33 bags of rice on PH and PL (average over all three types of rice), while it increases production by 1 unit on RG and RB. Given that PL is occasionally subjected to ﬂooding, production is best increased by placing fertilizers on PH. Fertilizer use, Figure A1: Players increasingly apply fertilizers to their plots over time, with no noticeable diﬀerence between men and women. Players were more likely to apply fertilizers to white rice compared to sticky rice, while no signiﬁcant diﬀerence was detected between improved rice and white rice. From the game mechanism, production increases the most by applying fertilizer to improved rice compared to either white rice or sticky rice. Fertilizers on improved rice increase production by 1.5 bag of rice (average during a good year over the four plot types), while it increases production by only 1 bag of rice when applied on white rice or sticky rice. However, in optimum conditions, white rice can be sold for 1.5 units of money per bag compared to only 1 unit of money per bag for sticky rice. Thus, applying fertilizer to a plot of improved rice or white rice is equally proﬁtable. Sequences of events in the game are presented in Table A1. • Flood (Round 4): The lower irrigated rice ﬁeld is ﬂooded. In consequence, this ﬁeld’s yield amounts to zero. Yields of ‘good’ and ‘bad’ weather years are presented in Tables A2 and A3, respectively. Table A2. Yield on the “good weather” year. Sticky/White Rice Improved Rice No Fertilizer Fertilizer No Fertilizer Fertilizer Rainfed—bad soil 2 3 3 4 Rainfed—good soil 3 4 4 5 Irrigated—high 3 4 4 6 Irrigated—low 3 4 4 6 Table A3. Yields on the “bad weather” year. Sticky/White Rice Improved Rice No Fertilizer Fertilizer No Fertilizer Fertilizer Rainfed—bad soil 1 1 2 2 Rainfed—good soil 2 2 3 3 Irrigated—high 3 4 4 6 Irrigated—low 3 4 4 6 Table A2. Yield on the “good weather” year. Table A3. Yields on the “bad weather” year. Sequences of events in the game are presented in Table A1. Table A1. Sequence of events. Round Round 1 Round 2 Round 3 Round 4 Round 5 Weather Good Bad Good Good Good Event Election of the Head of the Village Wedding Crisis Flood Trader can sell: F F-C-Fo F-C-Fo-DSM WR-IR F-C-Fo-DSM WR-IR F-C-Fo-DSM WR-IR Extensionist gives away: DSM-IR-F IR-F IR-F-IP+C IR-F-IP F: Fertilizer; C: Cattle; Fo: Fodder; DSM: Direct Seeding Machine; IR: Improved Rice; WR: White Rice; IP: Improved Pasture. 17 of 22 Sustainability 2020, 12, 6594 • Election of the head of the village (Round 1): Each farmer is assigned a colour. The head of the village is elected by putting a token of each colour in a bag. The game master will then randomly draw one. The farmer which was assigned the colour drawn will be elected head of the village. • Election of the head of the village (Round 1): Each farmer is assigned a colour. The head of the village is elected by putting a token of each colour in a bag. The game master will then randomly draw one. The farmer which was assigned the colour drawn will be elected head of the village. • Wedding (Round 2): The farmers draw a chance card. The one who draws the card with the couple on it will be getting married. He, therefore, draws a token out of the bag. The farmer of the colour drawn has to give one labour force to the one with the card. The player getting married has to pay for the wedding expenses amounting to two money tokens. • Wedding (Round 2): The farmers draw a chance card. The one who draws the card with the couple on it will be getting married. He, therefore, draws a token out of the bag. The farmer of the colour drawn has to give one labour force to the one with the card. The player getting married has to pay for the wedding expenses amounting to two money tokens. • Bad weather (Round 2): Bad weather occurs and lowers the productivity on the rainfed areas. The yield will be the same on these ﬁelds, with or without fertilizer. • Crisis (Round 3): It is the year of the global ﬁnancial crisis. The labour tokens sent to Thailand will not bring any money home. • Flood (Round 4): The lower irrigated rice ﬁeld is ﬂooded. In consequence, this ﬁeld’s yield amounts to zero. Appendix B. Additional Statistical Analysis and Figures Players were more likely to apply fertilizer on PH plots compared to PL and RB plots, with no marked diﬀerences between PH plots and RG plots. From the game mechanism, during good years (rounds 2, 3–5), adding fertilizer increases the production of rice by 1.33 bags of rice on PH and PL (average over all three types of rice), while it increases production by 1 unit on RG and RB. Given that PL is occasionally subjected to ﬂooding, production is best increased by placing fertilizers on PH. Players were more likely to apply fertilizer on PH plots compared to PL and RB plots, with no marked diﬀerences between PH plots and RG plots. From the game mechanism, during good years (rounds 2, 3–5), adding fertilizer increases the production of rice by 1.33 bags of rice on PH and PL (average over all three types of rice), while it increases production by 1 unit on RG and RB. Given that PL is occasionally subjected to ﬂooding, production is best increased by placing fertilizers on PH. 18 of 22 Sustainability 2020, 12, 6594 S t i bilit 2020 12 Figure B1. (a) Trends in the probability to apply fertilizers to a given plot property (rainfed with bad soil (RB), rainfed with good soil (RG), paddy highland (PH) and paddy lowland (PL)), rice type and gender. Each circle represents the observed fraction of fertilized plot given the number of farmed plots per plot property and rice type per game session. The size of the dot is proportional to the number of observations of the same values. The lines in colour are predictions from a logistic Figure A1. (a) Trends in the probability to apply fertilizers to a given plot property (rainfed with bad soil (RB), rainfed with good soil (RG), paddy highland (PH) and paddy lowland (PL)), rice type and gender. Each circle represents the observed fraction of fertilized plot given the number of farmed plots per plot property and rice type per game session. The size of the dot is proportional to the number of observations of the same values. The lines in colour are predictions from a logistic regression, while the grey ribbons represent the corresponding 95% conﬁdence intervals from the ﬁxed eﬀects. References 1. FAOSTAT. Lao People’s Democratic Republic. 2019. Available online: http://www.fao.org/countryproﬁles/index/ en/?iso3=LAO (accessed on 7 August 2020). 2. Castella, J.-C. Agrarian transition and farming system dynamics in the uplands of South-East Asia. In Conservation Agriculture and Sustainable Upland Livelihoods: Innovations for, With and by Farmers to Adapt to Local and Global Changes: Proceedings of the 3rd International Conference on Conservation Agriculture in Southeast Asia, Hanoi, Vietnam, 10–15 December 2012; CIRAD: Montpellier, France; NOMAFSI: Phu Ho Commune, Vietnam; University of Queensland: Brisbane, Australie, 2012; pp. 4–21. ISBN 978-2-87614-687-7. 3. Cook, S. Structural Change, Growth and Poverty Reduction in Asia: Pathways to Inclusive Development. Develop. Policy Rev. 2006, 6, s51–s80. 4. De Koninck, R. The Challenges of the Agrarian Transition in Southeast Asia. Labour Cap. Soc. 2004, 37, 285–288. 5. Humphrey, J. Prospects and Challenges for Growth and Poverty Reduction in Asia. Develop. Policy Rev. 2006, 24, s29–s49. 6. Nooteboom, G. Agrarian Change and the Governance of Poverty in Southeast Asia. In Proceedings of the 9th Euroseas Conference, Oxford, UK, 16–18 August 2017; University of Oxford: Oxford, UK, 2017. 7. Contzen, S.; Forney, J. Family farming and gendered division of labour on the move: A typology of farming-family conﬁgurations. Agric. Hum. Values 2017, 34, 27–40. [CrossRef] 8. Alexander, K.; Millar, J.; Lipscombe, N. Sustainable Development in the Uplands of Lao PDR. Sustain. Develop. 2010, 18, 62–70. . Alexander, K.; Case, P.; Jones, M.; Connell, J. Commercialising smallholder agricultural production in People’s Democratic Republic. Develop. Pract. 2017, 27, 965–980. [CrossRef] 0. Foppes, J. Knowledge Capitalization: Agriculture and Forestry Development at “Kum Ban” Village Cluster Lev Lao PDR; LEAP, NAFES: Vientiane, Laos, 2008. 11. FAO Laos at A Glance. Available online: http://www.fao.org/laos/fao-in-laos/laos-at-a-glance/en/ (accessed on 7 August 2020). 12. Alexander, K.S.; Greenhalgh, G.; Moglia, M.; Thephavanh, M.; Sinavong, P.; Larson, S.; Jovanovic, T.; Case, P. What is technology adoption? Exploring the agricultural research value chain for smallholder farmers in Lao PDR. Agric. Hum. Values 2020, 37, 17–32. [CrossRef] 13. Clarke, E.; Jackson, T.M.; Keoka, K.; Phimphachanvongsod, V.; Sengxua, P.; Simali, P.; Wade, L.J. Insights into adoption of farming practices through multiple lenses: An innovation systems approach. Develop. Pract. 2018, 28, 983–998. [CrossRef] 14. Moglia, M.; Alexander, K.S.; Thephavanh, M.; Thammavong, P.; Sodahak, V.; Khounsy, B.; Vorlasan, S.; Larson, S.; Connell, J.; Case, P. A Bayesian network model to explore practice change by smallholder rice farmers in Lao PDR. Agric. Syst. 2018, 164, 84–94. [CrossRef] 15. Appendix B. Additional Statistical Analysis and Figures (b) Coeﬃcients corresponding to trends (a) in the probability to apply fertilizers to a given plot property (rainfed with bad soil (RB), rainfed with good soil (RG), paddy highland (PH) and paddy lowland (PL)), rice type and gender. The regression coeﬃcient estimates (dots) with their corresponding 95% conﬁdence intervals (whiskers) of the variables of the model used for prediction in (a). Figure B1. (a) Trends in the probability to apply fertilizers to a given plot property (rainfed with bad soil (RB), rainfed with good soil (RG), paddy highland (PH) and paddy lowland (PL)), rice type and gender. Each circle represents the observed fraction of fertilized plot given the number of farmed plots per plot property and rice type per game session. The size of the dot is proportional to the number of observations of the same values. The lines in colour are predictions from a logistic Figure A1. (a) Trends in the probability to apply fertilizers to a given plot property (rainfed with bad soil (RB), rainfed with good soil (RG), paddy highland (PH) and paddy lowland (PL)), rice type and gender. Each circle represents the observed fraction of fertilized plot given the number of farmed plots per plot property and rice type per game session. The size of the dot is proportional to the number of observations of the same values. The lines in colour are predictions from a logistic regression, while the grey ribbons represent the corresponding 95% conﬁdence intervals from the ﬁxed eﬀects. (b) Coeﬃcients corresponding to trends (a) in the probability to apply fertilizers to a given plot property (rainfed with bad soil (RB), rainfed with good soil (RG), paddy highland (PH) and paddy lowland (PL)), rice type and gender. The regression coeﬃcient estimates (dots) with their corresponding 95% conﬁdence intervals (whiskers) of the variables of the model used for prediction in (a). 19 of 22 19 of 22 Sustainability 2020, 12, 6594 References Forecasting deﬁcit irrigation adoption using a mixed stakeholder assessment methodology. Technol. Forecast. Soc. Chang. 2014, 83, 183–193. 28. Feder, G.; Just, R.E.; Zilberman, D. Adoption of Agricultural Innovations in Developing Countries: A survey. Econ. Develop. Cult. Chang. 1985, 33, 255–298. 29. Hailu, B.K.; Abrha, B.K.; Weldegiorgis, K.A. Adoption and impact of agricultural technologies on farm income: Evidence from Southern Tigray, Northern Ethiopia. Int. J. Food Agric. Econ. 2014, 2, 91–106. 30. Knowler, D.; Bradshaw, B. Farmers’ adoption of conservation agriculture: A review and synthesis of recent research. Food Policy 2007, 32, 25–48. [CrossRef] 31. Pannell, D.J.; Marshall, G.R.; Barr, N.; Curtis, A.; Vanclay, F.; Wilkinson, R. Understanding and promoting adoption of conservation practices by rural landholders. Aust. J. Exp. Agric. 2006, 46, 1407–1424. 32. World Bank. Agricultural Innovations Systems—An Investment Source Book; World Bank: Washington, DC, USA, 2012; p. 658. 33. Asian Development Bank. Lao PDR: Accelerating Structural Transformation for Inclusive Growth: Country Diagnostic Study; ADB: Manila, Philippines, 2017; ISBN 978-92-9257-994-4. 34. Villamor, G.B.; Badmos, B.K. Grazing game: A learning tool for adaptive management in response to climate variability in semiarid areas of Ghana. Ecol. Soc. 2016, 21, 39. 35. Garcia, C.A.; Savilaakso, S.; Verburg, R.W.; Gutierrez, V.; Wilson, S.J.; Krug, C.B.; Sassen, M.; Robinson, B.E.; Moersberger, H.; Naimi, B.; et al. The Global Forest Transition as a Human Aﬀair. One Earth 2020, 2, 417–428. [CrossRef] 36. Vieira Pak, M.; Castillo Brieva, D. Designing and implementing a Role-Playing Game: A tool to explain factors, decision making and landscape transformation. Environ. Model. Softw. 2010, 25, 1322–1333. [CrossRef] decision making and landscape transformation. Environ. Model. Softw. 2010, 25, 1322–1333. [CrossRef] 37. Barreteau, O.; Le Page, C.; D’aquino, P. Role-playing games, models and negotiation processes. J. Artif. Soc. Soc. Simul. 2003, 6, 1–10. 37. Barreteau, O.; Le Page, C.; D’aquino, P. Role-playing games, models and negotiation processes. J. Artif. Soc. Soc. Simul. 2003, 6, 1–10. 38. D’Aquino, P.; Le Page, C.; Bousquet, F.; Bah, A. Using self-designed role-playing games and a multi-agent system to empower a local decision-making process for land use management: The selfCormas experiment in Senegal. J. Artif. Soc. Soc. Simul. 2003, 6, 3. 39. Castella, J.-C.; Trung, T.N.; Boissau, S. Participatory simulation of land-use changes in the northern mountains of Vietnam: The combined use of an agent-based model, a roleplaying game, and a geographic information system. Ecol. Soc. 2005, 10, 27. 40. References Philp, J.N.M.; Vance, W.; Bell, R.W.; Chhay, T.; Boyd, D.; Phimphachanhvongsod, V.; Denton, M.D. Forage options to sustainably intensify smallholder farming systems on tropical sandy soils. A review. Agron. Sustain. Develop. 2019, 39, 30. 16. FAO. Country Gender Assessment of Agriculture and the Rural Sector in Lao PDR; FAO Regional Oﬃce for Asia and the Paciﬁc Food and Agriculture Organization of the United Nations: Vientiane, Laos, 2018. 17. Akter, S.; Rutsaert, P.; Luis, J.; Htwe, N.M.; San, S.S.; Raharjo, B.; Pustika, A. Women’s empowerment and gender equity in agriculture: A diﬀerent perspective from Southeast Asia. Food Policy 2017, 69, 270–279. 18. Whitehead, A.; Kabeer, N. Living with Uncertainty: Gender, Livelihoods and Pro-Poor Growth in Rural Sub-Saharan Africa; Working Paper/Institute of Development Studies; IDS: Brighton, UK, 2001; ISBN 978-1-85864-366-3. 19. Kabeer, N. Gender Equality, Economic Growth, and Women’s Agency: The “Endless Variety” and “Monotonous Similarity” of Patriarchal Constraints. Fem. Econ. 2016, 22, 295–321. [CrossRef] 20. Whitehead, A. The gendered impacts of liberalization policies of African agricultural economies and rural livelihoods. In The Gendered Impacts of Liberalization: Towards “Embedded Liberalism”? Routledge: London, UK, 2009; ISBN 0-203-88403-5. 21. Bryceson, D.F. Gender and generational patterns of African deagrarianization: Evolving labour and land allocation in smallholder peasant household farming, 1980–2015. World Develop. 2019, 113, 60–72. [CrossRef] 21. Bryceson, D.F. Gender and generational patterns of African deagrarianization: Evolving labour and land allocation in smallholder peasant household farming, 1980–2015. World Develop. 2019, 113, 60–72. [CrossRef] Sustainability 2020, 12, 6594 20 of 22 20 of 22 22. SIDA. National Gender Proﬁle of Agricultural Households, Lao PDR 2010. Report Based on The Lao Expenditure and Consumptrion Surveys, National Agricultural Census and the National Population Census; Food and Agriculture Organization of the United Nations: Vientiane, Laos, 2010. 3. Phouxay, K.; Tollefsen, A. Rural–Urban Migration, Economic Transition, and Status of Female Indus Workers in Lao PDR. Popul. Space Place 2011, 17, 421–434. 24. Manivong, V.; Cramb, R.; Newby, J. Rice and Remittances: The Impact of Labour Migration on Rice Intensiﬁcation. In Proceedings of the Southern Laos Australian Agricultural and Resource Economics Society Conference (56th), Fremantle, Australia, 7–10 February 2012. 25. Rigg, J. Moving lives: Migration and livelihoods in the Lao PDR. Popul. Space Place 2007, 13, 163–17 26. Rigg, J. Unplanned Development: Tracking Change in South-East Asia; Zed Books: London, UK, 2012. 27. Alcon, F.; Tapsuwan, S.; Martínez-Paz, J.M.; Brouwer, R.; de Miguel, M.D. References Perrotton, A.; de Garine-Wichatitsky, M.; Valls-Fox, H.; Le Page, C. My cattle and your park: Codesigning a role-playing game with rural communities to promote multistakeholder dialogue at the edge of protected areas. Ecol. Soc. 2017, 22, 35. 41. Basco-Carrera, L.; Meijers, E.; Sarısoy, H.D.; ¸Sanli, N.O.; Co¸skun, S.; Oliemans, W.; van Beek, E.; Karaaslan, Y.; Jonoski, A. An adapted companion modelling approach for enhancing multi-stakeholder cooperation in complex river basins. Int. J. Sustain. Develop. World Ecol. 2018, 25, 747–764. [CrossRef] 42. Redpath, S.M.; Keane, A.; Andrén, H.; Baynham-Herd, Z.; Bunnefeld, N.; Duthie, A.B.; Frank, J.; Garcia, C.A.; Månsson, J.; Nilsson, L.; et al. Games as Tools to Address Conservation Conﬂicts. Trends Ecol. Evol. 2018, 33, 415–426. [CrossRef] 43. Villamor, G.B.; van Noordwijk, M. Social Role-Play Games Vs Individual Perceptions of Conservation and PES Agreements for Maintaining Rubber Agroforests in Jambi (Sumatra), Indonesia. Ecol. Soc. 2011, 16, art27. [CrossRef] 21 of 22 21 of 22 Sustainability 2020, 12, 6594 44. Ponta, N.; Cornioley, T.; Dray, A.; van Vliet, N.; Waeber, P.O.; Garcia, C.A. Hunting in Times of Change: Uncovering Indigenous Strategies in the Colombian Amazon Using a Role-Playing Game. Front. Ecol. Evol. 2019, 7, 1–19. 5. Garcia, C.; Dray, A.; Waeber, P. Learning Begins When the Game Is Over: Using Games to Embrace Comple in Natural Resources Management. GAIA—Ecol. Perspect. Sci. Soc. 2016, 25, 289–291. [CrossRef] 46. Garcia, C.A.; Vendé, J.; Konerira, N.; Kalla, J.; Nay, M.; Dray, A.; Delay, M.; Waeber, P.O.; Stoudmann, N.; Bose, A.; et al. Coﬀee, Farmers, and Trees—Shifting Rights Accelerates Changing Landscapes. Forests 2020, 11, 480. [CrossRef] 47. Speelman, E.N.; García-Barrios, L.E.; Groot, J.C.J.; Tittonell, P. Gaming for smallholder participation in the design of more sustainable agricultural landscapes. Agric. Syst. 2014, 126, 62–75. [CrossRef] 48. Bos, S.P.M.; Cornioley, T.; Dray, A.; Waeber, P.O.; Garcia, C.A. Exploring livelihood strategies of shifting cultivation farmers in assam through games. Sustainability 2020, 12, 2438. [CrossRef] 49. Ouariachi, T.; Elving, W.J.L.; Pierie, F. Playing for a sustainable future: The case of We Energy Game as an educational practice. Sustainability 2018, 10, 3639. [CrossRef] 50. Villamor, G.B.; Dah-gbeto, P.A.; Bell, A.; Pradhan, U.; van Noordwijk, M. Gender-speciﬁc spatial perspectives and scenario building approaches for understanding gender equity and sustainability in climate-smart landscapes. In Climate-Smart Landscapes: Multifunctionality in Practice; World Agroforestry Centre (ICRAF): Nairobi, Kenya, 2015; pp. 211–224. ISBN 978-92-9059-375-1. 51. Sims Feldstein, H.; Jiggins, J. References Tools for The Field: Methodologies Handbook for Gender Analysis in Agriculture; Feldstein, H.S., Jiggins, J., Eds.; Kumarian Press Library of Management for Development; Kumarian Press: West Hartford, CT, USA, 1994; ISBN 978-1-56549-028-4. 52. Villamor, G.B.; van Noordwijk, M.; Djanibekov, U.; Chiong-Javier, M.E.; Catacutan, D. Gender diﬀerences in land-use decisions: Shaping multifunctional landscapes? Curr. Opin. Environ. Sustain. 2014, 6, 128–133. [CrossRef] 53. Étienne, M. Companion Modelling: A Participatory Approach to Support Sustainable Development; Springer: Heidelberg, Germany, 2013; ISBN 978-94-017-8557-0. 54. Fisher, J.; Farré, I.; Dray, A.; Khimashia, N.; Perez, P. Serious games to explore uncertainty of future farms. In Proceedings of the 16th Australian Agronomy Conference, Armidale, New South Wales, Australia, 14–18 October 2012; Available online: http://www.regional.org.au/au/asa/2012/agriculture/8191_ﬁsherjs.htm (accessed on 7 August 2020). 55. Alexander, K.; Larson, S.; Moglia, M.; Greenhalgh, G.; Case, P.; Perez, P.; Jovanovic, T.; Giger-Dray, A. Smallholder Farmer Decision-Making and Technology Adoption in Southern Lao PDR: A Synthesis of Findings; Report for ACIAR ASEM/2014/052 project ‘Smallholder Farmer Decision-Making and Technology Adoption in Southern Laos: Opportunities and Constraints’; ACIAR: Canberra, ACT, Australia, 2017. 56. Alexander, K.; Larson, S. Smallholder Farmer Decision-Making and Technology Adoption in Southern Lao PDR: Opportunities and Constraints. Activity 1.5: Stakeholders Perceptions; Report for ACIAR ASEM/2014/052 Project ‘Smallholder Farmer Decision-Making and Technology Adoption in Southern Laos: Opportunities and Constraints’; ACIAR: Canberra, Australia, 2016. 57. Alexander, K.; Parry, L.; Thammavong, P.; Sacklokham, S.; Pasouvang, S.; Connell, J.; Jovanovic, T.; Moglia, M.; Larson, S.; Case, P. Rice farming systems in Southern Lao PDR: Interpreting farmers’ agricultural production decisions using Q methodology. Agric. Syst. 2018, 160, 1–10. 58. Larson, S.; Alexander, K. Smallholder Farmer Decision-Making and Technology Adoption in Southern Lao PDR: Opportunities and Constraints. Activity 1.4 Summary of the Secondary Livelihoods and Economic Data; Report for ACIAR ASEM/2014/052 project ‘Smallholder Farmer Decision-Making and Technology Adoption in Southern Laos: Opportunities and Constraints’; ACIAR: Canberra, Australia, 2016. 59. Eliste, P.; Santos, N.; Pravongviengkham, P. Lao People’s Democratic Republic Rice Policy Study; IRRI, WB: Rome, Italy, 2012. 60. Cramb, R.A.; Gray, G.D.; Gummert, M.; Haefele, S.M.; Lefroy, R.D.B.; Newby, J.C.; Stür, W.; Warr, P. Trajectories of Rice-Based Farming Systems in Mainland Southeast Asia; ACIAR Monograph No. 177; Australian Centre for International Agricultural Research (ACIAR): Canberra, Australia, 2015. 61. Stür, W.; Gray, G.D. Review of Rice-Based Farming Systems in Mainland Southeast Asia Working Paper 3. References PLoS ONE 2013, 8, e83700. [CrossRef] 70. Charness, G.; Rustichini, A. Gender diﬀerences in cooperation with group membership. Games Econ. Behav. 2011, 72, 77–85. [CrossRef] 70. Charness, G.; Rustichini, A. Gender diﬀerences in cooperation with group membership. Games Econ 2011, 72, 77–85. [CrossRef] 71. Balliet, D.; Li, N.P.; Macfarlan, S.J.; Van Vugt, M. Sex diﬀerences in cooperation: A meta-analytic review of social dilemmas. Psychol. Bull. 2011, 137, 881–909. [CrossRef] 72. Peshkovskaya, A.; Babkina, T.; Myagkov, M. Gender eﬀects and cooperation in collective action: A laboratory experiment. Ration. Soc. 2019, 31, 337–353. [CrossRef] 73. Croson, R.; Gneezy, U. Gender Diﬀerences in Preferences. J. Econ. Lit. 2009, 47, 448–474. [CrossRef] Croson, R.; Gneezy, U. Gender Diﬀerences in Prefere 74. Moglia, M.; Alexander, K.S.; Larson, S.; (Giger)-Dray, A.; Greenhalgh, G.; Case, P. Gendered roles in agrarian transition: A study of lowland rice farming in Lao PDR. Sustainability 2020, 12, 5403. 75. Deere, C.D. The Feminization of Agriculture? Economic Restructuring in Rural Latin America; Occasional Paper; United Nations Research Institute for Social Development: Geneva, Switzerland, 2005; ISBN 978-92-9085-049-6. 76. Doss, C.R.; Morris, M.L. How does gender aﬀect the adoption of agricultural innovations? The case of improved maize technology in Ghana. Agric. Econ. 2000, 25, 27–39. [CrossRef] 77. Ragasa, C. Improving Gender Responsiveness of Agricultural Extension. In Gender in Agriculture; Quisumbing, A.R., Meinzen-Dick, R., Raney, T.L., Croppenstedt, A., Behrman, J.A., Peterman, A., Eds.; Springer: Dordrecht, The Netherlands, 2014; pp. 411–430. ISBN 978-94-017-8615-7. 78. Peterman, A.; Behrman, J.A.; Quisumbing, A.R. A Review of Empirical Evidence on Gender Diﬀerences in Nonland Agricultural Inputs, Technology, and Services in Developing Countries. In Gender in Agriculture; Quisumbing, A.R., Meinzen-Dick, R., Raney, T.L., Croppenstedt, A., Behrman, J.A., Peterman, A., Eds.; Springer: Dordrecht, The Netherlands, 2014; pp. 145–186. ISBN 978-94-017-8615-7. 79. Bikketi, E.; Ifejika Speranza, C.; Bieri, S.; Haller, T.; Wiesmann, U. Gendered division of labour and feminisation of responsibilities in Kenya; implications for development interventions. Gender Place Cult. 2016, 23, 1432–1449. [CrossRef] cher, E.; Qaim, M. Gender, agricultural commercialization, and collective action in Kenya. Food Secur. 12, 4, 441–453. [CrossRef] 81. Ogunlana, E.A. The technology adoption behavior of women farmers: The case of alley farming in Renew. Agric. Food Syst. 2004, 19, 57–65. [CrossRef] 82. Doss, C.R. Designing Agricultural Technology for African Women Farmers: Lessons from 25 Years of Experience. World Develop. 2001, 29, 2075–2092. [CrossRef] 83. Castella, J.-C.; Bourgoin, J.; Lestrelin, G.; Bouahom, B. References Livestock in Smallholder Farming Systems of Mainland Southeast Asia; University of Queensland Australia and International Centre for Tropical Agriculture (CIAT): Hanoi, Vietnam, 2014. 22 of 22 Sustainability 2020, 12, 6594 62. Vote, C.; Newby, J.; Phouyyavong, K.; Inthavong, T.; Eberbach, P. Trends and perceptions of rural household groundwater use and the implications for smallholder agriculture in rain-fed Southern Laos. Int. J. Water Resour. Develop. 2015, 31, 558–574. [CrossRef] 3. World Bank (WB); International Rice Research Institute (IRRI). Lao People’s Democratic Republic Rice Po Study; Ministry of Agriculture and Forestry of the Lao People’s Democratic Republic: Vientiane, Laos, 2 63. World Bank (WB); International Rice Research Institute (IRRI). Lao People’s Democratic Republic Rice Policy Study; Ministry of Agriculture and Forestry of the Lao People’s Democratic Republic: Vientiane, Laos, 2012. 64. R Core Team. R: A Language and Environment for Statistical Computing; R Core Team: Vienna, Austria, 2019. Study; Ministry of Agriculture and Forestry of the Lao People s Democratic Republic: Vientiane, Laos, 2012. 64. R Core Team. R: A Language and Environment for Statistical Computing; R Core Team: Vienna, Austria, 2019. Study; Ministry of Agriculture and Forestry of the Lao People s Democratic Republic: Vientiane, Laos, 2012. 64. R Core Team. R: A Language and Environment for Statistical Computing; R Core Team: Vienna, Austria, 2019. 65. Bates, D.; Mächler, M.; Bolker, B.M.; Walker, S.C. Fitting linear mixed-eﬀects models using lme4. J. Stat. Softw. 2015. [CrossRef] 4. R Core Team. R: A Language and Environment for Statistical Computing; R Core Team: Vienna, Austria, 20 g g f p g 65. Bates, D.; Mächler, M.; Bolker, B.M.; Walker, S.C. Fitting linear mixed-eﬀects models using lme4. J. Stat. Softw. 2015. [CrossRef] 66. Barr, D.J.; Levy, R.; Scheepers, C.; Tily, H.J. Random eﬀects structure for conﬁrmatory hypothesis testing: Keep it maximal. J. Mem. Lang. 2013, 68, 255–278. [CrossRef] 67. Chatman, J.A.; O’Reilly, C.A. Asymmetric Reactions to Work Group Sex Diversity Among Men and Women. AMJ 2004, 47, 193–208. [CrossRef] 68. Hernán-González, R.; Kujal, P. Gender diﬀerences in cooperation and competition. In Experimental Economics I: Economic Decisions; Palgrave Macmillan: London, UK, 2015; pp. 154–168. 69. Molina, J.A.; Giménez-Nadal, J.I.; Cuesta, J.A.; Gracia-Lazaro, C.; Moreno, Y.; Sanchez, A. Gender Diﬀerences in Cooperation: Experimental Evidence on High School Students. PLoS ONE 2013, 8, e83700. [CrossRef] 69. Molina, J.A.; Giménez-Nadal, J.I.; Cuesta, J.A.; Gracia-Lazaro, C.; Moreno, Y.; Sanchez, A. Gender Diﬀerences in Cooperation: Experimental Evidence on High School Students. References A model of the science-practice-policy interface in participatory land-use planning: Lessons from Laos. Landsc. Ecol. Pract. 2014, 29, 1095–1107. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W2030807173	https://figshare.com/articles/journal_contribution/Planning_processes_for_advanced_manufacturing_technology_by_large_American_manufacturers/5085292/1/files/8619343.pdf	English	null	Planning processes for advanced manufacturing technology by large American manufacturers	Technovation	1,998	cc-by	7,699	ABSTRACT In response to an increasingly competitive environment, many manufacturers have invested in advanced manufacturing technology (AMT). While many studies regarding the use of AMT have been performed, little is known about how firms plan and manage their AMT investments. This is surprising given the evidence about the impact such investments can have on an organization. To address this, a survey of large American manufacturing firms was conducted. Respondents were asked, among other questions, why their firms made such investments and the fit with their business strategy; which functional areas were involved in idea generation, the planning process, and proposal assessment; and what the anticipated risks, difficulties and benefits were from making such investments. The results from this survey are provided along with an analysis of the responses. PLANNING PROCESSES FOR ADVANCED MANUFACTURING TECHNOLOGY BY LARGE AMERICAN MANUFACTURERS Robert Millen & Amrik S. Sonai Working Paper 45/98 June 1998 Advanced Manufacturing Technology, Planning, Implementation, USA, Australia MONASH UNIVERSITY FACULTY OF BUSINESS AND ECONOMICS PLANNING PROCESSES FOR ADVANCED MANUFACTURING TECHNOLOGY BY LARGE AMERICAN MANUFACTURERS Robert Millen & Amrik S. Sonai Working Paper 45/98 June 1998 MONASH UNIVERSITY FACULTY OF BUSINESS AND ECONOMICS 1. INTRODUCTION The competitive environment continues to become increasingly more demanding. The North American Free Trade Agreement (NAFTA), for example, has resulted in a considerable amount of pressure on manufacturers in the USA and Canada to introduce greater numbers of new products, to shorten time to market, to increase customer service levels, to provide higher levels of conformance and quality, and to offer more variety in their product lines. Reduction in trade barriers in other parts of the world have also presented new opportunities for North American manufacturers. In the current manufacturing environment characterised by low volumes and high variety, the adoption of advanced manufacturing technology (AMT) can provide many strategic benefits for the firm. The introduction of A M T has taken many forms and acronyms. AMTs refer to a family of technologies which includes computer-aided design (CAD), computer-aided engineering, (CAE), manufacturing resources planning (MRPII), automated materials handling systems, electronic data interchange (EDI) and computer-integrated manufacturing (CIM) systems. These technologies enable both economies and scale and economies of scope to be achieved without changing the hardware. The major strategic benefits that these technologies offer are the increased flexibility and the responsiveness, enabling an organisation to substantially improve its competitiveness in the marketplace. These new technologies require significant financial commitments in terms of hardware, software, support systems, and training. In addition, obtaining the greatest benefits generally requires alterations in other functional areas of an organization, such as marketing and maintenance. For example, firms that have achieved flexibility in their operations have entered market niches that in the past were neglected because the requisite capability to serve these niches effectively did not exist previously. Similarly, more emphasis must be placed on preventative maintenance policies since many of these AMTs account for a substantial proportion of the firm's investment in production technology and any amount of downtime can significantly increase operating costs. These competitive conditions and internal commitments make planning difficult for investments in AMT. Balancing the need to create a competitive advantage and the investment of time and money is required when A M T acquisition decisions are being made. This study explores the challenges associated with planning A M T investments and reports the experiences of 93 major American manufacturing organizations. In addition, results from the USA are compared with a similar study conducted in Australia. Keywords Advanced Manufacturing Technology, Planning, Implementation, USA, Australia 1. INTRODUCTION A number of studies regarding the use of A M T in various forms have been performed (see, for example, Boyer et. a!., (1996) for a comprehensive review.) Yet, little is known about how firms plan and manage their A M T investments. This is surprising given the evidence that exists about the impact such investments can have on an organization. To achieve the system wide benefits from A M T proffered by its exponents, there must be an alignment between corporate and manufacturing strategy. An inadequate appraisal of business needs and the exclusion of such needs in the assessment of A M T has prevented some firms from benefiting from their A M T investments (Zairi, 1992). Success will also depend on the ability of the organization to assimilate such new technology. It has been argued that firms that adopt AMTs without first redesigning organizational structures and processes risk institutionalizing bad practices (Duimering et. al. ,1993). Bessant (1994) also emphasises the many dimensions of the organisation which are impacted by AMT investments. At the work, management and inter-firm levels, he argues the need for parallel organisational change to accompany the investment. Similarly, Zammuto and O'Connor (1992) draw attention to the importance of organisational design and 2 culture in their discussion of the relationships among a company's culture, structure and the potential outcomes of an A M T investment. Clearly, companies need to rethink how their inter- and intra- organizational processes will be impacted as a result of A M T investments. culture in their discussion of the relationships among a company's culture, structure and the potential outcomes of an A M T investment. Clearly, companies need to rethink how their inter- and intra- organizational processes will be impacted as a result of A M T investments. Finally, A M T is not just about technology, organizational structures and business processes. The human element must also be addressed along three dimensions. First, A M T implementation failure has been attributed to the neglect of critical human resource factors (Saraph and Sebastian, 1992). Thomas and Wainwright (1994) also emphasise on the internal politics, especially the link between "people issues" and successfully obtaining the benefits of AMTs. Resistance to change is to be expected, but this resistance can be managed through effective communication programs and the involvement of individuals during the planning process. 1. INTRODUCTION Second, A M T is successful only when people want to use it (Markus and Keil, 1994). A number of factors can facilitate the acceptance of such investments through incentive schemes and the use of cross-functional teams. Third, most firms with successful A M T implementations enlist champions. These individuals provide a continual driving force throughout the initiative [Beatty (1992), Singh and Sohal (1995), and Sohal (1996)]. In light of the demonstrated importance of the above factors, we were able to identify only a few studies that have examined in detail the A M T planning process. In addition, only two of these few studies were conducted in the US. Currie (1989) found that the process of decision making for CAD systems was ad hoc in a study of twenty British firms. Technology was often introduced in a fragmented manner and was targeted to achieve narrow operational benefits in most of the firms examined by Currie. In studying how corporate managers evaluated new technologies, Gold (1988) found that the problems were not so much with the technologies as with the misconception of managers both in evaluating capabilities before adoption and in appraising efforts after implementation. Kaplan (1986) calls for more sophisticated financial appraisal of A M T investments and highlighted by Howell and Sourcy (1987) and Adler (1988) , the need to accommodate the less quantifiable effects of AMT. Small and Chen (1995) examined the impact of financial justification approaches and criteria at the plant level on the adoption of more integrated technologies. Their study provides insights into the formation of decision-making teams for the justification process. In a series of papers, Sohal et. al., have examined the A M T planning process by manufacturers in Australia [Sohal (1997), Sohal et. al. (1991)], New Zealand (Sohal et. al., 1996) and the United Kingdom (Sohal, 1994). This series of studies are more comprehensive than others in that the A M T idea generation, idea development, justification, implementation and post-implementations processes were investigated, and examined from the perspective of senior management of the firms. A question still remains, however, as to what processes are employed by major American manufacturing firms. In order to examine this question, a questionnaire survey was conducted amongst the largest American manufacturing companies using the survey instrument previously employed by Sohal, et al. in a number of other countries. The next section (Section 2) describes the research methodology. 1. INTRODUCTION This is followed by the main section of the paper (Section 3) which presents the overall findings from the American study. In Section 4, the key findings from the American study are discussed and a general comparison is made with the key findings from the 1993 Australian study. 2. RESEARCH METHODOLOGY The survey instrument used for the American study was based on the original questionnaire used by Sohal, et. al. (1991) to study the planning and implementation of AMT in Australia. The questionnaire comprised the following sections: • Company background details • A M T investment proposal generation 3 3 • Proposal assessment • A M T implementation • Post implementation study The main issues addressed by the research reported on in this article and the earlier article were: 1. Why do companies invest in AMT, and what effects do company factors (e.g., size) have on these decisions? 2. What are the size and nature of AMT investments? 3. Which functional areas generate investment ideas involving A M T and which functional areas are involved in the planning process? 4. How good was the fit between the A M T proposal and the business strategy? 5. Which functional areas were involved with the proposal assessment and which financial techniques were employed? 6. What were the anticipated benefits, risks and difficulties? The sample for the study was selected from Business Week's 1995 list of the 1,000 largest US companies. From this list, only manufacturing firms were selected. Furthermore, the major operating divisions of the large diversified companies (for example, General Electric's Locomotive, Airplane Engine, and Plastics divisions) were identified as well. The main issues addressed by the research reported on in this article and the earlier article were: 1. Why do companies invest in AMT, and what effects do company factors (e.g., size) have on these decisions? 2. What are the size and nature of AMT investments? 3. Which functional areas generate investment ideas involving A M T and which functional areas are involved in the planning process? 3. Which functional areas generate investment ideas involving A M T and which functional areas are involved in the planning process? 4. How good was the fit between the A M T proposal and the business strategy? 5. Which functional areas were involved with the proposal assessment and which financial techniques were employed? 5. Which functional areas were involved with the proposal assessment and which financial techniques were employed? 6. What were the anticipated benefits, risks and difficulties? The sample for the study was selected from Business Week's 1995 list of the 1,000 largest US companies. From this list, only manufacturing firms were selected. 2. RESEARCH METHODOLOGY Furthermore, the major operating divisions of the large diversified companies (for example, General Electric's Locomotive, Airplane Engine, and Plastics divisions) were identified as well. The name of a senior level executive in Manufacturing, Research and Development, Technology or Process Development was researched. A questionnaire was sent only when such an individual could be identified by name. This process resulted in a sample of 446 manufacturing organizations which represented a cross- section of most of the industrial sectors. Manufacturers using all types of production methods were represented in the sample. The questionnaire together with a covering letter and a pre-paid reply envelope were mailed out during the Fall of 1996. A total of 124 questionnaires were returned and no follow-up was conducted. Of this number, 93 contained useable responses. This resulted in a response rate of approximately 21% which is similar to response rates reported elsewhere for surveys of manufacturing practice [King and Grover (1991), King and Ramamurthy (1992), and Raho et. al. (1987)]. 3. SURVEY FINDINGS This section presents the overall findings from the American study. First, the profile of the responding companies is presented. Then, each of the six research issues presented earlier are discussed. Question 1: The purposes of AMT investments. Question 1: The purposes of AMT investments. Investment in A M T is one manner in which manufacturers believe they can improve their competitive position in world markets. This is reflected in the responses to the question concerning the reasons for investing in AMT. Respondents were provided with a list of five reasons for investing in AMT and asked to indicate the importance of each on a scale ranging from one (most important) to five (not at all important). The mean score of the responses was calculated for each reason, and the resulting values are provided in Table 2. Obtaining financial benefits and competitive advantage were the most important reasons given for investing in A M T with mean scores of 1.15 and 1.16, respectively. The next most important reason cited was to counter competitive threats, followed by countering skill deficiencies and enhancing the company's image. When analyzed by company size (measured both as annual sales revenues and number of employees), no statistically significant relationships were identified. Table 2: Motivations for AMT Investments Table 2: Motivations for AMT Investments Reasons Mean Score Obtaining financial benefits 1.15 Obtaining competitive advantage 1.16 Countering competitive threats 2.06 Countering skill deficiency 3.06 Enhancing company image 3.74 (1 = most important, 5 - not at all important) Reasons Mean Score Obtaining financial benefits 1.15 Obtaining competitive advantage 1.16 Countering competitive threats 2.06 Countering skill deficiency 3.06 Enhancing company image 3.74 Question 2: The size and nature of AMT investments Respondents Profile The number of employees and annual sales revenue for the sample are provided in Table 1. More than half of the sample had revenues in excess of $1 billion, and approximately half employed between 1,000 and 10,000 people. Flowline production is the principal type of manufacturing system used by 35% of the responding companies. Thirty-two percent employed continuous processes, 28% batch production and 5% job-shop type of manufacturing. 4 Table 1: Characterisitics of Responding Firms Annual Sales Revenue n % $100Mto$l,000M 44 47 >$1,000M 49 53 Employment n % 500 to 999 9 10 1,000 to 9,999 46 49 > 10.000 38 41 Principle Tvoe of Manufacturing n % Jobbing 5 5 Batch 26 28 Flowline 33 35 Continuous 29 31 Table 1: Characterisitics of Responding Firms Annual Sales Revenue n % $100Mto$l,000M 44 47 >$1,000M 49 53 Employment n % 500 to 999 9 10 1,000 to 9,999 46 49 > 10.000 38 41 Principle Tvoe of Manufacturing n % Jobbing 5 5 Batch 26 28 Flowline 33 35 Continuous 29 31 Table 1: Characterisitics of Responding Firms Table 1: Characterisitics of Responding Firms Question 2: The size and nature of AMT investments Question 2: The size and nature of AMT investments All respondents indicated that there had been more than one A M T investment in their company in the past three years. The percentage of firms investing in various types of A M T in the past three years is shown in Table 3. Although the percentages are not directly additive, it appears that the rank order of investments is computer hardware, computer software, and plant and equipment. For computer hardware, shop floor data capture, local area networks and on-line process instrumentation were the most popular areas of investments with 70%, 68% and 54% of the responding companies investing 5 in these technologies in the past three years, respectively. C A D / C A M , MRP II and quality control were the software invested in by the greatest number of respondent companies. Only 9% of the firms invested in OPT (Optimized Production Technology) in the past three years. The most common investments in plant and equipment were computer controlled testing equipment (41% of respondents, automated assembly (27%), and automated warehousing/order picking (36%). Table 4 provides data on the size of the largest investments made by these firms in the past three years. Whereas the largest investment for 24% of these firms was US$1 million or less, 36% spent more than US$10 million on their largest investment during this time period. Question 2: The size and nature of AMT investments Table 3: Nature of AMT Investments Table 3: Nature of AMT Investments Table 3: Nature of AMT Investments Computer Hardware Percentage of respondents PCs 100 Shop floor data capture 70 L A N 68 On-line process instrumentation 54 W A N 40 Minis 27 Graphics hardware 25 Micros 20 Mainframe 14 Computer Software Percentage of respondents C A D / C A M 79 MRP II 57 QC software 52 Data base mgmt systems 48 MRP 36 Expert systems 32 M A P 13 OPT 9 Plant and Equipment Percentage of respondents Computer controlled test equip 41 Automatic assembly 37 Automated warehousing/order picking 36 CNC M/Cs 30 Automatic testing equip 28 Laser measuring 25 AGVs 18 Flexible manufacturing systems 17 Laser cutting 14 Computer Hardware Percentage of respondents PCs 100 Shop floor data capture 70 L A N 68 On-line process instrumentation 54 W A N 40 Minis 27 Graphics hardware 25 Micros 20 Mainframe 14 Computer Software Percentage of respondents C A D / C A M 79 MRP II 57 QC software 52 Data base mgmt systems 48 MRP 36 Expert systems 32 M A P 13 OPT 9 Plant and Equipment Percentage of respondents Computer controlled test equip 41 Automatic assembly 37 Automated warehousing/order picking 36 CNC M/Cs 30 Automatic testing equip 28 Laser measuring 25 AGVs 18 Flexible manufacturing systems 17 Laser cutting 14 Plant and Equipment Percentage of respondents Computer controlled test equip 41 Automatic assembly 37 Automated warehousing/order picking 36 CNC M/Cs 30 Automatic testing equip 28 Laser measuring 25 AGVs 18 Flexible manufacturing systems 17 Laser cutting 14 6 6 Table 4: Size of the Largest AMT Investment Made in the Last Three Years Table 4: Size of the Largest AMT Investment Made in the Last Three Years Table 4: Size of the Largest AMT Investment Made in the Last Three Yea Investment ($ Million) Percentage of respondents <0.5 19 0.5-1 5 1- 2 18 2- 10 23 10-20 6 20-40 13 40+ 17 Questions 3 to 5: The investment planning process Questions 3 to 5: The investment planning process Idea generation. The idea to invest in A M T comes from many sources. Nearly all of the firms (98%) indicated that more than one functional area listed in Table 5 had provided such ideas. While many functional areas provided ideas, the areas that generated ideas in most companies were by personnel directly related with manufacturing (see Table 5). In more than 70% of the firms responding to this question, investment ideas were generated by Engineering, Production Engineering, and Production. Several other groups also generated A M T ideas in a substantial number of firms, including R&D, Data Processing and Suppliers. Table 5: Functional Areas Involved in Idea Generation for AMT Investments Table 5: Functional Areas Involved in Idea Generation for AMT Investments Functional Area Percentage of respondents Engineering 89 Production Engineering 76 Production 71 Research & Development 58 Data Processing 46 Suppliers 44 Consultants 27 Sales & Marketing 14 Personnel 12 Accounting & Finance 10 Functional Area Percentage of respondents Engineering 89 Production Engineering 76 Production 71 Research & Development 58 Data Processing 46 Suppliers 44 Consultants 27 Sales & Marketing 14 Personnel 12 Accounting & Finance 10 In a similar manner, 90% of the respondents indicated that more than one level of management was involved in A M T investment idea generation (see Table 6). The most frequently cited management levels where investment ideas were generated were Middle, Senior and Junior. Supervisory and Shop floor level managers were the source of ideas in a number of firms, and were more likely to generate such ideas than Top management. Training in the management of A M T was perceived to be needed more than training in the technologies themselves. Seventy percent of respondents indicated a perceived need for training in the management of A M T and 65% noted a perceived need for training in the technologies themselves. However, when asked to indicate whether an individual's firm had implemented such training, 57% indicated that training in the management of A M T had been implemented whereas 68% indicated that training in the technologies themselves had occurred. 7 Table 6: Management Involvement in Generating AMT Investment Ideas Management Levels Generating Amt Investment Ideas Percentage of Respondents Top 15 77 97 54 46 38 Senior Middle Junior Supervisory Shop Floor Number of management levels involved in A M T idea generation. Questions 3 to 5: The investment planning process Percentage of respondents One Two Three Four Five 10 25 31 29 5 Table 6: Management Involvement in Generating AMT Investment Ideas Proposal development. In 84% of the firms, a project team was established for developing the A M T proposal, and all respondents noted that these teams involved more than one of the functional areas listed in Table 6. Similar to the idea generation phase, those functions most commonly involved in the proposal generation were Production (94%), Production engineering (92%), and Engineering (89%). Proposal assessment. The financial evaluation techniques which respondents used in evaluating A M T proposals are provided in Table 7. Payback period and discounted cash flow were the most commonly employed techniques, although all four methods were utilized by almost three-quarters of the firms. In fact, all firms reported using more than one techniques with 45% reporting the use of all four listed in the table. Respondents were also asked to indicate the importance of each technique on a scale ranging from one (most important) to five (least important), and the results are also provided in Table 7. Of the responding companies, 55% claimed that the financial assessment had exceeded their normal investment criteria, while 45% claimed that it had just met their financial criteria. No firm indicated that the financial assessment failed to meet its normal investment criteria. This finding indicates that although most companies cited gaining competitive advantage as important, the normal financial investment criteria were applied. Table 7: Financial Evaluation Techniques Used Table 7: Financial Evaluation Techniques Used Table 7: Financial Evaluation Techniques Used Financial techniques Percentage of respondents Average level of importance Payback ROI (undiscounted) Discounted Cash Flow Sensitivity Analysis 89 84 89 74 1.59 1.56 2.18 3.36 (1 = most important, 5 = least important) (1 = most important, 5 = least important) 8 (1 = great importance, 5 = no importance) Question 6: Assessment of benefits and risks or difficulties A list of possible benefits resulting from the implementation of A M T was provided as part of the survey instrument. The benefits included in this list were those that had been frequently mentioned in the literature. Respondents were also provided the opportunity to identify additional benefits from their AMT investment under the "Other" category. Respondents were asked to indicate the importance of each item at the time of the assessment on a scale from one (of great importance) to five (of no importance). The resulting mean responses are provided in Table 8 where the items are listed in order of importance. The five items rated of greatest importance were: reduced costs, obtaining competitive advantage, improved quality, increased throughput, and increased sales. The five items rated of least importance were: enhanced company image, overcoming production management skill deficiencies, improved ability to respond to variations in suppliers' lead times, improved ability to respond to engineering changes, and better working relations. These responses suggests that most firms are focusing on improving costs and quality (now generally recognised as order qualifiers) and missing the opportunities in terms of improving flexibility and responsiveness (the order winners). Table 8: Expected Benefits from AMT Investment Table 8: Expected Benefits from AMT Investment Table 8: Expected Benefits from AMT Investment Benefits claimed at the time of assessment Importance Ratine Reduced costs 1.3 Obtaining competitive advantage 1.4 Improved quality 1.5 Increased throughput 1.7 Increased sales 1.8 Reduced change over/set up times 1.9 Improved response to variation in product volume 2.0 Increased flexibility 2.1 Improved response to variations in product mix 2.2 Better Mgmt control 2.3 Reduced work in process 2.3 Improved integration of manufacturing IS 2.4 Improved integration of IS across functions 2.6 Improved work force attitudes 2.6 Improved working environment 2.6 Widening product range 2.6 Improved ability to respond to variations in suppliers quality 2.6 Improved ability to implement engineering changes 2.7 Improved Mgmt attitudes 2.8 Reduced product development time 2.8 Improved ability to respond to engineering changes 2.9 Better working relations 2.9 Overcoming skill deficiencies 3.0 Improved ability to respond to variations in suppliers ' lead times 3.1 Overcoming production Mgmt skill deficiencies 3.2 Enhanced comnanv image 3.7 (1 = great importance, 5 = no importance) 9 As with the benefits, a list of potential risks or difficulties anticipated at the time of proposal assessment was also included in the questionnaire. Respondents were asked to indicate the importance of each item on a scale from one (of great importance) to five (of no importance). The mean responses to the items are provided in Table 9. The main risks or difficulties anticipated by the respondents were disruptions during implementation, failure to achieve financial targets, and an adverse affect on the workflow. Of lesser importance were opposition by staff/management, A M T skill deficiencies, and project management skill deficiencies. Concerns about obsolescence of technology and opposition by the workforce were rated as moderately important. It would appear from these responses that managers lack project management skills in effectively introducing and implementing AMTs into their organisations. Table 9: Anticipated Risks and Difficulties During Proposal Assessment Anticipated Importance Risks and Difficulties Ratine Disruptions during implementation 1.68 Failure to achieve financial targets 1.83 Adverse effect on workflow 1.88 Obsolescence of technology 2.33 Problems with interconnection of equipment 2.38 Opposition by workforce 2.67 Lack of integration of Manufacturing IS 2.77 Lack of integration across functions 2.94 Prod Mgmt skill deficiencies 3.00 A M T skill deficiencies 3.06 OpDosition bv staff/manaeement 3.07 4. SUMMARY AND DISCUSSION The previous section presented the overall findings from the American survey. In this section the key findings from this study are discussed and a general comparison made with the 1993 Australian survey (see Table 10). As previously mentioned, only a broad comparison will be made between the present American study and the earlier Australian study. The reasons for this are threefold. First, the three-year time difference between the two studies. Second, the Australian responding companies were smaller in size. For example, 86% employed less than 1,000 people compared with 10% for the American sample. In term of annual sales revenue, 53% of the American companies had sales revenue greater than $1 billion compared with only 12% of the Australian companies. Third, the difference in the type of production system used. Nearly two- thirds (64%) of the Australian sample used jobbing or batch production compared with one-third (33%) of the American sample. In spite of these differences a number of similarities were identified between the two countries. The main reasons for investing in A M T for both American and Australian manufacturers was to obtain a competitive advantage or obtain financial benefits. These reasons appeared independent of company size. For some firms, attaining competitive advantage can be achieved through lowered costs, whereas for other firms, enhanced quality will be needed. Improvements in both costs and quality can be achieved from effectively implemented AMTs. However, the analysis showed that benefits relating to increased flexibility and responsiveness were not being considered by the responding American or Australian organizations (see Table 8). These are the more strategic benefits offered by AMTs and companies must focus on these to 10 obtain a competitive advantage for the future as cost and quality have become prerequisites in many markets. 4. SUMMARY AND DISCUSSION Table 10: Comparing American and Australian AMT Practices American Study - 1996 Australian Study - 1993 Reasons for AMT Investments • Obtaining Financial Benefits • Obtaining Competitive Advantage • Obtaining Competitive Advantage • Obtaining Financial Benefit Size of AMT Investment in the past 3 Years <US$ l m : 24% US$ 1 - 10m: 40% >US$ 10m: 36% <A$ lm: 60% A$ 1 - 10m: 24% >A$ 10m: 16% Nature of AMT Investment Most Popular Computer Hardware Personal Computers On-line Process Instrumentation Least Popular Computer Hardware Mainframes / Micros Wide Area Networks Most Popular Computer Software C A D / C A M C A D / C A M Least Popular Computer Software OPT/MAP MAP/OPT Most Popular Plant & Equipment Computer - Controlled Test Equipment CNC Machines Least Popular Plant & Equipment Laser Cutting/FMS/AGV's AGV's/ Laser Cutting Project Team Established 84% of Responding Firms 83% of Responding Firms Main Source of AMT Investment Ideas Engineering/ Production Engineering / Production Production/ Engineering/ Production Engineering Number of Management Levels Involved in Idea Generation One Level Only: 10% Two Levels: 25% Three Levels: 31% Four/Five Levels: 34% One Level Only: 84 % Two Levels: 8 % Three Levels: 0 % Four/Five Levels: 3 % Main Functions Involved in AMT Proposal Development Production/Production Engineering/ Engineering Production/Accounting & Finance / Engineering Financial Evaluation Techniques Used Payback Period: 89% ROI (Undiscounted): 84% Discounted Cash Flow: 89 % Sensitivity Analysis: 74 % Payback Period: 69 % ROI (Undiscounted): 47 % Discounted Cash Flow: 45 % Sensitivity Analysis: 40 % Table 10: Comparing American and Australian AMT Practices 11 Table 10: Continued Table 10: Continued American Study -1996 Australian Study -1993 Perceived Need for Training Implemented Training In AMT: 65% in Production Mgmt: 70% In AMT: 57% In Production Mgmt.: 68% In A M T : 55% In Production Mgmt.: 68% In A M T : 60% In Production Mgmt. : 5 7% Financial Assessment of Investment Exceeded Criteria: 55% Just Met Criteria: 45% Failed to Meet Criteria: 0% Missing Data: 0% Exceeded Criteria: 45% Just Met Criteria: 39% Failed to Meet Criteria: 8% Missing Data: 8% Implementation Time Less Than 6 Months: 7% 6-12 Months: 48% More Than 12 Months: 45% Not Yet Completed 0% Less Than 6 Months: 27% 6-12 Months: 29% More Than 12 Months: 30% Not Yet Completed: 14% Anticipated Benefits - Most Important Anticipated Benefits - Least Important 1. Reduced Costs 2. 4. SUMMARY AND DISCUSSION In comparison, 24% of the American respondents made investments which were less than US$1 million and 36% made investments which were greater than US$10 million. A factor contributing to this difference was the worldwide slowdown in economic activity during the 1990 to 1993 period. A comparison of the 1993 and the 1989 Australian findings showed that in 1993 Australian companies were making smaller and fewer A M T investment compared to 1989 (Sohal, 1997). Another contributing factor is the smaller organizations in the Australian sample. This size difference may also account for the fewer management levels involved in the decision-making process in Australia. Only one level of Australian management was involved in 84% of the organizations whereas in nearly two-thirds of the American organizations, three or more levels of management were involved in A M T idea generation and development. In both countries ideas for A M T investments largely came from the Engineering, Production Engineering and Production functions and these were also the functions which were mainly involved in developing the proposal. In addition, a large number of organizations in both America (84%) and Australia (83%) established project teams to develop the A M T proposal. As shown in Table 10, just over one-quarter of the Australian organizations had completed their A M T implementation in less than six months. This compares with only 7% of the American firms. One reason for this could be the smaller size of the A M T investments made by the Australian organizations as well as the smaller-sized Australian organizations themselves. In this case decision-making may be quicker with fewer individuals having responsibility for a greater number of functional areas. Clearly, there are risks involved if the planning and implementation periods are too long. One risk is that competitors can adopt the same technology quickly and obtain the competitive advantage. The other risk is that the A M T becomes obsolete as the rate of technological advancement has increased significantly over recent years. Although many American and Australian respondents noted that training is needed in the relevant technology and production management issues, generally fewer companies implemented training. Clearly, workers must understand how to operate effectively new equipment, but to gain the business benefits referred to by Voss (1988), a detailed understanding of a broad range of managerial issues is required. 4. SUMMARY AND DISCUSSION Obtaining Competitive Advantage 3. Improved Quality 4. Increased Throughput 5. Increased Sales 1. Enhanced Company Image 2. Overcoming Production Management Skills 3. Improved Ability to Respond to Variation in Supplier's Lead Times 4. Improved Ability to Respond to Engineering Changes 5. Better Working Conditions 1. Reduced Costs 2. Obtaining Competitive Advantage 3. Improved Quality 4. Increased Throughput 5. Increased Sales 3. Overcoming Production Management Skills 4. Better Working Relationships 5. Reduced Product Development Time 6. Widening Product Range 7. Improved Ability to Respond to Variations in Supplier's Quality Anticipated Risks and Difficulties - Most Important Anticipated Risks and Difficulties - Least Important 1. Disruption During Implementation 2. Failure to Achieve Financial Targets 3. Adverse Effect on the Workflow 4. Opposition from Staff/Management 5. AMT Skill Deficiencies 6. Production Management Skills 1. Disruption During Implementation 2. Adverse Effect on the Workflow 3. Failure to Achieve Financial Targets 4. Opposition from Staff/Management 5. Obsolescence of Technology 6. Opposition by Workforce Although competitive advantage was the most important motivation, conventional financial investment criteria were applied. A higher proportion of American companies made use of the four different financial evaluation techniques listed in the questionnaire ROI DCF and Sensitivity Analysis were used by less than Although competitive advantage was the most important motivation, conventional financial investment criteria were applied. A higher proportion of American companies made use of the four different financial evaluation techniques listed in the questionnaire. ROI, DCF and Sensitivity Analysis were used by less than 12 half of the Australian sample. This compares with over three-quarters of the American sample making use of these techniques. A slightly higher proportion of American companies exceeded the normal investment criteria and none indicated failing to meet their normal investment criteria. In comparison, 8% of the Australian respondents said that they had failed to meet the normal investment criteria. There are close similarities between American and Australian manufacturers in terms of the types of A M T adopted. Computer-controlled machines/test equipment and C A D / C A M are the most popular AMTs adopted in both countries. However, the size of the A M T investments made by Australian manufacturers is significantly smaller than their American counterparts. For example, 60% of the Australian respondents made investments which were less than A $ l million and only 16% made investments which were greater than A$10 million. 4. SUMMARY AND DISCUSSION Case study based investigations of A M T implementations have shown that people are the most critical factor for success (Sohal, 1996). The involvement of shop floor personnel must be considered throughout all phases of planning and implementation of AMT. Mechanisms and procedures must be established so that experiences are recorded and shared amongst all employees. In terms of the anticipated benefits and risks and difficulties, the responses from the American organizations are identical to those given by their Australian counterparts. The more strategic benefits offered by AMTs (i.e. greater flexibility and responsiveness) are being ignored by both American and Australian 13 organizations and from the responses given to potential risks and difficulties, it appears that managers in both countries require enhanced project management skills. organizations and from the responses given to potential risks and difficulties, it appears that managers in both countries require enhanced project management skills. A major problem identified from both the American and the Australian surveys is that very few organizations undertake a post-implementation audit of their A M T projects. To become a "learning organization" firms must conduct such audits, and by doing so, establish a database which can be used for future A M T investments. Our studies suggests that organisational change in terms of the company's culture and structure, as suggested by Bessant (1994), and Zammutto and O'Connor (1992), is not taking place in organisations that have invested in AMT. This must happen if the strategic benefits of A M T are to obtained. 5. REFERENCES Adler, P. S., "Managing Flexible Automation", California Management Review, Vol. 30, No. 3, 1988, pp. 34-56. Beatty, C. A., "Implementing advanced manufacturing technologies: rules of the road", Sloan Management Review, Vol. 33, No. 4, 1992, pp. 49-60. Bessant, J., "Towards Total Integrated Manufacturing", International Journal of Production Economics, Vol. 34, No. 3, 1994, pp. 237-251. Boyer, K. K., Ward, P. T., and Leong, G. K., "Approaches to the factory of the future: an empirical taxonomy", Journal of Operations Management, Vol. 14, No. 4, 1996, pp. 297-313. Currie, W. L., "Investing in CAD: a case of ad hoc decision-making", Long Range Planning, Vol. 22, No. 6, 1989, pp. 85-91. Duimering, R., Safeyeni, F., and Purdy, L., "Integrated manufacturing: redesign the organization before implementing flexible technology", Sloan Management Review, Vol. 34, No. 4, 1993, pp. 47-55. Gold, B., "Charting a course to superior technology evaluation", Sloan Management Review, Vol. 28, No. 3, 1988, pp. 19-27. Howell, R. A. and Soucy, S. R., "Capital investment in new manufacturing environments", Management Accounting, November, 1987, pp. 26-32. Kaplan, R., "Must CIM be justified by faith alone?", Harvard Business Review, March/April, 1986, pp. 87- 95. King, W. R., and Grover, J., "The strategic use of information resources: an exploratory study", IEEE Transactions on Engineering Management, Vol. 38, No. 4, 1991, pp. 293-305. King, W. R., and Ramamurthy, K., "Do organizations achieve their objectives from computer based manufacturing technologies?", IEEE Transactions on Engineering Management, Vol. 39, No. 2, 1992, pp. 129-141. Markus, M . L., and Keil, M . , "If we build it, they will come: designing information systems that people want to use", Sloan Management Review, Vol. 35, No. 4, 1994, pp. 11-25. Raho, L. E., Belohar, J. A., and Fielder, K. D., "Assimilating technology into the organization: an assessment of the McFarland-McKenney model", MS Quarterly, Vol. 11, No. 1, 1987, pp. 47-57. 14 Saraph, J. V., and Sebastian, R. J., "Human resource strategies for effective introduction of advanced manufacturing technologies", Production and Inventory Management, Vol. 33., 1992, pp. 64-70. Singh, M . and Sohal, A. S., "The role of the engineer in technological change", Technology Management, Vol. 2, No. 5, 1995, pp. 210-215. Small, M.H. and Chen, I.J., "Investment justification of advanced manufacturing technology: An empirical analysis", Journal of Engineering and Technology Management, Vol. 12, 1995, pp. 27-55. Zammuto, R. F. and O'Connor, E. J., "Gaining advanced manufacturing technologies' benefits: The role of organisational design and culture", Academy of Management Review, Vol. 17, 1992, pp. 701-728. 1998 WORKING PAPERS 1/98 Alison Dean, "Issues and Challenges in Training HRM Practitioners by Distance Education" (January, pp.16). 2/98 Simon Moss, "Exposing Biased Samples: Derivation of the Sample-Subdivision Method" (January, pp.10). 3/98 Ian Roos, "Technical Regulation and Work Autonomy: Human Resource Management in a Specific Pathogen Free Animal Unit" (January, pp.15). 4/98 Loong Wong, "The State, Economic Growth and Environment in Malaysia" (January, pp.21). 5/98 Tim Haslett, "The Senge Archetypes: From Causal Loops to Computer Simulation" (January, pp.22). 6/98 Loong Wong, "Management Theory Meets the 'Other'" (January, pp. 15). 6/98 Loong Wong, "Management Theory Meets the 'Other'" (January, pp. 15). 7/98 Tim Haslett, "Implications of Systems Thinking for Research and Practice in Management" (January, pp. 19). 8/98 Jan Schapper, '"We had no Choice. It was Inevitable.' Some Thoughts on Parallel Processes Between Researcher and Researched in Response to Organizational Change" (January, pp. 17). 9/98 Tim Haslett, "The Dynamics of Garbage Collection: A Case Study of Privatization" (January, pp. 17). 10/98 Tim Haslett, Simon Moss, Charles Osbome and Paul Ramm, "The Application of Local Rules in Self Ordering Systems" (January, pp.17). 11/98 Ramanie Samaratunge, "Decentralisation and Development: Partners in the 21 s t Century?" (January, pp. 15). 12/98 Tim Haslett, "Teaching Systems Thinking and Theory: Management Education at Monash University" (January, pp.11). 13/98 Tim Haslett, "Building the Learning Organization - The Practical Steps" (January, pp. 10). 14/98 Mary Anderson and Daniel Moore "Classroom Globalization, "An Investigation of Teaching Methods to address the phenomemon of Students from Multiple National Cultures in business School Classrooms" (January, pp.7). 15/98 Judy H. Gray and Iain L. Densten, "Analysis of Latent and Manifest Variables in a Study of Small Business Strategy" (January, pp.13). 16/98 Kathryn M. Antioch, Chris Selby-Smith and Chris Brook, "Pathways to Cost Effective Prevention, Detection and Treatment of Prostrate Cancer in Australia: Achieving Goals for Australia's Health to 2000 and beyond" (January, pp.31). 17/98 Chris Selby-Smith, "The Impact of Vocational Education and Training Research on Policy, Practice and Performance in Australia" (January, pp. 17). 18/98 Mile Terziovski, Amrik Sohal and Simon Moss "Longhudunal Analysis of Quality Management Practices in Australian Organisations (January, pp.14). 19/98 Linda Brennan and Lynne Bennington, "Concepts in Conflict: Studies and Customers" (January, pp.15). 20/98 Dianne WaddelL "The Role Responsibilities Quality Managers" (January, pp.10). 21/98 Dianne WaddelL "Resistance to Change: A Company's Experience" (January, pp.13). 22/98 Iain L. Densten and Judy H. Gray, "Is Management-by-Exception a Single Factor? (January, pp.13). 5. REFERENCES Sohal, A.S., "Investing in Advanced Manufacturing Technology: Comparing Australia and the United Kingdom", Benchmarking for Quality Management and Technology: An International Journal, Vol. 1, No. 2, 1994, pp. 24-41. Sohal, A . S., "Assessing A M T implementations: an empirical field study", Technovation, Vol. 16, No. 8, 1996, pp. 377-384. Sohal, A., " A longitudinal study of planning and implementation of advanced manufacturing technologies", International Journal of Computer Integrated Manufacturing, Vol. 10 No. 1-4, 1997, pp. 281-295. Sohal, A., Maguire, W. A., and Putterill, M . S., " A M T investment in New Zealand: purpose, pattern, and outcomes", Integrated Manufacturing Systems, Vol. 7, No. 2, 1996, pp. 27-36. Sohal, A., Samson, D., and Weill, P., "Manufacturing and technology strategy: a survey of planning for AMT", Computer Integrated Manufacturing Systems, Vol. 4, No. 2, 1991, pp. 71-79. Thomas, P. and Wainwright, D., "Gaining the benefits of integrated manufacturing technology - just who benefits and how?", International Journal of Production Economics, Vol. 34, pp. 371-381. Voss, C. A., "Success and failure of advanced manufacturing technology", International Journal of Technology Management, Vol. 3, No. 3, 1988, pp. 285-297. Zairi, M . , "The role of consultants in A M T implementation", Journal of General Management, Vol. 18, 1992, pp. 63-73. Zammuto, R. F. and O'Connor, E. J., "Gaining advanced manufacturing technologies' benefits: The role of organisational design and culture", Academy of Management Review, Vol. 17, 1992, pp. 701-728. 15 1998 WORKING PAPERS 23/98 Mile Terziovski, "Best Predictors of High Performance Quality Organisations: Evidence from Australia and New Zealand" (March, pp.16). 24/98 Ronald W. Edwards and Peter J. Buckley, "Choice Ownership Mode and Entry Strategy: The Case of Australian Investors in the UK" (January, pp.18). 25/98 Tim Haslett and Charles Osborne, "Local Decision Rules: Complexity or Chaos?" (January, pp.14). 26/98 Ian Roos and T. Makela, "Employee Reactions to Controlled work Environments: The Dispensing of Anti-Cancer Drugs in Hospital Pharmacies" (January, pp.29). 27/98 Tim Haslett, Kosmas X. Smyrnios and Charles Osborne, "A Cusp Catastrophe Analysis of Anxiety Levels" (January, pp. 18). 28/98 Megan Seen and Anne Rouse, "Quality Certification: Lessons from Three Software Development Organisations" (March, pp.13). 29/98 E. Anne Bardoel and Tim Haslett, "The Use of Systems Thinking and Archetypes in Teaching Organisational Behavior" (March, pp.10). 30/98 Megan Seen and Anne Rouse, "The Effect of Quality Certification on the Software Development Process" (March, pp. 13). 31/98 Michael Morrison and Mile Terziovski, "The Relationship Between Quality Management Practices and Learning Outcomes: In the Australian Retail Hardware Sector" (March, pp.15). 32/98 Marjorie Jerrard, "Dinosaurs are not Dead - The Success of the AMIEU (QLD) in Coping with Industrial Relations Change and AWAS" (March, pp.20). 33/98 Lynne Bennington and James Cummane, "Customer Satisfaction, Loyalty and Public Services" (March, pp. 19 1998 WORKING PAPERS 34/98 Alison Dean, "Managing Quality Initiatives in Services: JIT Delivers but BPR Fails" (March, pp. 11). 35/98 Marjorie Jerrard, "A Surprising Struggle? The AMIEU(Qld) and the Fight for Equal Wages in the Meat Processing and Export Industry in the 1950s and 1960s" (March, pp.15). 36/98 Julie Wolfram Cox, Helen De Cieri and Marilyn Fenwick, "The Mapping of Strategic International Human Resource Management: Theory Development or Intellectual Imperialism?" (April, pp.23). 37/98 Max Coulthard and Timothy James Grogan, "The Impact of a Firm's Strategic Orientation on Environmental Scanning Practices in Two Australian Export Industries" (April, pp.13). 38/98 John W. Selsky, '"Even we are Sheeps': Cultural Displacement in Management Education" (April, pp.13) 39/98 Rowena Barrett, "Industrial Relations and Management Style in Small Firms" (April, pp. 18). 40/98 Loong Wong, "Why Jerry Maguire succeeds but not William Lomax: Management, Cultures and Postmodernism" (Apr pp.12). 41/98 Sarah Turberville. "The Nature of Employee Financial Participation: Evidence from the Australian Workplace" (June, pp.32). 42/98 Marilyn S. Fenwick, Helen L. De Cieri and Denice E. Welch "Cultural and Bureaucratic Control in MNEs: The Role of Expatriate Performance Management" (June, pp.16). 43/98 Stuart Orr and Amrik S. Sohal "Technology and Global Manufacturing: Some German Experiences" (June, PP-9). 44/98 Stuart Orr and Amrik S. Sohal "Global Manufacturing Issues: The Case of Siemens A G " (June, pp.12). 45/98 Robert Millen and Amrik S. Sohal "Planning Processes for Advanced Manufacturing Technology by Large American Manufacturers" (June, pp.15). 46/98 Amrik S. Sohal and Lionel Ng "The Role and Impact.of Information Technology in Australian Businesses" (June, PP-25). 47/98 Marcia Perry, Amrik S. Sohal and Peter Rumpf "Quick Response Supply Chain Alliances in the Australian Textiles, Clothing and Footwear Industry" (June, pp.16). 48/98 Andrea Howell and Amrik S. Sohal "Human Resources and training - The Core of Quality Improvement Initiatives" (June, pp.18). 49/98 John Gordon and Amrik S. Sohal "Assessing Manufacturing Plant Competitiveness: An Empirical Field Study" (June, PP-19). 50/98 Mile Terziovski and Danny Samson "Increasing and Sustaining Performance through an Integrated Quality Strategy" (June, pp.14).
https://openalex.org/W2063692230	https://europepmc.org/articles/pmc4437447?pdf=render	English	null	Differential transcriptional networks associated with key phases of ingrowth wall construction in trans-differentiating epidermal transfer cells of Vicia faba cotyledons	BMC plant biology	2,015	cc-by	12,238	Zhang et al. BMC Plant Biology (2015) 15:103 DOI 10.1186/s12870-015-0486-5 Zhang et al. BMC Plant Biology (2015) 15:103 DOI 10.1186/s12870-015-0486-5 Open Access © 2015 Zhang et al.; licensee BioMed Central. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0) which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Hui-Ming Zhang1, Simon Wheeler1, Xue Xia1, Ruslana Radchuk2, Hans Weber2, Christina E Offler1 and John W Patrick1* Hui-Ming Zhang1, Simon Wheeler1, Xue Xia1, Ruslana Radchuk2, Hans Weber2, Christina E Offler1 and John W Patrick1* * Correspondence: John.Patrick@newcastle.edu.au 1School of Environmental and Life Sciences, The University of Newcastle, Callaghan NSW 2308, Australia Full list of author information is available at the end of the article Background nematode infection [10]. Events downstream from the inductive signals largely have been identified by tran- scriptome analyses of developing cereal seeds and, in particular, for flange wall ingrowths formed in basal endosperm TCs (BETCs) of maize [11] and barley [8,12]. A more limited analysis has been reported for the for- mation of reticulate wall ingrowths in eudicots [10,13]. Transfer cells (TCs) trans-differentiate from a range of existing cell types belonging to the major tissue systems of dermal (e.g., epidermal cells), ground (e.g., endosperm cells; root cortical parenchyma cells) and vascular (phloem and xylem parenchyma cells; companion cells – [1]). Once TC development is completed, a 10- to 20- fold amplification of their plasma membrane surface area, containing high densities of solute transporters, confer these cells with an extraordinarily high capacity for nutrient exchange between apo- and symplasmic compartments located at bottlenecks for long-distance transport of nutrients throughout the plant body [1]. An intricate invaginated complex of cell wall ingrowths pro- vides structural scaffolding on which the amplified plasma membrane is arrayed to enhance nutrient flows. The tight coupling of this structure/function relationship is graphically illustrated by compromised seed filling of mutants in which TCs, located at the maternal/filial in- terfaces of both eudicots and monocots, exhibit an at- tenuated construction of their cell wall ingrowth complex [2,3]. Since a number of major crop species, including ce- reals and grain legumes, contain TCs positioned at the maternal/filial interfaces of their developing seeds [1], the shrivelled seed phenotype exhibited by TC mutants under- scores the important role these cells play in determining crop yields. While there are undoubtedly some shared features, given the disparate architectures of flange and reticulate wall ingrowths, we hypothesize that signalling pathways, cell wall biosynthesis and delivery of polysaccharides to the wall matrix will have features that are peculiar to each architectural type of ingrowth wall. This character- istic likely extends to distinctive elements responsible for constructing the uniform wall and wall ingrowth papil- lae. To this end, we used Vicia faba cotyledons, in which on transfer to culture, their adaxial epidermal cells spontan- eously undergo trans-differentiation to a TC-morphology [1]. This experimental system provided the opportunity to undertake a transcriptome analysis of TC-specific gene net- works [1]. Abstract Background: Transfer cells are characterized by intricate ingrowth walls, comprising an uniform wall upon which wall ingrowths are deposited. The ingrowth wall forms a scaffold to support an amplified plasma membrane surface area enriched in membrane transporters that collectively confers transfer cells with an enhanced capacity for membrane transport at bottlenecks for apo-/symplasmic exchange of nutrients. However, the underlying molecular mechanisms regulating polarized construction of the ingrowth wall and membrane transporter profile are poorly understood. Results: An RNAseq study of an inducible epidermal transfer cell system in cultured Vicia faba cotyledons identified transfer cell specific transcriptomes associated with uniform wall and wall ingrowth deposition. All functional groups of genes examined were expressed before and following transition to a transfer cell fate. What changed were the isoform profiles of expressed genes within functional groups. Genes encoding ethylene and Ca2+ signal generation and transduction pathways were enriched during uniform wall construction. Auxin-and reactive oxygen species-related genes dominated during wall ingrowth formation and ABA genes were evenly expressed across ingrowth wall construction. Expression of genes encoding kinesins, formins and villins was consistent with reorganization of cytoskeletal components. Uniform wall and wall ingrowth specific expression of exocyst complex components and SNAREs suggested specific patterns of exocytosis while dynamin mediated endocytotic activity was consistent with establishing wall ingrowth loci. Key regulatory genes of biosynthetic pathways for sphingolipids and sterols were expressed across ingrowth wall construction. Transfer cell specific expression of cellulose synthases was absent. Rather xyloglucan, xylan and pectin biosynthetic genes were selectively expressed during uniform wall construction. More striking was expression of genes encoding enzymes for re-modelling/degradation of cellulose, xyloglucans, pectins and callose. Extensins dominated the cohort of expressed wall structural proteins and particularly so across wall ingrowth development. Ion transporters were selectively expressed throughout ingrowth wall development along with organic nitrogen transporters and a large group of ABC transporters. Sugar transporters were less represented. Conclusions: Pathways regulating signalling and intracellular organization were fine tuned whilst cell wall construction and membrane transporter profiles were altered substantially upon transiting to a transfer cell fate. Each phase of ingrowth wall construction was linked with unique cohorts of expressed genes. Background In addition, because of the temporal sequence for the deposition of the uniform wall followed by wall in- growth papillae, gene cohorts associated with these two de- velopmental events could be distinguished. Hereafter, wall ingrowth papillae, which represent the initial stage of wall ingrowth construction [4], are referred to as wall ingrowths. The cell wall ingrowth complex is organized into one of two architectural types – flange or reticulate [1]. Flange ingrowths form as ribs or bands of wall material while the more commonly occurring reticulate wall in- growths arise as numerous wall papillae that develop at right angles to the original wall. The extent of their re- ticulation varies from cylindrical papillae alone to ones in which papillae branch and fuse to generate fenes- trated wall layers [4]. Immediately preceding construc- tion of reticulate, but not flange wall ingrowths [5], a structurally distinctive wall-layer, the so-called uniform wall, is rapidly laid down over the pre-existing primary wall of the trans-differentiating cell to a thickness that can extend up to 50% of that of the pre-existing primary wall [6]. Collectively, the uniform wall and reticulate wall ingrowths form the ingrowth wall [1]. Abstract Keywords: Transfer cell, Transcriptome, Cell wall construction, Uniform wall, Wall ingrowth, Transporters, Seed development, Vicia faba * Correspondence: John.Patrick@newcastle.edu.au 1School of Environmental and Life Sciences, The University of Newcastle, Callaghan NSW 2308, Australia Full list of author information is available at the end of the article Zhang et al. BMC Plant Biology (2015) 15:103 Page 2 of 13 Plant growth conditions, cotyledon culture and collection of tissue samples for sequencing V. faba L. (cv. Fiord) plants were raised under controlled environmental conditions [13]. Cotyledons of harvested pods were surgically excised and cultured aseptically on liquid MS medium for specified times (see below) before being fixed in 75% ethanol and 25% acetic acid for 1 h at 4°C. Peels of the adaxial epidermis and blocks of storage parenchyma cells (2 × 2 × 1 mm) were surgically re- moved from each fixed cotyledon. Collected tissues were immediately snap-frozen in liquid nitrogen and stored at -80°C until used for RNA extraction. Selection of times to obtain tissue samples from cul- tured cotyledons was based on the temporal pattern of uniform wall deposition preceding that of wall ingrowths (Figure 1 and, for more information, see Additional file 1). To this end, for the reference library, a representative bal- ance of genes induced to regulate uniform wall and wall ingrowth construction was obtained by collecting 6 mg of epidermal peels from each of freshly harvested cotyledons (0 h – reference to identify genes induced/switched off during trans-differentiation) and cotyledons cultured for 1 and 3 h (dominated by expression of genes regulating uni- form wall formation), 6 and 9 h (dominated by expression of genes regulating deposition of wall ingrowths as uni- form wall formation ceased). To identify expression of Despite the central role the ingrowth wall, and particu- larly its wall ingrowth component, plays in underpinning transport function of TCs, little is known about the mechanisms responsible for inducing and then orches- trating its construction. Transcriptional analyses of cells committed to a developmental pathway leading to form- ing a TC-morphology have identified up-regulated ex- pression of genes encoding components of signalling pathways for abscisic acid, auxin, ethylene, reactive oxy- gen species (ROS) and Ca2+ in developing seeds [1,7-9] and for auxin and ethylene in giant cells induced by Page 3 of 13 Zhang et al. BMC Plant Biology (2015) 15:103 Uniform wall Wall ingrowth EC Uniform wall enriched Wall ingrowth enriched Reference library Uniform wall enriched EC Wall ingrowth specific cDNA libraries for transcriptome analyses Relative deposition rate Sampling times O 3 6 9 12 15 Culture time (h) Figure 1 Tissue sampling times (vertical arrows) informed by the sequential temporal deposition of the uniform wall followed by wall ingrowth papillae. Plant growth conditions, cotyledon culture and collection of tissue samples for sequencing Note that across the period sampled, wall ingrowth papillae continue to be deposited and have not commenced to form the first fenestrated layer [4]. EC, epidermal cell. TC-specific genes regulating uniform wall and wall in- growth construction, epidermal peels (15 mg) and storage parenchyma tissue (30 mg) were sampled from replicate batches of freshly harvested cotyledons and cotyledons cultured for 3 h (dominated by uniform wall construction) and 12 h (wall ingrowth construction alone and for add- itional information see Additional file 1 and Results). Using Illumina CASAVA pipeline version 1.8.2, raw reads were trimmed with adaptor filtering and a read length cut-off of 50%. Thereafter, filtered reads with over 20% of their nucleotides having a Q score < 20 (probabil- ity of sequencing error > 0.01) or their sequences having a N reading over 5% were removed. To verify that each sampled population of cultured coty- ledons was trans-differentiating, percentages of epidermal cells forming wall ingrowths of a sub-set of cotyledons at 12 h of culture were scored (see Additional file 1 for Methods). Plant growth conditions, cotyledon culture and collection of tissue samples for sequencing Note that across the period sampled, wall ingrowth papillae continue to be deposited and have not commenced to form the first fenestrated layer [4]. EC, epidermal cell. sampled from each biological replicate of epidermal and storage parenchyma RNA extracts. Uniform wall Wall ingrowth EC Uniform wall enriched Wall ingrowth enriched Reference library Uniform wall enriched EC Wall ingrowth specific cDNA libraries for transcriptome analyses Relative deposition rate Sampling times O 3 6 9 12 15 Culture time (h) Total RNA quality was verified by determining the in- tegrity of the 25S and 18S RNA with an Agilent 2100 Bioanalyzer (Agilent, USA and see Additional file 2). The cDNA libraries were prepared from poly-A mRNA isolated from 1 μg of total RNA using a TruSeq® RNA v2 sample prep kit (Illumina, USA) according to manufac- turer’s instructions. cDNA quality was evaluated by de- termining size and purity using an Agilent 2100 bioanalyzer (see Additional file 3). cDNA fragments, ran- ging from 100 – 700 bp, were selected by agarose gel purification. For the reference library, selected cDNA fragments were 100 bp pair-end sequenced in a single lane on an Illumina HiSeq 2000 platform (Australian Genome Research Facility, Melbourne). Relative deposition rate To characterize gene cohorts linked with uniform wall and wall ingrowth deposition, only biological replicates with the anticipated 70% (or more) of their epidermal cells at 12 h of cotyledon culture containing detectable wall ingrowths [14] were processed for Illumina sequen- cing. In three compliant biological replicates for each specified harvest time, cDNA libraries (18 in total) were prepared from total RNA extracts of epidermal peels and storage parenchyma tissues as indicated above (and see Additional files 2 and 3 for RNA and cDNA quality checks). cDNA fragments were purified separately from each sample and indexed with unique nucleic acid iden- tifiers (Illumina TruSeq V2 index sequence). The indexed cDNA libraries were diluted to an average con- centration of 10 nM and pooled in equal volumes (10 μL of each library) to generate the final mixed cDNA pool for sequencing. The pool was then 100 bp pair-end se- quenced in two lanes on an Illumina HiSeq 2000 plat- form (Australian Genome Research Facility, Melbourne). Figure 1 Tissue sampling times (vertical arrows) informed by the sequential temporal deposition of the uniform wall followed by wall ingrowth papillae. De novo assembled transcriptome library for ingrowth wall deposition Mapping reads to the reference library, determining differentially expressed genes and GO-enrichment analysis Reads sequenced by the Illumina HiSeq 2000 platform from mRNA extracts of cotyledon epidermal peels and storage parenchyma tissues were analyzed by BGI. Raw reads were filtered as described for the reference tran- scriptome library. Clean reads were aligned to the refer- ence sequences (generated as described above) using the SOAPallgner/SOAP2 pipeline. No more than 5 mis- matches per read were allowed to ensure high quality alignment. Sequence coverage of unigenes in each sam- ple was calculated. During alignment, SOAPallgner/ SOAP2 reported the number of mapped reads per kilo base per million reads (RPKM) as a measure of tran- script abundance of each unigene. Transcripts with RPKMs < 0.45 were considered not to be expressed and were removed from the data sets. Sequencing of reference library cDNA fragments on an Illumina HiSeq 2000 platform generated 181,419,640 pair-end reads (100 bp). After filtering raw reads, 126,200,279 high-quality reads, in which 97% of nucleo- tides have a Phred quality score of ≥Q20 level (error probability ≤0.01) (Table 1) were de novo assembled into 131,279 contigs (>200 bp) with a N50 value of 980 bp and an average length of 423 bp. Clustering yielded 74,659 unigenes with 33,902 consensus sequences grouped into 11,083 distinct clusters and 41,567 single- tons (Table 1). This unigene population had a N50 value of 1723 bp and an average length of 1076 bp. High in- tegrity and accuracy of the assembly was indicated by 85.2% of filtered reads mapping to the assembled tran- scriptome with 41.4% of reads uniquely mapping to the transcriptome (Table 1). The uniquely mapped reads had an average coverage depth of 8.4 (12,620 bp per mRNA). Differentially expressed genes (DEGs) were determined using a computational algorithm based on digital gene expression profiles [16] to perform pairwise differential expression analysis. Multiple testing of the comparisons was corrected using the FDR method [17]. Sequences with a corrected FDR P value of no more than 0.05 were selected. After this correction, genes exhibiting a differ- ential expression of two-fold or more (P < 0.05) were identified as DEGs. Subcellular localization of selected proteins encoded by unigenes was predicted using the WoLF PSORT algorithm. BLASTX searches of publically available protein data- bases, using the putative unigenes as query sequences, yielded matches for 68.9% of the unigenes. De novo assembly, annotation and GO classification of transcriptome library Total RNA was extracted using Qiagen (USA) RNeasy plant mini kits. Contaminating genomic DNA was re- moved using DNase I. Filtered reads from the cDNA library were assembled de novo into contigs by Beijing Genomics Institution (BGI, Shenzhen, China) using Trinity software release- 20130225 [15] with a k-mer of 25 and a minimum k-mer threshold abundance of 2 (min_kmer_cov 2). The reads were then mapped back to contigs to assemble unigenes using Trinity Butterfly that filtered out transcriptional artifacts, misassembled transcripts and poorly supported transcripts. Potential redundant sequences were grouped, For preparation of an epidermal-specific TC transcrip- tome reference library, 1 μg of total RNA was aliquoted from the total RNA extracts of each specified harvested time point. The RNA aliquots were combined to gener- ate a temporally mixed RNA library (5 μg of total RNA). To identify genes responsible for uniform wall and wall ingrowth formation, 2 μg aliquots of total RNA were Zhang et al. BMC Plant Biology (2015) 15:103 Page 4 of 13 using the TGI Clustering Tool set with a minimum 40% sequence overlap and over 80% sequence identity. Assembled unigene sequences were annotated by alignment to the following publically available databases: NCBI nr (non-redundant, see http://www.ncbi.nlm.nih. gov/refseq/); Swissprot (http://www.uniprot.org/); KEGG (http://www.genome.jp/kegg/); COG (https://www.ncbi. nlm.nih.gov/COG/), using BLASTX with an e-value threshold of 1e-5. Unigenes with no hit in BLASTX were predicted using ESTScan. using the TGI Clustering Tool set with a minimum 40% sequence overlap and over 80% sequence identity. 95°C for 15 s, 60°C for 20s, 72°C for 30 s; steps 2 to 4 were repeated 50 times. High-resolution melting curves (72–95°C) following the final PCR cycle checked the specificity of the PCR products. For each cDNA sample, technical duplicates in each of three biological repli- cates were tested. Four housekeeping gene candidates were assessed using GeNorm. Relative expression levels of each unigene were determined using the two stand- ard curves method. Assembled unigene sequences were annotated by alignment to the following publically available databases: NCBI nr (non-redundant, see http://www.ncbi.nlm.nih. gov/refseq/); Swissprot (http://www.uniprot.org/); KEGG (http://www.genome.jp/kegg/); COG (https://www.ncbi. nlm.nih.gov/COG/), using BLASTX with an e-value threshold of 1e-5. Unigenes with no hit in BLASTX were predicted using ESTScan. De novo assembled transcriptome library for ingrowth wall deposition BLASTX- derived alignments predicted sequence orientation for 43,910 of the unigenes and identified 43,703 open read- ing frames. Taxonomic distribution of the annotated unigenes by Blast2GO demonstrated that over 85% uni- genes had their best hits (smallest e value) with their ho- mologs in other legume species (see Additional file 7). The raw reads and assembled unigene sequences are Transcriptome networks specifically expressed during uniform wall or wall ingrowth formation g The temporal sequence of uniform wall followed by wall ingrowth deposition (Figure 1; Additional file 1) pro- vided the opportunity to separate the TC-specific cohort of expressed genes into groups linked with the two phases of ingrowth wall construction and those shared between these phases as follows. The latter were identi- fied by manual inspection of the 3- and 12-h gene ex- pression profiles (Figure 2B). The remaining expressed genes were then specifically linked with either uniform wall or wall ingrowth formation and these were identified on the following grounds. At 12 h of cotyledon culture, uniform wall construction had ceased and TC-specific gene expression is solely committed to deposition of wall ingrowths (Figure 1; Additional file 1). Thus the expres- sion profile at 12 h contained transcripts specifically re- lated to constructing wall ingrowths alone (Figure 2B). In contrast, at 3 h of cotyledon culture, uniform wall con- struction is proceeding rapidly concurrent with a smaller population of cells starting to deposit wall in- growths (Figure 1, Additional file 1). Hence at 3 h of culture, genes involved in uniform wall and wall in- growth construction would be co-expressed. However, those exclusively directing deposition of wall ingrowths would be expressed at relatively lower levels. Thus sub- tracting the 12-h (wall ingrowths alone) from the 3-h (uniform wall plus wall ingrowths) gene expression pro- file identified those genes specifically expressed during uniform wall construction (Figure 2B). An estimated 27,244 unigenes were expressed in ad- axial epidermal cells of developing V. faba cotyledons (Figure 2A). Of these epidermal cell unigenes, 21,293 were shared with the underlying storage parenchyma cells leaving 5,951 unigenes that were epidermal cell specific (Figure 2A). Upon transiting to a TC fate, ex- pression of 2,805 unigenes in the precursor epidermal cells was switched off (Figure 2A). The combined tran- scriptomes of trans-differentiating TCs at 3 and 12 h of cotyledon culture totalling 33,472 unigenes approxi- mated those detected by RNAseq analyses of differenti- ating BETCs in barley grains (42,086 – [8]) and of nematode giant TCs in rice roots (42,756 unigenes at 7 days post infection; 41,179 genes at 14 days post in- fection – [18]). Seventy-three % of the epidermal TC A B Figure 2 Numbers of unigenes expressed preferentially in, or shared between, (A) cell types and (B) uniform wall and wall ingrowth formation occurring within adaxial epidermal cells of cultured V. faba cotyledons. Global gene expression patterns during early phases of trans-differentiation to a TC-morphology Total RNA was extracted from epidermal peels and stor- age tissue blocks at 0, 3 and 12 h of cotyledon culture and processed for RNAseq on an Illumina platform (see Methods). Between 19 – 37 million clean reads were generated for each biological replicate (see Additional file 8). An average of 85% of the clean reads mapped to expressed unigenes. Of these reads, 60% mapped with a perfect match and 40% with a mismatch less than 5 bp (Additional file 8). Total numbers of unigenes, detected within epidermal and storage parenchyma cells across cotyledon culture were temporally stable (Additional file 9). Some 70% of the unigenes exhibited expression levels > 0.45 RPKM within a cell type and these were dominated by unigenes with expression levels ranging from 0.5 to 10 RPKM (80%). Quantitative RT-PCR validation A collection of 15 unigenes with different expression patterns (see Additional files 4 and 5 for expression pat- tern information) was selected. cDNA was converted from the same RNA samples sent for Illumina sequen- cing using QuantiTect Reverse Transcription Kit (Qiagen, USA). Primers were designed using Primer 3 plus (Whitehead Institute for Biomedical Research, USA) and synthesized by Sigma-Aldrich Australia (see Additional file 6 for primer sequences). For each qRT-PCR reaction a 15 μL system containing 7.5 μL SYBR Green master mix (Qiagen, USA), 0.375 μL of forward and reverse primers (10 μM), 1.75 μL of nuclease free H2O and 5 μL cDNA was set up. The following PCR cycle was used: 95°C for 5 min, Table 1 Raw, filtered and mapped reads used for the de novo assembly of the reference transcriptome library for adaxial epidermal cells of freshly harvested and cultured V. faba cotyledons undergoing trans-differentiation to a TC morphology Category Number Percentage (compared with total number of reads) Raw reads 181,419,640 100% Clean reads 126,200,279 69.56% Mapped reads 107,093,666 59.03% Uniquely-mapped reads 44,339,744 24.44% Multiple-mapped reads 62,753,921 34.59% Table 1 Raw, filtered and mapped reads used for the de novo assembly of the reference transcriptome library for adaxial epidermal cells of freshly harvested and cultured V. faba cotyledons undergoing trans-differentiation to a TC morphology Table 1 Raw, filtered and mapped reads used for the de novo assembly of the reference transcriptome library for adaxial epidermal cells of freshly harvested and cultured V. faba cotyledons undergoing trans-differentiation to a TC morphology Page 5 of 13 Zhang et al. BMC Plant Biology (2015) 15:103 available at European Nucleotide Archive, accession number: PRJEB8906. available at European Nucleotide Archive, accession number: PRJEB8906. transcriptome was represented by ongoing expression of genes by their precursor epidermal cells whilst 27% were induced (Figure 2A). The percentage of induced genes was comparable to the proportion of up- regulated genes in nematode giant TCs [18]. Of the in- duced genes, 66% were TC specific and the remainder was shared with the underlying storage parenchyma cells (Figure 2A). Ten % of the epidermal cell genes were silenced in transiting to an epidermal TC identity and these were replaced 3-fold by induced genes (Figure 2A). TC specific genes were inspected manually to remove du- plicates and those with < 70% coverage to yield a total of 4,283 unigenes (Figure 2B). Transcriptome networks specifically expressed during uniform wall or wall ingrowth formation Temporal patterns of expression of selected genes from the targeted functional categories determined by qRT-PCR were consistent with those de- rived from Illumina sequencing (Additional file 15), BlastX searches, at a stringency of < e-5, established that 41% of the TC-specific genes and 34% of epidermal cell genes switched off shared homology with known genes listed in public databases (Table 2). The 59/66 % of un-known transcripts equates with a similar value reported for the transcriptome of developing barley BETCs [12]. Table 2 provides a summary of the changes in expres- sion of selected functional categories of TC-specific genes that are known to be central to TC development and function. For all categories, numbers of epidermal cell genes switched off in transiting to a TC fate were re- placed by increased numbers of TC-specific transcripts between 3- to 12-fold. Amongst these, expression of transcription/translation, receptor kinase/kinases, cyto- skeleton/vesicle trafficking, matrix polysaccharides and defense genes contributed to the two-fold greater num- ber of genes expressed during uniform wall compared to wall ingrowth construction (Table 2; Figure 2B). In addition, a portion of uniform wall associated expressed genes were linked with a co-occurring burst in cell div- ision/expansion (Additional file 10). Transcriptome networks specifically expressed during uniform wall or wall ingrowth formation BlastX searches, at a stringency of < that 41% of the TC-specific genes and 3 cell genes switched off shared homolo genes listed in public databases (Table 2). un-known transcripts equates with a simil for the transcriptome of developing barley Table 2 provides a summary of the ch sion of selected functional categories genes that are known to be central to T and function. For all categories, numbe cell genes switched off in transiting to a placed by increased numbers of TC-spe between 3- to 12-fold. Amongst these transcription/translation, receptor kinas skeleton/vesicle trafficking, matrix poly defense genes contributed to the two-fo ber of genes expressed during uniform w wall ingrowth construction (Table 2; addition, a portion of uniform wall asso genes were linked with a co-occurring b ision/expansion (Additional file 10). Gen Table 2 Numbers of annotated expressed g epidermal cells transiting to a TC fate and f to a TC morphology Functional category E sw Total number of expressed genes Total number of annotated genes DNA synthesis and modelling Transcription/translation Signalling Receptor kinase/kinase Hormonal ROS/Ca2+ Intracellular organization Cytoskeleton/vesicle trafficking Membrane microdomains Cell wall enzymes and structural proteins Cellulose Matrix polysaccharides Structural proteins Membrane transporters Defense Flavonoid synthesis and compartmentation Zhang et al. BMC Plant Biology (2015) 15:103 entire flavonol pathway were switched on during TC de- velopment (Table 2). deposition. Genes whose expression levels did not change accounted for 60% of the common cohort (Table 2). A more detailed analysis was undertaken of the tem- poral expression of TC-specific genes contained in func- tional groups that are at the heart of regulating development of the TC’s structural specialization of a polarized ingrowth wall that realises their functional capacity to support high rates of nutrient transport across their plasma membranes [1]. These categories in- cluded hormonal and ROS/Ca2+ signal generation and their signalling pathways, cytoskeleton/vesicle trafficking and lipid biosynthetic enzymes related to microdomain formation, cell wall biosynthetic and remodelling en- zymes along with cell wall structural proteins and mem- brane transporters (see Additional files 11, 12, 13 and 14 inclusive). These data are supplemented with estimates of their annotation and relative expression levels at spe- cified stages of ingrowth wall construction along with noting whether their homologues have been detected in transcriptomes of BETCs and nematode giant TCs (Additional file 5). TC specific genes separated into those genes expressed throughout uniform wall (UW) and wall ingrowth (WI) formation and those that are specific to each of these wall-building phases. Genes expressed throughout ingrowth wall formation are separated into groups depending on their differential expression patterns of no change, up-regulated during UW or WI formation (for more details, see Results). Transcriptome networks specifically expressed during uniform wall or wall ingrowth formation Genes encoding the Table 2 Numbers of annotated expressed genes ascribed to specified functional categories for genes switched off in epidermal cells transiting to a TC fate and for those specifically expressed in epidermal cells undergoing trans-differentiation to a TC morphology Functional category Epidermal switched-off genes Transfer cell specific expressed genes UW/WI no change UW up-regulated UW specific WI up-regulated WI specific Total number of expressed genes 1381 851 366 1898 214 958 Total number of annotated genes 468 364 217 647 138 407 DNA synthesis and modelling 17 8 1 13 2 13 Transcription/translation 32 27 11 31 3 19 Signalling Receptor kinase/kinase 15 17 20 37 8 16 Hormonal 4 6 6 13 8 14 ROS/Ca2+ 1 4 9 6 1 6 Intracellular organization Cytoskeleton/vesicle trafficking 10 6 2 11 1 7 Membrane microdomains 0 3 5 3 1 7 Cell wall enzymes and structural proteins Cellulose 3 0 1 3 1 1 Matrix polysaccharides 9 4 6 14 3 7 Structural proteins 2 3 3 3 5 5 Membrane transporters 14 13 9 12 1 13 Defense 4 10 21 12 1 6 Flavonoid synthesis and compartmentation 5 6 4 3 4 8 TC specific genes separated into those genes expressed throughout uniform wall (UW) and wall ingrowth (WI) formation and those that are specific to each of h ll b ildi h G d h h i h ll f i d i d di h i diff i l i f rs of annotated expressed genes ascribed to specified functional categories for genes switched off in transiting to a TC fate and for those specifically expressed in epidermal cells undergoing trans-differen Transcriptome networks specifically expressed during uniform wall or wall ingrowth formation Only unigenes with a RPKM value > 0.45 in all three biological replicates were defined as ‘expressed’ in one or more cell types/phases of ingrowth wall deposition. For (B), only unigenes with a coverage of > 70% were listed for further analysis. EC- adaxial epidermal cell; TC – epidermal TC trans-differentiated from adaxial epidermal cell; SP- storage parenchyma cell; UW - uniform wall; WI- wall ingrowth. A B A B The above exercise demonstrated that of the TC- specific expressed genes, 66% were selectively linked with uniform wall (44%) or wall ingrowth (22%) forma- tion with the remainder (34%) being expressed through- out ingrowth wall formation (Figure 2B and see Additional file 4 for their RPKM gene expression values). This latter cohort of genes was categorised according to their relative expression levels in the two phases of in- growth wall formation, that is, no change and up- regulated during uniform wall or wall ingrowth Figure 2 Numbers of unigenes expressed preferentially in, or shared between, (A) cell types and (B) uniform wall and wall ingrowth formation occurring within adaxial epidermal cells of cultured V. faba cotyledons. Only unigenes with a RPKM value > 0.45 in all three biological replicates were defined as ‘expressed’ in one or more cell types/phases of ingrowth wall deposition. For (B), only unigenes with a coverage of > 70% were listed for further analysis. EC- adaxial epidermal cell; TC – epidermal TC trans-differentiated from adaxial epidermal cell; SP- storage parenchyma cell; UW - uniform wall; WI- wall ingrowth. Figure 2 Numbers of unigenes expressed preferentially in, or shared between, (A) cell types and (B) uniform wall and wall ingrowth formation occurring within adaxial epidermal cells of cultured V. faba cotyledons. Only unigenes with a RPKM value > 0.45 in all three biological replicates were defined as ‘expressed’ in one or more cell types/phases of ingrowth wall deposition. For (B), only unigenes with a coverage of > 70% were listed for further analysis. EC- adaxial epidermal cell; TC – epidermal TC trans-differentiated from adaxial epidermal cell; SP- storage parenchyma cell; UW - uniform wall; WI- wall ingrowth. Zhang et al. BMC Plant Biology (2015) 15:103 Page 6 of 13 deposition. Genes whose expression change accounted for 60% of the c (Table 2). Discussion The ability to obtain peels of cotyledon epidermal cells permitted cell-specific transcriptomes of these cells to be identified as they trans-differentiated to a TC-morphology [13]. In addition, developmental stage-specific transcrip- tomes were isolated (Figure 2, Additional files 5, 11, 12, 13, and 14) by taking advantage of the temporal disjunc- tion between uniform wall and wall ingrowth construction (Figure 1, Additional file 1). We interrogated these TC- specific transcriptomes to discover genes encoding proteins of functional categories considered to contrib- ute to developing a transport capable TC (Figure 3 and see Additional files 5, 11, 12, 13, and 14). Transcrip- tional cascades are initiated by auxin and ethylene [13,19,20] along with ROS and Ca2+ serving as second- ary messengers and positional signals [6,9]. The sub- stantial numbers of expressed receptor kinases/kinases and transcription factors, point to these forming a sig- nificant component of the signalling network (Table 2 and see [12]). Re-organization of the cytoskeleton and vesicle trafficking together with membrane microdo- mains, under control of the inductive signalling Flavonoid synthesis and compartmentation TC specific genes separated into those genes expressed throughout uniform wall (UW) and wall ingrowth (WI) formation and those that are specific to each of these wall-building phases. Genes expressed throughout ingrowth wall formation are separated into groups depending on their differential expression patterns of no change, up-regulated during UW or WI formation (for more details, see Results). Page 7 of 13 Zhang et al. BMC Plant Biology (2015) 15:103 suggesting that the Illumina data are relatively reliable (Table 2, Additional files 4, 5, 11, 12, 13, and 14). suggesting that the Illumina data are relatively reliable (Table 2, Additional files 4, 5, 11, 12, 13, and 14). network, would be expected to direct targeted delivery of cargos containing cell wall biosynthetic enzymes and cell wall components to construct the polarized uniform wall and localized wall ingrowths (Figure 3 and Additional file 1). The latter provide a scaffold to support an amplified plasma membrane surface area to which membrane transporters are targeted through a polarized endomembrane secretory system to confer en- hanced TC transport function (Figure 3). In addition, tar- geted positioning of transporters is responsible for generating the polarized extracellular ROS [6] and cyto- solic Ca2+ [9] signals and possibly altering positioning of auxin transporters that may contribute to generating the auxin maximum ([13]; Figure 3). The following discussion focuses on expressed genes belonging to these three functional categories namely, signals and signalling pathways, intracellular organization and transport func- tion (Figure 3). The discussion starts with seeking out candidates responsible for generating the developmen- tal signals and downstream signalling pathways. In all cases, statements regarding gene identity and function should be considered as putative. Signals and signalling pathways Hormonal signals, acting in series with ROS and Ca2+ [1], likely activate expression of receptor kinases/kinases and transcription factors to induce TC trans-differenti- ation (Figure 3). ROS/Ca 2+ Hormones Receptor kinases/kinases Transcription factors SIGNALS & SIGNALLING PATHWAYS INTRACELLULAR ORGANIZATION Cytoskeleton/vesicle trafficking Membrane microdomains TRANSPORT FUNCTION CW biosynthesis/deposition pattern Membrane transporter/targeting pattern + + Figure 3 Regulatory interrelationships (red arrows) between the functional categories of expressed genes that contribute to developing a functional transfer cell. SIGNALS & SIGNALLING PATHWAYS Auxin, ethylene and abscisic acid form a complex of primary signals putatively regulating TC development Hormones Receptor kinases/kinases Transcription factors p y g p y g g p Consistent with previous findings for epidermal TCs [13,19] and those reported for nematode giant TCs (10), induction of trans-differentiation was accompanied by changes in auxin activity (as determined by expression of auxin-induced genes) increasing five-fold between uniform wall and wall ingrowth formation driven by an auxin maxima generated by enhanced auxin biosynthesis and disrupted polar transport (Additional file 11). For instance, auxin biosynthesis could be elevated by expres- sion of indole-3-acetate O-methyltransferase IAA, that catalyzes synthesis of a biologically more active IAA, methyl-indole-3-acetic acid [21] and indole-3-glycerol phosphate synthase that starts the tryptophan independ- ent IAA biosynthetic pathway [22]. Polar auxin transport was likely disrupted by de-localized distribution of expressed auxin efflux and influx carriers by co- expression of KAN and HD-ZIPIII [23]. As a point of distinction between mechanisms regulating flange and reticulate wall ingrowths, in contrast to the latter [10,13,19], differentially expressed auxin-related genes were not detected in developing BETCs [8,11]. Figure 3 Regulatory interrelationships (red arrows) between the functional categories of expressed genes that contribute to developing a functional transfer cell. Initiation of epidermal TC development is linked with a substantive burst in ethylene biosynthesis, regulated by Zhang et al. BMC Plant Biology (2015) 15:103 Page 8 of 13 an auxin-induced expression of ACC synthases [19], driving expression of ethylene response factors (ERFs) [19] and particularly so for uniform wall formation (Additional file 11). For example, ERF2.6 and SHINE ex- pression could invoke cell wall biosynthesis through re- spectively up-regulating callose and cellulose biosynthesis (Additional file 13 and see [24,25]). Strong expression of ethylene signalling genes during uniform wall formation corresponded with that reported for early phases of wall ingrowth construction in barley BETCs [8] and nematode giant TCs [10]. upon a spatio/temporal fine-tuning of ROS. For instance, antioxidant capacity could be down regulated by ascorbic acid oxidation catalyzed by an extracellular L-ascorbate oxidase ([33]; Additional file 11) whilst concurrent ROS degradation by peroxidation [32] could be mediated by carboxymethylenebutenolidase-like protein (Additional file 11), secreted to the cell wall [34]. In contrast to epider- mal TCs, differentially expressed peroxidases were not detected in barley BETCs [8] and were strongly down- regulated in nematode giant TCs [10]. Auxin, ethylene and abscisic acid form a complex of primary signals putatively regulating TC development Loci at which wall ingrowths are deposited are defined by a cytosolic Ca2+ signal spatially organized into narrow plumes shaped by co-operative activities of plasma membrane Ca2+-permeable channels organized into clusters surrounded by Ca2+-ATPases [9]. However, un- like the unknown DHP receptor Ca2+-permeable chan- nels responsible for generating the Ca2+ signal [9], ion conductances of the expressed SKOR, annexin and cyclic nucleotide gated channels are not inhibited by verapamil or nifedipine (e.g. [35]). This, together with channel ex- pression coinciding with uniform wall construction (Additional file 11), points to an additional role for Ca2+ signalling in transiting precursor epidermal cells to a TC fate. A conclusion that accommodates Ca2+ signalling genes being expressed in both flange [8] and reticulate forms [18,36] of developing TCs. Ca2+ signalling would appear to be mediated through Ca2+-dependent protein kinases during uniform wall formation while a calmodu- lin signalling route could operate when wall ingrowths are being deposited as suggested by expression of three IQ-domain containing proteins (Additional file 11) that bind calmodulins to facilitate their function. A switch to a specific ABA signalling pathway through- out ingrowth wall formation was suggested by expression of PYL6, an ABA receptor [26], LEC1, tightly connected with ABA signalling regulating AB13 and AB15 [27] and ERF RAP2.6 that acts through an AB14-mediated signal- ling pathway ([28]; Additional file 11). ABA also acts dur- ing wall ingrowth formation (Additional file 11). ABA action throughout ingrowth wall formation in epidermal TCs contrasts with that for barley BETCs where ABA in- duces a maturation program once wall ingrowth depos- ition ceases [8]. Three gibberellin biosynthetic genes with the bio- logical activities of their products being modulated by two gibberellin 2-oxidases and Short Internode Related during uniform wall formation are expressed (Additional file 11). These gibberellins likely regulate cell division and expansion (Additional file 10 and see [29]). An ab- sence of gibberellin activity during wall ingrowth con- struction was suggested by co-expression of Short Internode Related and gibberellin 2-beta-dioxygenase 8- like (Additional file 11; [29]). Residual cytokinin levels in the developing epidermal TCs would be depleted through degradation by the expressed cytokinin dehydro- genase [30] and by conjugation to an inactive form cata- lyzed by the five cytokinin-O-glucosyltransferases ([31]; Additional file 11). Auxin, ethylene and abscisic acid form a complex of primary signals putatively regulating TC development No components of the salicylic acid or brassinosteroid signalling pathways were detected in the TC-specific transcriptome (Additional files 4 and 11) indicating that these pathways likely did not play a role in trans-differentiation to a TC state. Receptor kinases/kinases represent a significant component of the signalling network TC ifi ki /ki d j TC-specific receptor kinases/kinases represented a major component of the signalling networks (59%) expressed within the epidermal TCs (Table 2). While their func- tions are unknown, similar to nematode giant TCs [10], members of a two component signalling system present in barley BETCs [12], was not detected in the epidermal TC cohort of receptor kinases/kinases. Thus, the pres- ence/absence of a two component signalling system could underpin the distinction between flange [12] and reticulate ([10]; this study) TC wall architectures. Reactive oxygen species and Ca2+ serve as secondary messengers directing ingrowth wall construction Extracellular ROS, produced by ethylene-induced re- spiratory burst oxidases (rbohs), regulate biosynthesis and polarized deposition of the uniform wall in epider- mal TCs [6,32]. Similarly, rboh expression co-occurs with wall ingrowth formation in barley BETCs [8]. In contrast, no differentially expressed rbohs were detected in nematode giant TCs [10]. Interestingly, ROS-related gene expression was dominated by those involved in extra- and intracellular ROS homeostasis (Additional file 11) suggesting that ingrowth wall formation depends TC transport function Expression of cell wall synthesizing and re-modelling en- zymes underpins construction of the ingrowth wall that supports the amplified plasma membrane in which are embedded high densities of membrane transporters to collectively confer an enhanced transport function (Figure 3). Intracellular organization Polarized accumulation of extracellular ROS regulating uniform wall synthesis [6] and cytosolic Ca2+-plumes deter- mining loci for wall ingrowth formation [9] respectively de- pend upon a finely-tuned intracellular distribution of rbohs and Ca2+-permeable channels. This requirement is un- doubtedly met through modifying cytoskeleton and vesicle trafficking combined with re-configuring microdomains in Page 9 of 13 Zhang et al. BMC Plant Biology (2015) 15:103 the plasma membrane abutting sites of ingrowth wall deposition (Figure 3). Formation of membrane microdomains may sub- compartmentalize the plasma membrane polar domain Sphingolipids and sterols form major components of membrane microdomains [47]. Indeed, genes encoding enzymes in their biosynthetic pathways are expressed during development of flange wall ingrowths in barley BETCs [8]. Sphingolipids consist of a polar head group linked to an amino alcohol long-chain base (LCB) with the amine group acylated with a fatty acid (ceramide) with C4 of LCB being (trihydroxy) or not being (dihy- doxy) hydroxylated. Bax inhibitor 1-like (Additional file 12) post-translationally activates the C4 hydrolase [48] thus promoting flux through this route for acyl- ation by very long-chain fatty acids (VLCFA; [49]). Significantly, synthesis of VLCFAs may be enhanced by expression of four 3-ketoacyl-CoA synthase iso- forms, a elongation of fatty acid protein and a long- chain-fatty-acid-CoA ligase (Additional file 12) that form part of the elongase complex [50]. An increased precursor flow into the VLCFA pathway is supported by up-regulated expression of biotin carboxyl carrier protein of acetyl-CoA carboxylase (Additional file 12), a sub-component of acetyl-CoA carboxylase catalyzing the first carboxylation step in, and rate limiting of, de novo fatty acid synthesis [51]. Thereafter, bax inhibitor 1-like may also increase activities of fatty acid hydroxylases and desaturases that further modify the ceramides [48] before synthesis of the final sphingolipid products of glucosylcer- amides or glycosyl inositolphosphoceramides [49]. Uni- form wall specific expression of an oxysterol-binding protein-related protein and a remorin (Additional file 12) respectively point to vesicle trafficking of sterols [52] and their enrichment in membrane microdomains [47]. Wall ingrowth deposition is linked with expression of putative sterol biosynthetic genes (Additional file 12) such as two epoxide hydrolases [53] and squalene epoxidase [54]. TC-specific expression of cytoskeleton and vesicle trafficking confers TC polarity Microtubule and actin-associated genes switched off on transiting to a TC fate were predominantly linked to cell division (Additional file 12) that declined rapidly from three to six hours (Additional file 10). Consistent with polarized cell wall deposition [4], TC- specific genes in- dicative of remodelling cytoskeletal elements [37] were expressed either across ingrowth wall deposition (kinesin motor domain and formin-like protein) or only during one phase of the process (Additional file 12). Uniform wall specific gene expression (Additional file 12) indi- cated substantial polymerization/reorganization of actin filaments (actin and actin-97-like) and polarization of transport vesicles and organelles (Myosin X1) [38] con- sistent with ROS-regulated polarization of the uniform wall [6]. Wall ingrowth specific expression of 65-kDa microtubule-associated protein 3-like and Villin-4 re- spectively suggests establishing a stable microtubule [39] and actin [40] network to orchestrate deposition of wall ingrowths at loci. In root hair growth of Arabidopsis, AtVLN4 regulated actin organization is Ca2+-dependent [41] pointing to an interaction between actin and Ca2+plumes that define wall ingrowth loci [9]. Rapid polarized/localized cell wall deposition during formation of the ingrowth wall (Additional file 1) is dependent upon exo/endocytotic activity [13]. Regula- tion of cycling of ER-derived proteins through assembly and directed transport of coat protein complex II (GTP- binding protein SAR1A and ER-derived vesicle protein ERV14) and ARFs (ADP-ribosylation factor GTPase- activating proteins) across ingrowth wall deposition is consistent with asymmetrical vesicle trafficking to effect polarity [42,43]. Expression of uniform wall- and wall ingrowth-specific exocyst complex components and v- and t-SNARES (Vesicle-associated protein 2-1-like, Vesicle- associated membrane protein and two, Syntaxin -112 genes), to direct targeted vesicle delivery to the plasma membrane [44], confirms the specialized nature of each phase of ingrowth wall deposition. Two of four genes encoding proteins involved in membrane trafficking (Dynamin –related protein 1C-like and Dynamin-2B- like) are potential regulators of localized endocytosis supporting wall ingrowth deposition. These two dyna- min genes represent two subfamilies of DRPs that inter- act and assemble with clathrin at discrete foci in the plasma membrane of Arabidopsis cells to regulate endocytosis [45] and cellulose deposition [46]. This profile of cytoskeleton and vesicle trafficking genes is comparable to that reported for BETCs [8] and nema- tode giant TCs [18]. Expression of cell wall biosynthesis and re-modelling genes construct the ingrowth wall Cellulose is present in the uniform wall and the inner, electron dense region, of wall ingrowths of epidermal TCs [55] but cellulose biosynthesis appeared not to de- pend on expression of TC-specific CesAs that were switched off and not replaced as the epidermal cells transited to a TC fate (Additional file 13). Rather, cellu- lose deposition could be regulated post-translationally Page 10 of 13 Zhang et al. BMC Plant Biology (2015) 15:103 conferred by co-expression of a pectin esterase inhibitor modulating catalytic activities of expressed pectin ester- ases (Additional file 13) that catalyze removal of methyl ester groups from homogalacturonan backbones to allow their cross-linking with Ca2+ [66]. This accounts for es- terified pectins being the predominate type present in the ingrowth wall [55]. Up-regulated and wall ingrowth specific expression of two beta-galactosidases could catalyze cleavage of terminal galactosyl residues from RGI [67]. Overall, uniform wall formation included pec- tin biosynthesis and re-modelling giving way to pectin re-modelling alone for wall ingrowth deposition. En- hanced pectin biosynthesis and re-modelling also charac- terizes construction of barley BETCs [8] and nematode giant TCs [36]. (Additional file 13). Thus during uniform wall formation, SHINE [56] and a GPI-anchored protein [57] could regu- late CesA activity with the extruded cellulose microfibrils being re-modelled by two extracellular endoglucanases [58]. Cellulose provides a structural scaffold to form wall ingrowths at right angles to the uniform wall [59]. Here, an acid endochitinase-like protein, known to associate with cellulose synthase complexes [25] and an extracellu- lar beta-glucosidase, may form part of the machinery redirecting extrusion of cellulose microfibrils (Additional file 13). Consistent with these observations, CesA expres- sion is down-regulated in nematode giant TCs [36] while in maize and barley BETCs, CesA expression is up- regulated during later stages of TC development [8,11]. In contrast to cellulose biosynthesis, there was a sub- stantial TC-specific induction of genes encoding en- zymes responsible for synthesizing and re-modelling cell wall matrix polysaccharides (Table 2) and, in particular, hemicelluloses (Additional file 13). To this end, a (1-3,1-4) β-D-glucan synthase, was switched off and replaced with enzymes catalyzing xylan biosynthesis (xylosyltransferases, glycotransferase family GT8 protein – [60]) and re- modelling (beta-D-xylosidase – [61]). Xylans in the uni- form wall are consistent with their presence in flange wall ingrowths as suggested by expression of a beta-D- xylosidase and a xylosyltransferase respectively in BETCs of barley [8] and maize [11]. Expression of cell wall biosynthesis and re-modelling genes construct the ingrowth wall Xyloglucan pres- ence in uniform walls [55] is consistent with expression of xylosyltransferases (biosynthesis - [62]) and endoxy- lotransglucosylase/hydrolases (chain re-modelling – [63]). Xyloglucan biosynthesis and re-modelling also dominates early developmental stages of barley BETCs [8] and nematode giant TCs [10]. Xyloglucan re-modelling coincides with TC division/expansion as indicated by co- expression of expansins ([8,10,11]; Additional files 10 and 13 in this study). In contrast, expansins expressed during wall ingrowth construction when cell division/expansion has ceased (Additional files 10 and 13), must serve a yet-to-be identified function. One possibility is to free xyloglucan chains from cellulose microfibrils rendering them available for co-operative re-modelling by alpha- xylosidases and a beta-glucosidase [64] and by two extracellular isoforms of beta-galactosidase ([58] – Additional file 13). Overall, uniform wall and wall ingrowth construction was characterized by intense xyloglucan re- modelling. Callose deposition is a balance between synthesis by callose synthases and degradation by β-1, 3-glucanases. Isoforms of both enzymes were switched off in epider- mal cells on transiting to a TC fate and were not replaced during uniform wall formation (Additional file 13). However, UDP-glucuronosyltransferase 1 (UGT1), a callose synthase complex component essential for cal- lose synthesis [68], was expressed during wall ingrowth construction accompanied by two beta-1, 3-glucanase isoforms (Additional file 13); an expression pattern consistent with callose being confined to the outer sheath enveloping each wall ingrowth [55]. This tem- poral profile of callose deposition corresponds with that reported for nematode giant TCs [36] but contrasts with callose synthesis/turnover being restricted to the cell division phase of barley BETC development [69]. Transcripts encoding enzymes of the phenylpropanoid pathway occurred across ingrowth wall formation (Additional file 13). Expressed ABC transporters (Additional file 14) could efflux the monolignol pre- cursors to the cell wall where class III peroxidases (Additional file 11) catalyze their polymerization to lignin [70]. Significantly, the full complement of phe- nylpropanoid pathway enzymes is expressed in nema- tode giant TCs [36], and a detailed histochemical study of maize BETCs [71], provides strong evidence for their lignification. Extensins are expressed in TCs forming flange [8,11] and reticulate [18] wall ingrowths. A large number (14) of extensin family members of hydroxyproline-rich gly- coproteins were expressed in developing epidermal TCs (Additional file 13), implying a key role played by exten- sins in shaping the ingrowth wall and, in particular, wall ingrowths. Conclusions In epidermal cells undergoing trans-differentiation to a TC-morphology, only a small proportion of their tran- scriptome (i.e., 10%) was found to be TC-specific. Within the cohort of TC-specific transcripts, those en- coding pathways regulating signalling and intracellular organization were fine-tuned whilst cell wall construc- tion and membrane transporter profiles were altered substantially upon transiting to a TC fate. Each phase of ingrowth wall construction was linked with an unique sub-cohort of expressed genes. Comparison of transcrip- tomes of cells forming flange or reticulate wall ingrowths detected subtle differences in their transcript profiles, with the major difference being the absence of a two- component signalling system from the latter. Availability of supporting data The cDNA sequence datasets of raw reads and assem- bled reference transcriptome library supporting the re- sults of this article are available in the repository of the European Nucleotide Archive (ENA) with the ENA ac- cession number: PRJEB8906. Expression of cell wall biosynthesis and re-modelling genes construct the ingrowth wall A key enzyme in the post-translational modi- fication pathway generating functional extensins, prolyl 4-hydroxylase [72], was expressed throughout ingrowth wall formation (Additional file 13). Upon being exocy- tosed into the cell wall space, extensin monomers are envisaged to cross-link to form networks, mediated by Pathways leading to pectin biosynthesis and modifica- tion altered on epidermal cells transiting to a TC fate (Additional file 13). During uniform wall construction, synthesis of the pectin rhamnogalacturon backbone (RG1) was likely catalyzed by expression of galactofuro- nosyl-transferase 13-like (GAUT13 - Additional file 13) that significantly is essential for polarized pollen tube tip growth [65]. Control over the pectin gel state is Zhang et al. BMC Plant Biology (2015) 15:103 Page 11 of 13 redox activity of type-III apoplasmic peroxidases ([72]; Additional file 11). Expression of a proline-rich- extensin-like receptor kinase (Additional file 3), regulat- ing extensin/pectin interactions [73], may generate supramolecular structures [72] to function as templates for ingrowth wall formation. Significantly, coincident with wall ingrowth construction, was the expression of two nodule-specific extensins that support polarized tip growth in elongating infection threads [74] and a DZ- HRGP homolog that contributes to polarized tip growth of pollen tubes [72]. ongoing expression of ABC transporters (Additional file 14) corresponded with expression of flavonoid bio- synthesis and defense genes (Table 2). Additional files Additional file 1: Figure S1, Table S1 and accompanying text. Temporal formation of the uniform wall and wall ingrowths. Additional file 1: Figure S1, Table S1 and accompanying text. Temporal formation of the uniform wall and wall ingrowths. Additional file 2: Figure S2. Quality assessment of total RNA extracted from adaxial epidermal and storage parenchyma cells of V. faba cotyledons. Additional file 3: Table S2. Quality assessment of total RNA extracted from, and cDNA libraries of V. faba cotyledons used for Illumina sequencing. Investment into nitrogen transporters was enhanced during uniform wall formation (Additional file 14). This included a chloride channel (CLC-7 – Additional file 14) that may function as a nitrate sensor to regulate develop- ment by modulating auxin transport [89]. Two additional transporters were recruited for potassium homeostasis; a SKOR channel (Additional file 11; [35]) and a potassium antiporter (Additional file 14). Most pronounced was the large number (12) of expressed ABC transporter genes (Additional file 14) linked with expression of flavonoid biosynthetic and defense genes (Table 2). However, along with effluxing xenobiotics, members of the ABCB trans- porter subfamily facilitate plasma membrane transport of IAA [78] contributing to the altered IAA homeostasis of epidermal TCs (see Additional file 11 and associated text). Additional file 4: Table S3. Annotation and expression levels of all genes specifically expressed in the epidermal cells of cultured V. faba cotyledons trans-differentiating to epidermal transfer cells. Additional file 5: Table S4. Annotation and expression levels (RPKM) of selected unigenes expressed specifically in trans-differentiating adaxial epidermal cells of V. faba cotyledons [80]. Additional file 6: Table S5. Primer sequences used for qRT-PCR determination of expression levels of specified genes to validate RNAseq data sets. Additional file 6: Table S5. Primer sequences used for qRT-PCR determination of expression levels of specified genes to validate RNAseq data sets. Additional file 6: Table S5. Primer sequences used for qRT-PCR determination of expression levels of specified genes to validate RNAseq data sets. Additional file 7: Figure S3. Species distribution of genes with highest sequence similarity to unigenes in the reference transcriptome library of trans-differentiating adaxial epidermal cells of cultured V. faba cotyledons. Additional file 7: Figure S3. Species distribution of genes with highest sequence similarity to unigenes in the reference transcriptome library of trans-differentiating adaxial epidermal cells of cultured V. faba cotyledons. Additional file 8: Table S6. A large array of membrane transporters is specifically expressed in TCs The large number of transporter genes switched off upon transiting to an epidermal TC fate (Additional file 14) sug- gests a substantial change in transport function. A striking feature of the 13 TC-specific membrane transporters, whose expression did not change during trans-differentiation, nine transport inorganic and or- ganic nitrogen compounds (Additional file 14). Five transport amino nitrogen compounds including two aquaglyceroporins (NIP1-2; nodulin26 – [75]; Additional file 14). Two nucleoside proton symporters, adenine/ guanine permease AZG1 [76] and equilibrative nucleo- side transporter 3 [77], likely function to retrieve apo- plasmic nucleosides released from degradation of endosperm and inner cell layers of seed coats crushed by expanding cotyledons. Collectively, expression of these transporters reflects a demand for nitrogenous precursors to support protein biosynthesis. Expression of a putative plasma membrane potassium transporter and tonoplast potassium proton antiporter (Additional file 14) ensure potassium homeostasis. Acknowledgements The research was supported by funding from the Australian Research Council (DP130101396) awarded to CEO and JWP together with funding from the German Research Foundation (DFG; grant GZ RA 2061/3-1) awarded to RR. XX and SW are grateful for the support of Faculty Research Higher Degree scholarships. We are indebted to Joseph Enwright for a continuous supply of healthy experimental plant material. The SEM and TEM observations were completed in The University of Newcastle EMX-ray Unit. 19. Zhou Y, Andriunas F, Offler CE, McCurdy DW, Patrick JW. An epidermal-specific ethylene signal cascade regulates trans-differentiation of transfer cells in Vicia faba cotyledons. New Phytol. 2010;185(4):931–43. 20. Andriunas FA, Zhang HM, Weber H, McCurdy DW, Offler CE, Patrick JW. Glucose and ethylene signalling pathways converge to regulate trans-differentiation of epidermal transfer cells in Vicia narbonensis cotyledons. Plant J. 2011;68:987–98. References 24. Ali MA, Abbas A, Kreil DP, Bohlmann H. Overexpression of the transcription factor RAP2.6 leads to enhanced callose deposition in syncytia and enhanced resistance against the beet cyst nematode Heterodera schachtii in Arabidopsis roots. BMC Plant Biol. 2013;13:47. 1. Andriunas FA, Zhang HM, Xia X, Patrick JW, Offler CE. Intersection of transfer cells with phloem biology-broad evolutionary trends, function, and induction. Front Plant Sci. 2013;4:221. 1. Andriunas FA, Zhang HM, Xia X, Patrick JW, Offler CE. Intersection of transfer cells with phloem biology-broad evolutionary trends, function, and induction. Front Plant Sci. 2013;4:221. 2. Borisjuk L, Wang TL, Rolletschek H, Wobus U, Weber H. A pea seed mutant affected in the differentiation of the embryonic epidermis is impaired in embryo growth and seed maturation. Development. 2002;129(7):1595–607. 2. Borisjuk L, Wang TL, Rolletschek H, Wobus U, Weber H. A pea seed mutant affected in the differentiation of the embryonic epidermis is impaired in embryo growth and seed maturation. Development. 2002;129(7):1595–607. 25. Doblin MS, Pettolino F, Bacic A. Plant cell walls: the skeleton of the plant world. Funct Plant Biol. 2010;37:357–81. 26. Miyakawa T, Fujita Y, Yamaguchi-Shinozaki K, Tanokura M. Structure and function of abscisic acid receptors. Trends Plant Sci. 2013;18(5):259–66. 3. Costa LM, Yuan J, Rouster J, Paul W, Dickinson H, Gutierrez-Marcos JF. Maternal control of nutrient allocation in plant seeds by genomic imprinting. Curr Biol. 2012;22(2):160–5. 27. Sreenivasulu N, Wobus U. Seed-development programs: a systems biology-based comparison between dicots and monocots. Annu Rev Plant Biol. 2013;64:189–217. 4. Talbot MJ, Franceschi VR, McCurdy DW, Offler CE. Wall ingrowth architecture in epidermal transfer cells of Vicia faba cotyledons. Protoplasma. 2001;215(1-4):191–203. 28. Zhu Q, Zhang J, Gao X, Tong J, Xiao L, Li W, et al. The Arabidopsis AP2/ARF transcription factor RAP2.6 participates in ABA, sat and osmotic stress responses. Gene. 2010;457:1–12. 5. Monjardino P, Rocha S, Tavares AC, Fernandes R, Sampaio P, Salema R, et al. Development of flange and reticulate wall ingrowths in maize (Zea mays L.) endosperm transfer cells. Protoplasma. 2013;250(2):495–503. 29. Hedden P, Thomas SG. Gibberellin biosynthesis and its regulation. Biochem J. 2012;444(1):11–25. p p 6. Andriunas FA, Zhang HM, Xia X, Offler CE, McCurdy DW, Patrick JW. Reactive oxygen species form part of a regulatory pathway initiating trans- differentiation of epidermal transfer cells in Vicia faba cotyledons. J Exp Bot. 2012;63(10):3617–29. 30. Li J, Nie X, Tan JL, Berger F. Competing interests h h d l h The authors declare that they have no competing interests. The authors declare that they have no competing interests. 15. Grabherr MG, Haas BJ, Yassour M, Levin JZ, Thompson DA, Amit I, et al. Full-length transcriptome assembly from RNA-Seq data without a reference genome. Nat Biotechnol. 2011;29(7):644–52. Author details 1 h l f 21. Zhao N, Guan J, Lin H, Chen F. Molecular cloning and biochemical characterization of indole-3-acetic acid methyltransferase from poplar. Phytochemistry. 2007;68(11):1537–44. 1School of Environmental and Life Sciences, The University of Newcastle, Callaghan NSW 2308, Australia. 2Leibniz Institute of Plant Genetics and Crop Plant Research (IPK), Corrensstrasse 3, D-06466, Gatersleben, Germany. 22. Mano Y, Nemoto K. The pathway of auxin biosynthesis in plants. J Exp Bot. 2012;63(8):2853–72. Received: 9 February 2015 Accepted: 1 April 2015 23. Huang T, Harrar Y, Lin C, Reinhart B, Newell NR, Talavera-Rauh F, et al. Arabidopsis KANADI1 acts as a transcriptional repressor by interacting with a specific cis-element and regulates auxin biosynthesis, transport, and signaling in opposition to HD-ZIPIII factors. Plant Cell. 2014;26(1):246–62. Authors’ contributions JWP and CEO initiated the research and designed the experimental approach. HMZ, XX and SW performed the experiments. HMZ processed the data and HMZ, JWP, CEO, RR and HW interpreted the results and prepared the manuscript. All authors have read and approved the final manuscript. 16. Audic S, Claverie JM. The significance of digital gene expression profiles. Genome Res. 1997;7(10):986–95. 17. Benjamini Y, Drai D, Elmer G, Kafkafi N, Golani I. Controlling the false discovery rate in behavior genetics research. Behav Brain Res. 2001;125(1-2):279–84. 18. Ji H, Gheysen G, Denil S, Lindsey K, Topping JF, Nahar K, et al. Transcriptional analysis through RNA sequencing of giant cells induced by Meloidogyne graminicola in rice roots. J Exp Bot. 2013;64(12):3885–98. Additional files Zhang HM, Imtiaz MS, Laver DR, McCurdy DW, Offler CE, van Helden DF, et al. Polarized and persistent Ca2+ plumes define loci for formation of wall ingrowth papillae in transfer cells. J Exp Bot. 2015;66:1179–90. 10. Cabrera J, Barcala M, Fenoll C, Escobar C. Transcriptomic signatures of transfer cells in early developing nematode feeding cells of Arabidopsis focused on auxin and ethylene signaling. Front Plant Sci. 2014;5:107. Additional file 13: Table S10. Genes encoding cell wall biosynthetic and remodelling enzymes and structural proteins switched off in epidermal cells transiting to a TC fate and those specifically expressed in epidermal cells undergoing trans-differentiation to a TC morphology. 11. Xiong Y, Li Q-B, Kang B-H, Chourey P. Discovery of genes expressed in basal endosperm transfer cells in maize using 454 transcriptome sequencing. Plant Mol Biol Rep. 2011;29(4):835–47. 12. Thiel J, Hollmann J, Rutten T, Weber H, Scholz U, Weschke W. 454 Transcriptome sequencing suggests a role for two-component signalling in cellularization and differentiation of barley endosperm transfer cells. PLoS One. 2012;7(7), e41867. Additional file 14: Table S11. Genes encoding membrane transporters switched off in epidermal cells transiting to a TC fate and those specifically expressed in epidermal cells undergoing trans-differentiation to a TC morphology. Additional file 14: Table S11. Genes encoding membrane transporters switched off in epidermal cells transiting to a TC fate and those specifically expressed in epidermal cells undergoing trans-differentiation to a TC morphology. 13. Dibley SJ, Zhou Y, Andriunas FA, Talbot MJ, Offler CE, Patrick JW, et al. Early gene expression programs accompanying trans-differentiation of epidermal cells of Vicia faba cotyledons into transfer cells. New Phytol. 2009;182(4):863–77. Additional file 15: Figure S5. Validation of expression patterns of unigenes obtained from Illumina sequencing using quantitative RT-PCR. 14. Wardini T, Wang XD, Offler CE, Patrick JW. Induction of wall ingrowths of transfer cells occurs rapidly and depends upon gene expression in cotyledons of developing Vicia faba seeds. Protoplasma. 2007;231(1-2):15–23. Additional files Reads generated from Illumina sequencing used to identify gene cohorts linked with uniform wall and wall ingrowth deposition in trans-differentiating epidermal cells of V. faba cotyledons. Additional file 9: Table S7. Temporal profile of unigene numbers expressed in adaxial epidermal or storage parenchyma cells of cultured V. faba cotyledons. With the onset of wall ingrowth deposition, appearance of two sugar transporter transcripts (Additional file 14) re- flects a growing dependence on an extracellular carbon supply [79]. Further investments were made in amino ni- trogen and potassium transporters (Additional file 14). An Additional file 10: Figure S4 and associated text. Cell division and expansion of adaxial epidermal cells of cultured V. faba cotyledons. Additional file 11: Table S8. Genes encoding proteins generating or transducing developmental signals switched off in epidermal cells Additional file 11: Table S8. Genes encoding proteins generating or transducing developmental signals switched off in epidermal cells Page 12 of 13 Zhang et al. BMC Plant Biology (2015) 15:103 8. Thiel J, Riewe D, Rutten T, Melzer M, Friedel S, Bollenbeck F, et al. Differentiation of endosperm transfer cells of barley: a comprehensive analysis at the micro-scale. Plant J. 2012;71(4):639–55. transiting to a TC fate and those specifically expressed in epidermal cells undergoing trans-differentiation to a TC morphology. transiting to a TC fate and those specifically expressed in epidermal cells undergoing trans-differentiation to a TC morphology. transiting to a TC fate and those specifically expressed in epidermal cells undergoing trans-differentiation to a TC morphology. Additional file 12: Table S9. Genes encoding proteins involved in regulation of the cytoskeleton and vesicle trafficking and sphingolipid/ sterol synthesis and transport that are switched off in epidermal cells transiting to a TC fate and those specifically expressed in epidermal cells undergoing trans-differentiation to a TC morphology. Additional file 12: Table S9. Genes encoding proteins involved in regulation of the cytoskeleton and vesicle trafficking and sphingolipid/ sterol synthesis and transport that are switched off in epidermal cells transiting to a TC fate and those specifically expressed in epidermal cells undergoing trans-differentiation to a TC morphology. Additional file 12: Table S9. Genes encoding proteins involved in regulation of the cytoskeleton and vesicle trafficking and sphingolipid/ sterol synthesis and transport that are switched off in epidermal cells transiting to a TC fate and those specifically expressed in epidermal cells undergoing trans-differentiation to a TC morphology. 9. References Xu J, Scheres B. Dissection of Arabidopsis ADP-RIBOSYLATION FACTOR 1 function in epidermal cell polarity. Plant Cell. 2005;17(2):525–36. 44. Zarsky V, Kulich I, Fendrych M, Pecenkova T. Exocyst complexes multiple functions in plant cells secretory pathways. Curr Opin Plant Biol. 2013;16(6):726–33. 70. Wang J, Nayak S, Koch K, Ming R. Carbon partitioning in sugarcane (Saccharum species). Front Plant Sci. 2013;4:201. 71. Rocha S, Monjardino P, Mendonça D, Da Câmara Machado A, Fernandes R, Sampaio P, et al. Lignification of developing maize (Zea mays L.) endosperm transfer cells and starchy endosperm cells. Front Plant Sci. 2014;20(5):102. 45. Fujimoto M, Ueda T. Conserved and plant-unique mechanisms regulating plant post-Golgi traffic. FrontPlant Sci. 2012;3:197. 46. Konopka CA, Bednarek SY. Comparison of the dynamics and functional redundancy of the Arabidopsis dynamin-related isoforms DRP1A and DRP1C during plant development. Plant Physiol. 2008;147(4):1590–602. 72. Velasquez M, Salter JS, Dorosz JG, Petersen BL, Estevez JM. Recent advances on the posttranslational modifications of EXTs and their roles in plant cell walls. Front Plant Sci. 2012;3:93. 47. Malinsky J, Opekarova M, Grossmann G, Tanner W. Membrane microdomains, rafts, and detergent-resistant membranes in plants and fungi. Annu Rev Plant Biol. 2013;64:501–29. 73. Gish LA, Clark SE. The RLK/Pelle family of kinases. Plant J. 2011;66(1):117–27. 74. Gucciardo S, Rathbun EA, Shanks M, Jenkyns S, Mak L, Durrant MC, et al. Epitope tagging of legume root nodule extensin modifies protein structure and crosslinking in cell walls of transformed tobacco leaves. Mol Plant Microbe Interact. 2005;18(1):24–32. 48. Nagano M, Ishikawa T, Ogawa Y, Iwabuchi M, Nakasone A, Shimamoto K, et al. Arabidopsis Bax inhibitor-1 promotes sphingolipid synthesis during cold stress by interacting with ceramide-modifying enzymes. Planta. 2014;240(1):77–89. 75. Wallace IS, Choi WG, Roberts DM. The structure, function and regulation of the nodulin 26-like intrinsic protein family of plant aquaglyceroporins. Biochim Biophys Acta. 2006;1758(8):1165–75. 49. Markham JE, Molino D, Gissot L, Bellec Y, Hematy K, Marion J, et al. Sphingolipids containing very-long-chain fatty acids define a secretory pathway for specific polar plasma membrane protein targeting in Arabidopsis. Plant Cell. 2011;23(6):2362–78. 76. Mansfield TA, Schultes NP, Mourad GS. AtAzg1 and AtAzg2 comprise a novel family of purine transporters in Arabidopsis. FEBS Lett. 2009;583(2):481–6. 50. Haslam TM, Kunst L. Extending the story of very-long-chain fatty acid elongation. Plant Sci. 2013;210:93–107. 77. Cornelius S, Traub M, Bernard C, Saizig C, Lang P, Möhimann T. References Integration of epigenetic and genetic controls of seed size by cytokinin in Arabidopsis. Proc Natl Acad Sci U S A. 2013;110(38):15479–84. 31. Kudo T, Makita N, Kojima M, Tokunaga H, Sakakibara H. Cytokinin activity of cis-zeatin and phenotypic alterations induced by overexpression of putative cis-zeatin-O-glucosyltransferase in rice. Plant Physiol. 2012;160(1):319–31. 7. Drea S, Leader DJ, Arnold BC, Shaw P, Dolan L, Doonan JH. Systematic spatial analysis of gene expression during wheat caryopsis development. Plant Cell. 2005;17(8):2172–85. 31. Kudo T, Makita N, Kojima M, Tokunaga H, Sakakibara H. Cytokinin activity of cis-zeatin and phenotypic alterations induced by overexpression of putative cis-zeatin-O-glucosyltransferase in rice. Plant Physiol. 2012;160(1):319–31. 7. Drea S, Leader DJ, Arnold BC, Shaw P, Dolan L, Doonan JH. Systematic spatial analysis of gene expression during wheat caryopsis development. Plant Cell. 2005;17(8):2172–85. Page 13 of 13 Zhang et al. BMC Plant Biology (2015) 15:103 Page 13 of 13 32. Xia X, Zhang HM, Andriunas FA, Offler CE, Patrick JW. Extracellular hydrogen peroxide, produced through a respiratory burst oxidase/superoxide dismutase pathway, directs ingrowth wall formation in epidermal transfer cells of Vicia faba cotyledons. Plant Signal Behav. 2012;7(9):1125–8. 56. Doblin MS, Johnson KL, Humphries J, Newbigin EJ, Bacic A. Are designer plant cell walls a realistic aspiration or will the plasticity of the plant’s metabolism win out? Curr Opin Biotechnol. 2014;26:108–14. 57. McFarlane HE, Doring A, Persson S. The cell biology of cellulose synthesis. Annu Rev Plant Biol. 2014;65:69–94. cells of Vicia faba cotyledons. Plant Signal Behav. 2012;7(9):1125– Annu Rev Plant Biol. 2014;65:69–94. 33. Fotopoulos V, Sanmartin M, Kanellis AK. Effect of ascorbate oxidase over- expression on ascorbate recycling gene expression in response to agents imposing oxidative stress. J Exp Bot. 2006;57(14):3933–43. 58. Minic Z. Physiological roles of plant glycoside hydrolases. Planta. 2008;227(4):723–40. 59. Talbot MJ, Wasteneys GO, Offler CE, McCurdy DW. Cellulose synthesis is required for deposition of reticulate wall ingrowths in transfer cells. Plant Cell Physiol. 2007;48(1):147–58. 34. Yano H, Kuroda M. Disulfide proteome yields a detailed understanding of redox regulations: a model study of thioredoxin-linked reactions in seed germination. Proteomics. 2006;6(1):294–300. 60. Rennie EA, Scheller HV. Xylan biosynthesis. Curr Opin Biotechnol. 2014;26:100–7. 35. Demidchik V, Maathius FJM. Physiological roles of non selective cation channels in plants: from salt stress to signaling and development. New Phytol. 2007;175(3):387–404. 61. Cleemput G, Hessing M, Van Oort M, Deconynck M, Delcour JA. Purification and characterization of a [beta]-D-xylosidase and an endo-xylanase from wheat flour. References Plant Physiol. 1997;113(2):377–86. 36. Portillo M1, Cabrera J, Lindsey K, Topping J, Andrés MF, Emiliozzi M, et al. Distinct and conserved transcriptomic changes during nematode-induced giant cell development in tomato compared with Arabidopsis: a functional role for gene repression. New Phytol. 2013;197(4):1276–90. 62. Oikawa A, Lund CH, Sakuragi Y, Scheller HV. Golgi-localized enzyme complexes for plant cell wall biosynthesis. Trends Plant Sci. 2013;18(1):49–58. 63. Eklof JM, Brumer H. The XTH gene family: an update on enzyme structure, function, and phylogeny in xyloglucan remodeling. Plant Physiol. 2010;153(2):456–66. 37. Bulbert RA, Offler CE, McCurdy DW. Polarized microtubule deposition coincides with wall ingrowth formation in transfer cells of Vicia faba L. cotyledons. Protoplasma. 1998;201(1-2):8–16. 64. Sampedro J, Pardo B, Gianzo C, Guitian E, Revilla G, Zarra I. Lack of alpha- xylosidase activity in Arabidopsis alters xyloglucan composition and results in growth defects. Plant Physiol. 2010;154(3):1105–15. 38. Tominaga M, Nakano A. Plant-specific myosin XI, a molecular perspective. Front Plant Sci. 2012;3:211. 39. Hamada T, Ueda H, Kawase T, Hara-Nishimura I. Microtubules contribute to tubule elongation and anchoring of endoplasmic reticulum, resulting in high network complexity in Arabidopsis. Plant Physiol. 2014;166(4):1869–76. 65. Wang L, Wang W, Wang YQ, Liu YY, Wang JX, Zhang XQ, et al. Arabidopsis galacturonosyltransferase (GAUT) 13 and GAUT14 have redundant functions in pollen tube growth. Mol Plant. 2013;6(4):1131–48. 40. Henty-Ridilla JL, Li J, Blanchoin L, Staiger CJ. Actin dynamics in the cortical array of plant cells. Curr Opin Plant Biol. 2013;16(6):678–87. 66. Jolie RP, Duvetter T, Van Loey AM, Hendrickx ME. Pectin methylesterase and its proteinaceous inhibitor: a review. Carbohydr Res. 2010;345(18):2583–95. 41. Zhang Y, Xiao Y, Du F, Cao L, Dong H, Ren H. Arabidopsis VILLIN4 is involved in root hair growth through regulating actin organization in a Ca2 +-dependent manner. New Phytol. 2011;190(3):667–82. 67. Martin I, Hernandez-Nistal J, Albornos L, Labrador E, Dopico B. BetaIII-Gal is involved in galactan reduction during phloem element differentiation in chickpea stems. Plant Cell Physiol. 2013;54(6):960–70. 42. Hanton SL, Chatre L, Matheson LA, Rossi M, Held MA, Brandizzi F. Plant Sar1 isoforms with near-identical protein sequences exhibit different localisations and effects on secretion. Plant Mol Biol. 2008;67(3):283–94. 68. De Storme N, Geelen D. Callose homeostasis at plasmodesmata: molecular regulators and developmental relevance. Front Plant Sci. 2014;5:138. 69. Thiel J, Müller M, Weschke W, Weber H. Amino acid metabolism at the maternal-filial boundary of young barley seeds: a microdissection-based study. Planta. 2009;230(1):205–13. 43. References Nucleoside transport across the plasma membrane mediated by equilibrative nucleoside transporter 3 influences metabolism of Arabidopsis seedlings. Plant Biology. 2012;14:696–705. 51. Saito K, Yonekura-Sakakibara K, Nakabayashi R, Higashi Y, Yamazaki M, Tohge T, et al. The flavonoid biosynthetic pathway in Arabidopsis: structural and genetic diversity. Plant Physiol Biochem. 2013;72:21–34. 78. Remy E, Duque P. Beyond cellular detoxification: a plethora of physiological roles for MDR transporter homologs in plants. Front Physiol. 2014;5:201. 52. Umate P. Oxysterol binding proteins (OSBPs) and their encoding genes in Arabidopsis and rice. Steroids. 2011;76(5):524–9. 79. Wardini T, Talbot MJ, Offler CE, Patrick JW. Role of sugars in regulating transfer cell development in cotyledons of Vicia faba seeds. Protoplasma. 2007;230:75–88. 53. Morisseau C. Role of epoxide hydrolases in lipid metabolism. Biochimie. 2013;95(1):91–5. 54. Rasbery JM, Shan H, LeClair RJ, Norman M, Matsuda SP, Bartel B. Arabidopsis thaliana squalene epoxidase 1 is essential for root and seed development. J Biol Chem. 2007;282(23):17002–13. 80. Silva-Sanchez C, Chen S, Zhu N, Li QB, Chourey PS. Proteomic comparison of basal endosperm in maize miniature1 mutant and its wild-type Mn1. Front Plant Sci. 2013;4:211. 55. Vaughn KC, Talbot MJ, Offler CE, McCurdy DW. Wall ingrowths in epidermal transfer cells of Vicia faba cotyledons are modified primary walls marked by localized accumulations of arabinogalactan proteins. Plant Cell Physiol. 2007;48(1):159–68. 55. Vaughn KC, Talbot MJ, Offler CE, McCurdy DW. Wall ingrowths in epidermal transfer cells of Vicia faba cotyledons are modified primary walls marked by localized accumulations of arabinogalactan proteins. Plant Cell Physiol. 2007;48(1):159–68.
https://openalex.org/W2003811480	https://zenodo.org/records/2494273/files/article.pdf	German	null	708. Zur Darstellung: des gelben Arsens	Berichte der Deutschen Chemischen Gesellschaft	1,904	public-domain	1,582	1) Eine Quecksilberchlorid-Verbinduog des Chelidonslurediathylesters wurde ') Ann. d. Chem. 144, 110 r18671. 3) Zeitschr. ffir anorgao. Chem. 32, 437 [1902]. ebenfalls von Willstiitter (loc. cit.) dargestellt. 4572 4572 achafc baaischen Sauerstoffs befinden, einen veratarkenden Einflusa aut die Basicitat der Subatanz auaiiben. Die Fahigkeit, Salze zu addiren, scheint aber nicht von der Basicitiit, mindeatens nicht von dieaer allein abzuhlingen. Offenbar ist ea specie11 die im Meaityloxyd, im P y r o n , in der Chelidonsliurel), nicht aber im M e t h y l h e p t e n o n und nicht im C i n e o l vorhandene Gruppe .CO.CH:, welche jene Additionsfiihig- keit begiinatigt. g g Damit stimmt iiberein, dasa auch Z i m m t a l d e h y d beim Schiitteln mit concentrirter Sublimatl8sun& aofort zu einer voluminosen Krystall- masae erstarrt. Die Yerbjndung krystallisirt ans Alkobol in aeide- glanzende Nadelu, welclie,von stud. F. P a t z s c h k e analysirt> wurden. CsHgO.HgCI1. Ber. H g 49.6. Gef. 49.2. )) C1 17.5, * 17.1. 0.5226 g Sbst.: 0.2995 HgS und 0.3625 AgCI. Demnach vereinigt sich auch in diesem Falle daa Quecksilber- chlorid in molekularer Menge mit der organischen Verbinducg. 708. A l f r e d Stook und Werner Siebert: Zur Daretellung des gelben Arsens. 708. A l f r e d Stook und Werner Siebert: Zur Daretellung des gelben Arsens. 708. A l f r e d Stook und Werner Siebert: Zur Daretellung des gelben Arsens. [Aus dem I. chernischen Institut der Universitst Berlin.! (Eingegaogen 28 November 1904 ) [Aus dem I. chernischen Institut der Universitst Berlin.! (Eingegaogen am 28. November 1904.) [Aus dem I. chernischen Institut der Universitst Berlin.! (Eingegaogen am 28. November 1904.) Die Existenz einer gelbeo Arsenmodification wurde i m Jahre 1867 von Bettendorfl) erkannt, nls er Arsendiimpfe schnell abkiihlte, E r und Audere, die eeiue Versnche wiederholten, erhielteu dabei nur ganz geriogfiigige Mengen des intere~santen Korpers; das meiste &sen schlagt sich megen der Unbestiindigkeit der gelben Form schwarz nieder. LZieungeu dea gelben Arsena i n Schwefelkohlenstoff sind vie1 bestiindiger; E r d m a n n und von U n r u h # ) hatten den gliick- lichen Oedanken, durch Condensation ron Arsendampf in Schwefel- kohlenstoff zunachzt eine Losung darzustellen, aus welcher dann durcb starke Abkiihlung krystallieirtes gelbes Arsen zu erhalten ist. Aber auch bei ihrer Methode geht ein erheblicher Theil des verdampften Araens wieder in die achwarze Form uber. Der nachstehend beschriebene Apparat erlaubt, ohne Anwendung eines Liisnngsmittels schwarzes Arsen - zunachst allerdings nur in 4573 geringer MeGge - q u a n t i t a t i v in gelbes iiberzufiihren. Dae Arsen wird im< tiefen Vacuum verdampft und der Dampf bei der Temperatur der fiusaigen Lnft condensirt; es entsteht dabei nur die gelbe Form. g ; g To das mit dern eeitlichen, zu einer Quecksilberluftpumpe fiihren- den Ansatzrohr I3 versebene gussere Glaegefass A (8. d. Zeichnung) rngt ein Rohr C aus Jenaer Glas von 9 mm Durchmesser binein, welches an seinem untereu Ende an kleinen OlasSsen ein mit diinnem Platindraht befestigtes Glasbecher- +E 3, chen D trfigt. Innen befindet sich eine etwa 2 cm hohe Schicht Kryptolmaese, in die zwei rnit dunnen Glasrohren isolierte, an ihren Enden ringfiirmig um- gebogene Kupferdrihte E+ und B- gesteckt sind. Sie sind rnit einer vierzelligen Accumulatorenbatterie verbunden, ein eingeschalteter Regulirwiderstand ge- stattet, fie 'Stromstarke zu regeln und das Kryptol auf belieb3iige Temperaturen zu erwarmen. 400° z. B \&den durch einen Strom von 2-3 AmpAre erreicbt. Zur Temperaturmessung ist in die Kryptol- mirsse eiq durch Glasrohr isolirtes Platin-Platin- rhodiumthermoelement eingefiibrt [in der Zeicbnung fortgela9~enl. 3 Sein lhrchrnesser betrug 4 crn. 708. A l f r e d Stook und Werner Siebert: Zur Daretellung des gelben Arsens. In den etwa 2 mm weiten Raum zwischen Qlas- becherclien D und Rolir C wird etwas durch Subli- mation im Vacuum frisch gereinigtes Arsen gegeben, C am oberen Ende von A mittelst Siegellack luft- dicht befestigt und aus dem Ranm zwischen A und C alle Luft mit Hiilfe der Quecksilberlnflpumpe entfernt. b a n n fiillt man in das tiussere durchsichtige W e i n h o l d - Gefass G filtrirte fliissige Luft. Er- wfrmt man nun die Kryptolmasse auf 425-450°, so verdampft das Arsen in dem Bechercben D, und nach kiirzer Zeit schlagt sich gelbes Arsen i m ' birnenfhnigen Theile von A') nieder. Weil der Dampf, ohne mit KBrpern mittlerer Temperatur in Berubrung zn kommen, aofort nuf die Temperatur der fliissigen Lult abgekiihlt wird, entsteht auaschliesslich die gelbe Modification. Sie erscheint entweder als hellgelber,' den Schwefelblumen gleichender Anflug oder iiberrieht - bei langsamerem Destilliren - A wie ein klar-dnrcbsichtiger gelber Lack, der im rothen Licht kaum zu bemerken ist. Die ganze Ope- Pation mu& namlich im Dunkelzimmer bei rothem Lichte vorgenommen werden; das auf diesem trocknen Wege dargestellte gelbe Arsen ist aueserordefitlich lichtempfindlich, wohl noch mehr als das grosskry- G 4574 stallinische, aus Schwefelkohlenstoff losung erhaltene. Die Strahleo eines in 10 cm Entfernung aufgestellten Auerbrenners oder einer elektriscben Gluhlampe verursachen schon nach wenigen Secunden beginnende Braunung, trotz der niedrigen Temperatur der fliissigen Luft. Schneller wirkt natiirlich Sonnenlicht. Entfernt man die fliissige Luft, sodass aich der Arsenbeechlag erwiirmt, so wird er momentan tiefschwarz; die damit verbundene VolumBnderung giebt sich durch Bildqng zahlloser Risse und Abspriogen der schwarzen Splitter von der Glaswand zu erkennen. WBhrend also die Um- wandlung durch Warme augenblicklich durch die gauze Masse erfolgt, ist das Bild ein anderes, wenn man Licht bei der Temperatur der fliissigen Lnft wirken lasst. Hier schreitet die Veranderung, sobald ein g e ~ i s s e r Grad der Dunkelfarbung erreicht ist, nur langsam weiter, die Arsenschicht bleibt selbst im Sonnenlicht, noch minutenlang roth- braun durchscheinend; durch die Braunung wird der wirksame Theil der Strahlung offenbar aufgehalten. Wir versuchten, ob bei der Umwandluug der gelben Arsen- modification in die schwarze etwa eine Leuchterscheinung zu bemerken ware. 4575 erlaubt aber, die Bildung des gelbin Arsens aus dem schwarzen uud seine .Riickuerwandlung in dieses als Vorlesungsvereuch vorzufiibreo. Uebrigens kam es uns bei unseren Versuchen weniger darauf an, gelbes Argen darzustellen, als den Apparat auszuprobiren; er sollte iins gewissertnaassen als Model1 dienen. Wir sind augenblicklich damit ' beschaftigt, ibn durch Verwendung anderen Mnteriales fiir grossere Temperaturuntetschiede verwendhar zu machen, und wollen dann ao die Lasung der Frage herangehen, ob sich mit ibm auch gelbes Antimon erhalten lasst. Vielleicht liefert una diese moderne Modification der D e v i l le'schen warm-kalten Rohre auch noch bei anderen Stoffen interessnnte Ergebnisse. 1) Zeits-hr. f8r physiol. Chem. 34, 207 [1901]. 708. A l f r e d Stook und Werner Siebert: Zur Daretellung des gelben Arsens. Es wurde i n der gescbilderten Weise der Ueberzug von gelbem Arsen erzeugt, ,d+r Apparrt durch Abscbmelzen des Rohres B von der Luftpumpe' getrennt und nach dem Liisclien des Lichtes, sobald das Auge sich an die v6llige Dunkelheit gewohnt hatte, pus der flussigen Luft heraus zur Erwiirmung in Alkohol yon Zimmertem- peralur gesteckt. Starkes Knistern des abspringenden schwarzen' Arsens bewies die Hefiigkeit der Umwandelung, irgeudwelche Licht- efscheinung zeigle sich aber nicht. Um uns zu uberzeugen, dass es sich bei unseren Versuchen auch wirklich um das schwefelkohlenstofflosliche Arsen handle, gaben wir in das Gefass A vor dem Abkiihlen mit Aussiger Luft einige ccni Schwefelkohlensloff. ( Nachdem sich das gelbe Arsen gebildet hntte,' erwlrmten wir den Apparat, der Schwefelkohlenstoff schmolz und loste beim Umschwenken das Arsen bis auf einen kleinen, durch die Temperaturerhohung geschwarzten Theil auf. Die entatandene Lo- sung eryies sich a19 yollkommen identisch mit eher nach der Vor- schrift von E r d m a n n und v o n U n r u h dargestellten. Die Menge gelbes Arsen, welche wir erhielten, war, wie schoo erwiihnt, nicht gross; sie betrug etwa '/lo g bei dem eiozelnen, '/a-'/4 Stupden wahrenden Versuche. Die Sublimation darf nicht zu sehr be- schleunigt werden, denn sonst kann es geschehen, dass sich die Innen- wand von A erwarmt und der gelbe Beechlag sich braunt. Fiir die Daratellung in grosserem Maassstabe ist das von E r d m a n n und von U n r u h ausgearbeitete Verfahren vorzuziehen. Unser Apparat 4575 709. Emil Fischer und U m e t a r o Suzuki: Synthese yon Polypeptiden. VII. Derivate des Cystins. [Aus dem I. Chemischen Institut der Universitiit Berlin.] (Eingegangeo am 15. November 1904.) (Eingegangeo am 15. November 1904.) 'NBch den schiinen Untersuchungen von K. A. H. Miirner') ist das @stin ein regelmassiger Beatandtheil der achwefelhaltigen Protei'n- stoffe, und da seine Menge in einzelnen Fallen bis fast zu 12 pCt. dee Oesammtgewichts ansteigt, so spielt es auch in dem Molekiil keines-' megs eine untergeordnete Rolle. Es schien deshalb von Interease, polypeptidartige Combinationen dieser geschwefelten Base mit ahderen Aminosauren herzoatellen, weil derartige Producte aller Wahrscheinlich- keit nach beim successiven Auf- und Abbnu der Protei'ne gebildet werden. In alkalischer Liisung lasst sich das Cystin in der That leicht rnit halogenhaltigen Saurechloriden vereinigen und nimmt davon zwei Mo- lekiile ,auf. Wir haben die Versuche mit Chloracetylcblorid, Brom- propionylbromid und a-Bromisocaprooylchlorid ausgeflihrt und in allen Fiilleo krystallinische Producte erhalten. Durch Behandlung mit Ammo- niak entstehen daraus die entsprechenden Polypeptide, das Diglycyl-, Dialanyl- und Dileucyl-Cystin. Die schiinsten Eigenschaften von ihneii besitzt das Alaninderivat, das aus Wssser krystallisirt, wiihqend die beiden anderen. Combinationen nur in amorpbem Zustand gewopnen wurden. Das POI^ uns benutzte Cystin war aus Rosshaar durch Kocben init Salzsiiure gewonneu und derart gereinigt, dass wit annehmen konnen, die reine, optisch active Base verarbeitet zu haben. Infolge- 1) Zeits-hr. f8r physiol. Chem. 34, 207 [1901].
https://openalex.org/W3209216156	http://cmhp.lenterakaji.org/index.php/cmhp/article/download/61/45	Indonesian	null	HUBUNGAN GENETIK DAN PSIKOLOGIS DENGAN KEJADIAN DEPRESI PADA REMAJA	Journal Of Community Mental Health and Public Policy	2,021	cc-by	4,892	Published online October, 11th, 2021 e-ISSN 2622-2655 (Journal Of Community Mental Health And Public Policy) ARTICLE INFO Depression is one of the most common types of mental disorders that can negatively affect a person's feelings, ways of thinking, and acting so that severe ones can lead to suicide. Depression's caused by various factors, which are genetic and psychological. This study was conducted to determine the relationship between genetic (gender and genetic history) and psychological (social support and psychosocial conditions) with the incidence of depression in adolescents (case study of XY High School students, Jember Regency). This research is an analytic observational study with a cross- sectional method. The sample in this study was 158 respondents who were taken using the simple random sampling method. Data were collected online with the Social Provisions Scale (SPS), Youth Pediatric Symptom-17 (Y PSC-17), and Beck Depression Inventory-II (BDI-II) instruments. The research data were analyzed using the Spearman Correlation Test. The results showed that there was no relationship between genetics which consisted of gender variable (p value = 0.958) and heredity history variable (p value = 0.061) with the incidence of depression in adolescents. While there is a relationship between psychology which consists of social support variables (p value = 0.003) and psychosocial condition variables (p value = 0.000) with the incidence of depression in adolescents. The results of this study can be used as a basis for making preventive, promotive, and rehabilitative efforts regarding depression in adolescents and for further research. Keywords: Depression, Genetic, Psychological, Adolescents. Genetic and Psychological Relationship with the Depression Event in Adolescents Lidya Isnaini Nuriyah 1, Ayik Mirayanti Mandagi 2, Xindy Imey Pratiwi 1 dya s a Nu ya , y aya t a dag , dy ey at w 1 Departemen Kesehatan Lingkungan, Fakultas Kesehatan Masyarakat, PSDKU Universitas Airlangga di Banyuwangi, Banyuwangi, Indonesia 2 2 Departemen Epidemiologi, Biostatistika Kependudukan, dan Promosi Kesehatan, Fakultas Kesehatan Masyarakat, PSDKU Universitas Airlangga di Banyuwangi, Banyuwangi, Indonesia lidya.isnaini.nuriyah-2017@fkm.unair.ac.id PENDAHULUAN terdapat banyak kemungkinan yang dapat menyebabkan depresi, seperti gangguan regulasi emosi, masalah kualitas hidup, harga diri yang rendah, pengalaman negatif hingga kurangnya dukungan keluarga (Marroquín & Nolen-Hoeksema, 2015; Orth et al., 2016; Schirda et al., 2015; Vardanyan, 2013). Depresi adalah salah satu jenis gangguan mental yang paling umum terjadi di dunia (World Health Organization, 2015). Depresi menjadi penyakit yang secara negatif memengaruhi perasaan, cara berpikir dan tindakan seseorang dengan menimbulkan rasa sedih dan menghilangkan minat pada aktivitas sehari- hari atau aktivitas yang digemari (Meridien Research, 2021). Pada tingkat yang parah depresi diketahui dapat menyebabkan bunuh diri (World Health Organization, 2015). Diketahui telah terjadi sebanyak 800.000 kasus bunuh diri dilaporkan di seluruh dunia setiap tahunnya dengan depresi sebagai penyebab utama (World Health Organization, 2015). Sementara di Indonesia diketahui bahwa kejadian depresi telah dialami oleh sebanyak 35 juta orang (Kementerian Kesehatan Republik Indonesia, 2016). y ) Para peneliti terdahulu menemukan bahwa genetik memengaruhi sebesar lebih dari 60% kejadian depresi berat pada seseorang (Purves et al., 2019). Sehingga ketika seseorang memiliki riwayat keturunan pernah mengalami depresi, ia memiliki kemungkinan lebih besar mengalami kejadian depresi. Selain itu jenis kelamin seseorang juga mempengaruhi kemungkinan ia dapat mengalami kejadian depresi. Beberapa peneliti terdahulu mengemukakan bahwa seseorang dengan jenis kelamin perempuan lebih mudah mengalami kejadi depresi (Fernandez- Pujals et al., 2015; Hyland et al., 2020). Hal tersebut dikarenakan hormon pada perempuan lebih tidak stabil dibandingkan pada laki-laki, sehingga ketika terdapat faktor yang mampu memicu depresi maka seseorang dengan jenis kelamin perempuan dapat dengan mudah mengalami kejadian depresi (Fernandez-Pujals et al., 2015). Beberapa peneliti terdahulu menyatakan bahwa awal kemunculan depresi terjadi sejak periode kehidupan remaja (Thapar et al., 2012; Williams et al., 2014). Data pada tahun 2017 menunjukkan bahwa terdapat hampir 2,3 juta remaja pada rentang usia 12 hingga 17 tahun di Amerika Serikat mengalami paling sedikit satu kali kejadian depresi mayor (National Institute of Mental Health, 2017). Di Indonesia terdapat data yang menunjukkan bahwa ada lebih dari 3 juta remaja usia mulai dari 10 hingga 19 tahun yang mengalami gangguan mental dengan depresi paling umum terjadi (Badan Pusat Statistik, 2014). Pada faktor biologis diketahui terjadinya peningkatan kadar sitokin yang berkombinasi dengan penurunan kadar kortisol dapat menyebabkan kejadian depresi (Brogan, 2014). Hasil penelitian pada sudut pandang lainnya menunjukkan bahwa hampir 33% kejadian depresi pada remaja disebabkan oleh kondisi psikososial (Azzahro et al., 2021). ABSTRAK Depresi adalah salah satu jenis gangguan mental paling umum terjadi, secara negatif mampu memengaruhi perasaan, cara berpikir dan bertindak seseorang sehingga pada tingkatan parah mampu menyebabkan bunuh diri. Depresi disebabkan oleh berbagai faktor, salah satunya adalah genetik dan psikologis. Penelitian ini dilakukan untuk mengetahui hubungan genetik (jenis kelamin dan riwayat keturunan) dan psikologis (dukungan sosial dan kondisi psikososial) dengan kejadian depresi pada remaja (studi kasus siswa SMA XY Kabupaten Jember). Penelitian ini merupakan penelitian observasional analitik dengan metode cross-sectional. Sampel dalam penelitian ini sebanyak 158 responden yang diambil menggunakan metode simple random sampling. Data diambil secara daring dengan instrumen Social Provisions Scale (SPS), Youth Pediatric Symptom-17 (Y PSC-17) dan Beck Depression Inventory-II (BDI-II). Data penelitian ini dianalisis menggunakan Uji Korelasi Spearman. Hasil penelitian menunjukkan bahwa tidak terdapat hubungan antara genetik yang terdiri dari variabel jenis kelamin (p value = 0.958) dan variabel riwayat keturunan (p value = 0.061) dengan kejadian depresi pada remaja. Sedangkan terdapat hubungan antara psikologis yang terdiri dari variabel dukungan sosial (p value = 0.003) dan variabel kondisi psikososial (p value = 0.000) dengan kejadian depresi pada remaja. Hasil penelitian ini dapat dijadikan dasar untuk pembuatan upaya preventif, promotif dan rehabilitatif perihal masalah depresi pada remaja serta menjadi landasan untuk penelitian lebih lanjut. Kata Kunci: Depresi, Genetik, Psikologis, Remaja. 26 CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja PENDAHULUAN Kondisi psikososial yang diketahui dapat menyebabkan kejadian depresi pada remaja berkaitan dengan dukungan atau keterikatan emosional, integrasi sosial, kepastian nilai, bantuan nyata, pengenalan dan kesempatan pengasuhan yang baik (Azzahro et al., 2021). Kejadian depresi pada remaja bukan hanya sekedar perasaan stres ataupun sedih, namun sudah menjadi keadaan serius yang dapat memengaruhi sifat, cara berpikir, dan berperilaku secara permanen sehingga membutuhkan penanganan yang serius (Azzahro et al., 2021). Penyebab depresi berdasarkan hasil penelitian terdahulu diketahui dikarenakan oleh beberapa faktor seperti genetik, biologis, dan kondisi psikososial (Dianovinina, 2018). Selain itu masih Dampak terjadinya depresi pada remaja di Indonesia dapat dilihat dalam bentuk peningkatan konsumsi rokok, penyalahgunaan zat terlarang, kasus bunuh 27 CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja diri, dan penurunan fungsi sosial serta prestasi (Praptikaningtyas et al., 2019). Pernyataan tersebut diperkuat dengan data yang menunjukkan bahwa penyalahgunaan narkotika pada kalangan remaja di Indonesia dewasa ini meningkat sebesar hampir 30% (Badan Narkotika Nasional Republik Indonesia, 2019). Berdasarkan pemaparan faktor penyebab dan dampak akibat kejadian depresi pada remaja, maka kejadian depresi perlu ditangani dengan cepat dan tepat agar tidak mengakibatkan permasalahan lebih lanjut bagi seseorang yang mengalaminya (Sulistyorini & Sabarisman, 2017). mengenai hubungan antara genetik dan psikologis dengan kejadian depresi. Penyampaian informasi tersebut diharapkan mampu mencegah masalah gangguan mental utamanya depresi secara lebih lanjut. HASIL Analisis hubungan antar variabel dilakukan menggunakan Uji Korelasi Spearman dengan signifikansi 0,05. Hasil uji korelasi Spearman ini diambil dengan dasar pengambilan keputusan melihat nilai korelasi koefisien dan signifikansi antar variabel. Selain itu analisis ini dilakukan untuk melihat hubungan antar variabel secara parsial. Hasil uji analisis dicantumkan dalam tabel 2. METODE PENELITIAN 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja Instrumen yang digunakan untuk mengukur faktor psikososial adalah instrumen terstandar yaitu Youth Pediatric Symptom-17 (Y-PSC-17) (Murphy et al., 2016). Y-PSC-17 dapat digunakan untuk mengukur faktor psikososial pada anak usia sekolah hingga usia remaja akhir yaitu pada usia 4 hingga 24 tahun. Sementara itu instrumen yang digunakan untuk meneliti kejadian depresi pada remaja adalah kuesioner Beck Depression Inventory-II (BDI-II) (Uin & Hidayatullah, 2018). Sumber: Data primer, 2020 Tabel 1. Distribusi Frekuensi Variabel Data Responden Tahun 2020. Variabel n % Jenis Kelamin Laki-Laki 37 23,4 Perempuan 121 76,6 Riwayat Keturunan Memiliki riwayat keturunan 11 7,0 Tidak memiliki riwayat Keturunan 147 93,0 Dukungan Sosial Baik 90 57,0 Buruk 68 43,0 Kondisi Psikososial Baik 79 50,0 Terganggu 79 50,0 Kejadian Depresi Depresi 72 45,6 Tidak Depresi 86 54,4 Tabel 1. Distribusi Frekuensi Variabel Data Responden Tahun 2020. y Pengolahan dan analisis data dilakukan menggunakan aplikasi SPSS 21 dengan analisis univariat dan analisis statistik non parametrik. Analisis univariat dalam penelitian ini dilakukan untuk mengetahui distribusi frekuensi data penelitian, sedangkan analisis statistik non parametrik dilakukan untuk menguji hubungan antar variabel. Analisis statistik non parametrik yang digunakan dalam penelitian ini adalah Uji Korelasi Spearman. METODE PENELITIAN Penelitian ini merupakan penelitian observasional analitik dengan metode cross-sectional. Penelitian ini dilakukan di beberapa Sekolah Menengah Atas (SMA) di Kabupaten Jember, Jawa Timur selama 3 bulan pada bulan April hingga Juni 2020. Populasi pada penelitian ini adalah seluruh siswa kelas 12 di lokasi penelitian dengan total sebanyak 350 orang responden. Penanganan dapat dilakukan sedari upaya pencegahan (preventive) hingga rehabilitatif. Peneliti terdahulu menunjukkan terdapat berbagai upaya yang dapat menurunkan kejadian depresi pada remaja seperti, pola asuh orang tua, efikasi diri, expressive writing therapy dan resiliensi individu (Danarti et al., 2018; Florensa et al., 2016; Safitri & Hidayati, 2013; Shintia & Maharani, 2021; Wahyuningsih & Mamnu’ah, 2015). Pemilihan pola asuh orang tua yang demokratif diketahui secara signifikan berhubungan dengan rendahnya tingkat depresi pada remaja (Safitri & Hidayati, 2013). Selain pemilihan pola asuh orang tua, peningkatan efikasi dan resiliensi diri juga berhubungan secara signifikan dengan penurunan depresi pada remaja (Florensa et al., 2016; Shintia & Maharani, 2021). Selain itu terdapat alternatif lain yang terbukti berpengaruh terhadap penurunan depresi pada remaja, yaitu expressive writing therapy (Danarti et al., 2018). y g p Besar sampel dalam penelitian ini dihitung menggunakan Rumus Slovin dan menghasilkan jumlah sampel sebanyak 158 orang responden. Metode pengambilan sampel yang digunakan dalam penelitian ini adalah simple random sampling. Pengambilan sampel dilakukan dengan memasukkan daftar nama seluruh siswa kemudian akan dipilih secara acak pada laman daring wheel of names yang digunakan karena kemudahan dan otomatisasi penyimpanan nama terpilih. Sehingga meminimalisir adanya nama terpilih yang terlewatkan. Nama responden yang terpilih kemudian akan dihubungi oleh peneliti dan apabila bersedia akan diberikan link pengisian kuesioner. Variabel bebas dalam penelitian ini antara lain genetik (jenis kelamin dan riwayat keturunan) dan psikologis (dukuangan sosial dan kondisi psikososial), sedangkan variabel terikatnya adalah kejadian depresi. Metode pengumpulan data pada penelitian ini adalah survei daring dengan melakukan pengisian kuesioner pada Formulir Google. Instrumen yang digunakan untuk meneliti genetik adalah pengisian data diri pada laman kuesioner. Instrumen yang digunakan untuk meneliti dukungan sosial menggunakan instrumen terstandar yaitu Social Provisions Scale (SPS) (Chiu et al., 2016). Pemaparan data di atas menjadi dasar pelaksanaan penelitian ini. Penelitian ini dilakukan dengan tujuan mengidentifikasi hubungan genetik (jenis kelamin dan genetik atau riwayat keturunan) dan psikologis (dukungan sosial dan kondisi psikososial) dengan kejadian depresi pada remaja. Urgensi penelitian ini adalah sebagai bentuk kontribusi dalam memberikan informasi bagi para remaja, keluarga atau pihak terkait lainnya 28 CMHP Oktober 2021: Vol. 4 No. Distribusi Frekuensi Variabel Data pada tabel 1 menunjukkan bahwa mayoritas responden berjenis kelamin perempuan sebesar 76,6% dan sisanya sebesar 23,4% adalah laki-laki. Diketahui pula bahwa mayoritas responden tidak memiliki riwayat keturunan pernah mengalami depresi (93,0%). Sebagian besar reponden juga memiliki tingkat dukungan sosial yang baik (57,0%). Meskipun gambaran kondisi psikososial pada responden seimbang yaitu pada kondisi baik maupun terganggu. Namun diketahui bahwa mayoritas responden tidak mengalami kejadian depresi (54,4%). Data tabulasi silang pada tabel 2 menunjukkan bahwa mayoritas responden perempuan yang mengalami kejadian depresi sejumlah 55 orang (34,8%). Sedangkan responden laki-laki yang mengalami kejadian depresi sejumlah 17 orang (10,8%). Pada sudut pandang riwayat keturunan diketahui mayoritas remaja yang mengalami kejadian depresi tidak memiliki riwayat keturunan orang tua pernah mengalami gangguan depresi sejumlah 64 orang (40,5%). Sedangkan remaja dengan riwayat keturunan yang mengalami kejadian depresi sejumlah 8 orang (5,0%). Berdasarkan interpretasi tersebut dari penelitian yang telah dilakukan maka dapat diketahui distribusi frekuensi data responden melalui tabel berikut: 29 CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja Tabel 2. Hubungan antara Genetik dan Psikologis dengan Kejadian Depresi pada Remaja Tahun 2020 Kejadian Depresi Total Koefisien Korelasi Sig. Depresi Tidak Depresi n % n % n % Genetik Jenis Kelamin Laki-Laki 17 10,8 20 12,8 37 23,4 -0,004 0,958 Perempuan 55 34,8 66 41,8 121 76,6 Riwayat Keturunan Ada 8 5,0 3 2,0 11 7,0 0,149 0,061 Tidak ada 64 40,5 83 52,5 147 93,0 Psikologis Dukungan Sosial Baik 32 20,3 58 36,7 90 57,0 -0,231 0,003 Buruk 40 25,3 28 17,7 68 43,0 Kondisi Psikososial Baik 20 12,7 59 37,3 79 50,0 -0,407 0,000 Terganggu 52 32,9 27 17,1 79 50,0 **Berhubungan secara signifikan pada α = 0.01 Sumber: Data primer, 2020 bel 2. Hubungan antara Genetik dan Psikologis dengan Kejadian Depresi pada Remaja hun 2020 Pada sudut pandang faktor psikologis diketahui bahwa mayoritas responden dengan dukungan sosial yang baik tidak mengalami kejadian depresi yaitu sebanyak 58 orang dari total 90 responden atau sebesar 36,7%. Sedangkan responden dengan dukungan sosial yang buruk mayoritas mengalami kejadian depresi dengan jumlah 40 orang dari total 68 responden atau sebesar 25,3%. Hal serupa juga terjadi pada data variabel psikososial yang menunjukkan sebanyak 59 orang dari total 79 responden dengan kondisi psikososial baik tidak mengalami kejadian depresi atau sebesar 37,3%. Distribusi Frekuensi Variabel Sementara sebanyak 52 orang dari total 79 responden dengan kondisi psikososial terganggu mengalami kejadian depresi atau sebesar 32,9%. diketahui bahwa genetik tidak memiliki hubungan dengan kejadian depresi pada remaja, sedangkan psikologis memiliki hubungan dengan kejadian depresi pada remaja. Pada tabel 2 juga ditunjukkan kuat dan nilai hubungan antar variabel. Variabel dukungan sosial diketahui berhubungan sangat lemah bernilai negatif (tidak searah) dengan kejadian depresi pada remaja. Berbeda dengan variabel dukungan sosial, varibel psikososial memiliki hubungan yang cukup kuat dengan kejadian depresi pada remaja meskipun sama-sama bernilai negatif (tidak searah). Hal tersebut bermakna bahwasannya apabila seorang remaja mendapatkan dukungan sosial dan psikososial yang semakin baik maka remaja tersebut semakin minim mengalami kejadian depresi. Hasil uji korelasi spearman pada tabel 2 menunjukkan bahwa variabel jenis kelamin (sig. 0,958) dan riwayat keturunan (sig. 0,061) tidak berhubungan secara signifikan dengan kejadian depresi pada remaja karena memiliki nilai signifikansi > 0,05. Sedangkan variabel dukungan sosial (sig. 0,003) dan psikososial (sig. 0,000) berhubungan secara signifikan dengan kejadian depresi pada remaja karena memiliki nilai signifikansi < 0,05. Sehingga PEMBAHASAN Hasil penelitian ini menunjukkan bahwa kejadian depresi yang dilihat dari sudut pandang jenis kelamin tidak berhubungan secara signifikan. Hal tersebut terlihat pada data bahwasannya baik remaja dengan jenis kelamin laki-laki maupun perempuan memilki probabilitas yang sama 30 CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja dalam mengalami kejadian depresi. Hasil ini menunjukkan bahwa jenis kelamin tidak menentukan seseorang akan lebih mungkin mengalami kejadian depresi. Pada variabel riwayat keturunan menunjukkan bahwa remaja yang memiliki riwayat keturunan depresi lebih berpotensi mengalami kejadian depresi. Namun ketika dilakukan uji analisis hubungan menunjukkan hasil bahwa riwayat keturunan tidak berhubungan secara signifikan dengan kejadian depresi. Sehingga meskipun seorang remaja tersebut berpotensi belum tentu akan benar-benar mengalami kejadian depresi. depresi berhubungan dengan kondisi psikososial yang dialami seseorang (Azzahro et al., 2021; Haryanto et al., 2015; Marsasina & Fitrikasari, 2016; Putri & Tobing, 2020; Santoso et al., 2018). Dukungan sosial dan kondisi psikososial diketahui sebagai variabel yang berjalan seiringan, dimana pemberian dukungan sosial yang positif pada seseorang dapat meningkatkan kondisi psikososialnya ke arah lebih baik dan mencegah terjadinya gangguan atau permasalahan psikososial (Fitria & Maulidia, 2018). Hal tersebut diperkuat dengan hasil penelitian lainnya yang menyatakan bahwa sebesar hampir 80% remaja membutuhkan dukungan sosial dari keluarga dalam bentuk dukungan emosional untuk lebih mempersiapkan diri dalam menghapi permasalahan yang sedang dihadapi (Fitria & Maulidia, 2018). Hasil uji hubungan antara genetik dengan kejadian depresi pada penelitin ini sejalan dengan penelitian terkini yang juga menyatakan bahwa riwayat keturunan tidak berhubungan dengan kejadian depresi yang dialami oleh para remaja utamanya siswa SMA (Herfinanda & Kaloeti, 2021). Hasil uji hubungan antara psikososial dengan kejadian depresi pada penelitian ini sejalan dengan hasil penelitian lainnya yang menyatakan bahwa dukungan sosial dan kondisi psikososial berhubungan dengan kejadian depresi pada remaja (Azzahro et al., 2021; Bintang & Mandagi, 2021; Rahmayanti & Rahmawati, 2018; Syahputra et al., 2018). Penelitian ini tentunya belum sepenuhnya sempurna, masih perlu dilakukan penelitian lebih lanjut terkait topik serupa bahkan dengan variabel yang lebih luas. Hal tersebut dikarenakan penelitian ini masih memiliki keterbatasan dalam proses pengambilan data. Pengambilan data yang dilakukan secara daring dari responden membuat peneliti tidak dapat mendampingi selama proses pengisian data. SIMPULAN DAN SARAN Kesimpulan dari sosialisasi ini adalah bahwa presentasi dan video animasi singkat dapat meningkatkan pengetahuan dan pemahaman remaja mengenai kecanduan gawai. Sebagai saran, program promosi mengenai penggunaan gawai yang sehat pada remaja dapat mencontoh metode ini. Studi berikutnya perlu melaporkan status kecanduan gawai dan hasil intervensi yang dilakukan untuk mencegah kecanduan gawai pada kelompok umur ini. Badan Narkotika Nasional Republik Indonesia. (2019). Penggunaan Narkotika di Kalangan Remaja Meningkat. https://bnn.go.id/. Badan Pusat Statistik. (2014). Data Remaja Teridentifikasi Depresi. www.bps.go.id. Bintang, A. Z., & Mandagi, A. M. (2021). Kejadian Depresi Pada Remaja Menurut Dukungan Sosial Di Kabupaten Jember. Journal of Community Mental Health and Public Policy, 3(2), 92–101. REFERENSI Adhikari, S. P., Meng, S., Wu, Y. J., Mao, Y. P., Ye, R. X., Wang, Q. Z., Sun, C., Sylvia, S., Rozelle, S., Raat, H., & Zhou, H. (2020). Epidemiology, Causes, Clinical Manifestation And Diagnosis, Prevention And Control Of Coronavirus Disease (COVID- 19) During The Early Outbreak Period: A Scoping Review. Infectious Diseases of Poverty, 9(1), 1–12. Maka dari itu dapat dilakukan penelitian serupa di berbagai lokasi lainnya di Indonesia maupun negara lainnya dengan responden yang memiliki kriteria sama. Meskipun demikian penelitian ini layak digunakan sebagai literatur atau pertimbangan pada penelitian selanjutnya melihat kuatnya teori atau penelitian sebelumnya yang melengkapi hasil penelitian ini. Azzahro, E. A., Dian, J., Sari, E., Masyarakat, F. K., Epidemologi, D., Kesehatan, F., & February, F. (2021). Faktor Psikososial Dengan Kejadian Depresi Pada Remaja (Studi pada Siswa Kelas 12 SMA XY Jember) Psychosocial Factors with The Incidence of Depression in Adolescents (Study at 12th Grade XY High Schools Students in Jember). 2655, 69–77. PEMBAHASAN Sehingga mengakibatkan beberapa responden mengalami kesulitan memahami pertanyaan dalam kuesioner. Dukungan sosial diketahui berhubungan dengan kejadian depresi secara signifikan bernilai negatif (tidak sejalan). Sehingga ketika terjadi peningkatan dukungan sosial maka terjadi penurunan kemungkinan kejadian depresi pada remaja (Bintang & Mandagi, 2021; Rahmayanti & Rahmawati, 2018; Syahputra et al., 2018). Mempertimbangkan kondisi pandemi COVID-19 saat dilakukannya penelitian ini juga dapat diketahui bahwa terdapat banyak faktor eksternal individu yang secara ekstrem lebih mempengaruhi kejadian depresi pada remaja (Adhikari et al., 2020; Brooks et al., 2020; Cornine, 2020; De Olivera F et al., 2020; Duan & Zhu, 2020; Hasanah et al., 2020; Huang et al., 2020; Hyland et al., 2020; Liang et al., 2020; Maia & Dias, 2020; Mazza et al., 2020; Sahu, 2020). Faktor eksternal tersebut meliputi, krisis global, pandemi dunia, tekanan psikologis, tekanan ekonomi, kekurangan sumber daya, kerugian sosial, kerugian finansial, pemberlakukan jaga jarak dan Kondisi psikososial diketahui memiliki hubungan yang signifikan bernilai negatif (tidak sejalan) dengan kejadian depresi pada remaja. Sehingga ketika kondisi psikososial seseorang semakin membaik maka akan terjadi penurunan kemungkinan kejadian depresi pada remaja (Azzahro et al., 2021). Hasil tersebut sejalan dengan beberapa penelitian sebelumnya yang juga menyatakan bahwa kejadian 31 CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja karantina mandiri (Adhikari et al., 2020; Brooks et al., 2020; Cornine, 2020; De Olivera F et al., 2020; Huang et al., 2020; Hyland et al., 2020; Liang et al., 2020; Mazza et al., 2020; Sahu, 2020). Sehingga keterbatasan variabel pengujian ini tidak dapat mengimbangi faktor eksternal lain yang juga mempengaruhi kejadian depresi pada responden penelitian. Penelitian ini juga hanya dilakukan pada satu kabupaten sehingga belum bisa menginterpretasikan kondisi di seluruh Indonesia. UCAPAN TERIMA KASIH Terima kasih kepada seluruh tim peneliti dan tim pengumpul data atas kontribusinya dalam penelitian ini dari awal hingga akhir. Sejak rumusan masalah, penyusunan proposal, instrumen penelitian dan perizinan hingga pelaporan hasil penelitian. Terima kasih juga kepada para responden yang bersedia mengikuti penelitian ini. Terima kasih juga kepada Universitas Airlangga yang telah membantu dan mendukung penelitian ini. Brogan, K. (2014). How Inflammation and Depression Gave Birth to New Medicine. https://kellybroganmd.com/. Brooks, S. K., Webster, R. K., Smith, L. E., Woodland, L., Wessely, S., Greenberg, N., & Rubin, G. J. (2020). The Psychological Impact Of Quarantine And How To Reduce 32 CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja It: Rapid Review Of The Evidence. The Lancet, 395(10227), 912–920. Study (GS: SFHS). PLoS ONE, 10(11), 1–18. Chiu, C.-Y., Motl, R. W., & Ditchman, N. (2016). Validation of the Social Provisions Scale in people with multiple sclerosis. Rehabilitation Psychology, 61(3), 297–307. Fitria, Y., & Maulidia, R. (2018). Hubungan Antara Dukungan Sosial Keluarga Dengan Depresi Pada Remaja Di Smpn Kota Malang (Relationship Between Social Support With Adolescent Depression in Junior High School at Malang). Prosiding Seminar Nasional Hasil Penelitian Dan Pengabdian Epada Masyarakat III, September, 270–276. Cornine, A. (2020). Reducing Nursing Student Anxiety in the Clinical Setting: An Integrative Review. Nursing Education Perspectives, 41(4), 229–234. Florensa, Keliat, B. A., & Wardani, I. Y. (2016). Peningkatan Efikasi Diri Dan Penurunan Depresi. Jurnal Keperawatan Indonesia, 19(3), 169– 175. Danarti, N. K., Sugiarto, A., & Sunarko. (2018). Pengaruh Expressive Writing Therapy Terhadap Penurunan Depresi, Cemas, dan Stres. Journal Ilmu Keperawatan Jiwa, 1(1), 48–61. Haryanto, Wahyuningsih, H. D., & Nandiroh, S. (2015). Sistem Deteksi Gangguan Depresi pada Anak - Anak dan Remaja. Jurnal Ilmiah Teknik Industri, 14(2), 142–152. De Olivera F, Abrantes L, Martins P, Bezerra C, & Rolim M. (2020). Impact Of Sars-Cov-2 And Its Reverberation In Global Higher Education And Mental Health. Psychiatry Research journal; 288: 1- 2. January. Hasanah, U., Fitri, N. L., Supardi, S., & PH, L. (2020). Depression Among College Students Due to the COVID-19 Pandemic. Jurnal Keperawatan Jiwa, 8(4), 421. Dianovinina, K. (2018). Depresi pada Remaja: Gejala dan Permasalahannya. Journal Psikogenesis, 6(1), 69–78. Herfinanda, R., & Kaloeti, V. S. (2021). UCAPAN TERIMA KASIH Family Resilience during the COVID-19 Pandemic : A Systematic Literature Study : Resiliensi Keluarga selama Pandemi COVID-19 : Studi Literatur Sistematik. Proceding of Inter-Islamic University Conference on Psychology, 1(1), 1–11. Duan, L., & Zhu, G. (2020). Psychological Interventions for People Affected by the COVID-19 Epidemic. The Lancet Psychiatry, 7(4), 300–302. h Fernandez-Pujals, A. M., Adams, M. J., Thomson, P., McKechanie, A. G., Blackwood, D. H. R., Smith, B. H., Dominiczak, A. F., Morris, A. D., Matthews, K., Campbell, A., Linksted, P., Haley, C. S., Deary, I. J., Porteous, D. J., MacIntyre, D. J., & McIntosh, A. M. (2015). Epidemiology And Heritability Of Major Depressive Disorder, Stratified By Age Of Onset, Sex, And Illness Course In Generation Scotland: Scottish Family Health Huang, C., Wang, Y., Li, X., Ren, L., Zhao, J., Hu, Y., Zhang, L., Fan, G., Xu, J., Gu, X., Cheng, Z., Yu, T., Xia, J., Wei, Y., Wu, W., Xie, X., Yin, W., Li, H., Liu, M., … Cao, B. (2020). Clinical Features Of Patients Infected With 2019 Novel Coronavirus in Wuhan, China. The Lancet, 395(10223), 497–506. 33 CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja Associated Factors. International Journal of Environmental Research and Public Health, 17(3165), 1–14. Hyland, P., Shevlin, M., McBride, O., Murphy, J., Karatzias, T., Bentall, R. P., Martinez, A., & Vallières, F. (2020). Anxiety And Depression In The Republic Of Ireland During The COVID-19 Pandemic. Acta Psychiatrica Scandinavica, 142(3), 249–256. Meridien Research. (2021). The Genetics of Depression. meridienresearch.net. Murphy, J. M., Bergmann, P., Chiang, C., Sturner, R., Howard, B., Abel, M. R., & Jellinek, M. (2016). The PSC- 17: Subscale Scores, Reliability, And Factor Structure In A New National Sample. Pediatrics, 138(3). Kementerian Kesehatan Republik Indonesia. (2016). Keperawatan Jiwa. Badan Pengembangan dan Pemberdayaan Sumber Daya Manusia. National Institute of Mental Health. (2017). Major Depression. www.nimh.nih.gov. Liang, L., Ren, H., Cao, R., Hu, Y., Qin, Z., Li, C., & Mei, S. (2020). The Effect of COVID-19 on Youth Mental Health. Psychiatric Quarterly, 91(3), 841–852. Orth, U., Robins, R. W., Meier, L. L., & Conger, R. D. (2016). Supplemental Material for Refining the Vulnerability Model of Low Self- Esteem and Depression: Disentangling the Effects of Genuine Self-Esteem and Narcissism. Journal of Personality and Social Psychology, 110(1), 133–149. Maia, B. R., & Dias, P. C. (2020). UCAPAN TERIMA KASIH Anxiety, Depression And Stress In University Students: The Impact of COVID-19. Estudos de Psicologia (Campinas), 37, 1–8. Marroquín, B., & Nolen-Hoeksema, S. (2015). Emotion Regulation And Depressive Symptoms: Close Relationships As Social Context And Influence. Journal of Personality and Social Psychology, 109(5), 836–855. Praptikaningtyas, A. A. I., Wahyuni, A. A. S., & Aryani, L. N. A. (2019). Hubungan Tingkat Depresi pada Remaja dengan Prestasi Akademis Siswa SMA Negeri 4 Denpasar. Jurnal Medika Udayana, 8(7), 1–5. Marsasina, A., & Fitrikasari, A. (2016). Gambaran Dan Hubungan Tingkat Depresi Dengan Faktor-Faktor Yang Mempengaruhi Pada Pasien Rawat Jalan Puskesmas (Studi Deskriptif Analitik Di Puskesmas Halmahera Semarang). Jurnal Kedokteran Diponegoro, 5(4), 440– 450. Purves, K., Coleman, J., Meier, S., Rayner, C., Davis, K., Choi, S. W., Deckert, J., Gaspar, H. A., Hettema, J., Hübel, C., Mattheisen, M., Nicodemus, K., Breen, G., & Eley, T. (2019). Major Role for Common Genetic Variation in Anxiety Disorders. European Neuropsychopharmacology, 29, S1196. Mazza, C., Ricci, E., Biondi, S., Colasanti, M., Ferracuti, S., Napoli, C., & Roma, P. (2020). A Nationwide Survey Of Psychological Distress Among Italian People During The Covid-19 Pandemic: Immediate Psychological Responses And Putri, K. F., & Tobing, D. L. (2020). Tingkat Resiliensi dengan Ide Bunuh Diri Pada Remaja. Jurnal Ilmiah Ilmu Keperawatan Indonesia, 10(01), 1–6. 34 CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja Rahmayanti, Y. E., & Rahmawati, T. (2018). Hubungan Dukungan Keluarga Dengan Kejadian Depresi Pada Remaja Awal. Jurnal Asuhan Ibu & Anak, 3(6), 8. Syahputra, A., Theresa, R. M., & Bustamam, N. (2018). Hubungan Dukungan Sosial Dengan Tingkat Depresi Pada Mahasiswa Fakultas Kedokteran Universitas Pembangunan Nasional “Veteran” Jakarta Angkatan 2018. Seminar Nasional Riset Kedokteran. Safitri, Y., & Hidayati, N. E. (2013). Hubungan Antara Pola Asuh Orang Tua Dengan Tingkat Depresi Remaja Di Smk 10 November Semarang. Jurnal Keperawatan Jiwa, 1(1), 11–17. Thapar, A., Collishaw, S., Pine, D. S., & Thapar, A. K. (2012). Depression in adolescence. The Lancet, 379(9820), 1056–1067. Sahu, P. (2020). Closure of Universities Due to Coronavirus Disease 2019 (COVID-19): Impact on Education and Mental Health of Students and Academic Staff. Cureus, 2019(4). Uin, S., & Hidayatullah, S. (2018). Uji Validitas Konstruk Beck Depression Inventory-II (BDI-II). Jurnal Pengukuran Psikologi Dan Pendidikan Indonesia, 4(1). Santoso, M. B., Siti Asiah, D. H., & Kirana, C. I. (2018). Bunuh Diri Dan Depresi Dalam Perspektif Pekerjaan Sosial. CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp UCAPAN TERIMA KASIH Prosiding Penelitian Dan Pengabdian Kepada Masyarakat, 4(3), 390. Vardanyan, A. (2013). Risk Factors And Prevalence Of Adolescent Depression In Yerevan, Armenia (A Cross-Sectional Study) Master of Public Health Integrating Experience Project Professional Publication. Schirda, B. L., Nicholas, J. A., & Prakash, R. S. (2015). Examining The Role Of Trait Mindfulness And Emotion Regulation In Quality Of Life In Multiple Sclerosis. Health Psychology, 34(11), 1107–1115. Wahyuningsih, S., & Mamnu’ah. (2015). Hubungan Faktor Keturunan Dengan Kejadian Gangguan Jiwa Di Desa Banaran Galur Kulor Progo Yogyakarta. Electronic Theses and Disertasi STIKES Aisyiyah Yogyakarta, 1–16. Shintia, & Maharani, W. (2021). Kemampuan Resiliensi Individu dalam Menghadapi Psychological Distress Siswa-Siswi SMA Jakarta di Masa Pandemi Covid-19. Prosiding Konferensi Nasional Universitas Nahdlatul Ulama Indonesia, 1(1), 45–54. Williams, J. M. G., Crane, C., Barnhofer, T., Brennan, K., Duggan, D. S., Fennell, M. J. V., Hackmann, A., Krusche, A., Muse, K., Von Rohr, I. R., Shah, D., Crane, R. S., Eames, C., Jones, M., Radford, S., Silverton, S., Sun, Y., Weatherley- Jones, E., Whitaker, C. J., … Russell, I. T. (2014). Mindfulness- Based Cognitive Therapy For Preventing Relapse In Recurrent Depression: A Randomized Dismantling Trial. Journal of Consulting and Clinical Psychology, 82(2), 275–286. Sholih, Rochani, Khairun, D. Y., & Alhakim, I. (2017). Meningkatkan Resiliensi Remaja Melalui Bibliocounseling. Jurnal Penelitian Bimbingan Dan Konseling, 2(1). Sulistyorini, W., & Sabarisman, M. (2017). Depresi : Suatu Tinjauan Psikologis. Sosio Informa, 3(2), 153–164. 35 CMHP Oktober 2021: Vol. 4 No. 1 (26-36) http://cmhp.lenterakaji.org/index.php/cmhp Lidya, et al. Hubungan Genetik dan Psikologis dengan Kejadian Depresi pada Remaja World Health Organization. (2015). Fact Sheets: Depression. www.who.int. World Health Organization. (2015). Fact Sheets: Depression. www.who.int. 36
https://openalex.org/W2771565478	https://europepmc.org/articles/pmc5730604?pdf=render	English	null	Synthetic integrin-binding immune stimulators target cancer cells and prevent tumor formation	Scientific reports	2,017	cc-by	8,946	Synthetic integrin-binding immune stimulators target cancer cells and prevent tumor formation Manuel Brehs1 André J G Pötgens2 Julia Steitz3 KarineThewes1 Janett Schwarz2 Received: 3 July 2017 Accepted: 28 November 2017 Published: xx xx xxxx Received: 3 July 2017 Accepted: 28 November 2017 Published: xx xx xxxx Immuno-oncology approaches mainly utilize monoclonal antibodies or protein-based scaffolds that bind with high affinity to cancer cells and can generate an immune response. Peptides can also bind with high affinity to cancer cells and are intermediate in size between antibodies and small molecules. They are also synthetically accessible and therefore easily modified to optimize their stability, binding affinity and selectivity. Here we describe the design of immune system engagers (ISErs), a novel class of synthetic peptide-based compounds that bind specifically to cancer cells and stimulate the immune system. A prototype, Y9, targets integrin α3, which is overexpressed on several cancer cells, and activates the immune system via a formyl methionine-containing effector peptide. Injection of Y9 leads to immune cell infiltration into tissue and prevents tumor formation in a guinea pig model. The anti- tumor activity and synthetic accessibility of Y9 illustrate that ISErs could be applied to a wide variety of targets and diseases. Antibodies (Abs), antibody-drug conjugates and their derivatives have become a significant part of state of the art cancer treatments1. More than 14 monoclonal antibodies (mAbs) are currently approved for cancer therapy and many more are under development2–4. These antibodies are raised against unique cell surface receptors or antigens, allowing specific targeting of tumor cells. g gi g g Immune-mediated mechanisms, including complement-dependent cytotoxicity (CDC), antibody-dependent cell-mediated cytotoxicity (ADCC) and other secondary immunological effects have been shown to play a cru- cial role in the therapeutic efficacy of mAbs2,5,6. The mechanism of action of mAbs was previously thought to involve blocking the physiological function of the target (e.g. a growth factor or cytokine receptor) by the Fab portion but recent studies have demonstrated stimulation of the immune system by the Fc portion; Fc gamma receptor-mediated activation of macrophages and natural killer cells leading to ADCC is necessary for the anti-tumor effects of Rituximab (anti-CD20) and Trastuzumab/Herceptin (anti-Her2/ErbB2)7. f Both biological and chemical approaches have been developed to overcome the intrinsic limitations of mAbs with respect to selectivity, large size, stability, limited scope for alterations and production costs. Several approaches based on full IgGs, antibody fragments or antibody mimics have been described8–19 and multispeci- ficity has been achieved via genetic fusion or by chemical cross-linking of different complementarity determining regions (CDRs)20,21. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Received: 3 July 2017 Accepted: 28 November 2017 Published: xx xx xxxx Synthetic integrin-binding immune stimulators target cancer cells and prevent tumor formation Manuel Brehs1 André J G Pötgens2 Julia Steitz3 KarineThewes1 Janett Schwarz2 A small molecule approach utilizes antibody-recruiting molecules (ARMs) that combine a target-binding moiety with one for antibody recruitment22. Recently, in a combination of biological and chem- ical methods, synthetic peptides have been linked to antibody scaffolds using site-selective reactions11,23 and a synthetic molecule with targeting and effector functions similar to those of antibodies has also been reported24. f We sought to design a fully synthetic molecule of intermediate size (~5 kDa) between small molecules and Abs that would harness the ability of Abs to recognize tumor cells and initiate an innate immune response against them. These immune system engagers (ISEr) comprise an immune stimulatory effector peptide and two binder peptides that bind selectively to cell-surface markers of tumor or tumor-associated cells, are synthetically acces- sible and do not activate cell-surface receptors (Fig. 1). The two binder peptides are linked to the effector peptide via chemically inert, non-immunogenic, monodisperse polyethylene glycol (PEG) chains. The PEG length (10 nm 1Institute of Biological Chemistry, Faculty of Chemistry, University of Vienna, 1090, Vienna, Austria. 2Syntab Therapeutics GmbH, 52074, Aachen, Germany. 3Institute for Laboratory Animal Science, University Hospital RWTH, Aachen, Germany. 4Department of Medicine III, University Hospital RWTH, Aachen, Germany. Manuel Brehs, André J.G. Pötgens, Julia Steitz, Karine Thewes and Janett Schwarz contributed equally to this work. Correspondence and requests for materials should be addressed to C.F.W.B. (email: christian.becker@univie.ac.at) Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 1 www.nature.com/scientificreports/ Figure 1. (A) Comparison of a typical IgG (left) and the synthetic immune system engager (ISEr) Y9 consisting of an effector (F) as well as two linker and binder (B) moieties. (B) Proposed mechanism of innate immune system activation by Y9. Upon specific binding to tumor cells, neutrophils, monocytes and macrophages are recruited to these cells and directly attack tumor cells displaying Y9. Further effects are based on cytokine release of macrophages and monocytes, recruiting other immune cells. Figure 1. (A) Comparison of a typical IgG (left) and the synthetic immune system engager (ISEr) Y9 consisting of an effector (F) as well as two linker and binder (B) moieties. (B) Proposed mechanism of innate immune system activation by Y9. Upon specific binding to tumor cells, neutrophils, monocytes and macrophages are recruited to these cells and directly attack tumor cells displaying Y9. Further effects are based on cytokine release of macrophages and monocytes, recruiting other immune cells. t Figure 1. Synthetic integrin-binding immune stimulators target cancer cells and prevent tumor formation Manuel Brehs1 André J G Pötgens2 Julia Steitz3 KarineThewes1 Janett Schwarz2 (A) Comparison of a typical IgG (left) and the synthetic immune system engager (ISEr) Y9 consisting of an effector (F) as well as two linker and binder (B) moieties. (B) Proposed mechanism of innate immune system activation by Y9. Upon specific binding to tumor cells, neutrophils, monocytes and macrophages are recruited to these cells and directly attack tumor cells displaying Y9. Further effects are based on cytokine release of macrophages and monocytes, recruiting other immune cells. per PEG27 25) was chosen to cover similar distances as the two paratopes in an antibody (Fig. 1A). By combining two binder peptides per ISEr, we aimed to avoid the fast dissociation and low retention times that can reduce the efficacy of even high-affinity monovalent binders under non-equilibrium physiological conditions26.f We selected an immune stimulatory effector peptide based on an N-formyl methionine containing peptide employed to activate the innate immune system via granulocytes, monocytes and macrophages (denoted as F1 or F2, Fig. 1). Such peptides are typically of bacterial origin and are well-known to elicit an innate immune response by interaction with various immune cell receptors such as three members of the N-formyl-peptide receptor fam- ily (FPR1–3)27–29. Previous experiments in which a formyl-methionyl-leucyl-phenylalanine peptide (fMLF) was covalently linked to an IgG antibody induced monocyte chemotaxis30 and a two-fold increase in macrophage infiltration of hepatomas and a decrease in tumor weight in guinea pigs31. At the same time no relevant toxicity in a human phase I clinical study was observed32.hfi p y The selected binder peptide (denoted as B9, Fig. 1) binds with high affinity to the integrin α3 chain and was identified by one-bead-one-compound (OBOC) screening33. The eight-residue binder consists of several non-proteinogenic amino acids and has been evolved as a peptidomimetic, cyclized via two terminal cysteine residues, using the one-bead-one-compound approach34. Here, we describe the synthesis and activity of a prototypical ISEr abbreviated Y9 (Y denotes the trimeric functionality of the molecule) that targets specific cell surface structures (here the integrin α3 chain) as well as engage the innate immune system (with an N-formyl effector). In a proof-of-concept study, this ISEr effectively prevented tumor formation in an allogenic adenocarcinoma model in guinea pigs. Results ff (C) Y9 was used for all in vitro and in vivo experiments in its oxidized, disulfide cyclized form as confirmed by ESI-MS (+de-convoluted spectrum, calculated MW of disulfide-cyclized Y9 is 5268 Da) and NMR experiments. These experiments confirmed the correct assembly of Y9 and formation of both disulfide bridges. Formation of the disulfide bond between Cys 1 and Cys 8 in the binder peptide of Y9 results in a change in the chemical shifts of the Cys Cβ resonances. An overlay of the 1H-13C HSQC NMR spectra of reduced (grey contours) and oxidized (black contours) Y9 is shown with Cys Cβ peaks labeled. Additional information on chemical shifts can be found in Table S2. diffusion of the larger molecule and potential occlusion of F1 by the PEG chains. Additional assays indicated that F1 and F1-2PEG are less potent immune activators in mice and guinea pigs (Figure S2C,D)36. In order to incorporate an effector suitable for activation of innate immune responses in humans, mice and guinea pigs, we selected fMIFL (F2), a sequence previously identified from Staphylococcus aureus37–39. F2 induces chemotaxis and superoxide production of human neutrophils starting at 10 pM concentration (Fig. 2A,B). Here, we also observe a loss of potency by 1–2 orders of magnitude when linked to the PEG chains (F2-2PEG, Figs 2A,B, S3). However, sufficiently high concentrations of effector to induce an effective immune response could still be reached, as demonstrated by subcutaneous application of F2-2PEG in mice and guinea pigs. Analysis of tissue sections of mouse skin samples 24 h after subcutaneous injection of 100 nmol F2-2PEG clearly showed infiltration by immune cells at the injection site in comparison to control injected animals (Fig. 3C). We could demonstrate that 6 h incubation of PBMCs with 100 nM F2-2PEG increased the levels of IL-1β secreted from human mono- cytes (Fig. 3D) similarly to bacterial lipopolysaccharides (LPS). Other cytokine levels such as IL-6, IL-8 and TNFα were not significantly increased by F2-2PEG while the positive control used here (LPS), induced increases of these cytokines (Fig. 3D). Specific binding to cancer cells. Our selected binder B9 was prepared by Fmoc-based SPPS and carefully analyzed with respect to binding to a panel of human, mouse and guinea pig tumor and blood cells (Tables 1 and S1). For the detection by streptavidin in flow cytometry, B9 was C-terminally labeled with biotin (Fig. 4A). Results ff Effector activation of the immune system. Separate testing of the two components comprising the ISEr Y9, the effector (F2) and the binder moiety (B9), were necessary to select suitable building blocks as well as to learn what changes of the individual properties occur upon incorporation into an ISEr (Figs 1A and 2A). We initially started with testing the well-known N-formyl peptide effector fMLF (denoted as F1, Figure S1) at concentrations between 1–10 nM. fMLF was selected based on prior reports of its use for inducing chemot- axis and immune reactions when linked to antibodies30,32. Our experiments confirmed that this short, N-formyl peptide prompts chemotaxis of human leukocytes and superoxide production in human granulocytes (dihy- dro-rhodamine (DHR) oxidation assay, Figure S2A,B)35. However, incorporation of F1 into a suitable scaffold to attach two binder moieties, consisting of a short glycine-glycine linker and two additional lysine residues linked to PEG27 via their ε-amino groups (F1-2PEG, Figure S1), reduced the potency to induce chemotaxis and neutro- phil activation by two orders of magnitude (Figure S2A,B). This behavior is best explained by the much slower Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 2 www.nature.com/scientificreports/ Figure 2. Synthesis and analytical data for Y9 (A) SPPS of Y9 comprising the fMIFL effector (F2) and two integrin α3β1 binder moieties (B9). (B) Analytical data of purified Y9 (RP-HPLC, ESI-MS and de-convoluted mass spectrum, calculated MW of fully reduced Y9 is 5271.0 Da). (C) Y9 was used for all in vitro and in vivo experiments in its oxidized, disulfide cyclized form as confirmed by ESI-MS (+de-convoluted spectrum, calculated MW of disulfide-cyclized Y9 is 5268 Da) and NMR experiments. These experiments confirmed the correct assembly of Y9 and formation of both disulfide bridges. Formation of the disulfide bond between Cys 1 and Cys 8 in the binder peptide of Y9 results in a change in the chemical shifts of the Cys Cβ resonances. An overlay of the 1H-13C HSQC NMR spectra of reduced (grey contours) and oxidized (black contours) Y9 is shown with Cys Cβ peaks labeled. Additional information on chemical shifts can be found in Table S2. Figure 2. Synthesis and analytical data for Y9 (A) SPPS of Y9 comprising the fMIFL effector (F2) and two integrin α3β1 binder moieties (B9). (B) Analytical data of purified Y9 (RP-HPLC, ESI-MS and de-convoluted mass spectrum, calculated MW of fully reduced Y9 is 5271.0 Da). Results ff Concentration-dependent flow cytometry measurements using this B9-Biotin demonstrated high affinity bind- ing to A431 tumor cells expressing integrin α3β1 (also known as VLA-3) on their surface (KD ~90 nM, Fig. 3 and S3). B9 competes with a commercially available anti-integrin α3 antibody (anti-CD49c) for the same bind- ing site, thereby unequivocally establishing the target of this peptide binder (Fig. 4C). Binding affinity depends on the addition of MnCl2 (at 2 mM) during incubation as Mn2+ induces the high-affinity state of the integrin (Fig. 4B40,41). In the absence of Mn2+ the apparent KD drops to 176 nM. Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 3 www.nature.com/scientificreports/ Figure 3. Analysis of effector F2. (A) Chemotaxis of human leukocytes upon stimulation with different concentrations of (PEGylated) effector. (B) Oxidative burst of human, murine and guinea pig leukocytes upon stimulation with (PEGylated) effector. In panels A and B, data of three independent experiments are averaged. (C) Immune cell infiltration in mouse skin (HE-stained sections) 24 h after subcutaneous injection of PBS buffer (top) and 100 nmol F2-2PEG (bottom) into mouse skin. Infiltrating immune cells are visible as small dark dots (scale bar 100 µm). Black arrows indicate matrigel depots under the skin. (D) Cytokine release from human monocytes upon stimulation with F2-2PEG or LPS (averages of two independent experiments with monocytes from different donors). Figure 3. Analysis of effector F2. (A) Chemotaxis of human leukocytes upon stimulation with different concentrations of (PEGylated) effector. (B) Oxidative burst of human, murine and guinea pig leukocytes upon stimulation with (PEGylated) effector. In panels A and B, data of three independent experiments are averaged. (C) Immune cell infiltration in mouse skin (HE-stained sections) 24 h after subcutaneous injection of PBS buffer (top) and 100 nmol F2-2PEG (bottom) into mouse skin. Infiltrating immune cells are visible as small dark dots (scale bar 100 µm). Black arrows indicate matrigel depots under the skin. (D) Cytokine release from human monocytes upon stimulation with F2-2PEG or LPS (averages of two independent experiments with monocytes from different donors). Additional flow cytometry analyses with a variety of primary cells such as macrophages, fibroblasts and endothelial cells (HUVECs) from humans, mice and guinea pigs were performed to exclude unspecific bind- ing of B9 (Tables 1 and S1). www.nature.com/scientificreports/ Flow cytometry based analysis of anti-CD49c, B9-Biotin and Biotin-Y9 binding to various human, mouse and guinea pig cell lines. ++ Strong binding, + medium binding, +/− weak binding, − no binding, nd not determined. Table 1. Flow cytometry based analysis of anti-CD49c, B9-Biotin and Biotin-Y9 binding to various human, mouse and guinea pig cell lines. ++ Strong binding, + medium binding, +/− weak binding, − no binding, nd not determined. Activation of the innate immune system by ISEr Y9. Y9 exhibits similar effector properties to F2-2PEG in chemotaxis and oxidative burst assays (Fig. 5A–C). Before using Y9 in any subsequent experiments we excluded endotoxin contamination using a commercially available test system based on the limulus amebocyte lysate (LAL) assay (Figure S6)44. In vivo experiments in mice and guinea pigs showed that local immune cell infiltration occurs upon subcutaneous injections (Figs 5C, S7). Immune cell infiltration into the injection site was most pronounced after the application of 100 nmol Y9 in guinea pigs. Immune infiltrate was also found (but was less intense) after the subcutaneous injection of 100 nmol Y9 into guinea pigs that were immune suppressed by four weekly injections of cyclophosphamide (CPA) as used in a subsequent efficacy study with the allogenic GPC-16 tumor model in guinea pigs (Figure S7A,B). To explore the possible use of the large variety of xenogeneic tumor models available in mice we also injected Y9 into immune-deficient Balb/cnu/nu mice. Here, dose-dependent increases in immune infiltrates between 200 nmol to 500 nmol could be observed (Fig. 5C). This finding indicates that higher dosages of Y9 are required for inducing an immune response in immune-deficient Balb/cnu/nu mice than for immune-suppressed guinea pigs. Immunohistochemical staining of the skin sections with antibodies against the myeloid cell marker myeloperoxidase (MPO) and the macrophage marker F4/80 demonstrated that the immune infiltrate consisted, at least in part, of granulocytes, monocytes and macrophages (Figure S7C–F). Specific integrin binding of ISEr Y9. Binding to integrin α3 was shown in vitro via surface plasmon res- onance (SPR) measurement using immobilized integrin α3. Here, a KD value of 60 nM for Y9 was determined (Fig. 5D). This value is in the same range as found for B9-Biotin in previous flow cytometry measurements of the binder moiety alone and agrees with measurements of biotinylated Y9 on cells. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Species Cells and cell lines CD49c B9 Y9 Human A431 ++ ++ ++ PC-3 ++ ++ ++ K562 − − − HUVEC + + ++ foreskin fibroblasts nd − nd primary lymphocytes − − − primary monocytes − − − primary granulocytes − − − primary monocytes 2 days − − nd primary macrophages 7 days +/− +/− +/− Mouse NIH-3T3 nd + + J774A.1 nd − nd 3LL-R nd + nd Hepa1–6 nd + nd primary hepatocytes nd − nd primary leukocytes nd nd − primary macrophages 7d nd − − Guinea pig GPC-16 ++ ++ ++ 104C1 ++ ++ nd JH4 clone 1 ++ ++ nd Table 1. Flow cytometry based analysis of anti-CD49c, B9-Biotin and Biotin-Y9 binding to various human, mouse and guinea pig cell lines. ++ Strong binding, + medium binding, +/− weak binding, − no binding, nd not determined. Species Cells and cell lines CD49c B9 Y9 Human A431 ++ ++ ++ PC-3 ++ ++ ++ K562 − − − HUVEC + + ++ foreskin fibroblasts nd − nd primary lymphocytes − − − primary monocytes − − − primary granulocytes − − − primary monocytes 2 days − − nd primary macrophages 7 days +/− +/− +/− Mouse NIH-3T3 nd + + J774A.1 nd − nd 3LL-R nd + nd Hepa1–6 nd + nd primary hepatocytes nd − nd primary leukocytes nd nd − primary macrophages 7d nd − − Guinea pig GPC-16 ++ ++ ++ 104C1 ++ ++ nd JH4 clone 1 ++ ++ nd Table 1. Flow cytometry based analysis of anti-CD49c, B9-Biotin and Biotin-Y9 binding to various human, mouse and guinea pig cell lines. ++ Strong binding, + medium binding, +/− weak binding, − no binding, nd not determined. Species Cells and cell lines CD49c B9 Y9 Human A431 ++ ++ ++ PC-3 ++ ++ ++ K562 − − − HUVEC + + ++ foreskin fibroblasts nd − nd primary lymphocytes − − − primary monocytes − − − primary granulocytes − − − primary monocytes 2 days − − nd primary macrophages 7 days +/− +/− +/− Mouse NIH-3T3 nd + + J774A.1 nd − nd 3LL-R nd + nd Hepa1–6 nd + nd primary hepatocytes nd − nd primary leukocytes nd nd − primary macrophages 7d nd − − Guinea pig GPC-16 ++ ++ ++ 104C1 ++ ++ nd JH4 clone 1 ++ ++ nd Table 1. Results ff No binding at concentrations below 1 µM was detected on most of the primary cell types, except for a weak, heterogeneous binding signal observed on human macrophages and HUVEC. Expression of integrin α3 on human, rabbit and guinea pig cells was verified with a mouse anti-human CD49c antibody and correlated with binding of B9 to these cell lines (Table 1). The CD49c specific antibody did not recognize mouse integrin α3, however binding of B9 to most murine tumor cell lines could be demonstrated (Tables 1 and S1). Synthesis and characterization of ISEr Y9. The ISEr Y9 combines F2 and B9, and was synthe- sized by Fmoc SPPS. N-terminal formylation of MIFLGGKK on resin was quantitatively achieved with p-nitrophenylformate42 and followed by Mtt removal from both lysine side chains onto which two PEG27 link- ers were coupled prior to generating identical binder peptides, also via stepwise SPPS. The complete molecule, denoted as Y9, was obtained in 20% yield based on the synthesis scale of 0.2 mmol and purified product (207 mg). Purification via RP-HPLC of reduced Y9 (Fig. 2B) was followed by oxidative cyclization of both binder moieties to give active Y9 with more than 98% purity (Fig. 2C). To characterize Y9 structurally, 2D NMR data were recorded in aqueous solution at pH 3.0 and chemical shifts were assigned for the reduced form of Y9 (Table S2). Most of the secondary Hα shifts of both the binder and effector peptides in Y9 were smaller than 0.1 ppm, showing that both peptides are in a predominantly random coil conformation (Figure S543). NMR data of the oxidized form of Y9 were compared with that of reduced Y9, confirming (in agreement with MS data) the presence of the disulfide bond. Most notably, the Cβ resonance of the cysteine residues in B9 shifted from ~28 ppm in reduced to ~41 ppm in oxidized Y9 (Fig. 2C). Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 4 www.nature.com/scientificreports/ Based on this data, we estimated a clearance half-life of ~2 hours for Y9. In all assays, no major degradation product was found that would point to design weaknesses. Stability and availability of ISEr Y9. The amount of effector available to attract and stimulate immune cells upon administering Y9 is not only affected by internalization and dissociation but also by its degrada- tion. To this end, we have tested stability of Y9 in mouse serum using extraction and HPLC-based quantifica- tion (Figure S11). Y9 remains largely intact for extended periods of time (>24 h) by combining the protective properties of the PEG chains against proteolysis as well as by the presence of several non-proteinogenic amino acids in the binder peptides and their cyclic structure (Fig. 2A). After 48 hours only 20% of Y9 were degraded. Bioavailability of Y9 in vivo was tested in sera taken from mice after subcutaneous application of Y9, using a sen- sitive dot-blot method and a custom-made antibody against B9 (detection limit ~1 nM, Figure S12). Upon local subcutaneous injection of 200 or 500 nmol of Y9, the binder peptide was still detected after 24 h. For injection of 200 nm Y9 additional measurements of serum concentrations were carried out after 1 h and 7 h showing serum concentrations of Y9 of 350 and 75 nM, respectively. Based on this data, we estimated a clearance half-life of ~2 hours for Y9. In all assays, no major degradation product was found that would point to design weaknesses. Effect of ISEr Y9 on tumor formation in guinea pigs. Antitumor efficacy of Y9 was demonstrated by administering a single dose of 200 nmol (~1 mg) together with GPC-16 tumor cells subcutaneously into guinea pigs. Based on our in vitro data and previous studies indicating successful stimulation of the innate immune system by fMLF-conjugated IgG, guinea pigs are a suitable animal model to test efficacy of Y932,45. We have estab- lished a guinea pig tumor model using guinea pig-derived colorectal adenocarcinoma cells (GPC-1646) injected subcutaneously into Dunkin Hartley guinea pigs. Since the genetic background of the guinea pig strain from which the GPC-16 tumor originated was unknown and due to the limited availability of specific-pathogen-free (SPF)-guinea pig strains, we established the tumor growth under immunosuppression with cyclophosphamide (CPA)47. www.nature.com/scientificreports/ To determine binding constants on integrin α3β1 expressing cells, we used biotinylated Y9 (biotin attached to the N-terminus of the effector) and determined an apparent KD of 244 nM for binding to A431 cells (Fig. 5E). This three-fold decrease in affinity when compared to B9-Biotin could be due to an occluding effect of the PEG spacers. To test this hypothesis we used our scaffold (fMIFLGGKK with two PEG27 chains) just carrying one B9 binder peptide. B9 competed with 200 nM B9-Biotin at an IC50 of 318 nM whereas the scaffold with only one B9 binder achieved the same effect only at 1533 nM. This confirmed our hypothesis that the PEG spacers have a negative impact on binding affinity. In competition experiments Y9 displaced B9-Biotin at a four-fold lower concentration than the monovalent binder (380 nM), which also demonstrated the positive effect of bivalency in Y9 (Figure S8). Further analysis of binding of B9 and Y9 under physiological conditions (37 °C) revealed that Y9 remains present on the surface of various cancer cell lines longer than B9 (Figure S9). B9-Biotin bound to PC-3 or A431 cells disappeared from cells quickly when incubated at 37 °C and no signal above background was detected by streptavidin staining after 5 min. However, biotinylated Y9 remained on cell membranes with t1/2 of at least 15 minutes (Figure S9). To exclude contributions by the effector F2 to binding of Y9 to A431 cells, staining with a formyl peptide receptor 1 (FPR-1)-specific antibody confirmed that no FPR-1 was present on these cells (Figure S10). Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 5 www.nature.com/scientificreports/ Figure 4. Analysis of binder B9. (A) Structure of B9-Biotin. (B) Concentration-dependent binding of B9- Biotin to A431 (epidermoid carcinoma cells) expressing integrin α3β1 in the presence and absence of Mn2+ to induce the high affinity state of the integrin. Data was fitted based on a one site-specific binding model and gave apparent KD values of 90 nM in the presence of Mn2+ and 176 nM in its absence. (C) Displacement of an anti-CD49c-PE conjugate by increasing concentrations of B9. No displacement of the antibody occurred with scrambled binder (sc-B9). An isotype control with mlgG1-PE was used to determine background signal. Figure 4. Analysis of binder B9. (A) Structure of B9-Biotin. www.nature.com/scientificreports/ (B) Concentration-dependent binding of B9- Biotin to A431 (epidermoid carcinoma cells) expressing integrin α3β1 in the presence and absence of Mn2+ to induce the high affinity state of the integrin. Data was fitted based on a one site-specific binding model and gave apparent KD values of 90 nM in the presence of Mn2+ and 176 nM in its absence. (C) Displacement of an anti-CD49c-PE conjugate by increasing concentrations of B9. No displacement of the antibody occurred with scrambled binder (sc-B9). An isotype control with mlgG1-PE was used to determine background signal. Figure 4. Analysis of binder B9. (A) Structure of B9-Biotin. (B) Concentration-dependent binding of B9- g y ( ) ( ) p g Biotin to A431 (epidermoid carcinoma cells) expressing integrin α3β1 in the presence and absence of Mn2+ to induce the high affinity state of the integrin. Data was fitted based on a one site-specific binding model and gave apparent KD values of 90 nM in the presence of Mn2+ and 176 nM in its absence. (C) Displacement of an anti-CD49c-PE conjugate by increasing concentrations of B9. No displacement of the antibody occurred with scrambled binder (sc-B9). An isotype control with mlgG1-PE was used to determine background signal. Stability and availability of ISEr Y9. The amount of effector available to attract and stimulate immune cells upon administering Y9 is not only affected by internalization and dissociation but also by its degrada- tion. To this end, we have tested stability of Y9 in mouse serum using extraction and HPLC-based quantifica- tion (Figure S11). Y9 remains largely intact for extended periods of time (>24 h) by combining the protective properties of the PEG chains against proteolysis as well as by the presence of several non-proteinogenic amino acids in the binder peptides and their cyclic structure (Fig. 2A). After 48 hours only 20% of Y9 were degraded. Bioavailability of Y9 in vivo was tested in sera taken from mice after subcutaneous application of Y9, using a sen- sitive dot-blot method and a custom-made antibody against B9 (detection limit ~1 nM, Figure S12). Upon local subcutaneous injection of 200 or 500 nmol of Y9, the binder peptide was still detected after 24 h. For injection of 200 nm Y9 additional measurements of serum concentrations were carried out after 1 h and 7 h showing serum concentrations of Y9 of 350 and 75 nM, respectively. www.nature.com/scientificreports/ Final efficacy studies were then performed under CPA treatment (200 mg/kg body weight) one day before tumor inoculation and thereafter once per week for 4 weeks. Two independent series of experiments showed that only in 3 out of 10 Y9-treated animals tumors were present after 35 days, whereas in the control group (GPC-16 cells in buffer) 8 out of 9 animals showed solid tumor formation after 35 days (Fig. 6) as verified by macroscopic and microscopic observations. None of the Y9-treated guinea pigs showed any undesired effects besides reddening at the injection site. Safety of ISEr Y9 in mice. No effect of Y9, at concentrations of up to 10 µM, on the viability and proliferation of different cancer cell lines was observed in MTT assays (Figure S13). To further confirm that Y9 can be safely applied in animals, we performed initial dose escalation studies testing acute toxicity after repeated application of Y9 in immune-competent Balb/c mice. Five animals per group were subcutaneously injected with 200, 400 or 800 nmol Y9 every two days for 14 days. A control group of ten mice received 0.9% NaCl. Four days after the last application of Y9, blood and serum parameters as well as relative organ weights were measured and tissue samples of organs were analyzed by histopathology. As expected Y9-treated animals exhibited local effects such as increased neutrophil and monocyte infiltrates at the injection site and, correlating with this observation, a mild but significant dose-dependent immune infiltration and lymphoid hyperplasia in the draining lymph nodes was found. No histopathological effects were detected in any of the organs tested (liver, kidney, pancreas, spleen, heart, lung, ovaries, brain, stomach, intes- tine, mesentery) after repeated 200 and 400 nmol Y9 injections (Figure S14). Only with the highest dose of 800 nmol moderate degenerative changes of the kidney, including distended tubules and protein casts, were detected. Measuring relative organ weights, a slightly enlarged liver weight was observed. However, no histopathological abnormalities or steatosis was detected in the liver (Figure S15). www.nature.com/scientificreports/ The analysis of serum and blood samples after the application of the highest dose of 800 nm Y9 revealed slightly but significant increased lipase and creatinine (CREA) levels and Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 6 www.nature.com/scientificreports/ significantly increased amylase, triglycerides, cholesterol and blood urea nitrogen (BUN) levels in serum (Figure S15 Hematological analysis including a differential hemogram of the white blood cells showed no significant change Figure 5. In vitro and in vivo analysis of Y9. (A) Chemotaxis of human and murine leukocytes upon stimulation wit different concentrations of Y9. (B) Oxidative burst of human, mouse and guinea pig leukocytes upon stimulation with Y9. (C) Immune cell infiltration in nude mice and guinea pig skin (HE stained sections, scale bars 100 µm) 24 h after s.c. injection of 100 nmol (guinea pig) or 500 nmol Y9 (Balb/cnu/nu mice), respectively. (D) SPR measurement of Y9 binding to immobilized integrin α3β1 (VLA-3). (E) Concentration-dependent binding of biotinylated Y9 to A43 cells in the presence or absence of MnCl2 (detected by incubation with PerCP-Cy5.5 labeled streptavidin). Figure 5. In vitro and in vivo analysis of Y9. (A) Chemotaxis of human and murine leukocytes upon stimulation w Figure 5. In vitro and in vivo analysis of Y9. (A) Chemotaxis of human and murine leukocytes upon stimulation with different concentrations of Y9. (B) Oxidative burst of human, mouse and guinea pig leukocytes upon stimulation with Y9. (C) Immune cell infiltration in nude mice and guinea pig skin (HE stained sections, scale bars 100 µm) 24 h after s.c. injection of 100 nmol (guinea pig) or 500 nmol Y9 (Balb/cnu/nu mice), respectively. (D) SPR measurement of Y9 binding to immobilized integrin α3β1 (VLA-3). (E) Concentration-dependent binding of biotinylated Y9 to A431 cells in the presence or absence of MnCl2 (detected by incubation with PerCP-Cy5.5 labeled streptavidin). significantly increased amylase, triglycerides, cholesterol and blood urea nitrogen (BUN) levels in serum (Figure S15). Hematological analysis including a differential hemogram of the white blood cells showed no significant changes. Overall, these results do not indicate any significant systemic toxic effect at these effective dosages. Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 7 www.nature.com/scientificreports/ Figure 6. Efficacy of Y9 tested in a GPC-16 tumor model in guinea pigs (A) Measurement of tumor size in GPC-16 inoculated, CPA-treated guinea pigs over 35 days. Discussion l Conventional antibodies, fragments thereof and other protein-based binders address a molecular weight range above 25 kDa, in contrast to small molecule drugs that are in the range of 300 to 1000 Da, which leaves the intermediate mass range open for new concepts such as the ISEr here. The trivalent ISEr Y9 provides a purely synthetic alternative to conventional antibodies and related biological techniques. It offers a versatile synthesis route to medium sized drug molecules (~5 kDa) that combine bivalent, specific target binding with activation of the innate immune system. Target specificity is achieved by attaching two binders (B9) to a generic scaffold carrying fMIFL (F2) as an effector for activating formyl peptide receptors (FPRs). This effector induces chem- otaxis and activation of human and guinea pig leukocytes in vitro and in vivo without undesired side effects.i f Formyl-peptides can also induce the expression of IL-1β, TNFα and IL-848, but we did not observe a significant induction of the latter two cytokines by F2. Our data thus indicates that treatment of tumors with F2-2PEG does not contribute to the expression of these potentially tumor-promoting cytokines49. Furthermore, the flexible synthesis route allows for the adaption of effectors to specific requirements with respect to immune stimulation and species.h pf pi q p p The design of Y9 allows access to more than one binding site on a cellular target such as in dimeric or clustered receptors, thereby improving binding to the targeted cells beyond simply increasing the local concentration of a receptor ligand towards more complex avidity effects. Such effects will help to accumulate Y9 at a concentration in tumor tissue sufficient to activate macrophages. Bispecific or multispecific molecules targeting two or more different cell-surface receptors can also be easily envisioned based on this design and synthesis scheme. Such molecules will combine the ease of chemical synthesis with well-known advantages of bi- or multispecific binders, as demonstrated for approved bispecific antibodies such as Catumaxomab50. The PEG chains reduced the affinity of Y9 for integrin α3-expressing A431 cells. Nevertheless, efficacy could still be demonstrated in vivo, possibly due to the role of PEG in protecting against biodegradation, improving solubility, and promoting the formation of monodisperse molecules in aqueous solution25. Y9 reduced the establishment of a GPC-16 based tumor in guinea pigs by 67% upon a single dose treat- ment during injection of tumor cells. www.nature.com/scientificreports/ Animals receiving 1.6–2 × 107 GPC-16 cells mixed with 200 nmol of Y9 and Matrigel before injection are compared with animals receiving GPC-16 cells premixed only with PBS and Matrigel as a control. With Y9 treatment only 3 out of 10 guinea pigs carried verifiable tumors whereas in the untreated control group 8 out of 9 animals carried tumors (statistics are based on t-test for the upper graph and on a Wilcoxon-test for the engraftment comparison). (B) Tumor growth at day 35 after inoculation in guinea pig skin (HE-sections) treated with PBS control (top) or Y9 (bottom); 50x magnification. Figure 6. Efficacy of Y9 tested in a GPC-16 tumor model in guinea pigs (A) Measurement of tumor size in GPC-16 inoculated, CPA-treated guinea pigs over 35 days. Animals receiving 1.6–2 × 107 GPC-16 cells mixed with 200 nmol of Y9 and Matrigel before injection are compared with animals receiving GPC-16 cells premixed only with PBS and Matrigel as a control. With Y9 treatment only 3 out of 10 guinea pigs carried verifiable tumors whereas in the untreated control group 8 out of 9 animals carried tumors (statistics are based on t-test for the upper graph and on a Wilcoxon-test for the engraftment comparison). (B) Tumor growth at day 35 after inoculation in guinea pig skin (HE-sections) treated with PBS control (top) or Y9 (bottom); 50x magnification. Online Methods f ff Synthesis of effector and binder peptides. Peptides were obtained via standard Fmoc SPPS. In brief, the first amino acid was DIC-activated and subsequently coupled to Wang resin (100–200 mesh, 0.9 mmol/g sub- stitution level) in the presence of DMAP as a catalyst. All following couplings were performed using HBTU as activator. To check for complete coupling, a Ninhydrin test was conducted. Fmoc was removed using 20% piperi- dine in DMF. Formylation was achieved using 3 eq p-nitrophenylformate in DMF for 3 h. The peptide was cleaved using TFA/TIS/H2O/DMS 92.5:2.5:2.5:2.5 (v/v) for 2 h, following peptide precipitation with cold diethyl ether and re-solubilization in ACN/H2O 1:1 (v/v) with 0.1% TFA. After freeze drying, peptides were dissolved in 6 M GndHCl, pH 4.7 and purified via RP-HPLC. Synthesis of Y9. Removal of the Mtt protecting group was achieved by flow washing the peptidyl-resin with DCM/TFA/TIS 98:1:1 (v/v) until the solution turned colorless. Each lysine side chain was PEGylated overnight using 1.12 equivalents of Fmoc-NH-(PEG)27-COOH and HATU as activator. Subsequently, the synthesis was continued using standard SPPS methods as described above. Surface plasmon resonance (SPR). Single cycle kinetic experiments were conducted on a Biacore 3000 system using a CM5 sensor chip at 25 and 37 °C with a flow rate of 10 µl/min. After coupling of the VLA-3 recep- tor (R&D Systems) via standard EDC/NHS immobilization chemistry (7 min activation with EDC/NHS followed by a 7 min flush with VLA-3) in 10 mM NaOAc buffer at pH 6 containing 1 mM MnCl2, a 1 mM solution of eth- anolamine was injected for 7 min to block all amine reactive sites. Subsequently the analyte was flushed over the chip surface at a flow rate of 10 µl/min in HEPES buffered saline supplemented with 1 mM MnCl2. As no method was found to regenerate the surface from Y9 without damaging the VLA-3, Y9 was injected in increasing concen- trations, taking care not to reach saturated binding (single cycle kinetics). Animals. All experiments were conducted in accordance with the German legislation governing animal studies. The Principles of Laboratory Animal Care (Guide for the Care and Use of Laboratory Animals: Eighth Edition. Washington, DC: The National Academies Press, 2011) were followed. The animal protocol was approved by the Governmental Animal Care and Use Committee (LANUV AZ. 87-51.04.2010.A278). All experiments were performed in the Institute for Laboratory Animal Science, a DIN ISO 9001/2008 certified facility. Human cells. Online Methods f ff Primary cells were obtained from healthy volunteers that served as donors at our blood bank, after written informed consent and approval by the local ethics committee of the Uniklinik RWTH Aachen, Germany. y For more information on experimental details about NMR measurements, binding and stability assays as well as on animal testing, please see Supplementary information. Discussion l As no direct toxic effects of Y9 on cancer cells were observed in vitro, we conclude that activation of immune cells interfered with tumor establishment. Clinical use of such a treatment can be envisioned in preventing tumor relapse or development of metastases after first line treatments such as surgery, chemotherapy and/or radiation. As the simultaneous injection of tumor cells and Y9 mimics a local administration of the drug, as for instance anticipated during chemoembolization or peri-surgical application, other therapeutic applications are possible as well. Considering that Y9 was still effective in cyclophosphamide treated guinea pigs, it might provide an effective treatment in combination therapies. No severe toxic effects other than the expected moderate local reactions to Y9 were observed in guinea pigs and mice. Only repetitive doses four-fold higher than used in the effective treatment led to mild pathological changes in kidneys and some serum and blood parameters. The moderate degenerative changes of the kidney including distended tubules and protein casts in the highest dosage group indicate renal clearance of Y9. In general, Y9 can be safely used within a dose range of 100 to 400 nmol (0.5–2 mg). Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 8 www.nature.com/scientificreports/ Overall, Y9 is an alternative to conventional antibody-based therapeutics and related protein-based approaches due to its entirely flexible and robust chemical synthesis. ISErs can be generated quickly, providing access to customized drugs targeting specific receptors not only relevant for cancer therapy but also for other disease areas requiring modulation of innate immune responses at distinct target sites such as inflammatory and autoimmune diseases. References y J p y py 2. Scott, A. M., Wolchok, J. D. & Old, L. J. Antibody therapy of cancer. Nat. Rev. Cancer 12, 278–87 (2012).h 3. Chames, P., Van Regenmortel, M., Weiss, E. & Baty, D. Therapeutic antibodies: successes, limitations and hopes for the future. Br. J Pharmacol. 157, 220–33 (2009). 4. Glassman, P. M. & Balthasar, J. P. Mechanistic considerations for the use of monoclonal antibodies for cancer therapy. Cancer Biol Med. 11, 20–33 (2014). 5. von Mehren, M., Adams, G. P. & Weiner, L. M. Monoclonal ant 6. Iannello, A. & Ahmad, A. Role of antibody-dependent cell-mediated cytotoxicity in the efficacy of therapeutic anti-cancer monoclonal antibodies. Cancer Metastasis Rev. 24, 487–99 (2005). 7. Dhodapkar, K. M. et al. Selective blockade of inhibitory Fc gamma receptor enables human dendritic cell maturation with IL production and immunity to antibody-coated tumor cells. Proc. Natl. Acad. Sci. USA 102, 2910–2915 (2005). p y y 8. Boder, E. T. & Jiang, W. Engineering antibodies for cancer therapy. Annu. Rev. Chem. Biomol. Eng. 2, 53–75 (2011). g g g y udson, P. J. Engineered antibody fragments and the rise of single do lliger, P. & Hudson, P. J. Engineered antibody fragments and the ris 10. Carter, P. J. Potent antibody therapeutics by design. Nat. Rev. Immunol. 6, 343–57 (2006).i 11. Doppalapudi, V. R. et al. Chemical generation of bispecific antibodies. Proc. Natl. Acad. Sci. USA 107, i an, L. Next generation of antibody therapy for cancer. Chin. J. Canc 13. Fischer, N. & Leger, O. Bispecific antibodies: molecules that enable novel therapeutic strategies. Pathobiology 74, 3–14 (2007). fbl l ffib d l l d f h d d b h l l l i 14. Lofblom, J. et al. Affibody molecules: engineered proteins for therapeutic, diagnostic and biotechnological applications. FEBS Lett. 584, 2670–80 (2010). 15. Baum, R. P. et al. Molecular imaging of HER2-expressing malignant tumors in breast cancer patients using synthetic 111In- or 68Ga-labeled affibody molecules. J. Nucl. Med. 51, 892–7 (2010). fi y ( ) 6. Skerra, A. Alternative binding proteins: anticalins - harnessing the structural plasticity of the lipocalin ligand pocket to engineer novel binding activities. FEBS J. 275, 2677–83 (2008). g J ( ) 17. Banta, S., Dooley, K. & Shur, O. Replacing antibodies: engineering new binding proteins. Annu. Rev. Biomed. Eng. 15, 93–113 (2013). Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 , , pi g g g py , ( ) 22. McEnaney, P. J., Parker, C. G., Zhang, A. X. & Spiegel, D. A. Antibody-recruiting molecules: an emerging paradigm for e immune function in treating human disease. ACS Chem. Biol. 7, 1139–51 (2012). References Biol Chem. 277, 19800–5 (2002). 1. Chen, J., Salas, A. & Springer, T. A. Bistable regulation of integrin adhesiveness by a bipolar metal ion cluster. Nat. Struct. Biol. 10 995–1001 (2003). 42. Martinez, J. & Laur, J. Active Esters of Formic-Acid as Useful Formylating Agents - Improvements in the Synthesis of Formyl-Amino Acid-Esters, N-Alpha-Formyl-Met-Leu-Phe-Oh, and Formyl-Met-Lys-Pro-Arg, a Phagocytosis Stimulating Peptide. Synthesis- Stuttgart, 979–981 (1982).t g ( ) 43. Wishart, D. S., Bigam, C. G., Holm, A., Hodges, R. S. & Sykes, B. D. 1H, 13C and 15N random coil NMR chemical shifts of the common amino acids. I. Investigations of nearest-neighbor effects. J. Biomol. NMR 5, 67–81 (1995). f 44. Iwanaga, S. Biochemical principle of Limulus test for detecting bacterial endotoxins. Proc. Jpn. Acad. Ser. B Phys. Biol. Sci. 83, 1 (2007). ( ) 45. Sandberg, A. L., Obrist, R. & Mergenhagen, S. E. In vitro and in vivo effects of anti-tumor antibody covalently coupled to a chemotactic peptide. Agents Actions Suppl. 12, 234–50 (1983). g pp 6. O’Donnell, R. W. & Cockerell, G. L. Establishment and biological properties of a guinea pig colonic adenocarcinoma cell line induced by N-methyl-N-nitrosourea. Cancer Res. 41, 2372–7 (1981).t 7. Wu, Y. J. et al. Cyclophosphamide enhances human tumor growth in nude rat xenografted tumor models. Neoplasia 11, 187–95 (2009). 48. Chen, L. Y. et al. Synergistic Induction of Inflammation by Bacterial Products Lipopolysaccharide and fMLP: An Important Microbial Pathogenic Mechanism. J. Immunol. 182, 2518–2524 (2009). g 9. Korkaya, H., Liu, S. & Wicha, M. S. Breast cancer stem cells, cytokine networks, and the tumor microenvironment. J. Clin. Invest 121 3804–9 (2011).h 50. Heiss, M. M. et al. The trifunctional antibody catumaxomab for the treatment of malignant ascites due to epithelial cancer: results of a prospective randomized phase II/IIItrial. Int. J. Canc. 127, 2209–2221 (2010). Author Contributions C.F.W.B., F.T. and A.P. designed research; M.Br., K.T., A.C.C., J.Sc. and A.P. performed research; J.St. designed and carried out animal studies; M.Ba. and F.T. contributed analytic tools; M.Br., K.T., A.C.C., J.Sc., J.St., A.P. and C.F.W.B. analyzed data; M.Br., K.T., A.C.C., J.Sc., F.T., J.S., A.P. and C.F.W.B. wrote the paper. References & Boulay, F. The N-formyl peptide receptors and the anaphylatoxin C5a receptors: An overview. Biochimie 89 1089–1106 (2007). 29. Showell, H. J. et al. Structure-Activity Relations of Synthetic Peptides as Chemotactic Factors and Inducers of Lysosomal Enzyme- Secretion for Neutrophils. J. Exp. Med. 143, 1154–1169 (1976). p p 30. Obrist, R. et al. Monocyte chemotaxis mediated by formyl-methionyl-leucyl-phenylalanine conjugated with monoclonal antibodies against human ovarian carcinoma. Int. J. Immunopharmacol. 5, 307–14 (1983). man ovarian carcinoma. Int. J. Immunopharmacol. 5, 307–14 (1983) g p 1. Obrist, R. & Sandberg, A. L. Enhancement of Macrophage Invasion of Tumors by Administration of Chemotactic Factor Anti Tumor Antibody Conjugates. Cell. Immunol. 81, 169–174 (1983). Tumor Antibody Conjugates. Cell. Immunol. 81, 169–174 (1983 y j g 32. Obrist, R., Schmidli, J., Muller, R., Gallati, H. & Obrecht, J. P. Acute and Subacute Toxicity of Chemotactic Conjugates between Monoclonal-Antibody and Fmet-Leu-Phe in Humans - a Phase-I Clinical-Trial. Cancer Immunol. Immun. 32, 406–408 (1991). 33. Yao, N. et al. Structure-activity relationship studies of targeting ligands against breast cancer cells. J. Med. Chem. 52, 6744–51 (2009). 4. Yao, N. et al. Discovery of targeting ligands for breast cancer cells using the one-bead one-compound combinatorial method. J. Med Chem. 52, 126–33 (2009).l 5. Henderson, L. M. & Chappell, J. B. Dihydrorhodamine 123: a fluorescent probe for superoxide generation? Eur. J. Biochem. 217 973–80 (1993). 36. He, H. Q., Troksa, E. L., Caltabiano, G., Pardo, L. & Ye, R. D. Structural determinants for the interaction of formyl peptide receptor 2 with peptide ligands. J. Biol. Chem. 289, 2295–306 (2014). p p g 7. Rot, A., Henderson, L. E., Copeland, T. D. & Leonard, E. J. A Series of 6 Ligands for the Human Formyl PeptideReceptor Tetrapeptides with High Chemotactic Potency and Efficacy. Proc. Natl. Acad. Sci. USA 84, 7967–7971 (1987). b l h d d d f l d d l ll fi 38. Rabiet, M. J., Huet, E. & Boulay, F. Human mitochondria-derived N-formylated peptides are novel agonists equally active on FPR and FPRL1, while Listeria monocytogenes-derived peptides preferentially activate FPR. Eur. J. Immunol. 35, 2486–95 (2005). 39. Southgate, E. L. et al. Identification of formyl peptides from Listeria monocytogenes and Staphylococcus aureus as p chemoattractants for mouse neutrophils. J. Immunol. 181, 1429–37 (2008). p 0. Mould, A. P. et al. Integrin activation involves a conformational change in the alpha 1 helix of the beta subunit A-domain. J. Acknowledgements g We are grateful to R.S. Goody for his initial input and continuous support of this project and to M. Felkl, K. Bäuml, S. Gentz for help with synthesis of binder and effector peptides. M. Novak and M. Hüffel are acknowledged for technical assistance with animal studies as well as R. Tolba for scientific support of animal studies, M. Afify for help with the pathology, K. Rohr for assistance with in vitro binding and effector studies, M. Sack for performing SPR studies, J. Van Ginderachter and D. Laoui for performing binding studies on murine tumor cell lines. This work was supported by the Bundesministerium für Bildung und Forschung (BMBF, KMU-innovativ-6, Förderkennzeichen 0315918) and by the European Union and NRW government (NRW-EU Ziel 2, (EFRE), 005- 0908-0112). The research leading to these results has received funding from the Mahlke-Obermann Stiftung and the European Union’s 7th FP under grant agreement no 609431. References 18 Sk A Alt ti tib d ff ld f l l iti C Opi Bi t h l 18 295 304 (2007) g 17. Banta, S., Dooley, K. & Shur, O. Replacing antibodies: engineering new binding proteins. Annu. Rev. Biomed. Eng. 15, 93–113 (2013). 18. Skerra, A. Alternative non-antibody scaffolds for molecular recognition. Curr. Opin. Biotechnol. 18, 295–304 (2007). g 17. Banta, S., Dooley, K. & Shur, O. Replacing antibodies: engineerif f 19. Oliveira, S., Heukers, R., Sornkom, J., Kok, R. J. & van Bergen En Henegouwen, P. M. Targeting tumors with nanobodies for cancer imaging and therapy. J. Control. Release 172, 607–17 (2013). g g py ( ) 0. Nagorsen, D. & Baeuerle, P. A. Immunomodulatory therapy of cancer with T cell-engaging BiTE antibody blinatumomab. Exp. Cel Res. 317, 1255–60 (2011).i 21. Baeuerle, P. A. & Reinhardt, C. Bispecific T-cell engaging antibodies for cancer therapy. Cancer Res. 69, 4941–4 (2009). 21. Baeuerle, P. A. & Reinhardt, C. Bispecific T-cell engaging antib 21. Baeuerle, P. A. & Reinhardt, C. Bispecific T-cell engaging antibodies for cancer therapy. Cancer Res. 69, 4941–4 (2009). 22. McEnaney, P. J., Parker, C. G., Zhang, A. X. & Spiegel, D. A. Antibody-recruiting molecules: an emerging paradigm for immune function in treating human disease. ACS Chem. Biol. 7, 1139–51 (2012). Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 9 www.nature.com/scientificreports/ 23. Rader, C. et al. Chemically programmed monoclonal antibodies for cancer therapy: adaptor immunotherapy based on a cov antibody catalyst. Proc. Natl. Acad. Sci. USA 100, 5396–400 (2003).f , y p g py p py antibody catalyst. Proc. Natl. Acad. Sci. USA 100, 5396–400 (2003). 24. McEnaney, P. J. et al. Chemically synthesized molecules with the targeting and effector functions of antibodies. J. Am. Chem. Soc. y y 24. McEnaney, P. J. et al. Chemically synthesized molecules with the targeting and effector functions of antibodies. J. Am. Chem 136, 18034–43 (2014). ( ) ermanson, G. T. Bioconjugation Techniques, (Academic Press, London 2008). 25. Hermanson, G. T. Bioconjugation Techniques, (Academic Press, London 2008). 6. Deyev, S. M. & Lebedenko, E. N. Multivalency: the hallmark of antibodies used for optimization of tumor targeting by design Bioessays 30, 904–18 (2008). y 27. Migeotte, I., Communi, D. & Parmentier, M. Formyl peptide receptors: a promiscuous subfamily of G protein-coupled receptors controlling immune responses. Cytokine Growth Factor Rev. 17, 501–19 (2006).h g p y 8. Rabiet, M. J., Huet, E. Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Additional Information Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-017-17627-0. Competing Interests: C.F.W.B. is a co-founder and partner of Syntab Therapeutics GmbH. Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 10 www.nature.com/scientificreports/ Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. fi Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2017 Scientific RePOrTS \| 7: 17592 \| DOI:10.1038/s41598-017-17627-0 11
https://openalex.org/W2166193955	https://traumamanagement.biomedcentral.com/track/pdf/10.1186/s13032-015-0023-4	English	null	Effects of a formalized collaboration between plastic and orthopedic surgeons in severe extremity trauma patients; a retrospective study	Journal of trauma management & outcomes	2,015	cc-by	7,462	© 2015 Sommar et al.; licensee BioMed Central. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Open Access Open Access Abstract Background: Severe trauma to the extremities often includes a combination of fractures and soft tissue injuries. Several publications support that the patient outcome is better when skeletal stabilization is followed by early soft-tissue coverage. In an effort to optimize the treatment of these patients, we established a formalized collaboration in 2008 between the Departments of reconstructive plastic surgery and orthopedics at the Karolinska University Hospital. Methods: A retrospective review was conducted for all patients who had suffered severe extremity trauma and received either a free or a pedicled flap for extremity reconstruction. We compared the management of patients 0–4 years before and 0–4 years after the collaboration started especially with respect to; choice of flap, time to flap coverage, number of operations/revisions, total in-hospital stay. Results: After initiation of the collaboration, the number of flaps increased from 13 flaps (5 free and 8 pedicled) to 44 flaps (21 free and 23 pedicled). Fewer postoperative revisions was seen, as well as shorter in-hospital stay. Conclusions: The present study highlights the importance of formalized collaboration between orthopedic and plastic surgeons in severe extremity trauma patients. The concept of an interdisciplinary approach has led to an increased number of trauma patients referred for plastic surgical consultation, an increased number of flaps, fewer postoperative revisions and shorter hospital stay. Keywords: Extremity trauma, Flaps, Orthoplastic collaboration Effects of a formalized collaboration between plastic and orthopedic surgeons in severe extremity trauma patients; a retrospective study Pehr Sommar1, Yamin Granberg2, Martin Halle1, Ann-Charlott Docherty Skogh1, Kalle T Lundgren1 and Karl-Åke Jansson3 Correspondence: pehr.sommar@karolinska.se 1Department of Molecular Medicine and Surgery, Section of Plastic Surgery, Karolinska Institutet and Karolinska University Hospital Stockholm, Stockholm, Sweden Full list of author information is available at the end of the article Sommar et al. Journal of Trauma Management & Outcomes (2015) 9:3 DOI 10.1186/s13032-015-0023-4 Sommar et al. Journal of Trauma Management & Outcomes (2015) 9:3 DOI 10.1186/s13032-015-0023-4 Introduction incidence rate, an orthoplastic approach has been sug- gested; i.e. a multidisciplinary collaboration involving both orthopedic and plastic surgery teams [7,8]. Severe extremity trauma is often associated with a com- bination of fractures and soft tissue injuries. It is gener- ally accepted that the outcome is better when skeletal stabilization is followed by early soft-tissue coverage [1-4]. Soft tissue flaps provide the protection and vascularization needed to resist infection and promote bone healing. Meling et al. reported the incidence of all open long bone fractures to be 13/105/year 2004–2007 [5]. Weiss et al. found a decreasing incidence of open tibial shaft fractures in the Swedish population, which was 2.3/105/year, 1998–2004 [6]. Because of the complexity of these fractures and the low Naique et al. demonstrated that severe open tibia frac- tures treated at dedicated trauma units with both ortho- pedic and plastic surgery services had lower complication rates and less need for revision surgery compared with those treated initially at hospitals without such combined services [8]. To ensure an orthoplastic approach, we formalized a collaboration between the Departments of reconstructive plastic surgery and orthopedics in April 2008 (Multidis- ciplinary collaboration MDC). The collaborative protocol agreed upon early attendance of plastic surgeons in ex- tremity trauma cases, preferably at the trauma room or at primary revisions of open fractures in order to expedite conjoint planning of further operations. An outpatient Page 2 of 8 Sommar et al. Journal of Trauma Management & Outcomes (2015) 9:3 chronic osteomyelitis, or patients referred to Karolinska University Hospital later than 14 days after the trauma were excluded from sub analysis. clinic with orthopedics and plastic surgeons for patients admitted from other hospitals was set up as well as multidisciplinary postoperative follow up of patients with trauma, and osteomyelitis following trauma to the extrem- ities. A contract was established to ensure equal contribu- tion of respective operating facilities and to ensure efficient rescheduling of elective surgery when emergent full day cases were to be planned with short notice. Material and methods All consecutive patients who had suffered extremity trauma and who received either a free or a pedicled flap for coverage of soft tissue defects, between April 2004 and April 2012, were included. A comparison was made between the management of patients 4 years before the formalized collaboration (21 April 2004 – 20 April 2008) and 4 years after the formalized collaboration (21 April 2008 – 20 April 2012). Only patients with acute extrem- ity trauma with exposed bone/open fracture or patients suffering from skin necrosis/infection after fracture sur- gery or chronic osteomyelitis following acute extremity trauma were selected for inclusion. Patients reconstructed after orthopedic tumor surgery or arthroplasty failures were not included. The included patients were identified by using the hospitals electronic operation planning system; Orbit (SYSteam Critical Care AB, Stockholm, Sweden). Patients were searched for using surgical pro- cedure codes for different types of free and pedicled flaps. Time to flap coverage was counted from trauma/ admission at Karolinska University Hospital, or in the secondary cases from occurrence of wounds, fistulation or exposed bone. In case of a failed flap with demand for a secondary flap, time to flap coverage was counted to the first flap. The time to complete soft tissue coverage after flap sur- gery was determined as no remaining skin wounds at clin- ical evaluation at follow up and accepted as stated in the respective patient charts. Fracture healing was determined by radiological and clinical evaluation retrieved from patient charts. The project was done in accordance with a protocol approved by the Ethical Committee at Karolinska Insti- tutet, Stockholm, Sweden. (Project 2013/307-31/2). The obtained results were statistically compared using a Mann–Whitney test in GraphPad Prism 5.0 (GraphPad Software Inc., CA). A p-value of ≤0.05 was considered statistically significant. Results were presented with median values. Introduction Operations/revisions included all surgery performed in an operating room on the patient due to the trauma and included; primary wound debridement, primary fixation with external fixation device, secondary fixation in those cases where it did not concur with flap surgery, wound re- visions/Topical negative pressure (TNP) therapy-change, reoperations due to flap failure, additional split skin graft- ing, extraction of external or internal fixation, bone graft- ing, conversion to circular external fixation (Taylor spatial frame™), corrections due to angulation/malrotation of fractures. TNP was only used as a dressing and not as a method to close the wound. To investigate the effects of this collaboration, a retro- spective review was conducted for all patients who had suffered severe extremity trauma and received either a free or a pedicled flap for coverage of soft tissue defects 0–4 years before and after the established collaboration. The main hypotheses were that this collaboration would decrease the time to soft tissue coverage as well as the in-hospital stay. The total in-hospital stay was calculated by adding all admissions for operations/revisions before and after flap coverage. Results In total, 52 patients (36 males, 16 females) treated with a free or pedicled flap were identified during the 8 year study period. One patient was included in both groups as she received one flap before and one after the start of the collaboration (53 cases). In total these patients re- ceived 57 flaps. 13 flaps were performed in the 4 years before the start of the collaboration, of which 5 were free flaps and 8 were pedicled flaps. 44 flaps were performed in the 4 years after the start of the collaboration, of which 21 were free flaps and 23 were pedicled flaps. Patients and choice of flaps is stated in Tables 1 and 2. All 5 free flaps performed before the collaboration survived. Total flap necrosis occurred in 2 out of 21 free flaps after the start of the collaboration. No partial necrosis was seen in free flaps. Total flap necrosis was seen in 1 pedicled flap before the start of the collaboration and 2 after the start of the collaboration. Partial flap necrosis was seen in 6 of the pedicled flaps (1 before and 5 after the start of the collab- oration). The median age at the time of flap reconstruc- tion before the start of the collaboration was 45 (21–64) Open fractures were classified according to Gustilo- Anderson [9]. Fractures were classified strictly after the status at admission. Fractures classified as GIIIA, which in secondary revisions developed into IIIB due to tissue necrosis or infection were not reclassified. In order to investigate the impact of the collaboration on time to flap surgery after admission to the hospital, and the impact on the postoperative treatment and in- hospital stay, only patients with acute extremity trauma in- cluding open fractures or soft tissue defects with exposed bone were selected for a sub analysis (Tables 1 and 2). Patients with skin necrosis/infection after fracture surgery, Sommar et al. Results Table 1 Patients treated before start of multidisciplinary collaboration Time to flap (d) Time to coverage (d) Patients treated before start of multidisciplinary collaboration with a free or pedicled flap for soft tissue reconstruction after lower extremity trauma. Patients excluded in the subgroup analysis of acute extremity trauma were patients recieving flaps due to skin necrosis/infection after fracture surgery, chronic osteomyelitis, or patients referred to Karolinska University Hospital later than 14 days after the trauma. Time to flap is counted from trauma/admission at Karolinska University Hospital, or in the secondary cases from occurrence of wounds, fistulation or exposed bone. Time to complete soft tissue coverage after flap surgery was determined as no remaining skin wounds at clinical evaluation at follow up. Patients treated before start of multidisciplinary collaboration with a free or pedicled flap for soft tissue reconstruction after lower extremity trauma. Patients excluded in the subgroup analysis of acute extremity trauma were patients recieving flaps due to skin necrosis/infection after fracture surgery, chronic osteomyelitis, or patients referred to Karolinska University Hospital later than 14 days after the trauma. Time to flap is counted from trauma/admission at Karolinska University Hospital, or in the secondary cases from occurrence of wounds, fistulation or exposed bone. Time to complete soft tissue coverage after flap surgery was determined as no remaining skin wounds at clinical evaluation at follow up. Most fractures were graded as GIIIB, but there were also severe soft tissue injuries with exposed bone without frac- tures. The median time to flap coverage of acute extremity trauma was 6 (0–51) days after the start of the collabor- ation, compared to 16.5 (0–220) days before the start of the collaboration (p = 0.283) (Figure 1). The median num- ber of revisions/operations prior to flap coverage of acute extremity trauma was 3 (0–7) after the start of the collab- oration compared to 2.5 (0–9) before the start of the col- laboration (p = 0.963) (Figure 2). The median number of revisions/operations following flap coverage of acute years. The median age after start of the collaboration was 47 (15–86) years. Most fractures were graded as GIIIB, but there were also severe soft tissue injuries with exposed bone without frac- tures. Results Journal of Trauma Management & Outcomes (2015) 9:3 Page 3 of 8 Page 3 of 8 Table 1 Patients treated before start of multidisciplinary collaboration Sex Age (y) Case Trauma Gustilo Flap Time to flap (d) Time to coverage (d) Free flaps included 1 M 59 Distal tibia fracture Excavator accident GIIIB Latissimus dorsi muscle 6 129 2 M 42 Proximal tibia fracture, fibula fracture, femoral fracture, humeral fracture MC accident GIIIB Fibula osteocutaneous flap (Medial gastro- cnemius flap) 49 (0) 425 Failed Medial gastrocnemius flap 3 M 48 Distal femoral fracture Bus accident GIIIC Latissimus dorsi muscle 7 Amputation 2 d after flap Free flaps excluded 4 M 54 Osteomyelitis calcaneal fracture 1 year earlier MC accident GIIIA Gracilis muscle 360 7 5 M 21 Osteomyelitis proximal tibia after fracture 5 years earlier Moped accident GIIIB Latissimus dorsi muscle 5 years 65 Pedicled flap included 1 M 26 Multitrauma, calcaneal fractures with skin necrosis Fall injury GIIIA Sural island flap 220 21 2 M 60 Diaphyseal tibia fracture Fall injury GIIIB Fasciocutaneous rotation flap 26 52 Pedicled flap excluded 3 M 64 Tibia condyle fracture and abundant soft tissue injury Outboard motor accident GIIIC Lateral gastrocnemius muscle 26 72 4 F 30 Tibia pilon fracture, secondary skin necrosis after surgery Fall injury - Sural island flap 38 4 5 F 54 Osteomyelitis, distal tibia fracture 2 years earlier, skin necrosis after secondary surgery Fall injury - Fasciocutaneous transposition flap 21 Partial flap necrosis Free flap 2011 Pat 15 Table 2 6 F 27 Multitrauma, proximal tibia fracture, admitted 1 month after trauma Car accident GIIIB Medial gastrocnemius muscle 51 57 7 M 26 Diaphyseal tibia fracture, admitted 1.5 months after trauma Climbing accident GIIIB Soleus muscle 51 334 Patients treated before start of multidisciplinary collaboration with a free or pedicled flap for soft tissue reconstruction after lower extremity trauma. Patients excluded in the subgroup analysis of acute extremity trauma were patients recieving flaps due to skin necrosis/infection after fracture surgery, chronic osteomyelitis, or patients referred to Karolinska University Hospital later than 14 days after the trauma. Time to flap is counted from trauma/admission at Karolinska University Hospital, or in the secondary cases from occurrence of wounds, fistulation or exposed bone. Time to complete soft tissue coverage after flap surgery was determined as no remaining skin wounds at clinical evaluation at follow up. Results The median time to flap coverage of acute extremity trauma was 6 (0–51) days after the start of the collabor- ation, compared to 16.5 (0–220) days before the start of the collaboration (p = 0.283) (Figure 1). The median num- ber of revisions/operations prior to flap coverage of acute extremity trauma was 3 (0–7) after the start of the collab- oration compared to 2.5 (0–9) before the start of the col- laboration (p = 0.963) (Figure 2). The median number of revisions/operations following flap coverage of acute In 6 out of 12 cases before the start of the collabor- ation and 25 out of 42 cases after the start of the collab- oration, flap treatment was due to acute extremity trauma with soft tissue defects. The rest were skin ne- crosis/infection after fracture surgery, chronic osteomye- litis, or patients referred to Karolinska University Hospital later than 14 days after the trauma. These patients were excluded from the analysis of treatment of acute extremity trauma to achieve a more uniform cohort for comparison. Sommar et al. Results Journal of Trauma Management & Outcomes (2015) 9:3 Page 5 of 8 Page 5 of 8 Table 2 Patients treated after start of multidisciplinary collaboration (Continued) 4 M 37 Proximal tibia fracture Bus accident GIIIA Medial gastrocnemius muscle 51 28 5 M 31 Multitrauma, diaphyseal tibia fracture Car accident GIIIB Fasciocutaneous rotation flap 4 16 6 M 65 Humeral fracture Gunshot GIIIB Latissimus dorsi muscle 3 35 7 F 64 Proximal tibia fracture Run over by a bus GIIIA Medial gastrocnemius muscle 16 65 8 F 15 Proximal tibia fracture Moped accident GIIIB Medial gastrocnemius muscle 1 21 9 F 43 Multitrauma, soft tissue defect over patella Train accident - Medial gastrocnemius muscle 34 10 10 M 55 Diaphyseal tibia fracture Bicycle accident GIIIB Soleus muscle 4 4 11 M 36 Radial fractures Gunshot GIIIC Fasciocutaneous transposition flap 0 15 Partial flap necrosis Pedicled flap excluded 12 F 59 Osteomyelitis, ankle fracture 5.5 years earlier Fall injury - Extensor digitorum brevis muscle 5.5 years 112 13 M 55 Distal tibia fracture 7 years earlier, secondary skin necrosis after surgery Fall injury - Extensor digitorum brevis muscle 7 years 49 14 M 47 Tibia pilon fracture, secondary skin necrosis after surgery Fall injury - Sural island flap 28 143 Partial flap necrosis 15 M 55 Osteomyelitis, tibia condyle fracture 17 months earlier secondary skin necrosis after surgery MC accident - Medial gastrocnemius muscle 156 8 Amputation 8 d after flap 16 F 71 Ulnar fracture, secondary skin necrosis after surgery Fall injury - Fasciocutaneous rotation flap 63 18 17 F 86 Proximal tibia fracture, secondary skin necrosis after surgery Fall injury - Medial gastrocnemius muscle 61 11 18 F 62 Distal tibia fracture 3 years earlier, wound infection after hardware removal Fall injury - Soleus muscle 239 17 Partial flap necrosis 19 M 63 Multitrauma, patellar fracture, secondary skin necrosis after surgery Car accident - Medial gastrochnemius muscle (Fasciocutaneous rotation flap) (103) 51 190 Partial flap necrosis rotation flap 20 F 67 Distal tibia fracture, secondary skin necrosis after surgery Fall injury - Propeller flap 135 Not healed by inclusion 21 M 76 Distal tibia fracture, secondary skin necrosis after plaster Wheelchair accident - Sural island flap 42 Flap failure, Amputation 70 d. after flap Patients treated after start of multidisciplinary collaboration with a free or pedicled flap for soft tissue reconstruction after lower extremity trauma. Results Journal of Trauma Management & Outcomes (2015) 9:3 Page 4 of 8 Page 4 of 8 Table 2 Patients treated after start of multidisciplinary collaboration Sex Age (y) Case Trauma Gustilo Flap Time to flap (d) Time to coverage (d) Free flap included 1 M 31 Traumatic arm amputation Work accident GIIIC Palmar free flap 0 75 2 F 33 Calcaneus fracture, ankle fracture, pelvic fracture Run over by a lorry GIIIB Latissimus dorsi muscle 4 203 3 M 55 Diaphyseal tibia fracture with bone defect Gunshot GIIIC Fibula osteo- cutaneous flap 4 100 Flap failure and amputation 4 M 42 Diaphyseal tibia fracture Car accident GIIIC Anterolateral thigh flap 11 192 5 M 33 Multitrauma, diaphyseal tibia fracture MC accident GIIIC Latissimus dorsi muscle 8 63 6 M 61 Distal tibia fracture MC accident GIIIC Latissimus dorsi muscle 3 87 7 F 46 Distal tibia/fibula fracture Riding accident GII Anterolateral thigh flap 7 31 8 F 31 Multitrauma, proximal tibia/fibula fracture Bicycle accident GIIIC Latissimus dorsi muscle 16 148 9 M 20 Multitrauma, distal tibia fracture Car accident GIIIB Latissimus dorsi muscle 6 23 Amputation 1 year after flap due to pain 10 M 27 Soft tissue defect tibia MC accident - Gracilis muscle 7 104 11 M 34 Distal tibia fracture MC accident GIIIA Gracilis muscle 29 35 12 M 66 Ankle fracture Work accident GIIIA Anterolateral thigh flap 22 16 13 M 36 Distal tibia fracture Car accident GIIIA Gracilis muscle 3 21 14 M 27 Multitrauma, tibia fracture with bone defect MC accident GIIIC Fibula osteo- cutaneous flap 5 10 Free flap excluded 15 F 60 Osteomyelitis, distal tibia fracture 6 years earlier Fall injury - Anterolateral thigh flap 5.5 years 40 Pedicled flap 2006 Pat 5 Table 1 16 M 32 Infection after distal tibia fracture, referred 2 months after trauma Car accident GIIIB Anterolateral thigh flap (Sural island flap) 120 (105) 42 Failed Sural island flap 17 F 53 Distal tibia fracture, secondary skin necrosis after surgery Fall injury - Anterolateral thigh flap 456 29 18 M 76 Trimalleolar ankle fracture, referred 2.5 months after trauma Fall injury GIIIA Gracilis muscle 105 128 19 M 44 Osteomyelitis, tibia fracture 15 months earlier Moped accident GIIIC Latissimus dorsi muscle 432 24 20 F 63 Osteomyelitis, tibia pilon fracture 11 months earlier Fall injury - Anterolateral thigh flap (Gracilis muscle) 295 (288) 134 Failed gracilis muscle Pedicled flap included 1 M 40 Proximal tibia fracture MC accident GIIIB Propeller flap and lateral gastrocnemius muscle 3 457 Partial flap necrosis 2 M 70 Soft tissue injury to elbow, and humeral fracture Lawn mower accident GIIIB Radial forearm flap 7 23 3 M 17 Soft tissue defect over knee Moped accident - Medial gastrocnemius muscle 11 9 Table 2 Patients treated after start of multidisciplinary collaboration Sex Age (y) Case Trauma Gustilo Flap Table 2 Patients treated after start of multidisciplinary collaboration Time to flap (d) Time to coverage (d) Sommar et al. Patients treated after start of multidisciplinary collaboration with a free or pedicled flap for soft tissue reconstruction after lower extremity trauma. Patients excluded in the subgroup analysis of acute extremity trauma were patients recieving flaps due to skin necrosis/infection after fracture surgery, chronic osteomyelitis, or patients referred to Karolinska University Hospital later than 14 days after the trauma. Time to flap is counted from trauma/admission at Karolinska University Hospital, or in the secondary cases from occurrence of wounds, fistulation or exposed bone. Time to complete soft tissue coverage after flap surgery was determined as no remaining skin wounds at clinical evaluation at follow up. Results Six patients before the start of the MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. 0.0464 Figure 3 Number of revisions/operations after flap surgery. Number of revisions/operations after flap surgery in acute extremity trauma patients 4 years before and after start of multidisciplinary collaboration (MDC). Dots and squares represent individual patients. Six patients before the start of the MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. Figure 3 Number of revisions/operations after flap surgery. Number of revisions/operations after flap surgery in acute extremity trauma patients 4 years before and after start of multidisciplinary collaboration (MDC). Dots and squares represent individual patients. Six patients before the start of the MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. Figure 1 Time to flap surgery. Days from admission with acute extremity trauma at Karolinska University Hospital to flap surgery 4 years before and after start of multidisciplinary collaboration (MDC). Dots and squares represent individual patients. Six patients before the start of the MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. start of the collaboration (p = 0.518). The median time in hospital after acute extremity trauma was reduced from 67.5 (16–88) days to 29 (8–121) days after the start of the collaboration (p = 0.044) (Figure 4). All included fractures healed. Time to fracture healing was 296 days before the start of the collaboration and 256 days after the start of the collaboration (p = 0.594). Only 5 out of 6 patients be- fore and 16 out of 25 patients after the start of the extremity trauma was 1 (1–10) after the start of the col- laboration compared to 4 (0–10) before the start of the collaboration (p = 0.046) (Figure 3).The median time for complete soft tissue coverage after flap surgery of acute extremity trauma was 33 (9–203) days after the start of the collaboration compared to 62 (2–425) days before the 0.0444 Figure 4 In-hospital stay. Results Days in hospital after admission with acute extremity trauma and treated with flaps in patients 4 years before and after start of multidisciplinary collaboration (MDC). Data includes all admissions, i.e. also secondary revisions. Dots and squares represent individual patients. Six patients before the start of MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. 0.0444 Figure 4 In-hospital stay. Days in hospital after admission with acute extremity trauma and treated with flaps in patients 4 years before and after start of multidisciplinary collaboration (MDC). Data includes all admissions, i.e. also secondary revisions. Dots and squares represent individual patients. Six patients before the start of MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. 0.0444 0.9633 Figure 2 Number of revisions/operations prior to flap surgery. Number of revisions/operations prior to flap surgery in acute extremity trauma patients 4 years before and after start of multidisciplinary collaboration (MDC). Dots and squares represent individual patients. Six patients before the start of the MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. 0.0444 Figure 4 In-hospital stay. Days in hospital after admission with acute extremity trauma and treated with flaps in patients 4 years before and after start of multidisciplinary collaboration (MDC). Data includes all admissions, i.e. also secondary revisions. Dots and squares represent individual patients. Six patients before the start of MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. 0.9633 Figure 2 Number of revisions/operations prior to flap surgery. Number of revisions/operations prior to flap surgery in acute extremity trauma patients 4 years before and after start of multidisciplinary collaboration (MDC). Dots and squares represent individual patients. Six patients before the start of the MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. 0.9633 Figure 2 Number of revisions/operations prior to flap surgery. Number of revisions/operations prior to flap surgery in acute extremity trauma patients 4 years before and after start of multidisciplinary collaboration (MDC). Dots and squares represent individual patients. Results Patients excluded in the subgroup analysis of acute extremity trauma were patients recieving flaps due to skin necrosis/infection after fracture surgery, chronic osteomyelitis, or patients referred to Karolinska University Hospital later than 14 days after the trauma. Time to flap is counted from trauma/admission at Karolinska University Hospital, or in the secondary cases from occurrence of wounds, fistulation or exposed bone. Time to complete soft tissue coverage after flap surgery was determined as no remaining skin wounds at clinical evaluation at follow up. Table 2 Patients treated after start of multidisciplinary collaboration (Continued) Patients treated after start of multidisciplinary collaboration with a free or pedicled flap for soft tissue reconstruction after lower extremity trauma. Patients excluded in the subgroup analysis of acute extremity trauma were patients recieving flaps due to skin necrosis/infection after fracture surgery, chronic osteomyelitis, or patients referred to Karolinska University Hospital later than 14 days after the trauma. Time to flap is counted from trauma/admission at Karolinska University Hospital, or in the secondary cases from occurrence of wounds, fistulation or exposed bone. Time to complete soft tissue coverage after flap surgery was determined as no remaining skin wounds at clinical evaluation at follow up. Page 6 of 8 Sommar et al. Journal of Trauma Management & Outcomes (2015) 9:3 0.0464 Figure 3 Number of revisions/operations after flap surgery. Number of revisions/operations after flap surgery in acute extremity trauma patients 4 years before and after start of multidisciplinary collaboration (MDC). Dots and squares represent individual patients. Six patients before the start of the MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. 0.2823 Figure 1 Time to flap surgery. Days from admission with acute extremity trauma at Karolinska University Hospital to flap surgery 4 years before and after start of multidisciplinary collaboration (MDC). Dots and squares represent individual patients. Six patients before the start of the MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. 0.2823 Figure 1 Time to flap surgery. Days from admission with acute extremity trauma at Karolinska University Hospital to flap surgery 4 years before and after start of multidisciplinary collaboration (MDC). Dots and squares represent individual patients. Discussion S The median time to flap coverage after the start of the collaboration was 6 days, which still is not satisfactory. Institutional factors such as available OR time and sup- port staff may make immediate reconstruction of ex- tremity injuries impossible. There can also be other concomitant injury which first has to be dealt with, and the extent of soft tissue injury is sometimes difficult to determine early in the process. Whereas Godina [1] clearly states the superiority of reconstruction within 3 days, other authors have shown excellent results with late coverage [11-13]. There was no difference in out- come or flap failure in our series when comparing sub- groups of patients; flap coverage < 3 days vs. > 3 days, or flap coverage < 6 days vs. > 6 days (data not shown). It is not possible to draw overreaching conclusions from this due to the relative small groups in comparison to the work by Godina [1]. One key to earlier flap coverage is to perform fewer revisions. After the start of the collabor- ation we still had a median number of 3 operations/revi- sions before flap coverage. Only one radical debridement performed by the most experienced surgeons before flap coverage has been set up as a future goal for the team. Severe extremity trauma with soft tissue defects is a dif- ficult task for both the patient and health care provider. In order to optimize the outcome for the patients we have improved our care of the patients. There is a grow- ing body of evidence that these complex extremity injur- ies are best handled by a multidisciplinary team of experienced plastic and orthopedic surgeons [8,10]. The present study clearly demonstrates an increase of flap re- construction in acute trauma patients and patients with secondary wounds following acute extremity trauma after the start of a multidisciplinary collaboration. This is hardly explained by natural variation, as the observa- tion time is 8 years, and it seems unlikely that variation could explain a 3.5-fold increase as in our series. In 2007 all high energy trauma in Stockholm County was region- alized to Karolinska University Hospital. This has re- sulted in a 0.75-fold increase of trauma patients. Patients with severe extremity trauma is now primary taken to Karolinska University Hospital. Prior to 2007 the refer- ence policy was less strict and hence not all cases were treated at Karolinska. Results Six patients before the start of the MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. Figure 4 In-hospital stay. Days in hospital after admission with acute extremity trauma and treated with flaps in patients 4 years before and after start of multidisciplinary collaboration (MDC). Data includes all admissions, i.e. also secondary revisions. Dots and squares represent individual patients. Six patients before the start of MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. Figure 4 In-hospital stay. Days in hospital after admission with acute extremity trauma and treated with flaps in patients 4 years before and after start of multidisciplinary collaboration (MDC). Data includes all admissions, i.e. also secondary revisions. Dots and squares represent individual patients. Six patients before the start of MDC and 25 after the start of MDC were treated with either free or pedicled flaps due to acute extremity trauma with soft tissue defects. Page 7 of 8 Page 7 of 8 Sommar et al. Journal of Trauma Management & Outcomes (2015) 9:3 Gustilo IIIA due to crush wounds on the heels. During the following days he developed skin necrosis on the right heal. Due to a more conservative attitude towards small wounds in extremity trauma before the start of MDC, there was an attempt of secondary healing. Today, this patient would probably have received a free flap in the near time period of the trauma. To reduce the effects of this “outlier” statistically we have chosen to use a nonparametric method, and present median values. This patient was not an outlier in comparisons other than time to flap. collaboration were possible to follow to fracture healing. One patient was amputated after flap treatment before the start of the collaboration. After the start of the collabor- ation, four patients had soft tissue defects with exposed bone without fractures, one had a primary traumatic am- putation, two were amputated prior to flap treatment and one patient was not healed at the time of inclusion in the study. One of the patients healed, but was amputated one year after flap coverage due to pain. Discussion S Journal of Trauma Management & Outcomes (2015) 9:3 are in concordance, showing favorable amputation rate and a more efficient planning of surgical workload [10]. 5. Meling T, Harboe K, Soreide K. Incidence of traumatic long-bone fractures requiring in-hospital management: a prospective age- and gender-specific analysis of 4890 fractures. Injury. 2009;40:1212–9. 6. Weiss RJ, Montgomery SM, Ehlin A, Al Dabbagh Z, Stark A, Jansson KA. Decreasing incidence of tibial shaft fractures between 1998 and 2004: information based on 10,627 Swedish inpatients. Acta Orthop. 2008;79:526–33. 7. Heitmann C, Levin LS. The orthoplastic approach for management of the severely traumatized foot and ankle. J Trauma. 2003;54:379–90. 8. Naique SB, Pearse M, Nanchahal J. Management of severe open tibial fractures: the need for combined orthopaedic and plastic surgical treatment in specialist centres. J Bone Joint Surg. 2006;88:351–7. 9. Gustilo RB, Mendoza RM, Williams DN. Problems in the management of type III (severe) open fractures: a new classification of type III open fractures. J Trauma. 1984;24:742–6. 10. Clover AJ, Rannan-Eliya S, Saeed W, Buxton R, Majumder S, Hettiaratchy SP, et al. Experience of an orthoplastic limb salvage team after the Haiti earthquake: analysis of caseload and early outcomes. Plast Reconstr Surg. 2011;127:2373–80. 11. Karanas YL, Nigriny J, Chang J. The timing of microsurgical reconstruction in lower extremity trauma. Microsurgery. 2008;28:632–4. 12. Hill JB, Vogel JE, Sexton KW, Guillamondegui OD, Corral GA, Shack RB. Re-evaluating the paradigm of early free flap coverage in lower extremity trauma. Microsurgery. 2013;33:9–13. 13. Francel TJ, Vander Kolk CA, Hoopes JE, Manson PN, Yaremchuk MJ. Microvascular soft-tissue transplantation for reconstruction of acute open tibial fractures: timing of coverage and long-term functional results. Plast Reconstr Surg. 1992;89:478–87. discussion 488–479. 14. Court-Brown CM, Rimmer S, Prakash U, McQueen MM. The epidemiology of open long bone fractures. Injury. 1998;29:529–34. 5. Meling T, Harboe K, Soreide K. Incidence of traumatic long-bone fractures requiring in-hospital management: a prospective age- and gender-specific analysis of 4890 fractures. Injury. 2009;40:1212–9. 6. 6. Weiss RJ, Montgomery SM, Ehlin A, Al Dabbagh Z, Stark A, Jansson KA. Decreasing incidence of tibial shaft fractures between 1998 and 2004: information based on 10,627 Swedish inpatients. Acta Orthop. 2008;79:526–3 Conclusion To conclude, the formalized collaboration has led to an increased number of flap coverage in extremity trauma patients. The acute cases have a shorter in-hospital stay, and fewer postoperative revisions after the start of the collaboration. However, if this has resulted in a decreased number of amputations and improved long-term outcome needs to be further analyzed in a larger cohort. Our col- laboration has also led to awareness among our ortho- pedic colleagues to establish an early contact with a plastic surgeon when a patient has an open fracture with soft tissue damage. The goal is that the plastic surgeon should be contacted from the emergency room in order to par- ticipate at the first operative revision to be able to plan for further surgery. Routines are also created for coordination of operative resources, and postoperative flap surveillance. Finally, this formal collaboration gives us better possibil- ities for quality control and evaluation. Our experience leads us to recommend other centers to establish a similar set-up for early multidisciplinary treatment of this patient group. p p 7. Heitmann C, Levin LS. The orthoplastic approach for management of the severely traumatized foot and ankle. J Trauma. 2003;54:379–90. 8. Naique SB, Pearse M, Nanchahal J. Management of severe open tibial fractures: the need for combined orthopaedic and plastic surgical treatment in specialist centres. J Bone Joint Surg. 2006;88:351–7. 8. Naique SB, Pearse M, Nanchahal J. Management of severe open tibial fractures: the need for combined orthopaedic and plastic surgical treatment in specialist centres. J Bone Joint Surg. 2006;88:351–7. 9. Gustilo RB, Mendoza RM, Williams DN. Problems in the management of type III (severe) open fractures: a new classification of type III open fractures. J Trauma. 1984;24:742–6. 12. Hill JB, Vogel JE, Sexton KW, Guillamondegui OD, Corral GA, Shack RB. Re-evaluating the paradigm of early free flap coverage in lower extremity trauma. Microsurgery. 2013;33:9–13. 13. Francel TJ, Vander Kolk CA, Hoopes JE, Manson PN, Yaremchuk MJ. Microvascular soft-tissue transplantation for reconstruction of acute open tibial fractures: timing of coverage and long-term functional results. Plast Reconstr Surg. 1992;89:478–87. discussion 488–479. g 14. Court-Brown CM, Rimmer S, Prakash U, McQueen MM. The epidemiology of open long bone fractures. Injury. 1998;29:529–34. Competing interests h h d l h Competing interests The authors declare that they have no competing interests. Authors’ contributions PS: Surgery, collection of data and writing of manuscript. YG: Collection of data and writing of manuscript. MH: Surgery, review of patient data and writing of manuscript. ADS: Surgery and writing of manuscript. KL: Surgery, review of patient data and writing of manuscript. KÅJ: Surgery and writing of manuscript. All authors read and approved the final manuscript. Discussion S The severity of orthopedic trauma has not changed during the years. Occasionally patients with open fractures are still treated in other hospitals in Stockholm, but they are usually referred to Karolinska after external fixation and primary revision. We have also seen an increase in both acute and secondary cases referred from other hospitals in the region as the collab- oration has become known. The regionalization only cannot explain a 3.5-fold increase of flaps. We believe that the main reason is an increased awareness among our orthopedic surgeons to refer extremity trauma pa- tients for plastic surgical consultation. The collaboration has also resulted in a more rapid communication, and quicker decisions. The introduction of a collaboration had a significant impact on the in hospital stay. The median time in hos- pital after acute extremity trauma with soft tissue defects decreased from 68.5 days to 29 days. In the 25 patients after the start of the collaboration, this would corres- pond to 988 hospital days, and a total cost reduction of 1.170 000 USD. In times where cost-analysis is crucial in public health care, this reduction which corresponds to 70% of one hospital bed/year is obviously beneficial. The reduction of hospital stay may be explained by the de- crease in revisions/operations after flap coverage, which may be related to a shorter time to coverage. The major limitation in this report is the small sample size. Since the patient group treated before start of the collaboration is small, the sub analysis of acute extremity trauma patients cannot provide accurate statistical com- parison, and the reader should interpret the significance between groups with care. One of the acute trauma pa- tients included before start of the MDC was not covered before 220 days after trauma. This patient had bilateral calcaneal fractures after a fall injury, classified as a The incidence of open tibia fractures in the Swedish population is 2.3/105/year [6], which corresponds to about 220 patients per year, of which 27% [5] have a Gustilo Type III fracture. Because of the low and declining inci- dence, centralization of these difficult fractures has been suggested [8,14]. We believe that these fractures should be dealt with in a center with a close collaboration between orthopedic and plastic surgeons. The experience after the Haiti earthquake with an orthoplastic limb salvage team Page 8 of 8 Page 8 of 8 Sommar et al. Author details 1 1Department of Molecular Medicine and Surgery, Section of Plastic Surgery, Karolinska Institutet and Karolinska University Hospital Stockholm, Stockholm, Sweden. 2Department of Orthopedic Surgery Västerås Hospital and Department of Molecular Medicine and Surgery, Karolinska Institutet Stockholm, Stockholm, Sweden. 3Department of Molecular Medicine and Surgery, Section of Orthopedics and Sports Medicine, Karolinska Institutet and Karolinska University Hospital Stockholm, Stockholm, Sweden. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: Received: 29 May 2014 Accepted: 13 March 2015 • Convenient online submission 12. Hill JB, Vogel JE, Sexton KW, Guillamondegui OD, Corral GA, Shack RB. Re-evaluating the paradigm of early free flap coverage in lower extremity trauma. Microsurgery. 2013;33:9–13. g y 13. Francel TJ, Vander Kolk CA, Hoopes JE, Manson PN, Yaremchuk MJ. Microvascular soft-tissue transplantation for reconstruction of acute open tibial fractures: timing of coverage and long-term functional results. Plast Reconstr Surg. 1992;89:478–87. discussion 488–479. Consent Written informed consent was obtained from the patients for the publication of this report. References 1. Godina M. Early microsurgical reconstruction of complex trauma of the extremities. Plast Reconstr Surg. 1986;78:285–92. 2. Tielinen L, Lindahl JE, Tukiainen EJ. Acute unreamed intramedullary nailing and soft tissue reconstruction with muscle flaps for the treatment of severe open tibial shaft fractures. Injury. 2007;38:906–12. 3. Gopal S, Majumder S, Batchelor AG, Knight SL, De Boer P, Smith RM. Fix and flap: the radical orthopaedic and plastic treatment of severe open fractures of the tibia. J Bone Joint Surg. 2000;82:959–66. 4 Levin LS Early versus delayed closure of open fractures Injury 2007;38:896 9 14. Court-Brown CM, Rimmer S, Prakash U, McQueen MM. The epidemiology of open long bone fractures. Injury. 1998;29:529–34. 11. Karanas YL, Nigriny J, Chang J. The timing of microsurgical re lower extremity trauma. Microsurgery. 2008;28:632–4. 2. Tielinen L, Lindahl JE, Tukiainen EJ. Acute unreamed intramedullary nailing and soft tissue reconstruction with muscle flaps for the treatment of severe open tibial shaft fractures. Injury. 2007;38:906–12. 5. Meling T, Harboe K, Soreide K. Incidence of traumatic long-bone fractures requiring in-hospital management: a prospective age- and gender-specific analysis of 4890 fractures. Injury. 2009;40:1212–9. 6. Weiss RJ, Montgomery SM, Ehlin A, Al Dabbagh Z, Stark A, Jansson KA. Decreasing incidence of tibial shaft fractures between 1998 and 2004: information based on 10,627 Swedish inpatients. Acta Orthop. 2008;79:526–33. 7. Heitmann C, Levin LS. The orthoplastic approach for management of the severely traumatized foot and ankle. J Trauma. 2003;54:379–90. 8. Naique SB, Pearse M, Nanchahal J. Management of severe open tibial fractures: the need for combined orthopaedic and plastic surgical treatment in specialist centres. J Bone Joint Surg. 2006;88:351–7. 9. Gustilo RB, Mendoza RM, Williams DN. Problems in the management of type III (severe) open fractures: a new classification of type III open fractures. J Trauma. 1984;24:742–6. 10. Clover AJ, Rannan-Eliya S, Saeed W, Buxton R, Majumder S, Hettiaratchy SP, et al. Experience of an orthoplastic limb salvage team after the Haiti earthquake: analysis of caseload and early outcomes. Plast Reconstr Surg. 2011;127:2373–80. 11. Karanas YL, Nigriny J, Chang J. The timing of microsurgical reconstruction in lower extremity trauma. Microsurgery. 2008;28:632–4. 12. Hill JB, Vogel JE, Sexton KW, Guillamondegui OD, Corral GA, Shack RB. Re-evaluating the paradigm of early free flap coverage in lower extremity trauma. Microsurgery. 2013;33:9–13. 13. Francel TJ, Vander Kolk CA, Hoopes JE, Manson PN, Yaremchuk MJ. Microvascular soft-tissue transplantation for reconstruction of acute open tibial fractures: timing of coverage and long-term functional results. Plast Reconstr Surg. 1992;89:478–87. discussion 488–479. 14. Court-Brown CM, Rimmer S, Prakash U, McQueen MM. The epidemiology of open long bone fractures. Injury. 1998;29:529–34. 1. Godina M. Early microsurgical reconstruction of complex trauma of the extremities. Plast Reconstr Surg. 1986;78:285–92. References 1. Godina M. Early microsurgical reconstruction of complex trauma of the extremities. Plast Reconstr Surg. 1986;78:285–92. g 2. Tielinen L, Lindahl JE, Tukiainen EJ. Acute unreamed intramedullary nailing and soft tissue reconstruction with muscle flaps for the treatment of severe open tibial shaft fractures. Injury. 2007;38:906–12. 3. Gopal S, Majumder S, Batchelor AG, Knight SL, De Boer P, Smith RM. Fix and flap: the radical orthopaedic and plastic treatment of severe open fractures of the tibia. J Bone Joint Surg. 2000;82:959–66. 4. Levin LS. Early versus delayed closure of open fractures. Injury. 2007;38:896–9.
https://openalex.org/W4225661610	https://www.researchsquare.com/article/rs-1234637/latest.pdf	English	null	The Paris System for reporting urinary cytology improves the negative predictive value of high-grade urothelial carcinoma	BMC urology	2,022	cc-by	3,916	The Paris System for Reporting Urinary Cytology Improves the Negative Predictive Value of High- grade Urothelial Carcinoma The Paris System for Reporting Urinary Cytology Improves the Negative Predictive Value of High- grade Urothelial Carcinoma Mari Yamasaki Department of Urology, Faculty of Medicine, Kagawa University Rikiya Taoka (  taoka.rikiya@kagawa-u.ac.jp ) Department of Urology, Faculty of Medicine, Kagawa University Kazuya Katakura Department of Diagnostic Pathology, University Hospital, Faculty of Medicine, Kagawa University Toru Matsunaga Department of Diagnostic Pathology, University Hospital, Faculty of Medicine, Kagawa University Naoya Kani Department of Urology, Faculty of Medicine, Kagawa University Tomoko Honda Department of Urology, Faculty of Medicine, Kagawa University Satoshi Harada Department of Urology, Faculty of Medicine, Kagawa University Yoichiro Tohi Department of Urology, Faculty of Medicine, Kagawa University Yuki Matsuoka Department of Urology, Faculty of Medicine, Kagawa University Takuma Kato Department of Urology, Faculty of Medicine, Kagawa University Homare Okazoe Department of Urology, Faculty of Medicine, Kagawa University Hiroyuki Tsunemori Department of Urology, Faculty of Medicine, Kagawa University Nobufumi Ueda Department of Urology, Faculty of Medicine, Kagawa University Reiji Haba Department of Diagnostic Pathology, University Hospital, Faculty of Medicine, Kagawa University Mikio Sugimoto Department of Urology, Faculty of Medicine, Kagawa University Mari Yamasaki Page 1/11 Research Article Keywords: urothelial carcinoma, high-grade urothelial carcinoma, urine cytology, The Paris System for Reporting Urinary Cytology, negative predictive value Posted Date: February 2nd, 2022 DOI: https://doi.org/10.21203/rs.3.rs-1234637/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Keywords: urothelial carcinoma, high-grade urothelial carcinoma, urine cytology, The Paris System for Reporting Urinary Cytology, negative predictive value Posted Date: February 2nd, 2022 DOI: https://doi.org/10.21203/rs.3.rs-1234637/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Keywords: urothelial carcinoma, high-grade urothelial carcinoma, urine cytology, The Paris System for Reporting Urinary Cytology, negative predictive value Posted Date: February 2nd, 2022 DOI: https://doi.org/10.21203/rs.3.rs-1234637/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Version of Record: A version of this preprint was published at BMC Urology on April 5th, 2022. See the published version at https://doi.org/10.1186/s12894-022-01005-8. Page 2/11 Abstract Background: The Paris System (TPS) for reporting urinary cytology differs from conventional systems (CS) in that it focuses on the diagnosis of high-grade urothelial carcinoma (HGUC). This study investigated the impact of TPS implementation on the diagnostic accuracy of HGUC by comparing it with CS. Methods: A total of 649 patients who underwent transurethral resection of bladder tumor (TURBT) between January 2009 and December 2020 were included in this study. Our institution adopted TPS to report urinary cytology in February 2020. The diagnostic accuracy of HGUC in preoperative urinary cytology was compared with the presence or absence of HGUC in resected specimens of TURBT before and after TPS implementation. Results: After implementing TPS in urinary cytology, 89 patients were reviewed and compared with 560 patients whose urinary cytology was diagnosed by CS. TPS and CS for detecting HGUC had 56.0% and 58.2% sensitivity, 97.8% and 91.2% specificity, and 93.3% and 87.9% positive predictive values, respectively. There were no significant differences between TPS and CS in terms of sensitivity, specificity, and positive predictive value for HGUC (P=0.83, 0.21, 1.00). On the other hand, the negative predictive value for HGUC using TPS was 80.0%, which was significantly higher than that of CS (66.4%, P=0.04) The multivariate logistic regression analysis indicated that not using TPS was one of the independent predictive factors associated with false-negative results for HGUC (odds ratio, 2.26; 95% confidence interval, 1.08–4.77; P=0.03). Results: After implementing TPS in urinary cytology, 89 patients were reviewed and compared with 560 patients whose urinary cytology was diagnosed by CS. TPS and CS for detecting HGUC had 56.0% and 58.2% sensitivity, 97.8% and 91.2% specificity, and 93.3% and 87.9% positive predictive values, Conclusion: In instances where urinary cytology is reported as negative for HGUC by TPS, there is a low probability of HGUC, indicating that TPS has a potential diagnostic benefit. Study population and design From January 2009 to December 2020, 755 patients underwent TURBT at Kagawa University Hospital. Of these, 21 patients who were pathologically diagnosed with non-urothelial malignancies and 85 patients whose urinary cytology had not been evaluated before TURBT were excluded. A total of 649 patients were included in the study. Our institution adopted TPS to report urinary cytology in February 2020. Before that, we used CS using the Papanicolaou-stained urinary cytology system. The included patients were divided into two groups of 89 patients and 560 patients each, who used TPS or CS to report urinary cytology, respectively. Urine specimens were used as voided urine samples on the day before TURBT. The CS has five classes: class 1, inadequate or absence of suspicious cells; class 2, atypical cells but not malignant; class 3, cells suspected of being malignant but not confidently; class 4, suspected malignant cells; and class 5, malignant cells. We defined the criteria that positive urinary cytology was greater than class 4, and negative urinary cytology was classified as class 1 or 2 in the CS. On the other hand, TPS requires four categories for reporting urinary cytology: negative for HGUC (NHGUC), atypical urothelial cell, suspicious for HGUC (SHGUC), and HGUC. This study defined SHGUC and HGUC as positive urinary cytology and NHGUC as negative urinary cytology in TPS. We retrospectively compared the positive or negative results of preoperative urinary cytology in the presence or absence of HGUC in resected specimens of TURBT. Background Based on the latest Global Cancer Incidence, Mortality And Prevalence data, bladder cancer (BC) is the 10th most common form of cancer worldwide, with an estimated 573,000 new cases in 2020 [1]. Approximately 75% of newly diagnosed BC cases are non-muscle-invasive BC (NMIBC) [2]. In clinical practice, NMIBC is treated with transurethral resection of bladder tumor (TURBT) followed by intravesical therapy, depending on the risk of recurrence and progression [3]. However, NMIBC recurs in approximately 50% of the cases [2]. Therefore, patients with NMIBC need surveillance using regular cystoscopy and urinary cytology for at least 5 years after the initial treatment [3]. Some histological types of BC contain pathologically different properties with varying clinical courses. In 2004, the World Health Organization/ International Society of Urological Pathology consensus classification system for papillary urothelial neoplasms of the urinary bladder was published [4]. The system classifies urothelial cancer (UC) into two main types: low-grade UC (LGUC) and high-grade UC (HGUC) [5]. Clinically, although LGUC has a low malignant potential, HGUC has a risk of disease Page 3/11 Page 3/11 progression and metastases, followed by death [6, 7]. Therefore, HGUC cannot be ignored in patients with BC. progression and metastases, followed by death [6, 7]. Therefore, HGUC cannot be ignored in patients with BC. Urinary cytology is a convenient screening tool for UC [3, 8]. However, the terminology for reporting urinary cytology has not been standardized. The Paris System (TPS) for reporting urinary cytology is a recently established international system for diagnosing urinary tract cytology specimens [9]. TPS differs from conventional systems (CS) in that it focuses on the diagnosis of HGUC. Effective detection of HGUC, which has the potential to progress and metastasize, is crucial for patients with suspected HGUC of the bladder. Therefore, TPS, which focuses on the diagnosis of HGUC, may benefit patients more than CS [10, 11]. However, there are few reports on the use TPS for HGUC detection. This study, therefore, aimed to reveal the impact of TPS implementation for urinary cytology diagnosis on the diagnostic accuracy of HGUC in real clinical situations by comparing it with CS. Data collection In addition to reports of histopathological diagnosis of the resected specimen and preoperative urinary cytology, age, sex, prior recurrence status, tumor appearance, tumor number, largest tumor diameter, tumor grade, pathological T-stage, and presence of carcinoma in situ (CIS) were retrospectively investigated. This study was conducted according to the principles outlined in the Declaration of Helsinki Page 4/11 Page 4/11 (as revised in Fortaleza, Brazil, October 2013), and these surveys were performed with the approval of the Ethics Committee of Kagawa University (permission number: 2020-070). The need for informed consent was waived by the Ethics Committee, Faculty of Medicine, Kagawa University because of the retrospective nature of this study. (as revised in Fortaleza, Brazil, October 2013), and these surveys were performed with the approval of the Ethics Committee of Kagawa University (permission number: 2020-070). The need for informed consent was waived by the Ethics Committee, Faculty of Medicine, Kagawa University because of the retrospective nature of this study. Statistical analysis The diagnostic accuracy of HGUC was calculated as the sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV). The Mann–Whitney U test or Fisher’s exact test was used to compare the clinical characteristics and diagnostic accuracy index between TPS and CS groups. Multivariate analysis with a logistic regression model was performed to determine an independent predictive factor for NPV errors. The number of variables incorporated into the multivariate analysis was determined based on the number of events. All statistical analyses were performed using SPSS for Windows version 12 (SPSS Inc., Chicago, IL, USA). Statistical significance was set at P<0.05. Table 2. Summary of cytological and histopathological diagnoses Patient characteristics Data from a total of 89 patients diagnosed with BC after implementing TPS were reviewed and compared with 560 patients whose urinary cytology was diagnosed by CS. Table 1 shows the patients’ characteristics in TPS and CS groups. The proportion of primary BC was higher in the TPS group than in the CS group (76.4% vs. 60.5%, P<0.01). In addition, the proportion of TURBT history within 90 days, including the second TURBT in the TPS group, was higher than that of the CS group (30.3% vs. 14.6%, P<0.01). Furthermore, the TPS group had a larger tumor size than the CS group (P=0.05). By contrast, there were no significant differences in age, sex, tumor appearance, tumor number, presence of high- grade tumor, pathological tumor stage, and presence of CIS between the two groups. Table 1. Patient characteristics Page 5/11 Page 5/11 Page 5/11 Page 5/11 Urinary cytology and histopathological diagnosis Urinary cytology and histopathological diagnosis Urinary cytology and histopathological diagnosis Table 2 shows a summary of the urinary cytology and histopathological diagnoses for each TPS and CS group. The proportion of patients who were diagnosed with SHGUC and HGUC in the TPS group was 16.9%, which was significantly lower than that of patients with positive urinary cytology in the CS group (26.6%, P=0.04). Of 560 patients in the CS group, 313 (55.9%) were histologically diagnosed with HGUC in the resected specimens. Of those, 94 (30.0%) patients who were diagnosed with negative urinary cytology were false-negative, and 131 (41.9%) patients with positive urinary cytology were truly positive. In contrast, of 89 patients in the TPS group, 42 (47.2%) were histologically diagnosed with HGUC. Of these, 11 (26.2%) patients whose urinary cytology was diagnosed as NHGUC by TPS had false-negative results, and 14 (33.3%) patients with positive urinary cytology were truly positive. Page 6/11 Page 6/11 Sensitivity, specificity, PPV, and NPV of urine cytology for HGUC Table 3 summarizes the detection of HGUC in TPS and CS groups. There were no significant differences between TPS and CS groups in terms of sensitivity, specificity, and PPV for HGUC. Contrastingly, the TPS group had a significantly higher NPV for HGUC than the CS group (80.0% vs. 66.4%, P=0.04). A total of 105 patients had false-negative urinary cytology results for HGUC. Table 3. The detection ability of high-grade urothelial carcinoma able 3. Discussion TPS was created as an international form to standardize the reporting of urinary cytology [9]. TPS differs from CS in that it focuses on the diagnosis of HGUC. However, there are few reports on the ability of TPS to detect HGUC. This study compared TPS and CS with regards to detecting HGUC in urine specimens before tumor resection and in resected tumor tissues. The results indicated that TPS was superior to CS in terms of its NPV for HGUC. HGUC is well known to be associated with a worse overall survival. Therefore, a high NPV for HGUC using TPS is clinically important for patients with suspected BC. Urinary cytology is typically used to screen for UC in two clinical situations: when new-onset UC is suspected, such as in patients with unexplained hematuria, and during surveillance after BC treatment [12]. In particular, clinical guidelines suggest that cystoscopy is needed for at least 5 years for patients with a history of BC because half of the patients with BC will have an intravesical recurrence [2, 3]. However, cystoscopy is associated with physical pain and mental burden [13] which some patients experience during urinary tract infections [14]. If the high NPV of TPS can deny the presence of HGUC, and reduce additional examinations such as cystoscopy, it may improve patient quality of life and cost- effectiveness. TPS defines standardized cytomorphologic and numerical criteria for its diagnostic categories, which stratify a patient’s risk for HGUC [9]. In other words, strict definitions of TPS may reduce the sensitivity of HGUC. In this study, 16.9% of patients were diagnosed with SHGUC and HGUC in the TPS group, which was significantly lower than that of patients with positive urinary cytology in the CS group. However, there was no difference in the sensitivity for HGUC between the TPS and CS groups. Recent reports demonstrated that the sensitivity for HGUC using TPS ranged from 83.3–87.1% [11, 15, 16]. In addition, the NPV was between 81.4% and 86.4%, which was similar to the value obtained in this study (80.0%). These data demonstrate that TPS can reduce the rate of unnecessary indeterminate diagnoses while maintaining excellent sensitivity for identifying HGUC. Multivariable logistic regression analysis indicated that TPS may reduce false-negative results for HGUC. At the same time, recurrent tumors also reduced false-negatives for HGUC. Lee et al. analyzed the causes of false-negative cytology in HGUC cases [17]. Patient characteristics The detection ability of high-grade urothelial carcinoma y g g Multivariate logistic regression analysis, which included prior recurrence status, prior TURBT within 90 days, and intraoperative factors such as tumor appearance, tumor number, and largest tumor diameter, revealed that not using TPS was an independent predictive factor associated with false-negative results for HGUC (odds ratio, 2.26; 95% confidence interval, 1.08–4.77; P=0.03; Table 4). Table 4. Multivariate analysis for the prediction of false-negatives for high-grade urothelial carcinoma Page 7/11 Page 7/11 Conclusions Our study revealed the impact of TPS implementation on the diagnostic accuracy of HGUC in real clinical situations by comparing it with CS. As a result, this study clearly indicates that, in instances where urinary cytology is reported as NHGUC by TPS, there is a low probability of HGUC. These results support the implementation of the TPS in clinical practice. Discussion The possible explanations were the overestimation of the grade by the pathologist and inflammation of the bladder. In this study, inflammation due to prior Page 8/11 Page 8/11 transurethral surgery and subsequent intravesical instillation therapy may have affected the diagnosis of urinary cytology. However, solid tumors were found to be at risk of false-negatives for HGUC. The relationship between solid tumors and false-negatives for HGUC may have been influenced by the smaller area of the tumor surface compared to papillary tumors. Although a certain number of false negatives for HGUC are expected, serial urinary cytology may decrease false-negative results and improve sensitivity for detection [18]. Therefore, in clinical practice, TPS should be used for urinary cytology because it yields fewer false-negative results than CS. One of this study’s limitations is that cytological findings were interpreted by a single pathologist. Of course, the accuracy of the diagnosis has been re-evaluated by another pathologist. However, differences in the diagnosis of urinary cytopathology between pathologists have been reported [19]. Therefore, further studies addressing interobserver variability should be conducted. In addition, urine collection methods, urine collection volumes, and urinary cell counts can affect the diagnostic quality of urinary cytology [9]. Since there was no information on urine specimens in this study, the effect of urine specimens on urinary cytology could not be investigated. The study’s retrospective nature was an additional limitation. However, to the best of our knowledge, this is the first study to consider various factors affecting the quality of urinary cytology, and demonstrate that the implementation of TPS in clinical practice may improve the NPV for HGUC. One of this study’s limitations is that cytological findings were interpreted by a single pathologist. Of course, the accuracy of the diagnosis has been re-evaluated by another pathologist. However, differences in the diagnosis of urinary cytopathology between pathologists have been reported [19]. Therefore, further studies addressing interobserver variability should be conducted. In addition, urine collection methods, urine collection volumes, and urinary cell counts can affect the diagnostic quality of urinary cytology [9]. Since there was no information on urine specimens in this study, the effect of urine specimens on urinary cytology could not be investigated. The study’s retrospective nature was an additional limitation. List Of Abbreviations CIS, carcinoma in situ; UC, urothelial carcinoma; TPS, the Paris System; LGUC, low-grade urothelial carcinoma; HGUC, high-grade urothelial carcinoma; CS, conventional systems; NPV, negative predictive value; PPV, positive predictive value; TURBT, transurethral resection of bladder tumor; SHGUC, suspicious for high-grade urothelial carcinoma; NHGUC, negative for high-grade urothelial carcinoma; BC, bladder cancer; NMIBC, non-muscle-invasive bladder cancer Consent for publication Not applicable. Authors’ contributions All the authors participated in the study design. MY and RT interpreted the data, performed statistical analyses, and wrote the manuscript. KK and TM contributed to the implementation of TPS and pathological diagnosis. RH and MS supported the original concept and design of this study. NK, TH, SH, YT, YM, TK, HO, HT, and NU critically revised the manuscript. All authors read and approved the final manuscript. Acknowledgments The authors would like to thank Editage (www.editage.jp) for English language editing. Funding No funding or grants were provided for this study. Availability of data and materials The dataset generated and/or analyzed during the current study is not publicly available due to identifiable patient information but is available from the corresponding author upon reasonable request. Ethics approval and consent to participate This study was conducted according to the principles outlined in the Declaration of Helsinki (as revised in Fortaleza, Brazil, October 2013), and these surveys were performed with the approval of the Ethics Committee of Kagawa University (permission number: 2020-070). The need for informed consent was Page 9/11 Page 9/11 waived by the Ethics Committee, Faculty of Medicine, Kagawa University because of the retrospective nature of this study. waived by the Ethics Committee, Faculty of Medicine, Kagawa University because of the retrospective nature of this study. Competing interests The authors declare that they have no competing interest References 1. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of incidence and mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71:209–49. 1. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of incidence and mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71:209–49. 2. Abraham J, Gulley JL. The Bethesda Handbook of Clinical Oncology. 5th ed. Philadelphia: Wolters Kluwer Health; 2019. 2. Abraham J, Gulley JL. The Bethesda Handbook of Clinical Oncology. 5th ed. Philadelphia: Wolters Kluwer Health; 2019. 3. Babjuk M, Burger M, Capoun O, Cohen D, Compérat EM, Dominguez Escrig JL, et al. European Association of Urology guidelines on non-muscle-invasive bladder cancer (ta, T1, and carcinoma in situ). Eur Urol. 2022;81:75–94. 3. Babjuk M, Burger M, Capoun O, Cohen D, Compérat EM, Dominguez Escrig JL, et al. European Association of Urology guidelines on non-muscle-invasive bladder cancer (ta, T1, and carcinoma in situ). Eur Urol. 2022;81:75–94. Page 10/11 4. Eble JN, Sauter G, Epstein JI. World Health Organization classification of tumours. Pathology and Genetics of Tumours of the Urinary System and Male Genital Organs. Lyon: IARC Press, 2004. Page 10/11 4. Eble JN, Sauter G, Epstein JI. World Health Organization classification of tumours. Pathology and Genetics of Tumours of the Urinary System and Male Genital Organs. Lyon: IARC Press, 2004. 5. Miyamoto H, Miller JS, Fajardo DA, Lee TK, Netto GJ, Epstein JI. Non-invasive papillary urothelial neoplasms: The 2004 WHO/ISUP classification system. Pathol Int. WHO. 2010;60:1–8. 6. Kamat AM, Hahn NM, Efstathiou JA, Lerner SP, Malmström PU, Choi W, et al. Bladder cancer. Lancet. 2016;388:2796–810. 7. Knowles MA, Hurst CD. Molecular biology of bladder cancer: New insights into pathogenesis and clinical diversity. Nat Rev Cancer. 2015;15:25–41. 8. Brown FM. Urine cytology. It is still the gold standard for screening? Urol Clin North Am. 2000;27:25– 37. 9. Rosenthal DL, Wojcik EM, Kurtycz DFI. The Paris System for Reporting Urinary Cytology. New York: Springer; 2016. 10. Cowan ML, Rosenthal DL, VandenBussche CJ. Improved risk stratification for patients with high- grade urothelial carcinoma following application of the Paris System for Reporting Urinary Cytology. Cancer Cytopathol. 2017;125:427–34. 11. Meilleroux J, Daniel G, Aziza J, d’Aure DM, Quintyn-Ranty ML, Basset CM, et al. References One year of experience using the Paris System for Reporting Urinary Cytology. Cancer Cytopathol. 2018;126:430– 6. 12. VandenBussche CJ. A review of the Paris system for reporting urinary cytology. Cytopathology. 2016;27:153–6. 13. Jung A, Nielsen ME, Crandell JL, Palmer MH, Bryant AL, Smith SK, et al. Quality of life in non-muscle- invasive bladder cancer survivors: A systematic review. Cancer Nurs. 2019;42:E21–33. 13. Jung A, Nielsen ME, Crandell JL, Palmer MH, Bryant AL, Smith SK, et al. Quality of life in non-muscle- invasive bladder cancer survivors: A systematic review. Cancer Nurs. 2019;42:E21–33. 14. Sullivan NM, Sutter VL, Mims MM, Marsh VH, Finegold SM. Clinical aspects of bacteremia after manipulation of the genitourinary tract. J Infect Dis. 1973;127:49–55. 14. Sullivan NM, Sutter VL, Mims MM, Marsh VH, Finegold SM. Clinical aspects of bacteremia after manipulation of the genitourinary tract. J Infect Dis. 1973;127:49–55. 15. Rai S, Lali BS, Venkataramana CG, Philipose CS, Rao R, Prabhu GL. A quest for accuracy: Evaluation of the Paris system in diagnosis of urothelial carcinomas. J Cytol. 2019;36:169–73. 15. Rai S, Lali BS, Venkataramana CG, Philipose CS, Rao R, Prabhu GL. A quest for accuracy: Evaluation of the Paris system in diagnosis of urothelial carcinomas. J Cytol. 2019;36:169–73. 16. Stanzione N, Ahmed T, Fung PC, Cai D, Lu DY, Sumida LC, et al. The continual impact of the Paris System on urine cytology, a 3-year experience. Cytopathology. 2020;31:35–40. 16. Stanzione N, Ahmed T, Fung PC, Cai D, Lu DY, Sumida LC, et al. The continual impact of the Paris System on urine cytology, a 3-year experience. Cytopathology. 2020;31:35–40. 17. Lee PJ, Owens CL, Lithgow MY, Jiang Z, Fischer AH. Causes of false-negative for high-grade urothelial carcinoma in urine cytology. Diagn Cytopathol. 2016;44:994–9. 17. Lee PJ, Owens CL, Lithgow MY, Jiang Z, Fischer AH. Causes of false-negative for high-grade urothelial carcinoma in urine cytology. Diagn Cytopathol. 2016;44:994–9. 18. VandenBussche CJ, Rosenthal DL, Olson MT. Adequacy in voided urine cytology specimens: The role of volume and a repeat void upon predictive values for high-grade urothelial carcinoma. Cancer Cytopathol. 2016;124:174–80. 18. VandenBussche CJ, Rosenthal DL, Olson MT. Adequacy in voided urine cytology specimens: The role of volume and a repeat void upon predictive values for high-grade urothelial carcinoma. Cancer Cytopathol. 2016;124:174–80. 19. Olson MT, Boonyaarunnate T, Aragon Han P, Umbricht CB, Ali SZ, Zeiger MA. A tertiary center's experience with second review of 3885 thyroid cytopathology specimens. References J Clin Endocrinol Metab. 2013;98:1450–7. Page 11/11
https://openalex.org/W4287547786	https://zenodo.org/records/4395793/files/7_semanTIK%20Vol%206%20no%202%20Hal43_50_handrizal.pdf	Indonesian	null	REVIEW EASEUS DATA RECOVERY WIZARD UNTUK DIGITAL FORENSIK	Zenodo (CERN European Organization for Nuclear Research)	2,020	cc-by	3,707	semanTIK, Vol.6, No.2, Jul-Des 2020, pp. 43-50 ISSN: 2502-8928 (Online) DOI : 10.5281/zenodo.4395793 Terakreditasi “Peringkat 4 (Sinta 4)” oleh Kemenristekdikti Website : http://ojs.uho.ac.id/index.php/semantik ◼ 43 semanTIK, Vol.6, No.2, Jul-Des 2020, pp. 43-50 ISSN: 2502-8928 (Online) DOI : 10.5281/zenodo.4395793 Terakreditasi “Peringkat 4 (Sinta 4)” oleh Kemenristekdikti Website : http://ojs.uho.ac.id/index.php/semantik ◼ 43 semanTIK, Vol.6, No.2, Jul-Des 2020, pp. 43-50 ISSN: 2502-8928 (Online) DOI : 10.5281/zenodo.4395793 Terakreditasi “Peringkat 4 (Sinta 4)” oleh Kemenristekdikti Website : http://ojs.uho.ac.id/index.php/semantik ◼ 43 semanTIK, Vol.6, No.2, Jul-Des 2020, pp. 43-50 ISSN: 2502-8928 (Online) DOI : 10.5281/zenodo.4395793 Terakreditasi “Peringkat 4 (Sinta 4)” oleh Kemenristekdikti Website : http://ojs.uho.ac.id/index.php/semantik ◼ 43 ◼ 43 43 Abstract This paper presents a review of an application called EaseUS Data Recovery Wizard used for recovering deleted data. This research was conducted to see the ability of this application in recovering deleted data on a USB flash drive and using the National Institute of Standard and Technology (NIST) method. The results of testing indicate that this application can work well in terms of finding data that has been deleted but is less reliable in recovering data that has been deleted. Test results on ten data of which only four (40%) data were recovered. Keywords: Data, Recovery, Forensic, EaseUS Abstrak Makalah ini menyajikan review sebuah aplikasi yang bernama EaseUS Data Recovery Wizard yang digunakan untuk pemulihan data yang sudah dihapus. Penelitian ini dilakukan untuk melihat kemampuan aplikasi ini dalam pemulihan data yang dihapus pada sebuah USB flash drive dan menggunakan metode National Institute of Standard and Technology (NIST). Hasil dari pengujian menunjukkan bahwa aplikasi ini dapat berkerja dengan baik dalam hal menemukan semua data yang sudah dihapus, akan tetapi kurang handal dalam memulihkan data yang sudah dihapus tersebut. Hasil pengujian terhadap sepuluh data yang hanya empat(40%) data yang berhasil di pulihkan. Kata kunci: Data, Recovery, Forensik, EaseUS y pp REVIEW EASEUS DATA RECOVERY WIZARD UNTUK DIGITAL FORENSIK Handrizal Program Studi S1 Ilmu Komputer Fakultas Ilmu Komputer dan Teknologi Informasi Universitas Sumatera Utara e-mail: Handrizal@usu.ac.id Keywords: Data, Recovery, Forensic, EaseUS 1.2. Digital Forensik Menurut Lazaridis[3], Digital forensik adalah ilmu yang membahas penemuan, validasi dan interpretasi bukti digital yang ditemukan pada perangkat elektronik yang sesuai dengan kejatahan komputer. Sedangkan menurut Sulianta[4], Digital forensik adalah pengaplikasian ilmu pengetahuan dalam mengidentifikasi, mengumpulkan, menguji, dan menganalisis data, kemudian menghadirkan informasi yang dapat diandalkan[5]. 2.1 Data yang digunakan Dalam penelitian ini penulis menggunakan data milik penulis sendiri, data tersebut terdiri file umum (pdf, docx, ppt dan lain-lain), seperti terlihat pada Tabel 1. 2.1 Data yang digunakan Dalam penelitian ini penulis menggunakan data milik penulis sendiri, data tersebut terdiri file umum (pdf, docx, ppt dan lain-lain), seperti terlihat pada Tabel 1. Tabel 1. Data yang digunakan No. Nama File Ekstensi File Ukuran File (kb) 1. abc .docx 13 2. abcd .doc 22 3. abcde .pdf 81 4. abcdef .html 21 5. abcdefg .xml 47 6. abcdefgh .pptx 34 7. abcdefghi .ppt 100 8. abcdefghij .xls 9 9. abcdefghijk .xps 87 10. abcdefghijkl .txt 1 Tabel 1. Data yang digunakan 1.3 EaseUS Data Recovery Wizard EaseUS Data Recovery Wizard merupakan software recovery file gratis untuk sistem operasi Windows dan Mac yang digunakan untuk memulihkan file yang dihapus oleh pengguna dari hard drive internal dan ekternal, serta perangkat USB, kartu memori, pemutar musik dan perangkat sejenisnya[6]. 1.1 Data Recovery Menurut Juju[1], Data recovery adalah proses penyelamatan data yang rusak, tidak bisa diakses, atau terformat dari media penyimpanan. Karena fungsinya adalah untuk menggembalikan data yang hilang maka proses data recovery ini bisa digunakan dalam konteks komputer forensik atau untuk mata- mata[2]. c) Analysis Tahap ini adalah proses menganalisis dan mengevaluasi kembali data yang ditemukan dari hasil examination. c) Analysis Tahap ini adalah proses menganalisis dan mengevaluasi kembali data yang ditemukan dari hasil examination. d) Reporting Tahap reporting merupakan proses pelaporan hasil analisis dari data yang ditemukan. d) Reporting Tahap reporting merupakan proses pelaporan hasil analisis dari data yang ditemukan. 1. PENDAHULUAN sasarannya adalah akses ilegal (mengelabui kontrol akses), malware dan serangan DoS. Contoh kejahatan cybercrime di mana komputer sebagai tempatnya adalah penipuan identitas. Sedangkan contoh kejahatan tradisional dengan komputer sebagai alatnya adalah pornografi anak dan judi online. S alah satu dampak negatif yang timbulkan dari kemajuan teknologi adalah penyalahgunaan teknologi tersebut untuk kejahatan. Kejahatan yang berkaitan dengan penggunaan komputer pada media tersebut biasanya dikenal dengan nama cybercrime. Istilah ini juga digunakan untuk kegiatan kejahatan tradisional di mana komputer atau jaringan komputer digunakan untuk mempermudah atau memungkinkan kejahatan itu terjadi. Contoh kejahatan cybercrime di mana komputer sebagai alat adalah spamming dan kejahatan terhadap hak cipta serta kekayaan intelektual. Contoh kejahatan cybercrime di mana komputer sebagai S Perilaku cybercrime sudah tentu sangat merugikan korbannya dan bertentangan dengan hukum. Untuk memberi hukuman kepada pelaku cybercrime ini pihak berwajib biasanya akan mencari beberapa alat bukti. Salah satu alat bukti adalah komputer yang digunakan oleh pelaku. Data yang ada didalam komputer akan diambil sebagai alat bukti dalam menghukum pelaku cybercrime. Dalam prakteknya data didalam komputer tersebut Review EaseUS Data Recovery Wizard untuk Digital Forensik 44 0 digital. Pada penelitian ini penulis menggunakan sebuah perangkat keras flash disk Kingston 512 MB. sudah dihapus oleh pelaku sebelum komputer tersebut disita oleh pihak berwajib. Dalam hal inilah diperlukan aplikasi untuk mencari data yang sudah dihapus tersebut. Proses pencarian data yang sudah dihapus ini disebut dengan istilah digital forensik. b) Examination Examination merupakan proses pengambilan data pada barang bukti menggunakan tool forensik. Pada penelitian ini penulis menggunakan tool forensik Easeus Data Recovery Wizard b) Examination Examination merupakan proses pengambilan data pada barang bukti menggunakan tool forensik. Pada penelitian ini penulis menggunakan tool forensik Easeus Data Recovery Wizard 3.1. Implementasi Penerapan aplikasi EaseUS Data Recovery Wizard dilakukan pada sistem operasi Windows 7. Aplikasi ini adalah software yang bisa di download secara gratis. Setelah software tersebut di download kemudian diinstall, tampilan awal untuk aplikasi EaseUS Data Recovery Wizard seperti pada Gambar 1. g py 4. Hapus semua data didalam flash drive, seperti terlihat pada Gambar 4. g py 4. Hapus semua data didalam flash drive, seperti terlihat pada Gambar 4. Gambar 1. Tampilan EaseUS Data Recovery Wizard seperti terlihat pada Gambar 4. Gambar 1. Tampilan EaseUS Data Recovery Wizard Gambar 4. Proses menghapus semua file Gambar 4. Proses menghapus semua file 5. Kosongkan recycle bin, seperti terlihat pada Gambar 5. 3. HASIL DAN PEMBAHASAN 3. Copy sepuluh buah file dari drive D ke flash drive, seperti terlihat pada Gambar 3. Gambar 3. Proses mengcopy file 4. Hapus semua data didalam flash drive, seperti terlihat pada Gambar 4. Gambar 4. Proses menghapus semua file 5. Kosongkan recycle bin, seperti terlihat pada Gambar 5. Gambar 5. Proses penghapus recycle bin 3. Copy sepuluh buah file dari drive D ke flash drive, seperti terlihat pada Gambar 3. flash drive, seperti terlihat pada Gambar 3. 3.2. Pengujian Pengujian aplikasi ini dilakukan untuk mengetahui bagaimana kinerja aplikasi dalam pencarian data yang sudah dihapus didalam sebuah flash drive. Dalam pengujian ini akan dilihat hasilnya berdasarkan banyaknya jumlah data yang dapat discan dan jumlah data yang dapat dipulihkan. p Gambar 5. Proses penghapus recycle bin Gambar 5. Proses penghapus recycle bin 3.3. Pengujian EaseUS Data Recovery Wizard Untuk pengujian dengan EaseUS Data Recovery Wizard dilakukan dengan National Institute of Standard and Technology (NIST) 3.3. Pengujian EaseUS Data Recovery Wizard Untuk pengujian dengan EaseUS Data Recovery Wizard dilakukan dengan National Institute of Standard and Technology (NIST) Gambar 5. Proses penghapus recycle bin 2. METODE PENELITIAN Metodologi yang digunakan untuk penelitian ini adalah National Institute of Standard and Technology (NIST) yang dibagi menjadi empat tahap yakni, Collection, Examination, Analysis, dan Reporting[7], Masing-masing fase yang berbeda ini dijelaskan lebih lanjut: a) Collection Collection merupakan proses identifikasi barang bukti yang digunakan berupa perangkat keras yang akan diambil datanya untuk digunakan sebagai bukti Handrizal 45 ◼ 3. Copy sepuluh buah file dari drive D ke flash drive, seperti terlihat pada Gambar 3. Gambar 3. Proses mengcopy file 4. Hapus semua data didalam flash drive, seperti terlihat pada Gambar 4. Gambar 4. Proses menghapus semua file 3.3.2 Examination Pada tahap ini dilakukan dengan langkah-langkah berikut ini: Pada tahap ini dilakukan dengan langkah-langkah berikut ini: 3.3.1 Collection Pada tahap ini dilakukan dengan langkah-langkah berikut ini: 3.3.1 Collection Pada tahap ini dilakukan dengan langkah-langkah berikut ini: 3.3.1 Collection Pada tahap ini dilakukan dengan langkah-langkah berikut ini: 1. Jalankan aplikasi EaseUS Data Recovery Wizard. Pada langkah ini akan didapatkan tampilan pada layar aplikasi seperti Gambar 6 1. Masukkan USB flash drive ke port USB. 2. Format USB flash drive, seperti terlihat pada Gambar 2. pada Gambar 2. Gambar 2. Format USB flash drive pada Gambar 2. Gambar 2. Format USB flash drive Gambar 6. Gambar 6. Proses scanning semua drive Gambar 6. Gambar 6. Proses scanning semua drive Gambar 2. Format USB flash drive Gambar 6. Proses scanning semua drive Review EaseUS Data Recovery Wizard untuk Digital Forensik 46 ISSN: 197 2. Untuk memilih drive tertentu, klik pada drive yang di inginkan, dalam uji coba ini di pilih drive dengan nama AAN(F), kemudian akan tampil seperti Gambar 7. 20 6. Langkah selanjutnya beri tanda ceklist pada file yang akan direcovery, seperti pada Gambar 10. 20 6. Langkah selanjutnya beri tanda ceklist pada file yang akan direcovery, seperti pada Gambar 10. Gambar 10. Pemilihan file yang akan direcovery Gambar 10. Pemilihan file yang akan direcovery Gambar 7. Tampilan proses scan drive F Gambar 7. Tampilan proses scan drive F Gambar 10. Pemilihan file yang akan direcovery Gambar 7. Tampilan proses scan drive F 7. Langkah selanjutnya klik “Recovery” dan tentukan tempat penyimpan file yang akan direcovery, seperti pada Gambar 11. Gambar 11. Proses recovery file 8. Setelah proses recovery selesai akan tampil seperti Gambar 12. Gambar 12. Proses recovery selesai 9. Setelah proses recovery selesai, langkah selanjutnya melihat file tersebut pada USB flash drive, seperti pada Gambar 13. 7. Langkah selanjutnya klik “Recovery” dan tentukan tempat penyimpan file yang akan direcovery, seperti pada Gambar 11. 3. Setelah proses scan drive selasai akan tampil seperti Gambar 8. Gambar 8. Hasil proses scan drive 3. Setelah proses scan drive selasai akan tampil seperti Gambar 8. y, p p Gambar 11. Proses recovery file Gambar 11. Proses recovery file 8. Setelah proses recovery selesai akan tampil seperti Gambar 12. Gambar 12. Proses recovery selesai 8. Setelah proses recovery selesai akan tampil seperti Gambar 12. 8. Setelah proses recovery selesai akan tampil seperti Gambar 12. 4. Pada gambar 8 terdapat tiga buah folder yaitu 5. Gambar 16. Isi file hasil recovery Pada Gambar 16 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 17. 3.3.2 Examination Untuk memilih folder tertentu, klik pada folder yang di inginkan, dalam uji coba ini di pilih folder dengan nama Delete Files, kemudian akan tampil seperti Gambar 9. Gambar 9. Tampilan isi folder Delete Files 5. Untuk memilih folder tertentu, klik pada folder yang di inginkan, dalam uji coba ini di pilih folder dengan nama Delete Files, kemudian akan tampil seperti Gambar 9. Gambar 12. Proses recovery selesai 9. Setelah proses recovery selesai, langkah selanjutnya melihat file tersebut pada USB flash drive, seperti pada Gambar 13. 9. Setelah proses recovery selesai, langkah selanjutnya melihat file tersebut pada USB flash drive, seperti pada Gambar 13. Gambar 9. Tampilan isi folder Delete Files ◼ 47 Handrizal Handrizal IJCCSISSN: 1978-1520 Gambar 13. File hasil recovery Gambar 13. File hasil recovery 3.3.3 Analysis Pada tahap ini dilakukan dengan langkah-langkah berikut ini: 1. Langkah selanjutnya melihat masing- masing isi file tersebut pada USB flash drive, dalam penelitian ini di coba melhat isi file yang pertama seperti pada Gambar 14. 14. Gambar 14. Isi file hasil recovery Gambar 17. Isi file hasil sebelum di delete 3. Langkah selanjutnya melihat isi file yang ketiga seperti pada Gambar 18. Gambar 18. Isi file hasil recovery Pada Gambar 18 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 19. Gambar 19. Isi file hasil sebelum di delete Gambar 17. Isi file hasil sebelum di delete 3. Langkah selanjutnya melihat isi file yang ketiga seperti pada Gambar 18. Gambar 18. Isi file hasil recovery Titl f i t i h t d l i li h lt (Fi t A th ) Gambar 14. Isi file hasil recovery Pada Gambar 14 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 15. Gambar 15. Isi file hasil sebelum di delete 2. Langkah selanjutnya melihat isi file yang kedua seperti pada Gambar 16. Gambar 17. Isi file hasil sebelum di delete 3. Langkah selanjutnya melihat isi file yang ketiga seperti pada Gambar 18. Gambar 18. Isi file hasil recovery Pada Gambar 18 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 19. Gambar 19. Isi file hasil sebelum di delete Gambar 17. Isi file hasil sebelum di delete f 3. Langkah selanjutnya melihat isi file yang ketiga seperti pada Gambar 18. ketiga seperti pada Gambar 18. Gambar 18. Isi file hasil recovery Gambar 14. Isi file hasil recovery Pada Gambar 14 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 15. Gambar 15. Isi file hasil sebelum di delete Gambar 18. Isi file hasil recovery Gambar 18. Isi file hasil recovery Pada Gambar 18 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 19. Pada Gambar 18 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 19. p Gambar 19. Isi file hasil sebelum di delete Gambar 15. Isi file hasil sebelum di delete 2. 3.3.3 Analysis Langkah selanjutnya melihat isi file yang kedua seperti pada Gambar 16. Gambar 19. Isi file hasil sebelum di delete ◼ Review ISSN: 1978-1 48 4. Langkah selanjutnya melihat isi file yang keempat seperti pada Gambar 20. Gambar 20. Isi file hasil recovery Pada Gambar 20 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 21. Gambar 21. Isi file hasil sebelum di delete 5. Langkah selanjutnya melihat isi file yang kelima seperti pada gambar 22. Gambar 22. Isi file hasil recovery Pada Gambar 22 dapat dilihat bahwa file tersebut bisa dibuka dan hasilnya sama dengan isi file aslinya seperti terlhat pada Gambar 23. Review EaseUS Data Recovery Wizard untuk Digital Forensik 48 20 Gambar 23. Isi file hasil sebelum di delete ISSN: 197 4. Langkah selanjutnya melihat isi file yang keempat seperti pada Gambar 20. ISSN: 197 4. Langkah selanjutnya melihat isi file yang keempat seperti pada Gambar 20. Gambar 20. Isi file hasil recovery Gambar 23. Isi file hasil sebelum di delete Gambar 23. Isi file hasil sebelum di delete 6. Langkah selanjutnya melihat isi file yang keenam seperti pada Gambar 24. Gambar 24. Isi file hasil recovery 6. Langkah selanjutnya melihat isi file yang keenam seperti pada Gambar 24. 6. Langkah selanjutnya melihat isi file yang keenam seperti pada Gambar 24. Gambar 20. Isi file hasil recovery Pada Gambar 20 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 21. p Gambar 21. Isi file hasil sebelum di delete Gambar 24. Isi file hasil recovery Gambar 24. Isi file hasil recovery Pada Gambar 24 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 25. Pada Gambar 24 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 25. Gambar 21. Isi file hasil sebelum di delete p Gambar 25. Isi file hasil sebelum di delete 7. Langkah selanjutnya melihat isi file yang ketujuh seperti pada Gambar 26. Gambar 26. Isi file hasil recovery Gambar 25. Isi file hasil sebelum di delete 5. Langkah selanjutnya melihat isi file yang kelima seperti pada gambar 22. 5. Langkah selanjutnya melihat isi file yang kelima seperti pada gambar 22. p p g Gambar 22. 3.3.3 Analysis Isi file hasil recovery Gambar 25. Isi file hasil sebelum di delete Gambar 25. Isi file hasil sebelum di delete Gambar 25. Isi file hasil sebelum di delete f 7. Langkah selanjutnya melihat isi file yang ketujuh seperti pada Gambar 26. 7. 7. Langkah selanjutnya melihat isi file yang ketujuh seperti pada Gambar 26. Gambar 26. Isi file hasil recovery Gambar 22. Isi file hasil recovery Pada Gambar 22 dapat dilihat bahwa file tersebut bisa dibuka dan hasilnya sama dengan isi file aslinya seperti terlhat pada Gambar 23. Gambar 26. Isi file hasil recovery Gambar 26. Isi file hasil recovery Handrizal IJCCSISSN: 1978-1520 ◼ Ti l f i i h d l i li h l (Fi A h ) 49 Pada Gambar 26 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 27. Gambar 27. Isi file hasil sebelum di delete 8. Langkah selanjutnya melihat isi file yang kedelapan seperti pada Gambar 28. Gambar 28. Isi file hasil recovery Pada Gambar 28 dapat dilihat bahwa file tersebut bisa dibuka dan hasilnya sam dengan isi file aslinya seperti terlhat pada Gambar 29. Gambar 29. Isi file hasil sebelum di delete 9. Langkah selanjutnya melihat isi file yang kesembilan seperti pada Gambar 30. Gambar 30. Isi file hasil recovery Pada Gambar 30 dapat dilihat bahwa file tersebut bisa dibuka dan hasilnya sama dengan isi file aslinya seperti terlhat pada Gambar 31. Gambar 31. Isi file hasil sebelum di delete 10. Langkah selanjutnya melihat isi file yang kesepuluh seperti pada Gambar 31. Gambar 32. Isi file hasil recovery Pada Gambar 32 dapat dilihat bahwa file tersebut bisa dibuka dan hasilnya sama dengan isi file aslinya seperti terlihat pada Gambar 33. Gambar 33. Isi file hasil sebelum di delete 49 ◼ Handrizal Pada Gambar 26 dapat dilihat bahwa file tersebut tidak bisa dibuka dan hasilnya tidak sesuai dengan isi file aslinya seperti terlhat pada Gambar 27. Gambar 30. Isi file hasil recovery p Gambar 27. Isi file hasil sebelum di delete Gambar 27. Isi file hasil sebelum di delete Gambar 30. Isi file hasil recovery Gambar 30. Isi file hasil recovery Pada Gambar 30 dapat dilihat bahwa file tersebut bisa dibuka dan hasilnya sama dengan isi file aslinya seperti terlhat pada Gambar 31. 4. KESIMPULAN Berdasarkan penelitian yang telah dilakukan dapat ambil kesimpulan, antara lain: [6] [6] T. Fisher, “A Full Review of EaseUS Data Recovery Wizard a free file undelete tool.” 2020, [Online]. Available: https://www.lifewire.com/easeus-data- recovery-wizard-review-2622879. 1. Aplikasi EaseUS Data Recovery Wizard dapat menemukan semua file yang sudah dihapus dalam sebuah flash drive dan sudah dikosong dari recycle bin. 2. Aplikasi EaseUS Data Recovery Wizard hanya dapat memulihkan 4 file (40%) yang sudah dihapus dalam sebuah flash drive. 2. Aplikasi EaseUS Data Recovery Wizard hanya dapat memulihkan 4 file (40%) yang sudah dihapus dalam sebuah flash drive. [7] I. Riadi, S. Sunardi, and Sahiruddin, “Perbandingan Tool Forensik Data Recovery Berbasis Android Menggunakan Metode NIST,” J. Teknol. Inf. dan Ilmu Komput., vol. 7, no. 1, pp. 197–204, 2020, doi: 10.25126/jtiik.202071921. 3.3.3 Analysis Pada Gambar 30 dapat dilihat bahwa file tersebut bisa dibuka dan hasilnya sama dengan isi file aslinya seperti terlhat pada Gambar 31. Gambar 27. Isi file hasil sebelum di delete y p p Gambar 31. Isi file hasil sebelum di delete y p p Gambar 31. Isi file hasil sebelum di delete 8. Langkah selanjutnya melihat isi file yang kedelapan seperti pada Gambar 28. Gambar 28. Isi file hasil recovery 8. Langkah selanjutnya melihat isi file yang kedelapan seperti pada Gambar 28. Gambar 31. Isi file hasil sebelum di delete Gambar 31. Isi file hasil sebelum di delete 10. Langkah selanjutnya melihat isi file yang kesepuluh seperti pada Gambar 31. 10. Langkah selanjutnya melihat isi file yang kesepuluh seperti pada Gambar 31. Gambar 28. Isi file hasil recovery p p p Gambar 32. Isi file hasil recovery Pada Gambar 28 dapat dilihat bahwa file tersebut bisa dibuka dan hasilnya sam dengan isi file aslinya seperti terlhat pada Gambar 29. Gambar 29. Isi file hasil sebelum di delete Gambar 32. Isi file hasil recovery Pada Gambar 32 dapat dilihat bahwa file tersebut bisa dibuka dan hasilnya sama dengan isi file aslinya seperti terlihat pada Gambar 33. Gambar 29. Isi file hasil sebelum di delete y p p Gambar 33. Isi file hasil sebelum di delete 9. Langkah selanjutnya melihat isi file yang kesembilan seperti pada Gambar 30. Gambar 33. Isi file hasil sebelum di delete Gambar 33. Isi file hasil sebelum di delete 50 Review EaseUS Data Recovery Wizard untuk Digital Forensik 3.3.4 Reporting Dari pengujian yang sudah dilakukan menggunakan USB flash drive, diperoleh hasil seperti terlihat pada Tabel 2. [1] D. Juju, “Data Recovery, Pulihkan Data dengan Tool Sederhana dan Mudah.” Elex Media Komputindo, 2008. Tabel 2. Hasil Pengujian EaseUS Data Recovery Wizard No Parameter Hasil 1 Jumlah data yang berhasil di scan 10 2 Jumlah data yang berhasil di recovery 4 Tabel 2. Hasil Pengujian EaseUS Data Recovery Wizard [2] B. Mathew, "File Data Recovery : PC Hard drive Data Recovery, USB Data Recovery, Mac Data Recovery, Android Data Recovery, Data Recovery Services". CreateSpace Independent Publishing Platform, 2014. [3] [3] I. Lazaridis, T. Arampatzis, and S. Pouros, “Evaluation of Digital Forensics Tools on Data Recovery and Analysis,” Third Int. Conf. Comput. Sci. Comput. Eng. Soc. Media, pp. 67– 71, 2016. Berdasarkan Tabel 2 diketahui bahwa aplikasi EaseUS Data Recovery Wizard yang digunakan dapat menemukan semua file yang sudah dihapus, tetapi hanya mampu memulihkan kembali 4 file (40%) yang sudah dihapus tersebut, hal ini kemungkinan terjadi karena perangkat flash disk yang digunakan sudah diformat dan data pada recycle bin juga dikosongkan, sehingga datanya menjadi susah untuk dipulihkan. [4] F. Sulianta, “Komputer Forensik.” PT Elex Media Komputindo, Jakarta, 2008. [5] [5] V. Singh, A. Tarannum, and V. Saran, “Efficiency of open source tools for Recovery of Unconventional deleted data : A Review,” no. August, pp. 19– 24, 2015. 5. SARAN 1. Untuk mengetahui lebih jauh mengenai kemampuan aplikasi ini, disarankan untuk melakukan pengujian terhadap media penyimpan yang lain. 2. Selain aplikasi yang sudah diuji dalam penelitian ini, masih banyak aplikasi yang lainnya. Untuk itu disarankan agar melakukan penelitian dengan menggunakan aplikasi yang lain.
https://openalex.org/W3040121687	https://www.nature.com/articles/s41467-020-17086-8.pdf	English	null	Molecular resolution imaging by post-labeling expansion single-molecule localization microscopy (Ex-SMLM)	Nature communications	2,020	cc-by	14,917	1 Department of Biotechnology and Biophysics, Biocenter, University of Würzburg, Am Hubland, 97074 Würzburg, Germany. 2 Department of Cell Biology, Sciences III, University of Geneva, Geneva, Switzerland. 3 School of Chemistry, Monash University, Clayton, VIC 3800, Australia. ✉email: virginie.hamel@unige.ch; paul.guichard@unige.ch; m.sauer@uni-wuerzburg.de ARTICLE Molecular resolution imaging by post-labeling expansion single-molecule localization microscopy (Ex-SMLM) Fabian U. Zwettler1, Sebastian Reinhard1, Davide Gambarotto 2, Toby D. M. Bell3, Virginie Hamel 2✉, Paul Guichard 2✉& Markus Sauer 1✉ Fabian U. Zwettler1, Sebastian Reinhard1, Davide Gambarotto 2, Toby D. M. Bell3, Virginie Hamel 2✉, Paul Guichard 2✉& Markus Sauer 1✉ Expansion microscopy (ExM) enables super-resolution ﬂuorescence imaging of physically expanded biological samples with conventional microscopes. By combining ExM with single- molecule localization microscopy (SMLM) it is potentially possible to approach the resolution of electron microscopy. However, current attempts to combine both methods remained challenging because of protein and ﬂuorophore loss during digestion or denaturation, gela- tion, and the incompatibility of expanded polyelectrolyte hydrogels with photoswitching buffers. Here we show that re-embedding of expanded hydrogels enables dSTORM imaging of expanded samples and demonstrate that post-labeling ExM resolves the current limitations of super-resolution microscopy. Using microtubules as a reference structure and centrioles, we demonstrate that post-labeling Ex-SMLM preserves ultrastructural details, improves the labeling efﬁciency and reduces the positional error arising from linking ﬂuorophores into the gel thus paving the way for super-resolution imaging of immunolabeled endogenous proteins with true molecular resolution. 1 Department of Biotechnology and Biophysics, Biocenter, University of Würzburg, Am Hubland, 97074 Würzburg, Germany. 2 Department of Cell Biology, Sciences III, University of Geneva, Geneva, Switzerland. 3 School of Chemistry, Monash University, Clayton, VIC 3800, Australia. ✉email: virginie.hamel@unige.ch; paul.guichard@unige.ch; m.sauer@uni-wuerzburg.de 1 NATURE COMMUNICATIONS \| (2020) 11:3388 \| https://doi.org/10.1038/s41467-020-17086-8 \| www.nature.com/naturecommunications ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 B y linking a ﬂuorophore or a protein of interest into a dense, cross-linked network of a swellable polyelectrolyte hydrogel, biological specimen can be physically expanded allowing for magniﬁed imaging with subdiffraction-resolution on conventional microscopes. Since its introduction in 20151, expansion microscopy (ExM) has shown impressive results including the magniﬁed visualization of pre- or postexpansion labeled proteins and RNAs with ﬂuorescent proteins (FPs), antibodies, and oligonucleotides, respectively, in isolated organelles, cells, pathogens, tissues, and human clinical specimen2–4. In addition, various protocols have been developed to anchor proteins or RNA into charged poly- acrylamide hydrogels5–8. Using 2.5% (w/w) acrylamide and 8.55% sodium acrylate with 0.15% (w/w) of the cross-linker N-N ′-methylenebisacrylamide accomplishes a ~4.5x linear expansion of biological specimens. Decreasing the cross-linker concentration usually permits higher gel expansion factors of up to 10x but also increases proportionally the linkage error deﬁned by the afﬁnity reagent, linker and ﬂuorophore and leads to greater gel instability9. Results Re-embedding of expanded samples enables Ex-SMLM in photoswitching buffer. A major problem of Ex-SMLM is the shrinking of the expanded hydrogels in photoswitching buffer due to ionic interactions between ions of the buffer and the ionic side groups of the gel. Therefore, we re-embedded expanded charged hydrogels in an uncharged polyacrylamide gel as recently introduced for ExM of RNA6. We started using pre- labeling ExM in combination with standard immunostaining using unmodiﬁed primary and ﬂuorophore labeled secondary antibodies to realize Ex-SMLM (Supplementary Fig. 1). We used microtubules as reference structure to investigate the expansion factor, spatial resolution, structural distortions, and the labeling density. Microtubules are assembled from αß tubulin hetero- dimers, which stack head-to-tail into polar protoﬁlaments with a periodicity of 8 nm, with ~13 protoﬁlaments associating laterally in parallel to form a hollow, polar cylinder (Fig. 1a)23,24. As previously measured by transmission electron microscopy (TEM), microtubules are hollow tubes with an outer diameter of 25 nm and 60 nm, respectively, after immunostaining with pri- mary and secondary antibodies22. This results in a linkage error deﬁned by the size of the primary and secondary antibody of 17.5 nm (Fig. 1a). y To further enhance the resolution, ExM has been combined with structured illumination microscopy (SIM)13,14 and stimu- lated emission depletion (STED) microscopy2,15. By careful optimization of the expansion protocol U-ExM demonstrated that even ultrastructural details of multiprotein complexes such as centrioles can be truthfully preserved2. Combining ExM with SMLM methods (Ex-SMLM) can then potentially further improve the spatial resolution to enable true molecular resolution and bridge the gap to the electron microscopy regime. However, despite these apparent advantages, attempts to combine ExM with SMLM have remained rare and unoptimized due to several challenges5,16. There are two major determinants that control the resolution of SMLM, the localization precision and the localiza- tion density11,12. The localization precision remains unaltered by sample expansion and therefore allows achieving an improved resolution depending on the expansion factor. The localization density is arguably the more important determinant for SMLM on expanded samples. According to information theory, the required density of ﬂuorescent probes has to be sufﬁciently high to satisfy the Nyquist–Shannon sampling theorem17. At its most basic level, the theorem states that the mean distance between neighboring localized ﬂuorophores (the sampling interval) must be at least twice as ﬁne as the desired resolution. Molecular resolution imaging by post-labeling expansion single-molecule localization microscopy (Ex-SMLM) It is also possible to expand samples in series enabling gel expansion factors of 20x and higher with a demonstration of 53x expansion of microtubules10. However, ﬂuorescence imaging of such greatly enlarged samples is complicated by the dilution of ﬂuorescent labels and dramatic increase in the physical separation between the ﬂuorophore and its target due to the linkage error. Nevertheless, ExM with lower expansion factors enables confocal diffraction- limited ﬂuorescence imaging with spatial resolutions comparable to that of super-resolution microscopy methods11,12. B y b to form polymers. However, free radicals also have the potential to react with the ﬂuorophores which can irreversibly destroy them1,3,5. Therefore, the ﬂuorophore density will be further diluted in ExM protocols that use pre-expansion labeling and consequently reduce the structural resolution. The extent of irreversible ﬂuorophore destruction during gelation varies across ﬂuorophores. Unfortunately, the best suited dyes for dSTORM, the carbocyanine dyes Cy5 and Alexa Fluor 64719,20, are almost completely destroyed during gelation1,3,5 Here, post-expansion labeling approaches (post-labeling ExM) offer acceptable solutions2,7,8, though they require preservation of protein epi- topes during expansion. An additional beneﬁt of post-labeling ExM is improved epitope accessibility for antibodies and a reduction of the linkage error proportional to the expansion factor compared to pre-labeling ExM2. For instance, after 4x expansion, the immunolabeling linkage error of 17.5 nm deﬁned by the primary and secondary antibodies22, would reduce to 4.4 nm, which is the size of a tubulin monomer23. Combining SMLM with post-labeling ExM reduces the linkage error by the expansion factor and could thus enable ﬂuorescence imaging with molecular resolution. Here, we set out to develop post-labeling Ex-SMLM with organic ﬂuor- ophores with minimal linkage error. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 The data were obtained from n=35 microtubule segments in 12 cells and 3 independent experiments. Scale bars, 2 µm (b, f), 500 nm (c, g). Fig. 1 Re-embedding enables Ex-dSTORM. a Model of microtubules with an outer diameter of 25 nm stained with conventional primary (pab) and ﬂuorescently labeled secondary IgG antibodies (sab) results in a total diameter of 60 nm with a linkage error (deﬁned by the size of the primary and secondary antibody) of 17.5 nm22. b dSTORM image of pre-labeled proExM expanded and re-embedded Cos-7 cells stained with primary antibodies against α-tubulin and secondary Alexa Fluor 532 conjugated antibodies (Al532). The small logo in the upper left corner symbolizes that microtubules have been immunolabeled before expansion (pre-labeled). c Zoom in on highlighted region in (b). d Averaged cross-sectional proﬁle of nine microtubule segments with a total length of 29.1 µm (segment lengths range from 2.1-4.5 µm) measured in two cells from 1 expanded sample. e Histogram of peak-to- peak distances with normalized normal distribution curve (red) determined by bi-Gaussian ﬁtting of the data analyzed in (c) with an average distance of 137.1 ± 10.1 nm (mean ± sd). The data were obtained from n = 9 microtubule segments in 2 cells from 1 expanded sample. f Unexpanded Cos-7 cells labeled with an anti α-tubulin primary antibody and Alexa Fluor 532 (Al532) conjugated IgG secondary antibodies. The small logo in the upper left corner symbolizes that microtubules have been immunolabeled and not expanded. g Zoom in of the white boxed region in (f). h Average intensity proﬁle of 35 microtubule segments with a length between 1.1 and 5.8 µm (mean = 2.0 µm) and a total length of 69.6 µm analyzed in 12 dSTORM images. i Histogram of peak-to-peak distances with normalized normal distribution curve (red) determined by bi-Gaussian ﬁtting of cross-sectional proﬁles of the analyzed microtubule segments in (h) with a mean peak-to-peak distance of 36.2 ± 5.4 nm (mean ± sd). The data were obtained from n=35 microtubule segments in 12 cells and 3 independent experiments. Scale bars, 2 µm (b, f), 500 nm (c, g). two-dimensional (2D) projection of the ﬂuorescence intensity distribution measured from microtubule ﬁlaments show a bimodal distribution, the peak-to-peak distance can then be ﬁtted with a sum of two Gaussians and used as an estimate of the spatial resolution. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 α-tubulin β-tubulin 8 nm sab sab pab pab a b c proExM AI532 137.1 ± 10.1 nm pre post 36.2 ± 5.4 nm Unexpanded AI532 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –400 –200 –200 400 0 Distance [nm] 120 130 140 150 160 Peak-to-peak distance [nm] 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –200 200 0 Distance [nm] 20 25 30 35 40 45 50 55 Peak-to-peak distance [nm] g f d e h i Fig. 1 Re-embedding enables Ex-dSTORM. a Model of microtubules with an outer diameter of 25 nm stained with conventional primary (pab) and ﬂuorescently labeled secondary IgG antibodies (sab) results in a total diameter of 60 nm with a linkage error (deﬁned by the size of the primary and secondary antibody) of 17.5 nm22. b dSTORM image of pre-labeled proExM expanded and re-embedded Cos-7 cells stained with primary antibodies against α-tubulin and secondary Alexa Fluor 532 conjugated antibodies (Al532). The small logo in the upper left corner symbolizes that microtubules have been immunolabeled before expansion (pre-labeled). c Zoom in on highlighted region in (b). d Averaged cross-sectional proﬁle of nine microtubule segments with a total length of 29.1 µm (segment lengths range from 2.1-4.5 µm) measured in two cells from 1 expanded sample. e Histogram of peak-to- peak distances with normalized normal distribution curve (red) determined by bi-Gaussian ﬁtting of the data analyzed in (c) with an average distance of 137.1 ± 10.1 nm (mean ± sd). The data were obtained from n = 9 microtubule segments in 2 cells from 1 expanded sample. f Unexpanded Cos-7 cells labeled with an anti α-tubulin primary antibody and Alexa Fluor 532 (Al532) conjugated IgG secondary antibodies. The small logo in the upper left corner symbolizes that microtubules have been immunolabeled and not expanded. g Zoom in of the white boxed region in (f). h Average intensity proﬁle of 35 microtubule segments with a length between 1.1 and 5.8 µm (mean = 2.0 µm) and a total length of 69.6 µm analyzed in 12 dSTORM images. i Histogram of peak-to-peak distances with normalized normal distribution curve (red) determined by bi-Gaussian ﬁtting of cross-sectional proﬁles of the analyzed microtubule segments in (h) with a mean peak-to-peak distance of 36.2 ± 5.4 nm (mean ± sd). The data were obtained from n=35 microtubule segments in 12 cells and 3 independent experiments. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 Scale bars, 2 µm (b, f), 500 nm (c, g). b pre post a c c proExM AI532 137.1 ± 10.1 nm 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –400 –200 –200 400 0 Distance [nm] 120 130 140 150 160 Peak-to-peak distance [nm] d e d f 36.2 ± 5.4 nm Unexpanded AI532 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –200 200 0 Distance [nm] 20 25 30 35 40 45 50 55 Peak-to-peak distance [nm] g h i g g 1 0h Fig. 1 Re-embedding enables Ex-dSTORM. a Model of microtubules with an outer diameter of 25 nm stained with conventional primary (pab) and ﬂuorescently labeled secondary IgG antibodies (sab) results in a total diameter of 60 nm with a linkage error (deﬁned by the size of the primary and secondary antibody) of 17.5 nm22. b dSTORM image of pre-labeled proExM expanded and re-embedded Cos-7 cells stained with primary antibodies against α-tubulin and secondary Alexa Fluor 532 conjugated antibodies (Al532). The small logo in the upper left corner symbolizes that microtubules have been immunolabeled before expansion (pre-labeled). c Zoom in on highlighted region in (b). d Averaged cross-sectional proﬁle of nine microtubule segments with a total length of 29.1 µm (segment lengths range from 2.1-4.5 µm) measured in two cells from 1 expanded sample. e Histogram of peak-to- peak distances with normalized normal distribution curve (red) determined by bi-Gaussian ﬁtting of the data analyzed in (c) with an average distance of 137.1 ± 10.1 nm (mean ± sd). The data were obtained from n = 9 microtubule segments in 2 cells from 1 expanded sample. f Unexpanded Cos-7 cells labeled with an anti α-tubulin primary antibody and Alexa Fluor 532 (Al532) conjugated IgG secondary antibodies. The small logo in the upper left corner symbolizes that microtubules have been immunolabeled and not expanded. g Zoom in of the white boxed region in (f). h Average intensity proﬁle of 35 microtubule segments with a length between 1.1 and 5.8 µm (mean = 2.0 µm) and a total length of 69.6 µm analyzed in 12 dSTORM images. i Histogram of peak-to-peak distances with normalized normal distribution curve (red) determined by bi-Gaussian ﬁtting of cross-sectional proﬁles of the analyzed microtubule segments in (h) with a mean peak-to-peak distance of 36.2 ± 5.4 nm (mean ± sd). Results In real samples, however, the relationship between localization density and reso- lution is far more complex18. Empirically, it seems that for a given resolution the distance between neighboring localizations should be signiﬁcantly less than that indicated by a naive application of the Nyquist limit19. g We used the proExM protocol, in which proteins are directly anchored to the hydrogel using the succinimidyl ester of 6- ((acryloyl)amino)hexanoic acid (AcX)5. To minimize ﬂuorophore loss during gelation in pre-labeling ExM methods, we used the rhodamine derivative Alexa Fluor 532, which retains ~50% of its pre-gelation brightness after expansion1,3,5. To prevent shrinking of the hydrogel upon addition of photoswitching buffer, expanded hydrogels were re-embedded in acrylamide for serial staining of the expanded specimen6. Hydrogels were incubated twice in 10% AA, 0.15% bis-acrylamide, 0.05% APS, 0.05% TEMED in 5 mM Tris (pH 8.9) for 30 min each and subsequently transferred onto coverslips functionalized with acrydite via glass silanization to minimize lateral drift of expanded samples (Supplementary Fig. 1). After polymerization of the re- embedding gel, hydrogels were immersed in photoswitching buffer containing 100 mM mercaptoethylamine (MEA) in PBS. The expansion factor was determined by comparing the post-expansion and post re-embedding ﬂuorescence images with These considerations illustrate the challenges Ex-SMLM is confronted with. First, the ﬂuorophore density is considerably diluted in expanded samples9,10, which often results in unclear views of biological structures and complicates SMLM data interpretation. For example, a 4x expansion in three dimensions effectively lowers the labeling density 64-fold. Second, addition of a thiol-containing phosphate-buffered saline (PBS) photoswitch- ing buffer as required for dSTORM20,21 to a swellable polyelec- trolyte hydrogel with hydrophilic ionic side groups results in substantial shrinking of the gel in the worst case down to its initial size. Finally, ExM protocols use free-radical polymerization URE COMMUNICATIONS \| (2020) 11:3388 \| https://doi.org/10.1038/s41467-020-17086-8 \| www.nature.com/naturecommunications 2 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 NATURE COMMUNICATIONS \| (2020) 11:3388 \| https://doi.org/10.1038/s41467-020-17086-8 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 To ensure an objective evaluation and comparison of the spatial resolution achieved, we developed “Line Proﬁler”, an automated image processing software. Line Proﬁler automatically evaluates potential regions of interest along ﬁlamentous structures to generate cross-sectional proﬁles that can be ﬁt by a sum of two Gaussians to determine the peak-to-peak distance between the sidewalls of the ﬁlamentous structure (Supplementary Fig. 4). pre-expansion ﬂuorescence images. The results showed a low distortion introduced by the re-embedding process and a reduction in gel size of ~20% from ~3.9x before to ~3.1x after re-embedding (Supplementary Figs. 2 and 3). g pp y g A caveat of imaging expanded samples is that super-resolution imaging methods, and in particular SMLM, are most effective when used on thin samples located within a few micrometers above the coverslip surface. However, expanded specimen can be easily located several tens of micrometers above the coverslip. In addition, expanded specimens are transparent because they consist largely of water. Hence, the use of oil-immersion objectives and total internal reﬂection ﬂuorescence (TIRF) microscopy as used in most SMLM applications to achieve a higher signal-to-background ratio is in this case not the best choice. Therefore, we decided to use a water- immersion objective and epi-illumination in all SMLM experi- ments. The corresponding dSTORM images of pre-labeled expanded microtubules showed homogeneously labeled ﬁlaments with some labeling gaps reﬂecting ﬂuorophore and protein loss during polymerization and enzymatic digestion, respectively (Fig. 1b, c). In addition, we imaged unexpanded microtubules by dSTORM under identical experimental conditions (Fig. 1f, g). In order to compare the experimentally measured peak-to-peak distances of different expansion protocols, we simulated trans- verse proﬁles of unexpanded and expanded microtubules using a cylindrical distribution function to describe the hollow annular structure of microtubules (Fig. 2a and Supplementary Fig. 5) similar to the approach used recently for iterative expansion10. The resulting peak-to-peak distances were used to determine the molecular expansion factor of expanded immunolabeled microtubules considering the inﬂuence of the label size on the broadening of the microtubule diameter (Supplementary Table 1 and Supplementary Note 1). To examine the achieved spatial resolution, cross-sectional proﬁles of selected microtubule areas are often consulted21. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 If the 3 3 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 a b c g pre post pre post Antibodies unexpanded (32.0 nm) Antibodies pre-labeling (proExM/GA) (143.0 nm) DNA Cy5 unexpanded (41.5 nm) DNA Cy5 expanded (202.0 nm) DNA AI532 expanded (226.5 nm) 1.0 0.5 0.0 –150 –100 –50 0 Distance [nm] Intensity [a.u.] 50 100 150 f j k 133.8 ± 13.2 nm 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –400 –200 –200 400 0 Distance [nm] 120 110 100 130 140 150 170 160 Peak-to-peak distance [nm] d e 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –400 –200 –200 400 0 Distance [nm] 195 180 165 210 225 240 Peak-to-peak distance [nm] 133.8 ± 13.2 nm 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –200 200 0 Distance [nm] Peak-to-peak distance [nm] h i Antibody-AI532 (GA) Unexpanded DNA-Cy5 DNA-Cy5 DNA-AI532 30 35 40 45 50 55 o n 201.0 ± 12.9 nm p q 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –400 –200 –200 400 0 Distance [nm] 210 195 180 225 240 255 270 Peak-to-peak distance [nm] 226.7 ± 15.3 nm l m Fig. 2 Pre-labeling Ex-dSTORM. a Simulated intensity proﬁles using a cylindrical distribution function to describe unexpanded or 3.2x expanded immunostained microtubules (labeled with IgG antibodies or DNA modiﬁed IgG antibodies pre-expansion) and resulting peak-to-peak distances of th cross-sectional proﬁles. b dSTORM image of expanded and re-embedded α- and β-tubulin pre-labeled with secondary Alexa Fluor 532 IgG antibodies (Al532) using the MA-NHS/GA method6, i.e. antibodies are cross-linked with glutaraldehyde (GA) into the hydrogel (Antibody-Al532 (GA)). c Zoom in white boxed region in (b). d Averaged cross-sectional proﬁle of 8 microtubule segments with a length between 1.5–6.4 µm and 28.6 µm in total measur in 4 expanded cells. e Histogram of peak-to-peak distance distribution with normalized normal curve (red) of microtubule segments analyzed in (d) at n 8 microtubule segments in 4 cells from 1 expansion experiment with a mean distance of 133.8 ± 13.2 nm (mean ± sd). f Unexpanded dSTORM image ssDNA-Cy5 secondary antibody hybridized with Cy5 bearing oligonucleotides pre-expansion (DNA-Cy5 protocol). g Magniﬁed view of white boxed regi in (f). h Average cross-sectional proﬁle of 7 microtubule segments with a length between 1.0–1.8 µm and 8.7 µm in total. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 d Averaged cross-sectional proﬁle of 8 microtubule segments with a length between 1.5–6.4 µm and 28.6 µm in total measured in 4 expanded cells. e Histogram of peak-to-peak distance distribution with normalized normal curve (red) of microtubule segments analyzed in (d) at n = 8 microtubule segments in 4 cells from 1 expansion experiment with a mean distance of 133.8 ± 13.2 nm (mean ± sd). f Unexpanded dSTORM image of ssDNA-Cy5 secondary antibody hybridized with Cy5 bearing oligonucleotides pre-expansion (DNA-Cy5 protocol). g Magniﬁed view of white boxed region in (f). h Average cross-sectional proﬁle of 7 microtubule segments with a length between 1.0–1.8 µm and 8.7 µm in total. i Histogram of peak-to-peak distances with normalized normal distribution curve (red) of the data analyzed in (h) along n = 7 microtubule segments in 2 cells from 1 experiment with a mean distance of 43.9 ± 3.7 nm (mean ± sd). j Expanded dSTORM image of microtubules labeled with α-tubulin and dsDNA (DNA-Al532) conjugated secondary antibodies exhibiting a methacryloyl group to crosslink the DNA with ﬂuorophores pre-expansion into the hydrogel (original ExM trifunctional label concept)1. k Zoom-in of white boxed region in (j). l Average intensity proﬁle of 26 microtubule segments with a length of 2.4–10.7 µm and 118.6 µm in total. m Histogram of peak-to-peak distances with normalized normal distribution curve (red) determined from n = 26 microtubule segments in 4 cells from 1 expanded sample showing a mean distance of 226.7 ± 15.3 nm (mean ± sd). n dSTORM image of α- and β-tubulin expanded according to the DNA- Cy5 protocol strategy with labels at Cy5-bearing oligonucleotides introduced post-re-embedding. o Zoom in of white boxed region in (n). p Average intensity proﬁle of 15 microtubule segments with a length between 1.6–25.1 µm and a total length of 126.0 µm in 1 expanded sample. q Histogram of peak- to-peak distances with normalized normal distribution curve (red) determined by ﬁtting the cross-sectional proﬁles analyzed in (p) along n = 22 microtubule segments in 4 cells from 1 expanded sample showing a mean distance of 201.0 ± 12.9 nm (mean ± sd). The small logos in the upper left corner symbolize the labeling method, e.g. pre- and post-immunolabeled with or without DNA-linker, respectively. Scale bars, 2 µm (b, f, j, n), 500 nm (c, g, k, o). Pre-labeling Ex-SMLM. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 h Average cross-sectional proﬁle of 7 microtubule segments with a length between 1.0–1.8 µm and 8.7 µm in total. i Histogram of peak-to-peak distances with normalized normal distribution curve (red) of the data analyzed in (h) along n = 7 microtubule segments in 2 cells from 1 experiment with a mean distance of 43.9 ± 3.7 nm (mean ± sd). j Expanded dSTORM image of microtubules labeled with α-tubulin and dsDNA (DNA-Al532) conjugated secondary antibodies exhibiting a methacryloyl group to crosslink the DNA with ﬂuorophores pre-expansion into the hydrogel (original ExM trifunctional label concept)1. k Zoom-in of white boxed region in (j). l Average intensity proﬁle of 26 microtubule segments with a length of 2.4–10.7 µm and 118.6 µm in total. m Histogram of peak-to-peak distances with normalized normal distribution curve (red) determined from n = 26 microtubule segments in 4 cells from 1 expanded sample showing a mean distance of 226.7 ± 15.3 nm (mean ± sd). n dSTORM image of α- and β-tubulin expanded according to the DNA- Cy5 protocol strategy with labels at Cy5-bearing oligonucleotides introduced post-re-embedding. o Zoom in of white boxed region in (n). p Average intensity proﬁle of 15 microtubule segments with a length between 1.6–25.1 µm and a total length of 126.0 µm in 1 expanded sample. q Histogram of peak- to-peak distances with normalized normal distribution curve (red) determined by ﬁtting the cross-sectional proﬁles analyzed in (p) along n = 22 microtubule segments in 4 cells from 1 expanded sample showing a mean distance of 201.0 ± 12.9 nm (mean ± sd). The small logos in the upper left corner symbolize the labeling method, e.g. pre- and post-immunolabeled with or without DNA-linker, respectively. Scale bars, 2 µm (b, f, j, n), 500 nm (c, g, k, o). Fig. 2 Pre-labeling Ex-dSTORM. a Simulated intensity proﬁles using a cylindrical distribution function to describe unexpanded or 3.2x expanded immunostained microtubules (labeled with IgG antibodies or DNA modiﬁed IgG antibodies pre-expansion) and resulting peak-to-peak distances of the cross-sectional proﬁles. b dSTORM image of expanded and re-embedded α- and β-tubulin pre-labeled with secondary Alexa Fluor 532 IgG antibodies (Al532) using the MA-NHS/GA method6, i.e. antibodies are cross-linked with glutaraldehyde (GA) into the hydrogel (Antibody-Al532 (GA)). c Zoom in of white boxed region in (b). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 i Histogram of peak-to-pea distances with normalized normal distribution curve (red) of the data analyzed in (h) along n = 7 microtubule segments in 2 cells from 1 experiment with mean distance of 43.9 ± 3.7 nm (mean ± sd). j Expanded dSTORM image of microtubules labeled with α-tubulin and dsDNA (DNA-Al532) conjugated secondary antibodies exhibiting a methacryloyl group to crosslink the DNA with ﬂuorophores pre-expansion into the hydrogel (original ExM trifunction label concept)1. k Zoom-in of white boxed region in (j). l Average intensity proﬁle of 26 microtubule segments with a length of 2.4–10.7 µm and 118.6 µm total. m Histogram of peak-to-peak distances with normalized normal distribution curve (red) determined from n = 26 microtubule segments in 4 ce from 1 expanded sample showing a mean distance of 226.7 ± 15.3 nm (mean ± sd). n dSTORM image of α- and β-tubulin expanded according to the DN Cy5 protocol strategy with labels at Cy5-bearing oligonucleotides introduced post-re-embedding. o Zoom in of white boxed region in (n). p Average intensity proﬁle of 15 microtubule segments with a length between 1.6–25.1 µm and a total length of 126.0 µm in 1 expanded sample. q Histogram of pea to-peak distances with normalized normal distribution curve (red) determined by ﬁtting the cross-sectional proﬁles analyzed in (p) along n = 22 microtubule segments in 4 cells from 1 expanded sample showing a mean distance of 201.0 ± 12.9 nm (mean ± sd). The small logos in the upper left corn symbolize the labeling method, e.g. pre- and post-immunolabeled with or without DNA-linker, respectively. NATURE COMMUNICATIONS \| (2020) 11:3388 \| https://doi.org/10.1038/s41467-020-17086-8 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 Scale bars, 2 µm (b, f, j, n), 500 nm (c, g, k, o b pre post a Antibodies unexpanded (32.0 nm) Antibodies pre-labeling (proExM/GA) (143.0 nm) DNA Cy5 unexpanded (41.5 nm) DNA Cy5 expanded (202.0 nm) DNA AI532 expanded (226.5 nm) 1.0 0.5 0.0 –150 –100 –50 0 Distance [nm] Intensity [a.u.] 50 100 150 c 133.8 ± 13.2 nm 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –400 –200 –200 400 0 Distance [nm] 120 110 100 130 140 150 170 160 Peak-to-peak distance [nm] d e d 1.0 0.5 0.0 –150 –100 –50 0 Distance [nm] Intensity [a.u.] 50 100 150 f g 133.8 ± 13.2 nm 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –200 200 0 Distance [nm] Peak-to-peak distance [nm] h i Unexpanded DNA-Cy5 30 35 40 45 50 55 g 1.0h pre post j k DNA-AI532 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –400 –200 –200 400 0 Distance [nm] 210 195 180 225 240 255 270 Peak-to-peak distance [nm] 226.7 ± 15.3 nm l m k n n o n 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –400 –200 –200 400 0 Distance [nm] 210 195 180 225 240 255 270 Peak-to-peak distance [nm] 226.7 ± 15.3 nm l m l q p Fig. 2 Pre-labeling Ex-dSTORM. a Simulated intensity proﬁles using a cylindrical distribution function to describe unexpanded or 3.2x expanded immunostained microtubules (labeled with IgG antibodies or DNA modiﬁed IgG antibodies pre-expansion) and resulting peak-to-peak distances of the cross-sectional proﬁles. b dSTORM image of expanded and re-embedded α- and β-tubulin pre-labeled with secondary Alexa Fluor 532 IgG antibodies (Al532) using the MA-NHS/GA method6, i.e. antibodies are cross-linked with glutaraldehyde (GA) into the hydrogel (Antibody-Al532 (GA)). c Zoom in of white boxed region in (b). d Averaged cross-sectional proﬁle of 8 microtubule segments with a length between 1.5–6.4 µm and 28.6 µm in total measured in 4 expanded cells. e Histogram of peak-to-peak distance distribution with normalized normal curve (red) of microtubule segments analyzed in (d) at n = 8 microtubule segments in 4 cells from 1 expansion experiment with a mean distance of 133.8 ± 13.2 nm (mean ± sd). f Unexpanded dSTORM image of ssDNA-Cy5 secondary antibody hybridized with Cy5 bearing oligonucleotides pre-expansion (DNA-Cy5 protocol). g Magniﬁed view of white boxed region in (f). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 Unexpanded, microtubules labeled with primary and secondary IgG antibodies exhibit a diameter of ~60 nm with a linkage error (deﬁned by the size of the primary and secondary antibody) of 17.5 nm22. For example, for 3.3x expansion this translates into a microtubule diameter of 3.3 × 25 nm = 82.5 nm whereas the diameter of the immunolabeled microtubule is substantially broadened to ~198 nm because of the linkage error of 3.3 × 17.5 nm = 57.75 nm introduced by the primary and secondary antibody that has to be added to both sides of the microtubule (Supplementary Fig. 5). In other words, even though SMLM achieves high localization precisions12,13, a linkage error of > 50 nm undoes much, or even all, of the gain in resolution. Next, we tested the original ExM protocol with trifunctional DNA-modiﬁed secondary antibodies1, which can be labeled with dye-functionalized complementary oligonucleotides that contain an acrydite linker modiﬁcation. Alternatively, anti- bodies can be modiﬁed with a single stranded DNA that is incorporated directly into the hydrogel. Antisense dye-labeled oligonucleotides can then be hybridized after re-embedding of the hydrogel, which enables the use of ﬂuorophores that would not survive the radical polymerization process. Since the linkage error is mainly determined by the IgG antibodies and the 40 bases long DNA strand (Supplementary Table 2) both methods still belong to the pre-labeling Ex-SMLM method (Supplementary Fig. 1). Pre- versus post-labeling Ex-SMLM. In order to reduce the linkage error, we next explored post-labeling Ex-SMLM. It has been shown that the ﬂuorescence signals from some genetically encoded FPs as well as conventional ﬂuorescently labeled sec- ondary antibodies and streptavidin that are directly anchored to the gel are at least partially preserved by proExM even when subjected to the strong nonspeciﬁc proteolytic digestion used in the original ExM protocol1,5. Therefore, we anticipated that pro- tein epitopes might survive the proExM protocol26. To compare the labeling density of pre- and post-labeling Ex-SMLM we immunostained microtubules with Alexa Fluor 532 before and additionally after expansion. Intriguingly, combining pre- with post-labeling resulted in a substantial shortening of the average peak-to-peak distance of the sidewalls of microtubules to 79.5 ± 6.6 nm (mean ± s.d.) determined from dSTORM images (Fig. 3). We speculated that the protease digestion step may destroy a large fraction of the pre-labeled antibody complexes but to our surprise, the majority of tubulin epitopes survives this critical step. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 Toge- ther with the increased accessibility of tubulin epitopes for pri- mary antibodies and primary antibody epitopes for secondary antibodies after expansion this results in peak-to-peak distances undistinguishable from solely post-labeled microtubules. First, we tested the approach on unexpanded microtubules and obtained peak-to-peak distances of 43.9 ± 3.7 nm (mean ± s.d.) (Fig. 2f–i) and 37.0 ± 4.8 nm (mean ± s.d.) (Supplementary Fig. 6) for labeling with Cy5- and Alexa Fluor 532-modiﬁed oligonu- cleotides, respectively. These values are in good agreement with the theoretically expected value of 41.5 nm for immunolabeling with 42 bases long trifunctional oligonucleotide-modiﬁed sec- ondary antibodies (Fig. 2a). Due to the additional modiﬁcation of the secondary antibodies, the peak-to-peak distances should be a few nanometers larger than the value measured for standard immunolabeled microtubules of 36.2 ± 5.4 nm (mean ± s.d.) (Fig. 1f–i). g y To examine more quantitatively epitope survival and increased epitope accessibilities, we simulated the cross-sectional proﬁles expected for pre- and post-labeled microtubules. Here we assumed a ~10-fold signal dilution (3.22) for the 2D projection of the ﬂuorescence signals of 3.2x expanded pre-labeled antibodies (Fig. 3e and Supplementary Fig. 5). Hence, the cross- sectional microtubule proﬁles show the superposition of the proﬁle calculated for the 3.2x expansion of 25 nm diameter microtubules post-immunolabeled and 60 nm diameter micro- tubules pre-immunolabeled. The resulting superposition proﬁle shows a peak-to-peak distance of 79.5 nm (Fig. 3e) emphasizing the advantage of post-labeling Ex-SMLM. Post-labeling Ex- SMLM using the proExM protocol5 provides an improved labeling efﬁciency and a reduced linkage error. In fact, the immunolabeling linkage error of ~58 nm for pre-labeling reduces to ~5 nm for post-labeling considering a 3.2x expansion factor and thus improves the effective achievable resolution (Supplementary Fig. 8). g If the oligonucleotide-modiﬁed secondary antibodies are labeled with complementary Alexa Fluor 532-modiﬁed oligonucleotides prior to expansion, we measured a peak-to-peak distance of 226.7 ± 15.3 nm (mean ± s.d.) from dSTORM images (Fig. 2j–m). Since Cy5 does not survive gelation1,3,5, we performed an additional experiment labeling the oligonucleotide-modiﬁed secondary anti- bodies after expansion with complementary Cy5-modiﬁed oligo- nucleotides, performed dSTORM imaging in photoswitching buffer and determined a slightly shorter peak-to-peak distance of 201.0 ± 9.3 nm (mean ± s.d.) (Fig. 2n–q). Both values are in excellent agreement with the theoretical peak-to-peak distance of 226.5 nm and 202 nm, respectively (Fig. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 dSTORM images of unexpanded and expanded microtubules clearly showed a bimodal signal distribution along the ﬁlaments, similar to that of previous super-resolution microscopy studies (Fig. 1c, d and 1g, h)21,25. When the cross- sectional proﬁle of unexpanded microtubules was ﬁt with a sum of two Gaussians, the peak-to-peak distance between the sidewalls showed a mean value of 36.2 ± 5.4 nm (mean ± s.d.) analyzed over several microtubule ﬁlament segments (Fig. 1i). This value is expected for the projection of a 25 nm inner diameter cylinder that has been broadened by primary and secondary antibodies on both sides by 17.5 nm22 (Fig. 1a) and corresponds well to the simulated peak-to-peak distance of 32.0 nm for unexpanded microtubules (Fig. 2a and Supplementary Note 1). The mean peak-to-peak dis- tance of proExM treated and expanded microtubules was deter- mined to 137.1 ± 10.1 nm (mean ± s.d.) (Fig. 1e). This value corresponds to an expansion factor of 3.1x determined from 4 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 ~14.3 nm) and, interestingly conformational differences between single and double-stranded DNA linkers. simulation of expanded microtubules pre-labeled with primary and secondary IgG antibodies (Supplementary Table 1). y g pp y We next used the post-stain linker-group functionalization method (MA-NHS/GA method)7 as an alternative pre-labeling Ex-SMLM method. Here, the entire sample was functionalized with polymer-linking groups after conventional immunostaining. The resulting dSTORM images showed a peak-to-peak distance between the microtubule sidewalls of 133.8 ± 13.2 nm (mean ± s. d.) (Fig. 2b–e) corresponding to a simulated expansion factor of 3.0x (Supplementary Table 1). The determined peak-to-peak distance is in good agreement with the peak-to-peak distance determined from proExM expanded microtubules (Fig. 1b–e). Variations in the measured peak-to-peak distances can be well explained by varying initial expansion factors of hydrogels which are typically in the range of ~4.0–4.5x for the used ExM gel composition. Considering a ~20% reduction in gel size caused by re-embedding of the hydrogel, an ultimate expansion factor of ~3.1–3.6x can be expected which ﬁts well with the determined molecular expansion factors. g Noteworthy is that the total size of an expanded sample is not only determined by the biomolecule of interest, e.g. microtubules, but also by the ﬂuorescent probe, e.g. primary and secondary antibodies, used to label the biomolecule of interest. NATURE COMMUNICATIONS \| (2020) 11:3388 \| https://doi.org/10.1038/s41467-020-17086-8 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 f Histogram of peak-to-peak distances with normalized normal curve (red) of ﬁtted proﬁles analyzed in (e) with an average distance of 79.5 ± 6.6 nm (mean ± sd) analyzed along n = 11 microtubule segments in 2 cells from 1 expanded sample. g Image projection of the xz-axes averaged along two microtubule ﬁlaments (iv) and (v) shown in (b) (red dotted lines) using the “z projection analysis” of the software “Line Proﬁler”. h Cross-sectional proﬁle (blue dots) of the xz-projection shown in (g). Using a bi-Gaussian ﬁt (red) the peak-to-peak distance is determined to 81.2 nm. Scale bars, 10 µm (a), 5 µm (b), 1 µm (c), 500 nm (d), 100 nm (g). d iii a b c pre 630 0 V iV i i ii ii z [nm] y x z x post d f h 79.5 ± 6.6 nm 81.2 nm 1.0 0.5 Frequency [a.u] 0.0 1.0 0.5 Intensity [a.u.] 0.0 –150 –100 –50 100 50 150 0 Distance [nm] 70 75 80 85 90 95100 65 60 Peak-to-peak distance [nm] f 79.5 ± 6.6 nm 1.0 0.5 Frequency [a.u] 0.0 70 75 80 85 90 95100 65 60 Peak-to-peak distance [nm] f e Intensity [a.u.] 1.0 0.5 0.0 –200 –100 100 200 0 Distance [nm] e h 81.2 nm 1.0 0.5 Intensity [a.u.] 0.0 –150 –100 –50 100 50 150 0 Distance [nm] Peak to peak distance [nm] h g Fig. 3 3D post-labeling Ex-dSTORM. SMLM image of re-embedded and post-expansion labeled microtubules. a 3D dSTORM image of re-embedded Cos-7 cells expanded according to the Protein-Retention protocol (proExM)4 pre-labeled with anti α- and β-tubulin antibodies and additionally post-labeled with anti α-tubulin. The secondary antibodies were labeled with Alexa Fluor 532 (proExM Al532). The small logo in the upper left corner symbolizes the labeling method, e.g. pre- and post-immunolabeling with Al532 secondary antibodies. b Magniﬁed view of highlighted region in (a). c xz-side view cross-sections (white lines) (i) and (ii) shown in (b) revealing the hollow structure of microtubules. d Magniﬁed view of highlighted region (white box) in (b). Since post- labeling dominates the signal, the method is termed proExM Al532 (post-labeled). e Averaged cross-sectional proﬁle (blue) of 11 analyzed microtubule segments along a total of 28.2 µm ﬁlament (2.1–5.5 µm segments) of one expanded sample. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 2a), simulated for 3.2x expanded microtubules taking into account the length of the 42 base pair trifunctional oligonucleotide, the position of ﬂuoro- phores within the DNA strand and its spatial orientation (Supplementary Fig. 7 and Supplementary Note 1). The slightly shorter peak-to-peak distance measured in the Cy5-experiment where the dye-labeled complementary strand was hybridized after expansion can be explained most likely by coiling of the single- stranded trifunctional oligonucleotide during gelation (Supple- mentary Fig. 7). These results indicate that Ex-SMLM can resolve linker differences of 42 DNA base pairs (corresponding to pp y g Therefore, dSTORM images of Alexa Fluor 532 labeled microtubules clearly revealed the hollow cylinder of micro- tubules (Fig. 3c) using a water-immersion objective and epi- illumination, similar to recently published results obtained by DNA-PAINT TIRF microscopy and experimental point spread function ﬁtting27. The average distance between the sidewalls of 5 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 a b e f g h d c pre 630 0 V iV i i ii ii iV/V iii z [nm] y x z x z x post 79.5 ± 6.6 nm 81.2 nm 1.0 0.5 Frequency [a.u] 0.0 Intensity [a.u.] 1.0 0.5 0.0 –200 –100 100 200 0 Distance [nm] 1.0 0.5 Intensity [a.u.] 0.0 –150 –100 –50 100 50 150 0 Distance [nm] 70 75 80 85 90 95100 65 60 Peak-to-peak distance [nm] Fig. 3 3D post-labeling Ex-dSTORM. SMLM image of re-embedded and post-expansion labeled microtubules. a 3D dSTORM image of re-embedded Cos-7 cells expanded according to the Protein-Retention protocol (proExM)4 pre-labeled with anti α- and β-tubulin antibodies and additionally post-labeled with anti α-tubulin. The secondary antibodies were labeled with Alexa Fluor 532 (proExM Al532). The small logo in the upper left corner symbolizes the labeling method, e.g. pre- and post-immunolabeling with Al532 secondary antibodies. b Magniﬁed view of highlighted region in (a). c xz-side view cross-sections (white lines) (i) and (ii) shown in (b) revealing the hollow structure of microtubules. d Magniﬁed view of highlighted region (white box) in (b). Since post- labeling dominates the signal, the method is termed proExM Al532 (post-labeled). e Averaged cross-sectional proﬁle (blue) of 11 analyzed microtubule segments along a total of 28.2 µm ﬁlament (2.1–5.5 µm segments) of one expanded sample. The simulated cross-sectional proﬁle for 3.2x proExM expanded pre- and post-labeled microtubule assuming a pre- to post-labeling ratio of 0.1 is shown in red. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 4h) post-labeling 3D Ex-SMLM clearly visualized the centriole as a bundle of nine microtubule triplets. SMLM images of expanded isolated Chlamydomonas centrioles showed the 9-fold symmetry of the procentrioles (Fig. 4b, f) with tubulin diameters of ~220 nm in agreement with previous studies2,30. Even in side views of centrioles imaged by 3D Ex- dSTORM the neighboring microtubule triplets are clearly separated (Fig. 4c). Furthermore, 3D Ex-dSTORM allowed us to resolve ring-like sub-structures of centrioles indicating hollow cylinders of microtubule triplets (Supplementary Fig. 11). According to these results, re-embedding of the sample and dSTORM in photoswitching buffer provides currently the best Ex-SMLM performance. Since microtubule triplets separated by 15–20 nm30 are very well resolved in the expanded images post-labeling Ex- SMLM exhibits a spatial resolution that is way below 15–20 nm reaching the structural resolution required to resolve the molecular architecture of centrioles. Post-labeling Ex-SMLM of centrioles. Motivated by the results, we set out to explore the molecular organization of centriole organelles by Ex-SMLM. We used isolated Chlamydomonas cen- trioles, which have a characteristic 9-fold microtubule triplet-based symmetry, forming a polarized cylinder ~ 500 nm long and ~ 220 nm wide28 (Supplementary Fig. 9). Recently2, U-ExM has been developed as an extension of ExM that allows for near-native expansion of organelles and multiprotein complexes and visuali- zation of preserved ultrastructures by optical microscopy. When combined with STED microscopy, details of the ultrastructural organization of isolated centrioles such as the 9-fold symmetry and centriolar chirality could be visualized2. Advantageously, U-ExM uses post-labeling to improve the epitope accessibility after expansion. Here, we used U-ExM treated centrioles in combination with post-labeling with Alexa Fluor 647 secondary antibodies to enable dSTORM imaging, which has previously be impossible due to shrinking of expanded hydrogels in photoswitching buffer. Therefore, samples were re-embedded and transferred onto coverslips functionalized with acrydite via glass silanization to minimize lateral drift. This allowed us to perform post-labeling 3D Ex-dSTORM on ~3.4x expanded centrioles (Fig. 4a–d and Supplementary Fig. 10). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 The simulated cross-sectional proﬁle for 3.2x proExM expanded pre- and post-labeled microtubule assuming a pre- to post-labeling ratio of 0.1 is shown in red. f Histogram of peak-to-peak distances with normalized normal curve (red) of ﬁtted proﬁles analyzed in (e) with an average distance of 79.5 ± 6.6 nm (mean ± sd) analyzed along n = 11 microtubule segments in 2 cells from 1 expanded sample. g Image projection of the xz-axes averaged along two microtubule ﬁlaments (iv) and (v) shown in (b) (red dotted lines) using the “z projection analysis” of the software “Line Proﬁler”. h Cross-sectional proﬁle (blue dots) of the xz-projection shown in (g). Using a bi-Gaussian ﬁt (red) the peak-to-peak distance is determined to 81.2 nm. Scale bars, 10 µm (a), 5 µm (b), 1 µm (c), 500 nm (d), 100 nm (g). the xz-projection of a 6.5 µm long microtubule ﬁlament was determined to 81.2 nm (Fig. 3g–h) highlighting the high spatial resolution of pre-labeling 3D Ex-dSTORM. the xz-projection of a 6.5 µm long microtubule ﬁlament was determined to 81.2 nm (Fig. 3g–h) highlighting the high spatial resolution of pre-labeling 3D Ex-dSTORM. improved the blinking characteristics of HMSiR but reduced the expansion factor to ~2x, which limits the spatial resolution of the SMLM experiments (Fig. 4d, g). In contrast to 3D dSTORM images of unexpanded centrioles (Fig. 4h) post-labeling 3D Ex-SMLM clearly visualized the centriole as a bundle of nine microtubule triplets. SMLM images of expanded isolated Chlamydomonas centrioles showed the 9-fold symmetry of the procentrioles (Fig. 4b, f) with tubulin diameters of ~220 nm in agreement with previous studies2,30. Even in side views of centrioles imaged by 3D Ex- dSTORM the neighboring microtubule triplets are clearly separated (Fig. 4c). Furthermore, 3D Ex-dSTORM allowed us to resolve ring-like sub-structures of centrioles indicating hollow cylinders of microtubule triplets (Supplementary Fig. 11). According to these results, re-embedding of the sample and dSTORM in photoswitching buffer provides currently the best Ex-SMLM performance. Since microtubule triplets separated by 15–20 nm30 are very well resolved in the expanded images post-labeling Ex- SMLM exhibits a spatial resolution that is way below 15–20 nm reaching the structural resolution required to resolve the molecular architecture of centrioles. improved the blinking characteristics of HMSiR but reduced the expansion factor to ~2x, which limits the spatial resolution of the SMLM experiments (Fig. 4d, g). In contrast to 3D dSTORM images of unexpanded centrioles (Fig. Discussion l Electron microscopy has been the only viable method able to reveal the ultrastructure of organelles and molecular complexes for decades because of the diffraction limit of optical microscopy. Super-resolution microscopy offers up to ~10x higher resolution than conventional diffraction-limited microscopy11,12. Improved super-resolution microscopy methods can now localize single emitters with a precision of a few nanometers31–33, but limitations in labeling efﬁciency and linkage error have thus far prevented the translation of high localization precision into sub-10-nm spatial resolution. Therefore, the spatial resolution provided by all these inventive methods is currently still too low Alternatively, we used the spontaneously blinking Si-rhodamine dye HMSiR29 that enables SMLM in the absence of photoswitching buffer and does thus not require re-embedding. Using double- deionized water, we achieved a molecular expansion factor of ~4x (Fig. 4d–f and Supplementary Fig. 9). Unfortunately, since the pH of double-deionized water is below 7.0, HMSiR does not exhibit optimal blinking characteristics29. Addition of PBS buffer, pH 7.4 NATURE COMMUNICATIONS \| (2020) 11:3388 \| https://doi.org/10.1038/s41467-020-17086-8 \| www.nature.com/naturecommunications 6 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 0.8 25–75% IQR 1.5IQR Mean Median (IQR-Interquartile range) p<0.02 Diameter [μm] 0.7 0.6 0.5 0.4 0.3 AI647 MEA HMSiR PBS HMSiR Water a b c d e f g h Fig. 4 Ex-SMLM of U-ExM expanded centrioles. a-c 3D dSTORM image of U-ExM expanded and re-embedded Chlamydomonas centrioles stained post re- embedding with anti α-tubulin primary antibody and Alexa Fluor 647 conjugated secondary antibodies measured in MEA buffer. b Zoom-in on highlighted region in (a) revealing the 9-fold symmetry of the shown procentriole. c Side view of two mature centrioles with clearly separated triplets. The inlet shows the cross-sectional proﬁle along the centriole (white box) showing ﬁve distinct peaks of microtubule triples (marked with arrow heads). d Comparison of the diameters determined from expanded centrioles measured using different protocols (re-embedded and labeled with Alexa Fluor 647, and imaged in MEA photoswitching buffer, labeled with HMSiR 647 and imaged in double-deionized water or in pH optimized PBS (1x) buffer with pH 7.4). Mean values are 657 ± 90 nm (mean ± sd) for Alexa Fluor 647 in MEA buffer (n = 12 centrioles), 428 ± 74 nm (mean ± sd) for HMSiR in PBS (n = 7 centrioles), and 750 ± 34 nm (mean ± sd) for HMSiR 647 in water (n = 8 centrioles). Data from n = 2 independent experiments for each condition. Discussion l Divided by the previously analyzed diameter of α-tubulin labeled centriole expansion factors translates into ~3.4x, ~2.2x, and ~3.9x for expanded centrioles labeled with Alexa Fluor 647 in MEA buffer, HMSiR in PBS (1x), and HMSiR 647 in water, respectively. Statistical signiﬁcance was assessed by one-way ANOVA: the mean values of the diameters are signiﬁcantly different with p<0.02 (F = 3.80) (P ≤0.05, P ≤0.01). e–g 2D dSTORM image of U-ExM expanded centrioles labeled with HMSiR 647 imaged in water (e–f) or PBS(1x) (g) f Zoom-in on highlighted region in (e). h 3D dSTORM image of unexpanded isolated Chlamydomonas centrioles immunostained with antibodies against glutamylated tubulin and Alexa Fluor 647 conjugated secondary antibodies. Scale bars, 1 µm (a, e), 500 nm (b, f, g), 1.5 µm (c), 250 nm (h). 0.8 25–75% IQR 1.5IQR Mean Median (IQR-Interquartile range) p<0.02 Diameter [μm] 0.7 0.6 0.5 0.4 0.3 AI647 MEA HMSiR PBS HMSiR Water d c c d b a AI647 MEA HMSiR PBS HMSiR Water h g f e f h g e Fig. 4 Ex-SMLM of U-ExM expanded centrioles. a-c 3D dSTORM image of U-ExM expanded and re-embedded Chlamydomonas centrioles stained post re- embedding with anti α-tubulin primary antibody and Alexa Fluor 647 conjugated secondary antibodies measured in MEA buffer. b Zoom-in on highlighted region in (a) revealing the 9-fold symmetry of the shown procentriole. c Side view of two mature centrioles with clearly separated triplets. The inlet shows the cross-sectional proﬁle along the centriole (white box) showing ﬁve distinct peaks of microtubule triples (marked with arrow heads). d Comparison of the diameters determined from expanded centrioles measured using different protocols (re-embedded and labeled with Alexa Fluor 647, and imaged in MEA photoswitching buffer, labeled with HMSiR 647 and imaged in double-deionized water or in pH optimized PBS (1x) buffer with pH 7.4). Mean values are 657 ± 90 nm (mean ± sd) for Alexa Fluor 647 in MEA buffer (n = 12 centrioles), 428 ± 74 nm (mean ± sd) for HMSiR in PBS (n = 7 centrioles), and 750 ± 34 nm (mean ± sd) for HMSiR 647 in water (n = 8 centrioles). Data from n = 2 independent experiments for each condition. Discussion l By linking a protein of interest into a cross-linked network of a swellable polyelectrolyte hydrogel, biological speci- mens can be physically expanded allowing for sub-diffraction resolution imaging on conventional microscopes1–10. However, even in combination with super-resolution microscopy techni- ques, spatial resolutions below ~20 nm have so far proven to be very difﬁcult to achieve by ExM16. Here, we have shown that re- embedding of expanded hydrogels enables the use of standard photoswitching buffers and dSTORM imaging of ~3.2x expanded samples. Our results demonstrate that post-labeling ExM using the proExM protocol5 or U-ExM2 provides solutions for the two major limiting problems of improved super-resolution micro- scopy, the labeling efﬁciency and linkage error. First, as shown for microtubules, expansion of the sample increases the epitope accessibility and thus the labeling efﬁciency. Comparison experiments demonstrated that post-labeling outperforms pre- labeling several times in this regard (Fig. 3). Second, post-labeling ExM reduces the linkage error proportionally to the expansion factor. Hence, post-immunolabeling of 3.2x expanded micro- tubules reduces the linkage error from 17.5 nm22 to ~5 nm (Fig. 3). Since the linkage error also inﬂuences the localization accuracy and thus the effective achievable resolution (Supple- mentary Figs. 8 and 9)34,35 our ﬁndings are highly relevant. Very to unravel the composition and molecular architecture of protein complexes or dense protein networks. Expansion microscopy (ExM) represents an alternative approach to bypass the diffrac- tion barrier. By linking a protein of interest into a cross-linked network of a swellable polyelectrolyte hydrogel, biological speci- mens can be physically expanded allowing for sub-diffraction resolution imaging on conventional microscopes1–10. However, even in combination with super-resolution microscopy techni- ques, spatial resolutions below ~20 nm have so far proven to be very difﬁcult to achieve by ExM16. Here, we have shown that re- embedding of expanded hydrogels enables the use of standard photoswitching buffers and dSTORM imaging of ~3.2x expanded samples. Our results demonstrate that post-labeling ExM using the proExM protocol5 or U-ExM2 provides solutions for the two major limiting problems of improved super-resolution micro- scopy, the labeling efﬁciency and linkage error. First, as shown for microtubules, expansion of the sample increases the epitope accessibility and thus the labeling efﬁciency. Comparison experiments demonstrated that post-labeling outperforms pre- labeling several times in this regard (Fig. 3). Second, post-labeling ExM reduces the linkage error proportionally to the expansion factor. Discussion l Divided by the previously analyzed diameter of α-tubulin labeled centriole expansion factors translates into ~3.4x, ~2.2x, and ~3.9x for expanded centrioles labeled with Alexa Fluor 647 in MEA buffer, HMSiR in PBS (1x), and HMSiR 647 in water, respectively. Statistical signiﬁcance was assessed by one-way ANOVA: the mean values of the diameters are signiﬁcantly different with p<0.02 (F = 3.80) (P ≤0.05, **P ≤0.01). e–g 2D dSTORM image of U-ExM expanded centrioles labeled with HMSiR 647 imaged in water (e–f) or PBS(1x) (g) f Zoom-in on highlighted region in (e). h 3D dSTORM image of unexpanded isolated Chlamydomonas centrioles immunostained with antibodies against glutamylated tubulin and Alexa Fluor 647 conjugated secondary antibodies. Scale bars, 1 µm (a, e), 500 nm (b, f, g), 1.5 µm (c), 250 nm (h). recently36,37, trifunctional linkers have been introduced that are inert to polymerization, digestion and denaturation, and enable direct covalent linking of target molecules and functional groups to the hydrogel. Therefore, trifunctional linkers can retain a high number of labels and ﬂuorescence molecules available for post- expansion imaging. However, since the target molecules are labeled with primary and secondary antibodies or enzymatic tags (e.g. SNAP-tags) functionalized with the trifunctional anchor before expansion, linkage errors remain. The improved labeling efﬁciency of post-labeling Ex-dSTORM in combination with small (1.5 × 2.5 nm) camelid antibodies (“nanobodies”)38,39 and 10−20x expansion factors9,10 can thus pave the way for true molecular resolution imaging of endogenous proteins with 1–5 nm spatial resolution. On the other hand, at such a small length scale, distortions of the structure may occur. To realize more homogeneous gel network structures, a new gelation method based on a highly homogeneous expansion microscopy polymer composed of tetrahedron-like monomers has been introduced40. The new tetra-gel polymer chemistry may introduce fewer spatial errors than earlier versions, and enable molecular resolution post- labeling Ex-dSTORM with reduced distortion. Nevertheless, already ~3x Ex-SMLM can resolve small linker length and con- formational differences between labeling approaches as shown here for oligonucleotide-functionalized secondary antibodies (Fig. 2). In addition, we have shown that post-labeling 3D Ex- dSTORM exhibits excellent structure preservation and already to unravel the composition and molecular architecture of protein complexes or dense protein networks. Expansion microscopy (ExM) represents an alternative approach to bypass the diffrac- tion barrier. Methods Reagents. Acetic acid (A6283, Sigma), Acrylamide (AA, 40%, A4058, Sigma), Acryloyl-X, SE, 6-((acryloyl)amino)hexanoic Acid, Succinimidyl Ester (A20770, Thermo Fisher), Agarose (A9539, Sigma), Ammonium persulfate (APS, A3678, Sigma), Bind Silane (GE17-13330-01, Sigma), Bovine Serum Albumin (BSA, A2153, Sigma), Cysteamine hydrochloride (MEA, 6500, Sigma), Dextran sulfate (D8906, Sigma), DMEM/HAM’s F12 with L-glutamine (Sigma, D8062), Ethanol (absolute, ≥99.8%, 32205, Sigma), Ethylenediaminetetraacetic acid (E1644, Sigma) h l l l b ( h l h ) ′ ′ d ( Ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA, 03777, Sigma), 10% FBS (Sigma, F7524), Formaldehyde (FA, 36.5-38%, F8775, Sigma), Glutaraldehyde (GA, 25%, G5882, Sigma), Guanidine hydrochloride (50933, Sigma), 2-(N-morpholino)ethanesulfonic acid (MES, M3671, Sigma), N,N ′-methylen-bisacrylamide (BIS, 2%, M1533, Sigma), N,N,N′,N y y ( g ) ′-Tetramethylethylenediamine (TEMED, T7024, Sigma), Poly-D-lysine hydropromide (P6407, Sigma), Polyoxyethylene (20) sorbitan monolaurate solution (Tween-20, 10%, 93774, Sigma), Potassium hydroxide (P5958, Sigma). P t i K (P4850 Si ) S li di it t b ff (SSC 20 15557 y y g ′-Tetramethylethylenediamine (TEMED, T7024, Sigma), Poly-D-lysine hydropromide (P6407, Sigma), Polyoxyethylene (20) sorbitan monolaurate solution (Tween-20, 10%, 93774, Sigma), Potassium hydroxide (P5958, Sigma). P i K (P Si ) S li di i b ff (SSC Proteinase K (P4850, Sigma), Saline sodium citrate buffer (SSC,20x, 15557, Thermo Fisher), Sodium acrylate (SA, 97-99%, 408220, Sigma), Sodium chloride (NaCl, S7653, Sigma), Sodium dodecyl sulfate (SDS, L3771, Sigma), streptomycin (0.1 mg/ml) (Sigma, R8758), Tris base (T6066, Sigma), Triton X-100 Surfact-Amps Detergent Solution (10% (w/v), 28313, Thermo Fisher), Yeast tRNA (AM7119, Thermo Fisher). Proteinase K (P4850, Sigma), Saline sodium citrate buffer (SSC,20x, 15557, Thermo Fisher), Sodium acrylate (SA, 97-99%, 408220, Sigma), Sodium chloride (NaCl, S7653, Sigma), Sodium dodecyl sulfate (SDS, L3771, Sigma), streptomycin (0.1 mg/ml) (Sigma, R8758), Tris base (T6066, Sigma), Triton X-100 Surfact-Amps Detergent Solution (10% (w/v), 28313, Thermo Fisher), Yeast tRNA (AM7119, Thermo Fisher). Protein Retention protocol (proExM protocol). Blocking and immunostaining were performed as described under “Immunostaining of unexpanded Cos-7 cells” incubating anti-α-tubulin antibody (ab1825, diluted 1:500, cend = 2 µg/ml) and anti-ß-tubulin (T8328, diluted 1:200, cend = 10 µg/ml) simultaneously in blocking buffer as primary antibodies and Alexa Fluor 532 IgG antibodies (A-11002 and A- 1109, each diluted 1:200 to cend = 10 µg/ml) diluted in blocking buffer as secondary antibodies.For copolymerization of amine groups into the hydrogel, cells were treated with the amine reactive agent Acryloyl X-SE (0.1 mg/ml) in PBS. Methods The agent was freshly prepared from desiccated stock aliquots kept at −20 °C, incubated overnight in a humidiﬁed chamber, and subsequently washed twice for 15 min each in PBS (1x). Hydrogel formation, Proteinase K digestion and expansion in water were performed as described under “ExM protocol”. After re-embedding of expanded hydrogels as described in section “Bind-silane treatment and re- embedding”, samples were labelled with α-tubulin primary antibody solution (ab1825, diluted 1:500, cend = 2 µg/ml) in 2% BSA for 3 h at 37 °C and then washed twice with 0.01% Tween in PBS for 20 min each and twice with PBS (1x) for 10 min each. Secondary antibodies were incubated for 3 h at 37 °C and washed twice with 0.01% Tween in PBS for 30 min each and twice with PBS (1x) for 30 min followed by a washing step over night in PBS (1x). Antibodies and labeling reagents. Rabbit anti α-tubulin antibody (ab18251, abcam), Mouse anti β-tubulin antibody (T8328, Sigma), Mouse anti poly- glutamylated tubulin, mAb (GT335) (Adipogen, AG-20B-0020), Alexa Fluor 647 F (ab‘)2 of goat anti rabbit IgG (A-21246, Thermo Fisher), Alexa Fluor 647 F(ab‘)2 of goat anti mouse IgG (A-21235, Thermo Fisher, Al532-Goat anti Rabbit IgG (H+L) (A-11009, Thermo Fisher), Al532-Goat anti Rabbit IgG (H +L) (A-11002, Thermo Fisher), HMSiR 647 (A208-01, MoBiTec) conjugated to goat anti rabbit IgG F(ab‘)2 (SAB3700946, Sigma), TetraSpeck Microspheres (0.1 µm, T27279, Thermo Fisher). Chlamydomonas reinhardtii centriole isolation. Centrioles were isolated from the cell wall-less Chlamydomonas reinhardtii strain CW15 by centrifugation at 600g for 10 min in 50 ml conical tubes41. Isolated centrioles were thawed on ice and diluted with cold K-Pipes 10 mM pH 7.2. Centrioles were then loaded in a 15 ml Corex tube with a homemade adaptor and concentrator, and spun onto a 12 mm Poly-D- lysine coated coverslip through centrifugation at 10,000g for 10 min with a JS- 13.1 swinging bucket rotor (Beckman) at 4 °C. Coverslips were then processed for immunostaining and expansion microscopy. ExM protocol with glutaraldehyde linker (ExM-GA protocol). Blocking and immunostaining were performed as described under “Immunostaining of unex- panded Cos-7 cells” using α- (ab1825, diluted 1:500, cend = 2 µg/ml) and ß-tubulin (T8328, diluted 1:200, cend = 10 µg/ml) antibodies as primary antibodies and a mixture of Alexa Fluor 532 IgG secondary antibodies (A-11002 and A-1109, each diluted 1:200 to cend = 10 µg/ml) in blocking buffer. Methods Cells were then ﬁxed and permeabilized simultaneously incubating a primary ﬁxative solution of 0.3% glutaraldehyde and 0.25% Triton X-100 in CB-buffer for 90 s followed by a second ﬁxation using 2% glutaraldehyde in CB-buffer for 10 min. Fixation was stopped by a 7 min reduction step with freshly prepared 0.5% NaBH4 in PBS. Specimen were then washed three times with PBS (1x) for 5 min each and treated differently depending on subsequent expansion method described below. l h d ll b d DNA label with Cy5 (DNA-Cy5 protocol). Blocking and immunostaining were performed as described under “ExM protocol” with a mixture of primary α- and ß-tubulin antibodies (ab1825 diluted 1:500 with 2 µg/ml and T8328 diluted 1:200 with 10 µg/ml) and DNA conjugated secondary antibodies “Antibody B Cy5” or “Antibody C Cy5” in hybridization buffer that were then directly incorporated into the hydrogel. Hydrogel formation, proteinase K digestion and expansion were performed as described under “ExM protocol”. After re-embedding on 24-mm silanized round coverslips samples were incubated over night with Cy5 antisense oligos with a DNA concentration of 0.5 ng/µl for each oligo in hybridization buffer. y p g q p Unless otherwise stated all incubations were carried out at room temperature in the following protocols. Immunostaining was either performed pre-gelation (referred to as pre-labeling), post-expansion (post-labeling) or post-re-embedding (post re-embedding labeling). Sequences and modiﬁcations of DNA labels are listed in Supplementary Table 2. A list of primary and secondary antibodies used for immunostaining in the corresponding Figures is provided in Supplementary Table 3 and Supplementary Table 4 with details about the expansion protocol used. Ultrastructure expansion microscopy (U-ExM). Twelve millimeters cover glasses with isolated2 centrioles were placed in a solution containing 0.7% FA, 1% AA diluted in PBS (1x). Next, 35 µl of pre-chilled U-ExM monomer solution (19% (w/ w) SA, 10% (w/w) AA, 0.1% (w/w) BIS) supplemented with 0.5% APS and 0.5% TEMED in PBS for 1 min on a paraﬁlm coated plate put on ice. Gelation proceeded for 1 h at 37 °C in a humidiﬁed chamber. Samples were placed in denaturation buffer (200 mM SDS, 200 mM NaCl in 50 mM Tris (pH 9.0)) for 15 min and then gels were carefully removed from the cover glasses and transferred to 1.5 ml cen- trifuge tubes ﬁlled with denaturation buffer. ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 3.4x expansion using standard protocols can provide a sufﬁcient structural resolution to resolve details of the molecular archi- tecture of centrioles (Fig. 4 and Supplementary Fig. 11). ExM protocol using DNA trifunctional labels (ExM protocol). After blocking with 5% BSA in PBS for 1 h, cells were incubated with anti-alpha tubulin primary antibody (ab1825, diluted 1:500, cend = 2 µg/ml) in blocking buffer (5% BSA in PBS) for 1 h, followed by three washes in PBS (1x) for 5 min each and incubation of “Antibody B” DNA-labeled secondary antibodies (10 µg/ml) in hybridization buffer (2x saline sodium citrate (SSC), 10% dextran sulfate, 1 mg/ml yeast tRNA, 5% BSA) for 3 h. Antisense DNA B1-Alexa Fluor 532 and DNA B2-Alexa Fluor 532 oligos were hybridized simultaneously at a total DNA concentration of 1.0 ng/ µl for 3 h in hybridization buffer. Then samples were washed three times with PBS (1x) for 10 min each.Gelation was performed on the lid of a 4-well cell culture plate put on ice and covered with paraﬁlm that served as a ﬂat hydrophobic gelation surface. 18 mm cover glasses with cells facing down were placed on top of 90 µl pre-chilled ExM monomer solution (8.625% (w/w) SA, 20% (w/w) AA, 0.15% (w/w) BIS, 2 M NaCl in PBS) supplemented with 0.2% APS and 0.2% TEMED. Samples were then carefully transferred to a humidiﬁed chamber and incubated for 1.5 h at 37 °C for chemical crosslinking of acrylic monomers and trifunctional labels. After gelation samples were treated with 8 U Proteinase K in digestion buffer (50 mM Tris (pH 8.0), 0.5% TritonX-100, 0.8 M guanidine hydrochloride, 1 mM EDTA) and then expanded in double-deionized water. Water was exchanged several times until the maximum expansion factor of the hydrogel was reached. The expansion factor was determined by measuring the diameter of the gel using a calipser. When the expansion factor did not change within three water exchanges this factor was assumed as maximum expansion of the hydrogel. Methods After washing with PBS (1x), cells were incubated with 0.25% GA in PBS for 10 min and washed thrice in PBS (1x) for 5 min each before proceeding with gelation of the samples. Gelation, digestion, and expansion was performed as described under “ExM protocol”. Cell culture of mammalian cells. COS-7 monkey kidney cells (purchased from CLS Cell Line Servie GmbH) were cultured at 37 °C and 5% CO2 in DMEM/HAM’s F12 medium with L-glutamine containing FBS (10%) and peni- cillin (100 U/ml) and streptomycin (0.1 mg/ml). 20–30,000 cells per well were seeded on round 18 mm high precision cover glasses (No 1.5) in 12-well culture plates (Techno Plastic Products, 92012) and grown for 24 h prior to ﬁxation. Sample preparation. For ﬁxation, all solutions were pre-warmed to 37 °C and ﬁxation was conducted on a heating plate set to 37 °C. Right before ﬁxation samples were rinsed once with pre-warmed Cytoskeleton buffer (CB-buffer, 10 mM MES, 150 mM NaCl, 5 mM EGTA, 5 mM glucose and 5 mM MgCl2, pH 6.1). Cells were then ﬁxed and permeabilized simultaneously incubating a primary ﬁxative solution of 0.3% glutaraldehyde and 0.25% Triton X-100 in CB-buffer for 90 s followed by a second ﬁxation using 2% glutaraldehyde in CB-buffer for 10 min. Fixation was stopped by a 7 min reduction step with freshly prepared 0.5% NaBH4 in PBS. Specimen were then washed three times with PBS (1x) for 5 min each and treated differently depending on subsequent expansion method described below. Unless otherwise stated all incubations were carried out at room temperature in the following protocols. Immunostaining was either performed pre-gelation (referred to as pre-labeling), post-expansion (post-labeling) or post-re-embedding (post re-embedding labeling). Sequences and modiﬁcations of DNA labels are listed in Supplementary Table 2. A list of primary and secondary antibodies used for immunostaining in the corresponding Figures is provided in Supplementary Table 3 and Supplementary Table 4 with details about the expansion protocol used. Sample preparation. For ﬁxation, all solutions were pre-warmed to 37 °C and ﬁxation was conducted on a heating plate set to 37 °C. Right before ﬁxation samples were rinsed once with pre-warmed Cytoskeleton buffer (CB-buffer, 10 mM MES, 150 mM NaCl, 5 mM EGTA, 5 mM glucose and 5 mM MgCl2, pH 6.1). Discussion l Hence, post-immunolabeling of 3.2x expanded micro- tubules reduces the linkage error from 17.5 nm22 to ~5 nm (Fig. 3). Since the linkage error also inﬂuences the localization accuracy and thus the effective achievable resolution (Supple- mentary Figs. 8 and 9)34,35 our ﬁndings are highly relevant. Very 7 TURE COMMUNICATIONS \| (2020) 11:3388 \| https://doi.org/10.1038/s41467-020-17086-8 \| www.nature.com/naturecommunications 7 ARTICLE Therefor ﬂuorescent marker stock supension (0.1 µm, ~1.8 × 1011 particles/mL, TetraSpeck Microspheres, Thermo Fisher) was vortexed for ~1 min and then diluted 1:50 in the corresponding monomer solution. After adding TEMED and APS in the appropriate concentrations, the bead-gel solution was vortexed again for ~ 20 s, polymerized, and expanded as described under the respective expansion protocol (omitting the digestion or denaturation step). The expanded gels were then transferred on poly-L-lysine (0.1%) coated coverslips and additionally embedded in 1% (w/v) Agarose. 4 µm z-stacks of several ﬂuorescent markers dispersed in the hydrogel ~50–400 µm above the coverslip were recorded and used to generate 3D calibration curves as described below. The software Micro-manager 1.4 was used for image acquisition and to control the piezo driven stage. Re-embedding of expanded hydrogels (Re-embedding protocol). To avoid shrinking caused by dSTORM photoswitching buffer and to prevent drifting of the hydrogel during image acquistion an uncharged acrylamide gel was crosslinked throughout the hydrogel while chemically binding it on Bind-silane treated cover glasses. Round 24-mm cover glasses (high precision) were sonicated successively in double-deionized water, absolute ethanol and 5 M potassium hydroxide for 20 min each and washed with deionized water between every sonication step and ﬁnally oven dried at 100 °C. 200 µl of Bind-silane working solution (5 µl Bind-Silane in 8 ml absolute ethanol, 200 µl glacial acetic acid, 1.8 ml double deionized water) were distributed evenly on cleaned 24-mm cover glasses and left for around 1 h until the solution was fully evaporated. Cover glasses were then rinsed with doubly deio- nized water and air-dried. Glasses were prepared shortly before use. For re- embedding expanded hydrogels were placed in 6-well cell culture plates and each sample was covered with 3 ml of freshly prepared Re-embedding solution (10% acrylamide, 0.15% bis-acrylamide, 0.05% APS, 0.05% TEMED in 5 mM Tris (pH 8.9)). Samples were incubated on a platform shaker twice with freshly prepared solution for 30 min each. Shaking of the Re-embedding solution is crucial in this step as it brings oxygen into the solution that prevents it from gelling to early. The stirring speed should be adjusted so that the liquid is in motion but the gels are not damaged. After the second incubation, samples were transferred on silanized coverglasses while carefully removing excess solution from the hydrogels using laboratory wipes. ARTICLE Furthermore, a deformation vector ﬁeld of pre-expansion and post re- embedding RCM images was created using elastix and transformix6. Elastix and transformix code were executed in Wolfram Mathematica 11.2. Microscopes. Single-molecule localization microscopy (SMLM) image acquisition was performed on a custom-built setup with an inverted Zeiss Axio Observer Z1 (Carl Zeiss Microscopy) microscope equipped with a Deﬁnite Focus autofocusing system. For excitation of different ﬂuorescent molecules the setup provides three iBeam smart diode lasers with 405 nm (100 mW output power), 488 nm (200 mW output power) and 640 nm (150 mW output power) and a DPSS (diode pumped solid state) 532 nm laser (gem532, Laserquantum). Lasers were ﬁltered with laser clean-up ﬁlters according to the speciﬁc wavelength and focused on the back focal plane of the objective to achieve a wide ﬁeld illumination. To match the aqueous refractive index of expanded samples a water-immersion objective (LD C- p p j ( Apochromat 63x/1.15 W Corr M27, Carl Zeiss Microscopy) is implemented in the microscope. The excitation light passes a quad-band dichroic beam splitter (Di01- R405/488/532/635-25×36, BrightLine) combined with a quad-band rejection ﬁlter (ZET405/488/532/642 m, Chroma). For recording the emission of excited ﬂuor- ophores the setup is equipped with two Andor Ixon Ultra 897 EMCCD (electron- multiplying charge-coupled device) cameras at the side port of the microscope. The software Andor Solis (Version 4.28.30014) was used to control the EMCCD cameras. The ﬂuorescence light is parallelized through a 160 mm achromatic lens (Thorlabs) and can be spectrally separated by a 630 DCXR (Chroma) dichroic beam splitter. In this conﬁguration, two different emission wavelengths can be focussed on two cameras arranged perpendicular to each other. For all dSTORM measurement in this work the beam splitter was removed and the emission light was directed to one camera. Suitable emission light ﬁlters were placed in front of the camera depending on the detected ﬂuorescent wavelength. For 3D imaging, an additional achromatic cylindrical lens (f = 250 mm, Thorlabs) was placed in the detection path close to the imaging plane before the relay system. Rescanning confocal imaging (RCM) was performed on a Nikon TiE inverted microscope equipped with an RCM unit (Confocal.nl) that is based on the image scanning principle42. The setup was operated by the microscope software NIS-Elements (version 4.6). Analysis of microtubule transversal proﬁles. ARTICLE Another coverglass that was not silanized was placed on top of the hydrogels during the following steps. The whole setup was transferred to a humidiﬁed container equipped with gas injection holes. To accelerate gelation oxygen was extracted from the container by purging the chamber with nitrogen for 15 min. The samples were then incubated at 37 °C for 2 h. After polymerization of the re-embedding gel samples were washed at least thrice for 30 min in double deionized water. Coverglasses on top of the hydrogel come off themselves during washing or can be detached carefully after the ﬁrst washing steps. Re-embedded gels were then placed in imaging buffer or staining buffer depending on the sub- sequent protocol. Image processing. For 2D and 3D dSTORM image reconstruction super- resolution images were analyzed, post-processed and visualized using the analysis platform SMAP (Superresolution Microscopy Analysis Platform) with the GPU based 3D ﬁtter ﬁt3Dcspline27 and the ImageJ plugin ThunderSTORM43. The respective integrated calibration tools were used for generating 3D astigmatism calibration curves. Localizations were further corrected for drift using the cross- correlation method, ﬁltered for molecules with poor precision and grouped to one localization when molecules appeared in several consecutive images. Expansion factor determination. Centriole diameters of U-ExM expanded sam- ples were determined by averaging peak-to-peak distances of two cross-sectional proﬁles that were drawn through the center of the ninefold-symmetrical α-tubulin signal using the line proﬁle tool of Fiji44. Peaks were then determined by using the peakﬁnder minitool implemented in the analyse software Origin (OriginLab, Northampton, MA). To determine the expansion factor post-expansion and post- re-embedding, Cos-7 cells were labeled with a-tubulin and ß-tubululin and expanded according to the “proExM protocol”. An additional post-expansion immunostaining for α-tubulin was performed using the same primary and sec- ondary antibodies. RCM images of the same cells were acquired before gelation, after expansion and after re-embedding in different imaging buffers. Images were then registered via rigid (similarity) and non-rigid registration (B-spline) using the open source, command-line program elastix6. The transform parameters of the similarity transformation of pre- and post-expansion RCM images were used to determine the initial expansion factor of the sample. RCM images acquired post-re- embedding in PBS (1x) and cysteamine hydrochloride as well as a dSTORM image in photoswitching buffer of the same area were registered in the same way using elastix to determine the expansion factor after re-embedding in different imaging buffers. ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 1:200, cend = 10 µg/ml) diluted in 2% BSA were incubated 3 h at 37 °C followed by two washing steps in 0.01% Tween in PBS for 30 min each and two washes with PBS (1x) for 30 min. Before imaging gels were washed once more overnight in PBS (1x). For imaging of unexpanded centrioles the primary antibody anti poly- glutamylated tubulin (Adipogen, 1:500) was diluted in 5% BSA in PBS and incu- bated for 1 h at room temperature, washed thrice in PBS for 5 min each, followed by incubation with secondary Alexa Fluor 647 F(ab‘)2 antibodies (A-21246, 1:200, cend = 10 µg/ml) diluted in 2% BSA for 1 h. The samples were then washed twice in 0.01% Tween in PBS and once in PBS for 10 min each. 1:200, cend = 10 µg/ml) diluted in 2% BSA were incubated 3 h at 37 °C followed by two washing steps in 0.01% Tween in PBS for 30 min each and two washes with PBS (1x) for 30 min. Before imaging gels were washed once more overnight in PBS (1x). For imaging of unexpanded centrioles the primary antibody anti poly- use. The hydrogel was incubated in photoswitching buffer twice for 20 min each before imaging. U-ExM treated samples labeled with the spontaneously blinking Si- rhodamine dye HMSiR were immobilized on Poly-L-lysine (0.1%) coated 24-mm high-precision cover glasses and additionally embedded in 1% (w/v) Agarose. As imaging buffer, double deionized water or pH adjusted PBS buffer (1x, pH 7.4) was used. For unexpanded dSTORM imaging samples were placed in 100 mM MEA in PBS adjusted to pH 7.5 (with KOH) for DNA-Cy5 and Streptavidin-Alexa Fluor 647 and pH 7.9 (with KOH) when using Alexa Fluor 532. g g p p y y p y glutamylated tubulin (Adipogen, 1:500) was diluted in 5% BSA in PBS and incu- bated for 1 h at room temperature, washed thrice in PBS for 5 min each, followed by incubation with secondary Alexa Fluor 647 F(ab‘)2 antibodies (A-21246, 1:200, cend = 10 µg/ml) diluted in 2% BSA for 1 h. The samples were then washed twice in 0.01% Tween in PBS and once in PBS for 10 min each. 3D dSTORM calibration. To obtain 3D calibration curves, ﬂuorescent beads were mixed in U-ExM or ExM monomer solution for 3D image acquisitions of U-ExM or ExM samples, respectively. Methods Hydrogels were then incubated for 30 min at 95 °C and then expanded in double deionized water until the maximum expansion of the gels were reached. After re-embedding on Bind-silane treated 24-mm cover glasses, centrioles were labelled with anti alpha-tubulin primary antibodies (ab1825, cend = 2 µg/ml) diluted 1:500 in 2% BSA in PBS for 3 h at 37 °C, washed twice with 0.01% Tween in PBS for 20 min each and twice with PBS (1x) for 10 min each. Next, secondary Alexa Fluor 647 F(ab‘)2 antibodies (A-21246, Immunostaining of unexpanded Cos-7 cells. Cells were placed in blocking buffer (5% BSA in PBS) for 1 h and then incubated for 1 h with anti-alpha tubulin primary antibody solution (ab1825, diluted 1:500, ﬁnal concentration cend = 2 µg/ ml) diluted in blocking buffer. Samples were washed thrice in PBS (1x) for 5 min each and incubated with secondary Alexa Fluor 532 IgG antibody solution in blocking buffer (A-11002, diluted 1:200, cend = 10 µg/ml) for 1 h followed by three washes in PBS (1x) for 10 min each. NATURE COMMUNICATIONS \| (2020) 11:3388 \| https://doi.org/10.1038/s41467-020-17086-8 \| www.nature.com/naturecommunications 8 NATURE COMMUNICATIONS \| (2020) 11:3388 \| https://doi.org/10.1038/s41467-020-17086-8 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 (optimal for proﬁles containing a dip). (optimal for proﬁles containing a dip). 14. Wang, Y. et i 14. Wang, Y. et al. Combined expansion microscopy with structured illumination microscopy for analyzing protein complexes. Nat. Protoc. 13, 1869–1895 (2018). ð3Þ TriGaussian : y ¼ h1e ðx c1Þ 2w2 1 þ h2e ðx c2Þ 2w2 2 þ h3e ðx c3Þ 2w2 3 þ (optimal for proﬁles exhibiting a dip and high background signal). 15. Gao, M. et al. Expansion stimulated emission depletion microscopy (ExSTED). ACS Nano 12, 4178–4185 (2018). Cylinder : y ¼ h ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ r2 2 ðx cÞ2 q ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ r2 1 ðx cÞ2 q ; if x k k<r1 h ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ r2 2 ðx cÞ2 q ; if x k k ≥r1; x k k<r2 0; else 8 > > > > > < > > > > > : 16. Xu, H. et al. Molecular organization of mammalian meiotic chromosome axis revealed by expansion STORM microscopy. Proc. Natl Acad. Sci. USA 116, 18423–18428 (2019). ð4Þ 17. Shannon, C. E. Communication in the presence of noise. Proc. IEEE Inst. Electr. Electron Eng. 37, 10–21 (1949). g 18. Baddeley, D. & Bewersdorf, J. Biological insight from super-resolution microscopy: what we can learn from localization-based images. Annu. Rev. Biochem 87, 965–989 (2018). (y describes the theoretical intensity proﬁle of microtubules where r1 and r2 denote the inner and outer cylinder radius. The quality of the ﬁt strongly depends on the initial estimation of the parameters, due to the nonlinearity of the cylinder function.) ( ) 19. Legant, W. R. et al. High-density three-dimensional localization microscopy across large volumes. Nat. Methods 13, 359–365 (2016). 20. Heilemann, M. et al. Subdiffraction-resolution ﬂuorescence imaging with conventional ﬂuorescent probes. Angew. Chem. Int. Ed. 47, 6172–6176 (2008). Multi Cylinder : y ¼ cylði1; c; 25ex=2 2a; 25ex=2 aÞ þ cyl ði2; c; 42:5ex=2; 42:5ex=2 þ aÞ þ cylði3; c; 25ex=2 þ a; 25ex=2 þ 2aÞ þ b ð5Þ ð5Þ 21. Dempsey, G. T., Vaughan, J. C., Chen, K. H., Bates, M. & Zhuang, X. Evaluation of ﬂuorophores for optimal performance in localization-based super-resolution imaging. Nat. Methods 8, 1027–1036 (2011). (includes the theoretical dimensions of microtubules leaving less degrees of freedom. Might result in a better ﬁt). Data availability p g 25. Vaughan, J. C., Jia, S. & Zhuang, X. Ultrabright photoactivatable ﬂuorophores created by reductive caging. Nat. Methods 9, 1181–1184 (2012). All data that support the ﬁndings described in this study are available within the manuscript, the related supplementary information or deposited at https://doi.org/ 10.6084/m9.ﬁgshare.12415787.v1. Additional information is available from the corresponding authors upon reasonable request. 26. Cahoon, C. K. et al. Superresolution expansion microscopy reveals the three- dimensional organization of the Drosophila synaptonemal complex. Proc. Natl Acad. Sci. USA 114, E6857–E6866 (2017). 27. Li, Y. et al. Real-time 3D single-molecule localization using experimental point spread functions. Nat. Methods 15, 367–369 (2018). References 1. Chen, F., Tillberg, P. W. & Boyden, E. S. Expansion microscopy. Science 347, 543–548 (2015). 33. Gwosch, K. C. et al. MINFLUX nanoscopy delivers 3D multicolor nanometer resolution in cells. Nat. Methods 17, 217–224 (2020). 2. Gambarotto, D. et al. Imaging cellular ultrastructures using expansion microscopy (U-ExM). Nat. Methods 16, 71–74 (2019). 34. Deschout, D. et al. Precisely and accurately localizing single emitters in ﬂuorescence microscopy. Nat. Methods 11, 253–266 (2014). Ehemals 30. py 3. Wassie, A. T., Zhao, Y. & Boyden, E. S. Expansion microscopy: principles and uses in biological research. Nat. Methods 16, 33–41 (2019). 35. Virnat, D. et al. A peptide tag-speciﬁc nanobody enables high quality labeling for dSTORM imaging. Nat. Commun. 9, 930 (2018). uses in biological research. Nat. Methods 16, 33–41 (2019) g 4. Kunz, T. C., Götz, R., Sauer, M. & Rudel, T. Detection of Chlamydia 4. Kunz, T. C., Götz, R., Sauer, M. & Rudel, T. Detection of Chlamydia developmental forms and secreted effectors by expansion microscopy. Front. Cell Infect. Microbiol. 9, 276 (2019). 36. Wen, G. et al. Evaluation of direct grafting strategies via trivalent anchoring for enabling linpid membrane and cytoskeleton staining in expansion microscopy. ACS Nano https://doi.org/10.1021/acsnano.9b09259 (2020). developmental forms and secreted effectors by expansion microscopy. Front. Cell Infect. Microbiol. 9, 276 (2019). f 5. Tillberg, P. W. et al. Protein-retention expansion microscopy of cells and tissues labeled using standard ﬂuorescent proteins and antibodies. Nat. Biotechnol. 34, 987–992 (2016). 37. Shi, X. et al. Label-retention expansion microscopy. Preprint at https://doi.org/ 10.1101/687954 (2019). 6. Chen, F. et al. Nanoscale imaging of RNA with expansion microscopy. Nat. Methods 13, 679–684 (2016). 38. Ries, J., Kaplan, C., Platonova, E., Eghlidi, H. & Ewers, H. A simple, versatile method for GFP-based super-resolution microscopy via nanobodies. Nat. Methods 9, 582–584 (2012). ( ) 7. Chozinski, T. J. et al. Expansion microscopy with conventional antibodies and ﬂuorescent proteins. Nat. Methods 13, 485–488 (2016). 39. Mikhaykova, M. et al. Resolving bundled microtubules using anti-tubulin nanobodies. Nat. Commun. 6, 7933 (2015). 8. Ku, T. et al. Multiplexed and scalable super-resolution imaging of three- dimensional protein localization in size-adjustable tissues. Nat. Biotechnol. 34, 973–981 (2016). 40. Gao, R. et al. A highly homogeneous expansion microscopy polymer composed of tetrahedron-like monomers. Preprint at https://doi.org/10.1101/ 814111 (2019). 9. Truckenbrodt, S., Sommer, C., Rizzoli, S. O. & Danzl, J. G. A practical guide to optimization in X10 expansion microscopy. Nat. Protoc. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 Note that the ﬁt intensity (h) gives a good estimation for the relative labeling density. 22. Weber, K., Rathke, P. C. & Osborn, M. Cytoplasmic microtubular images in glutaraldehyde-ﬁxed tissue culture cells by electron microscopy and by immunoﬂuorescence microscopy. Proc. Natl Acad. Sci. USA 75, 1820–1824 (1978). Using the splines ﬁtted to the maximum intensity projection we constructed xz- proﬁle projections of microtubules, by taking line proﬁles in each z-stack of the 3D image. Averaging the aligned line proﬁles in a layer yields the intensity values for the corresponding row of the xz-projection. 23. Zhang, R., Alushin, G. M., Brown, A. & Nogales, E. Mechanistic origin of microtubule dynamic instability and its modulation by EB proteins. Cell 162, 849–859 (2015). Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article. 24. Zwettler, F. et al. Synaptonemal Complex Line Proﬁler. 10.5281/zenodo 2643214 https://doi.org/10.5281/zenodo.2643214 (2019). Received: 14 April 2020; Accepted: 5 June 2020; 30. Pigino, G. et al. Cryoelectron tomography of radial spokes in cilia and ﬂagella. J. Cell Biol. 195, 673–687 (2011). ( ) 31. Gu, L. et al. Molecular resolution imaging by repetitive optical selective exposure. Nat. Methods 16, 1114–1118 (2019). 31. Gu, L. et al. Molecular resolution imaging by rep exposure. Nat. Methods 16, 1114–1118 (2019). 32. Cnossen, J. et al. Localization microscopy at doubled precision with patterned illumination. Nat. Methods 17, 59–63 (2020). ARTICLE To analyze and compare the dif- ferent expansion protocols we developed a home written software that detects ﬁber like structures and automatically determines the transversal proﬁle along these structures in reconstructed SMLM images. In detail the SMLM images are ﬁrst convolved with a Gaussian blur compensating for noise discontinuity or holes. A thresholding algorithm45 then converts the image from grayscale to binary. Using Lees algorithm46 the expanded lines are reduced to one pixel width. The pixel coordinates from all still connected lines are then retrieved and tested for continuity. Points of discontinuity are used as breakpoints and all following coordinates are connected to a new line. Lines, shorter than the minimum required length are discarded. An univariate spline of degree 3 (c-spline) is ﬁtted to each line. Note that shape and gradient of the line depend on the smoothing parameter. The result is a table containing the spline coordinates and the local derivatives. Perpendicular to the derivative a line proﬁle is extracted from the original image at each coordinate point. The averaged proﬁles for each spline are ﬁtted with the following functions (Eqs. (1–5)): Gaussian : y ¼ he ðxcÞ2 2w2 þ b ð1Þ Gaussian : y ¼ he ðxcÞ2 2w2 þ b ð1Þ Mounting and SMLM image conditions. Re-embedded hydrogels immobilized on 24-mm cover glasses were immersed in photoswitching buffer consisting of 100 mM cysteamine hydrochloride (MEA) in PBS with optimized pH (adjusted with KOH) depending on the utilized ﬂuorescent dye. For Alexa Fluor 647 and Cy5 ﬂuorophores, the pH of the imaging buffer was adjusted to pH 7.7 and to pH 7.9 when using Alexa Fluor 532, respectively. The buffer was prepared freshly before (where h is the intensity, c the center, b the offset, and w the variance of the distribution. Optimal for single proﬁles). (where h is the intensity, c the center, b the offset, and w the variance of the distribution. Optimal for single proﬁles). BiGaussian : y ¼ h1e ðxc1Þ2 2w2 1 þ h2e ðxc2Þ2 2w2 2 þ b ð2Þ 9 ARTICLE Code availability 28. Hamel, V. et al. Identiﬁcation of chlamydomonas central core centriolar proteins reveals a role for human WDR90 in ciliogenesis. Curr. Biol. 27, 2486–2498 (2017). The automated image processing software Line Proﬁler is available at https://line-proﬁler. readthedocs.io/en/latest/. The automated image processing software Line Proﬁler is available at https://line-proﬁler. readthedocs.io/en/latest/. 29. Uno, S. N. et al. A spontaneously blinking ﬂuorophore based on intramolecular spirocyclization for live-cell super-resolution imaging. Nat. Chem. 6, 681–689 (2014). Received: 14 April 2020; Accepted: 5 June 2020; Received: 14 April 2020; Accepted: 5 June 2020; Author contributions F.U.Z., S.R., D.G., T.D.M.B., V.H., P.G., and M.S. conceived and designed the project. M.S, V.H., and P.G supervised the project. F.U.Z. performed all Ex-SMLM experiments. S.R. developed Line Proﬁler and analyzed the data together with F.U.Z., D.G. provided the centriole samples. All authors wrote and revised the ﬁnal manuscript. Acknowledgements The authors thank P. Gessner and L. Behringer-Pliess for assistance in immunocy- tochemistry and cell culture preparation. This work was supported by the German Research Foundation (DFG, TRR 166 ReceptorLight, project A04) and the European Regional Development Fund (EFRE project “Center for Personalized Molecular Immu- notherapy”). This work is supported by the European Research Council (ERC; StG 715289 ACCENT to P.G.) and the Swiss National Science Foundation (SNSF) PP00P3_187198 to P.G. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-020-17086-8 Additional information Correspondence and requests for materials should be addressed to V.H., P.G. or M.S. Correspondence and requests for materials should be addressed to V.H., P.G. or M.S. 46. Lee, T. C., Kashyap, R. L. & Chu, C. N. Building skeleton models via 3-D medial surface/axis thinning algorithms. Computer Vis. Graph. Image Process. 56, 462–478 (1994). Peer review information Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available. Reprints and permission information is available at http://www.nature.com/reprints References 14, 832–863 (2019). 41. Klena, N. et al. Isolation and ﬂuorescence imaging for single-particle reconstruction of Chlamydomonas centrioles. J. Vis. Exp. 2018, e58109 (2018). 10. Chang, J. B. et al. Iterative expansion microscopy. Nat. Methods 14, 593–599 (2017). 11. Sauer, M. & Heilemann, M. Single-molecule localization microscopy in eukaryotes. Chem. Rev. 117, 7478–7509 (2017). 42. Luca, G. M. Rde et al. Conﬁgurations of the re-scan confocal microscope (RCM) for biomedical applications. J. Microsc. 266, 166–177 (2017). y 12. Schermelleh, L. et al. Super-resolution microscopy demystiﬁed. Nat. Cell Biol. 21, 72–84 (2019). 43. Ovesný, M., Křížek, P., Borkovec, J., Švindrych, Z. & Hagen, G. M. ThunderSTORM: a comprehensive ImageJ plug-in for PALM and STORM data analysis and super-resolution imaging. Bioinformatics 30, 2389–2390 (2014). 13. Halpern, A. R., Alas, G. C. M., Chozinski, T. J., Paredez, A. R. & Vaughan, J. C. Hybrid structured illumination expansion microscopy reveals microbial cytoskeleton organization. ACS Nano 11, 12677–12686 (2017). 10 NATURE COMMUNICATIONS \| (2020) 11:3388 \| https://doi.org/10.1038/s41467-020-17086-8 \| www.nature.com/naturecommunicatio Additional information 44. Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682 (2012). 44. Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682 (2012). Supplementary information is available for this paper at https://doi.org/10.1038/s41467- 020-17086-8. 45. Otsu, N. A threshold selection method from grey level histograms. IEEE Trans. Syst. Man Cybernetics 9, 62–66 (1979). Competing interests © The Author(s) 2020 The authors declare no competing interests. 11 NATURE COMMUNICATIONS \| (2020) 11:3388 \| https://doi.org/10.1038/s41467-020-17086-8 \| www.nature.com/naturecommunications
https://openalex.org/W4309965631	https://zenodo.org/record/3928589/files/19720-37624-1-PB-Wavelet%20based.pdf	English	null	Wavelet based performance analysis of AC transmission systems with unified power flow controller under power quality issues	Zenodo (CERN European Organization for Nuclear Research)	2,019	cc-by	4,049	International Journal of Applied Power Engineering (IJAPE) Vol.8, No.3, December 2019, pp. 299~308 ISSN: 2252-8792 DOI: 10.11591/ijape.v8.i3.pp299-308 International Journal of Applied Power Engineering (IJAPE) Vol.8, No.3, December 2019, pp. 299~308 ISSN: 2252-8792 DOI: 10.11591/ijape.v8.i3.pp299-308  299  299 299 Corresponding Author: Corresponding Author: M. Nagaraju, Department of Electrical and Electronics Engineering, Audisankara Institute of Technology, Nellore, Andhra Pradesh 524405, India. Email: nag707@gmail.com M. Nagaraju, Department of Electrical and Electronics Engineering, Audisankara Institute of Technology, Nellore, Andhra Pradesh 524405, India. Email: nag707@gmail.com Wavelet based performance analysis of AC transmission systems with unified power flow controller under power quality issues M. Nagaraju1, V. V. K. Reddy 2, M. Sushama3 1Electronics Engineering, Audisankara Institute of Technology, India 2Department of Electrical and Electronics Engineering, NBKR Institute of Science and Technology, India 3Department of Electrical and Electronics Engineering, JNT University, Hyderabad, India ABSTRACT Article history: Received Feb 9, 2019 Revised Apr 25, 2019 Accepted May 12, 2019 The developments in power quality are fast and difficult to predict. The majority of power quality issues experienced by industrial customers can be attributed to momentary interruptions, voltage sags or swells, transients, harmonic distortion, electrical noise, and flickering lights, among others. A new device may be invented tomorrow solving power quality problems. The FACTS devices could provide fast control of active and reactive power through a transmission line. The unified power-flow controller (UPFC) is a member of the FACTS family with very attractive features. This device can independently control many parameters, so it is the combination of the properties of a static synchronous compensator (STATCOM) and static synchronous series compensator (SSSC).The performance of AC Transmission system with Unified power flow controller under various power quality problems analysis described. The proposed system is formulated and research work is done by wavelet multi resolution analysis using Bior1.5 mother wavelet with MATLAB/SIMULINK software. It is observed that the effectiveness of AC power transmission through Unified power flow controller under power quality problems of sag, swell, transient, temporary faults and capacitive switching. Keywords: FACTS Devices Power quality UPFC Voltage source converter wavelet analysis Copyright © 2019 Institute of Advanced Engineering and Science. All rights reserved. Copyright © 2019 Institute of Advanced Engineering and Science. All rights reserved. Corresponding Author: M. Nagaraju, Department of Electrical and Electronics Engineering, Audisankara Institute of Technology, Nellore, Andhra Pradesh 524405, India. Email: nag707@gmail.com Corresponding Author: 1. INTRODUCTION Power quality is an issue that is becoming increasingly important to electricity consumers at all levels of power system network. Sensitive and non-linear loads are common place in both the industrial and the domestic applications [1]. Developments in the power quality area will take long time, so that these problems will be around for at least several more years. Power quality problems related to long and short interruptions of power supply. Long interruptions are the most serious voltage quality disturbance. Short interruptions are shown to be due to a combination of automatic reclosing and a system design aimed at limiting the number of reclosers.Automatioc reclosing makes that a long interruption becomes a short interruption and such a mitigation method. Voltage sag is short duration reduction in voltage, caused by short circuits, overloads, and starting of induction motors. As long as voltage is sinusoidal, it does not matter whether rms voltage, fundamental voltage, or peak voltage is used to obtain the sag magnitude. The interface between the system and the equipment is the most common place to mitigate sag and interruptions. Most of the mitigation techniques are based on the injection of active power, thus Journal homepage: http://iaescore.com/journals/index.php/IJAPE  300 ISSN: 2252-8792 compensating the loss of active power supplied by the system. All modern techniques are based on power electronic devices, with the voltage source converters being the building block. A voltage source converter is a power electronic device which can generate a sinusoidal voltage at any required frequency, magnitude, and phase angle. The voltage source converter technology is used for “Flexible AC Transmission systems” and for mitigation of harmonic distortion and voltage fluctuations. compensating the loss of active power supplied by the system. All modern techniques are based on power electronic devices, with the voltage source converters being the building block. A voltage source converter is a power electronic device which can generate a sinusoidal voltage at any required frequency, magnitude, and phase angle. The voltage source converter technology is used for “Flexible AC Transmission systems” and for mitigation of harmonic distortion and voltage fluctuations. The UPFC consists of shunt and series compensation device is one of the flexible AC transmission system [2]. It can control bus voltage and power flows of a network. 2. PROPOSED UPFC CONNECTED TRANSMISSION MODEL The 48-pulse voltage source converter is composed of three-phase controllable bridges to produce current that is injected into a transmission line using a series transformer. In this UPFC is designed by using series and shunt voltage sourced converters can control active and reactive power flows in a transmission line. The UPFC is a combination of a static synchronous series compensator (SSSC) and static synchronous compensator (STATCOM) coupled through a common DC voltage linking injected voltage in quadrature with current within the power system [5]. The proposed UPFC connected transmission model illustrated in Figure 1. Figure 1. The proposed model of a UPFC connected transmission system Figure 1. The proposed model of a UPFC connected transmission system The series VSC blocks are connected in series with the transmission line by a series coupling transformer and it can provide either capacitive or inductive voltage compensation [6]. The converter is operated under fundamental frequency for the main bridges and six time fundamental frequency for the auxiliary circuit. The converter switching loses and switching device dynamic voltage stress is reduced significantly [7]. 1. INTRODUCTION This paper presents the complete digital simulation of the configurations of UPFC within the power system is performed in the MATLAB/Simulink environment to handle the power quality issues. The paper proposes a full model comprising of voltage source converter based UPFC is constructed for digital simulation to investigate the performance of the controller on various power quality problems and observed that considerable improvement is identified on 48-pulse Voltage source converter. The control scheme has the fast dynamic response and hence is adequate for improving transient behaviour of power system after transient conditions. When no UPFC is installed, real and reactive power through the transmission line cannot be controlled. A control system which enables the UPFC to follow the changes in reference values like AC voltage, DC voltage and angle order of the series voltage source converter is simulated. In past, the power quality problems are identified by total harmonic distortions, but this method of approach is failed under micro analysis. To overcome the previous analysis, the proposed research work is carried wavelet based performance analysis of ac transmission systems with unified power flow controller under power quality issues. The sum of the detailed coefficients of current and voltage signals are calculated by make use of bior1.5 mother wavelet at each terminal [3, 4]. The analysis can be carried under various power quality problems and it is found that the analysis is superior to existing methods. 3. DIGITAL SIMULATION MODEL OF UPFC The single line diagram representing proposed system model for 60-pulse the UPFC connected transmission system under study is illustrated in Figure 2. The sample system with the Unified power Flow Controller with shunt as well as series VSC controller and its control scheme is connected to four bus system. The feeder network is connected at Bus1 with a 25 kV,100MVA source and 3MW injected load connected at Bus2 , at injected load of 5MW and 2Mvar at bus 3. Int J Appl Power Eng, Vol. 8, No. 3, December 2019: 299 – 308 ISSN: 2252-8792 301  301  3  Int J Appl Power Eng Figure 2. Single line diagram of proposed model with the 48-pulse UPFC between Bus B2 and B3 Figure 2. Single line diagram of proposed model with the 48-pulse UPFC between Bus B2 and B3 The UPFC located at the left end of the 10 km feeder from B2. It consists of three 100-MVA, three- level, 48-pulse GTO-based converters, connected as shunt at bus B2. The shunt and series converter can exchange active and reactive power through a common DC bus. The reactive power variation is obtained by varying the DC bus voltage. The digital simulation model comprises of four 12-pulse GTO-converters, phase-shiftedby 〖7.5〗^0from each other and model, can provide the full 48-pulse converter operation. The 48-pulse converter can be used in high-voltage and power applications operation there is no need of any ac filters to control harmonic distortion content on the ac side [8, 9]. The output voltage has normal harmonics n=48r▁(+)1, where r=0, 1, 2, 3. 4. RESULTS AND ANALYSIS 4. RESULTS AND ANALYSIS The variation of three phase currents and voltages are shown in Figure 3 and Figure 4 under switching transient condition. The current and voltage transients are compensated with UPFC as shown in Figure 5 and Figure 6. The variation of three phase currents and voltages are shown in Figure 3 and Figure 4 under switching transient condition. The current and voltage transients are compensated with UPFC as shown in Figure 5 and Figure 6. Figure 3. Absolute and detailed index of transient current signal of uncompensated transmission system Wavelet based performance analysis of AC transmission systems with unified power flow … (M. Nagaraju) Figure 3. Absolute and detailed index of transient current signal of uncompensated transmission system Figure 4. Absolute and detailed index of transient voltage signal of uncompensated transmission system Figure 3. Absolute and detailed index of transient current signal of uncompensated transmission system Wavelet based performance analysis of AC transmission systems with unified power flow … (M. Nagaraju) Figure 4. Absolute and detailed index of transient voltage signal of uncompensated transmission system Wavelet based performance analysis of AC transmission systems with unified power flow … (M. Nagaraju Figure 4. Absolute and detailed index of transient voltage signal of uncompensated transmission system  ISSN: 2252-8792 302 Figure 5. Absolute and detailed index of transient current signal of UPFC Compensated transmission system igure 5. Absolute and detailed index of transient current signal of UPFC Compensated transmission system Figure 6. Absolute and detailed index of transient voltage signal of UPFC compensated transmission system Figure 6. Absolute and detailed index of transient voltage signal of UPFC compensated transmission system The variation of three phase currents and voltages are shown in Figure 7 and Figure 8 under capacitive reactive load in uncompensated the transmission system, current and voltage distortions are suppressed by Unified power flow condition compensator can be observed and illustrated in Figure 9 and Figure 10. Figure 7. Absolute and detailed index of current signal of uncompensated transmission system under leading reactive power injection Figure 7. Absolute and detailed index of current signal of uncompensated transmission system under leading reactive power injection Int J Appl Power Eng, Vol. 8, No. 3, December 2019: 299 – 308 ISSN: 2252-8792 303  Int J Appl Power Eng Figure 8. Absolute and detailed index of voltage signal of uncompensated transmission system under leading reactive power injection Figure 8. Wavelet based performance analysis of AC transmission systems with unified power flow … (M. Nagaraju) 4. RESULTS AND ANALYSIS Absolute and detailed index of voltage signal of uncompensated transmission system under leading reactive power injection Figure 9. Absolute and detailed index of current signal of UPFC compensated transmission system under leading reactive power injection Figure 9. Absolute and detailed index of current signal of UPFC compensated transmission system under leading reactive power injection Figure 10. Absolute and detailed index of voltage signal of UPFC compensated transmission system under leading reactive power injection Figure 10. Absolute and detailed index of voltage signal of UPFC compensated transmission system under leading reactive power injection The variation of three phase currents and voltages are shown in Figure 11 and Figure 12 under sudden load in uncompensated the transmission system, current and voltage distortions are suppressed by Unified power flow condition compensator can be observed and illustrated in Figure 13 and Figure 14. Wavelet based performance analysis of AC transmission systems with unified power flow … (M. Nagaraj ISSN: 2252-8792  304  ISSN: 2252 8792 304 Figure 11. Absolute and detailed index of current signal of uncompensated transmission system under sudden load Figure 12. Absolute and detailed index of voltage signal of uncompensated transmission system under sudden load Figure 13. Absolute and detailed index of current signal of UPFC compensated transmission system under sudden load Figure 11. Absolute and detailed index of current signal of uncompensated transmission system under sudden load Figure 11. Absolute and detailed index of current signal of uncompensated transmission system under sudden load Figure 12. Absolute and detailed index of voltage signal of uncompensated transmission system under sudden load Figure 12. Absolute and detailed index of voltage signal of uncompensated transmission system under sudden load Figure 13. Absolute and detailed index of current signal of UPFC compensated transmission system under sudden load Figure 13. Absolute and detailed index of current signal of UPFC compensated transmission system under sudden load Figure 14. Absolute and detailed index of voltage signal of UPFC compensated transmission system under sudden load Figure 14. Absolute and detailed index of voltage signal of UPFC compensated transmission system under sudden load Int J Appl Power Eng, Vol. 8, No. 3, December 2019: 299 – 308 Int J Appl Power Eng, Vol. 8, No. 3, December 2019: 299 – 308  305 305 ISSN: 2252-8792  Wavelet based performance analysis of AC transmission systems with unified power flow … (M. Nagaraju) Int J Appl Power Eng The variation of three phase currents and voltages are shown under single line to ground fault and double line to ground faults are described on uncompensated and compensated system are shown from Figure15 to Figure 22. The variation of three phase currents and voltages are shown under single line to ground fault and double line to ground faults are described on uncompensated and compensated system are shown from Figure15 to Figure 22. Figure 15:Absolute and detailed index of current signal of uncompensated transmission system under temporary single line to ground fault Figure 15:Absolute and detailed index of current signal of uncompensated transmission system under temporary single line to ground fault Figure 16. Absolute and detailed index of voltage signal of uncompensated transmission system under temporary single line to ground fault Figure 16. Absolute and detailed index of voltage signal of uncompensated transmission system under temporary single line to ground fault Figure 17. Absolute and detailed index of current signal of UPFC compensated transmission system under temporary single line to ground fault Figure 17. Absolute and detailed index of current signal of UPFC compensated transmission system under temporary single line to ground fault Figure 17. Absolute and detailed index of current signal of UPFC compensated transmission system under temporary single line to ground fault Wavelet based performance analysis of AC transmission systems with unified power flow … (M. Nagaraj ISSN: 2252-8792  306 Figure 18. Absolute and detailed index of voltage signal of UPFC compensated transmission system under temporary single line to ground fault Figure 18. Absolute and detailed index of voltage signal of UPFC compensated transmission system under temporary single line to ground fault Figure 19. Absolute and detailed index of current signal of uncompensated transmission system under temporary double line to ground fault Figure 19. Absolute and detailed index of current signal of uncompensated transmission system under temporary double line to ground fault Figure 20. Absolute and detailed index of voltage signal of uncompensated transmission system under temporary double line to ground fault Figure 20. Absolute and detailed index of voltage signal of uncompensated transmission system under temporary double line to ground fault Figure 21. Absolute and detailed index of current signal of UPFC compensated transmission system under temporary double line to ground fault Figure 21. 5. CONCLUSION The increasing complexities of large inter connected networks had fluctuations in reliability of power supply, which resulted in system instability, difficult to control the power flow, existence of voltage sag, voltage swell, interruptions, harmonic distortions, transients and security problems are analysed. The multiline transmission systems with UPFC is a VSC based controller is developed using the MATLAB/Simulink. For detecting and characterizing disturbances in the transmission system is carried wavelet multi resolution analysis technique is implemented. An operating current signals are identified and then sum of the detailed coefficients are calculated by make use of bior1.5 mother wavelet at each terminal. This is compared to various power quality problems. The proposed model is tested under various power quality problems and found that voltage source converter based UPFC can control power quality issues effectively. Int J Appl Power Eng Absolute and detailed index of current signal of UPFC compensated transmission system under temporary double line to ground fault Int J Appl Power Eng, Vol. 8, No. 3, December 2019: 299 – 308 ISSN: 2252-8792 Int J Appl Power Eng  307 Figure 22. Absolute and detailed index of voltage signal of UPFC compensated transmission system under temporary double line to ground fault Figure 22. Absolute and detailed index of voltage signal of UPFC compensated transmission system under temporary double line to ground fault The wavelet approach is more prominent method to analyse the reactive power compensating transmission system with as well as without UPFC under various power quality issues [10]. The wavelet approach is more prominent method to analyse the reactive power compensating transmission system with as well as without UPFC under various power quality issues [10]. Wavelet based performance analysis of AC transmission systems with unified power flow … (M. Nagaraj REFERENCES He started his career as Assistant Professor in the Department EEE, N.B.K.R.Institute of science Technology Nellore dist,AP, in the year 1981.. He worked Associate Professor for 13 years and as professor 8 years in the Department of Electrical Electronics Engineering, N.B.K.R.Institute of science Technology Nellore dist, AP . Presently is working as Professor & Director in N.B.K.R.Institute of science Technology Nellore dist, A India. He had 28 years of teaching & 10 years of research experience.He has published vario international and national conference papers as well as journal papers. Dr. M.Sushama, born on 8th Feb 1973, in Nalgonda , a small town near Nagarjuna Sagar, A. India .Obtained her B.Tech degree in 1993 and M.Tech degree in 2003 with a specialization Electrical Power Systems from Jawaharlal Nehru Technological University, India. She obtain her Ph.D. from JNTU Hyderabad, India in 2009 in the area of “Power Quality” using Wave Transforms. She started her career as Assistant Professor in the Department of EEE, JNT College of Engg., Anantapur, in the year 1995. She worked as Associate Professor for 8 years the Department of Electrical & Electronics Engineering, JNTUH College of Engg., Hyderaba Presently she is working as Professor in Electrical & Electronics Engineering, JNTUH Colle of Engineering, Kukatpally, Hyderabad. She had 20 years of teaching & 9 years of resear experience. She has published 17 international conference papers in various IEEE sponsor conferences, 15 International journal papers and one article in “Electrical INDIA”. Her resear interests include Power Quality, Wavelet Transforms, Neural & Fuzzy expert Systems. She currently guiding 5 Ph.D students. She is a life member of ISTE, Systems Society of India (SS & IETE. M. Nagaraju, born on 4th june 1978, in Naidupet,Nellore district, A.P., India. Obtained his B.Tech degree in 2002 and M.Tech degree in 2008 with a specialization in electrical power systems from Jawaharlal Nehru Technological University Hyderabad , India and pursuing PhD in Electrical power system at JNT University, Hyderabad. Presently working as Associate Professor department of Electrical &Electronics Engineering, Audisankara Institute Of Technology, Nellore Dist, AP, India. His area of interest includes Power Quality, Facts controllers, Wavelets and computer applications to power systems. V. Vijaya Kumar Reddy, born on 1st July 1955, in Nellore, a small town near gudur A.P., India .Obtained his B.Tech degree in 1979 and M.Tech degree in 1981 with a specialization in Electrical Machines & Industrial Drives from REC, Warangal, India. Int J Appl Power Eng, Vol. 8, No. 3, December 2019: 299 – 308 REFERENCES [1] Roger C. Dugan, Mark f. Mcgranaghan, Dr. Surya Santoso and H. Wayne Beaty., "Electrical power systems quality," Tata McGraw Hills publications, 2002. [2] Narain G. Hingorani, Laszlo Gyugyi, "Understanding FACTS: Concepts and Technology of Flexible AC Transmission Systems," Wiley-IEEE Press, 2000. ] O. Rioul and M. Vetterli, "Wavelets and signal processing," in IEEE Signal Processing Magazine, vol. 8, no. 4, p 14-38, Oct. 1991. [4] J. Barros, R.I. Diego, M.D. Apraiz, "Applications of wavelets in electric power quality: Voltage events," Electric Power Systems Research, vol. 88, pp. 130– 136, 2012. y pp [5] Abdelouahed Touhami, Zidi Sid Ahmed, "Modeling and Transient Simulation of Unified Power Flow Controllers (UPFC) in Power System". [6] Eskandar Gholipur and Shahrokh Saadate, "Improving of Transient Stability is Power Systems Using UPFC" IEEE Trans. Power Del., vol. 20, no. 2, pp. 1677-1682, Apr. 2005. [7] G.Ravi kumar, Shaik Abdul Gafoor, S.S.Tulasiram, "Fuzzy –Wavelet Based Double Transmission Line Transmission System Protection Scheme in The presence of UPFC" Electrical Power and Energy Systems, vol. 70, pp. 91-98, 2015. pp ] Collins C., Watson N. and Wood A., "UPFC Modeling in Harmonic Domain," IEEE Trans. on Power Deliver vol. 21, no. 2, pp. 933-938, 2006. [9] Yan Zhang, and Jovica V. Milanovic, "Global Voltage Sag Mitigation with FACTS Based Devices," IEEE Transactions on Power Delivery, vol. 25, no. 4, pp. 2842-2850, 2010. y, , , pp , [10] S. Muthukrishna and A. Nirmalkumar, "Enhancement of power quality in 14- bus system using UPFC," Research Journal of Applied sciences, Engineers and Technology, vol. 2, no. 4, pp. 356-361, 2010.  ISSN: 2252-8792 308 BIOGRAPHIES OF AUTHORS M. Nagaraju, born on 4th june 1978, in Naidupet,Nellore district, A.P., India. Obtained h B.Tech degree in 2002 and M.Tech degree in 2008 with a specialization in electrical pow systems from Jawaharlal Nehru Technological University Hyderabad , India and pursuing Ph in Electrical power system at JNT University, Hyderabad. Presently working as Associa Professor department of Electrical &Electronics Engineering, Audisankara Institute Technology, Nellore Dist, AP, India. His area of interest includes Power Quality, Fac controllers, Wavelets and computer applications to power systems. V. Vijaya Kumar Reddy, born on 1st July 1955, in Nellore, a small town near gudur A.P., Ind .Obtained his B.Tech degree in 1979 and M.Tech degree in 1981 with a specialization Electrical Machines & Industrial Drives from REC, Warangal, India. He obtained his Ph.D. fro NIT Warangal, India in 2008. REFERENCES He obtained his Ph.D. from NIT Warangal, India in 2008. He started his career as Assistant Professor in the Department of EEE, N.B.K.R.Institute of science Technology Nellore dist,AP, in the year 1981.. He worked as Associate Professor for 13 years and as professor 8 years in the Department of Electrical & Electronics Engineering, N.B.K.R.Institute of science Technology Nellore dist, AP . Presently he is working as Professor & Director in N.B.K.R.Institute of science Technology Nellore dist, AP, India. He had 28 years of teaching & 10 years of research experience.He has published various international and national conference papers as well as journal papers. Dr. M.Sushama, born on 8th Feb 1973, in Nalgonda , a small town near Nagarjuna Sagar, A.P., India .Obtained her B.Tech degree in 1993 and M.Tech degree in 2003 with a specialization in Electrical Power Systems from Jawaharlal Nehru Technological University, India. She obtained her Ph.D. from JNTU Hyderabad, India in 2009 in the area of “Power Quality” using Wavelet Transforms. She started her career as Assistant Professor in the Department of EEE, JNTU College of Engg., Anantapur, in the year 1995. She worked as Associate Professor for 8 years in the Department of Electrical & Electronics Engineering, JNTUH College of Engg., Hyderabad. Presently she is working as Professor in Electrical & Electronics Engineering, JNTUH College of Engineering, Kukatpally, Hyderabad. She had 20 years of teaching & 9 years of research experience. She has published 17 international conference papers in various IEEE sponsored conferences, 15 International journal papers and one article in “Electrical INDIA”. Her research interests include Power Quality, Wavelet Transforms, Neural & Fuzzy expert Systems. She is currently guiding 5 Ph.D students. She is a life member of ISTE, Systems Society of India (SSI) & IETE.
https://openalex.org/W2797574258	https://europepmc.org/articles/pmc5890979?pdf=render	English	null	Psychosocial and demographic factors influencing pain scores of patients with knee osteoarthritis	PloS one	2,018	cc-by	5,215	RESEARCH ARTICLE Editor: Carlos M. Isales, Augusta University, UNITED STATES Editor: Carlos M. Isales, Augusta University, UNITED STATES Received: April 4, 2017 Accepted: March 17, 2018 Published: April 9, 2018 Copyright: © 2018 Eberly et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Methods The pain-evaluation instrument was a 0- to 10-point rating scale. Data obtained retrospec- tively from the patients’ medical records were demographic characteristics, body mass index (BMI), concomitant disorders, illicit and prescription drug use, alcohol use, smoking, knee OA treatment, and severity of knee OA indicated by Kellgren-Lawrence (KL) radio- graphic grade. Univariate and multivariate analyses were performed to determine whether these variables correlated with reported pain scores. Copyright: © 2018 Eberly et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Lauren Eberly1☯, Dustin Richter1☯, George Comerci2☯, Justin Ocksrider1☯, Deana Mercer1☯, Gary Mlady3☯, Daniel Wascher1☯, Robert Schenck1☯ Lauren Eberly1☯, Dustin Richter1☯, George Comerci2☯, Justin Ocksrider1☯, Deana Mercer1☯, Gary Mlady3☯, Daniel Wascher1☯, Robert Schenck1☯ Lauren Eberly1☯, Dustin Richter1☯, George Comerci2☯, Justin Ocksrider1☯, Deana Mercer1☯, Gary Mlady3☯, Daniel Wascher1☯, Robert Schenck1☯ 1 Department of Orthopaedics & Rehabilitation, The University of New Mexico Health Sciences Center, Albuquerque, New Mexico, United States of America, 2 Department of Internal Medicine, The University of New Mexico Health Sciences Center, Albuquerque, New Mexico, United States of America, 3 Department of Radiology, The University of New Mexico Health Sciences Center, Albuquerque, New Mexico, United States of America a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 ☯These authors contributed equally to this work. DLRichter@salud.unm.edu Background Citation: Eberly L, Richter D, Comerci G, Ocksrider J, Mercer D, Mlady G, et al. (2018) Psychosocial and demographic factors influencing pain scores of patients with knee osteoarthritis. PLoS ONE 13(4): e0195075. https://doi.org/10.1371/journal. pone.0195075 Pain levels in patients with osteoarthritis (OA) of the knee are commonly assessed by using a numeric scoring system, but results may be influenced by factors other than the patient’s actual physical discomfort or disease severity, including psychosocial and demographic var- iables. We examined the possible relation between knee-pain scores and several psychoso- cial, sociodemographic, disease, and treatment variables in 355 patients with knee OA. ☯These authors contributed equally to this work. * DLRichter@salud.unm.edu Funding: The authors received no specific funding for this work. Our results in a large, ethnically diverse group of patients with knee OA suggest that psycho- social and sociodemographic factors may be important determinants of pain levels reported by patients with knee OA. Competing interests: The authors have declared that no competing interests exist. Competing interests: The authors have declared that no competing interests exist. Psychosocial and demographic determinants of pain in knee osteoarthritis Introduction More than nine million people in the United States have symptomatic knee osteoarthritis (OA) [1], with pain being the primary reason patients seek care and the leading cause of dis- ability from the disease [2,3]. Patients who present with OA are usually asked to describe their level of pain, often by referencing a numeric rating scale. However, a discrepancy between patients’ reports of their pain level and OA disease severity as assessed by radiographic studies has often been observed [4–6]. Therefore, it has been suggested that demographic and psycho- social factors may influence pain reports. Factors that have been assessed for their possible relation to reports of OA-associated pain include age, sex, body mass index (BMI), race/ethnicity, substance abuse, and psychological variables such as depression, hopelessness, overall well-being, and social stress [4,7–10]. The results of such correlation studies have varied, however, and some factors that may affect pain-level reports have not been examined. With the goal of increasing under- standing of determinants of pain reports in patients with knee OA, we conducted a retro- spective study of the possible correlation between pain score and several demographic, psychosocial, disease, and treatment characteristics, as well as between pain score and the severity of OA as assessed radiologically with use of the Kellgren-Lawrence (KL) classifica- tion system [11]. In addition, because little research has been done on a possible link between patients’ reported pain levels and clinical decision making, we examined whether pain scores were correlated with a recommendation by the surgeon that the patient undergo surgical rather than nonoperative treatment of OA. Our hypothesis was that demographic, psychosocial, and overall health-status variables would correlate with patients’ reported lev- els of pain. Results On univariate analysis, higher pain scores were significantly associated with Native Ameri- can or Hispanic ethnicity; a higher BMI; current prescription for an opioid, antidepressant, or gabapentinoid medication; depression; diabetes mellitus; fibromyalgia; illicit drug use; lack of health insurance; smoking; previous knee injection; and recommendation by the cli- nician that the patient undergo knee surgery. Neither the patient’s sex nor the KL grade showed a correlation. On multivariate analysis, depression, current opioid prescription, and Native American or Hispanic ethnicity retained a significant association with higher pain scores. Data Availability Statement: Data were obtained from The University of New Mexico Hospital Electronic Medical Record database. Due to the regulations of HIPAA in protecting patient privacy, some of the potentially identifiable data is not available for use. Deidentified data from the study is available from the corresponding author by contacting him via email at DLRichter@salud.unm. edu, or from Sahar Freedman at SaharF@salud. unm.edu (505-925-0614). 1 / 11 PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 Conclusions and implications Funding: The authors received no specific funding for this work. Patients With approval from the Human Research Review Committee of The University of New Mexico Health Sciences Center (HRRC#13–523), we reviewed the medical records of all 611 new patients who presented to our orthopaedics clinic and received a primary diagnosis of knee OA between 1st January 2013 and 31st December 2013. In each case, the diagnosis was confirmed by an evaluation of the patient’s medical history, a physical examination, radio- graphic studies, or a combination of these methods. Patients with a concurrent ligamentous injury, inflammatory arthritis, or bilateral knee OA were excluded from the study; therefore, 355 were enrolled. Institutional review board waived the need for patient consent, and data was accessed anonymously with no patient identifiers. At their initial presentation, all patients had been asked to describe their pain level with respect to a number from 0 to 10, with 0 representing “no pain” and 10 “the worst pain imag- inable.” This numeric scale is commonly used to assess arthritic pain [12]. Aside from pain score, the record review obtained the following data for each patient: age; sex; race or ethnicity PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 2 / 11 Psychosocial and demographic determinants of pain in knee osteoarthritis (self-reported by patients’ checking a box on intake form); BMI; current tobacco, alcohol, and illicit drug use (both illegal drugs and overuse of legal drugs); current prescription for an opi- oid, gabapentinoid, or antidepressant agent; current diagnosis of depression, fibromyalgia, or posttraumatic stress disorder (PTSD); health-insurance status (yes or no); previous injection of a corticosteroid agent or hyaluronic acid in the affected knee; and whether the patient’s cli- nician recommended surgical treatment of the knee OA during the initial presentation. Infor- mation on smoking, alcohol, and illicit drug use was obtained with a self-report intake form that did not allow specification of the level of use or the type of illicit drug. BMI was calculated from patients’ most recently documented height and weight measured at that clinic visit (S1 Dataset). At the time of the medical-record review, radiographs obtained during the initial patient visit were assessed by a musculoskeletal radiologist who was blinded to all patient information in the records, including pain scores. The radiologist assigned each image a KL grade of 0, 1, 2, 3, or 4 on the basis of the extent of degenerative changes observed. Statistical analysis In estimating the appropriate sample size for the study, we assumed that most clinicians would consider a difference in pain score of 2 to be clinically important and we wanted to limit the type I error to 0.05 or lower and achieve a power of 90%. Because we anticipated simultaneous analysis of up to 22 independent variables, we used a type I error of 0.002 (on the basis of the Bonferroni inequality). Under these assumptions, we calculated that the study should include about 310 patients. Standard summary statistics were calculated for all variables. Univariate analysis was used initially to assess the possible relation between pain score and each independent binary vari- able studied. The Student t test or Mann-Whitney U test was applied as appropriate. Analysis of variance was used to examine the possible relation between pain score and each of the fol- lowing: age, BMI, race or ethnic group, and KL grade. A p value of <0.05 was considered to indicate a statistically significant difference. Multivariate analysis was performed by using a general linear model algorithm and maxi- mum-likelihood estimation. Variables that were significantly associated with pain score in the univariate analysis were sequentially fitted into the model with use of adjusted R2 analysis. The variables were removed sequentially until only those that had a significant relation with pain score remained in the model. All statistical analyses were performed with Statgraphics Centu- rion XV software (StatPoint, Herndon, VA). PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 Patient characteristics Table 1 shows characteristics of the 355 patients in the study. Most of the patients were white women, but our cohort was also diverse in that Native American and Hispanic patients each comprised more than 10% of the total. The patients ranged in age from twenty-four to ninety years. The average BMI was 31.0 kg/m2 (range, 19.1–61.9 kg/m2). The overall mean (standard deviation [SD]) pain score was 5.0 (2.9). The mean pain scores in white, Native American, His- panic, “other,” and black patients were 4.5 (3.0), 6.3 (2.7), 6.4 (2.5), 4.7 (3.1), and 5.8 (2.8), respectively. The “other” category included two patients who indicated that they were Asian and twenty-seven who reported that they were “other” than white, Native American, Hispanic, or black. The most common KL score was 3, indicating moderate OA. 3 / 11 PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 Psychosocial and demographic determinants of pain in knee osteoarthritis Psychosocial and demographic determinants of pain in knee osteoarthritis Table 1. Demographic and clinical variables for 355 patients with knee osteoarthritisa. Variable Value Age, y, mean ± standard deviation 58.6 ± 11.8 Male 139 (39) Female 216 (61) Body mass index, kg/m2, mean ± standard deviation 31.2 ± 7.3 Ethnicityb White 191 (64.7) Native American 35 (11.9) Hispanic 30 (10.2) Other 29 (9.8) Black 10 (3.4) Comorbidities/insurance/history Current smoking 56 (15.9) Alcohol use 157 (46.6) Illicit drug use 18 (5.4) Opioid-agent prescription 75 (21.6) Gabapentinoid prescription 43 (12.4) Antidepressant prescription 84 (24.2) Depression 94 (27.1) Fibromyalgia 22 (6.4) Posttraumatic stress disorder 7 (2.0) Health insurance 318 (91.4) Previous knee injection 110 (32.3) Surgery recommended by clinician 59 (16.8) Kellgren-Lawrence gradec 0 9 (2.8) 1 46 (14.2) 2 82 (25.4) 3 110 (34.1) 4 76 (23.5) Values represent number of patients (percentage of the indicated group), unless otherwise stated. Values represent number of patients (percentage of the indicated group), unless otherwise stated. cA Kellgren-Lawrence grade of 0 (no osteophytes or joint-space narrowing) indicates no osteoarthritis; a grade of 1 (questionable osteophytes), possible osteoarthritis; a grade of 2 (definite osteophytes, no joint-space narrowing), mild osteoarthritis; a grade of 3 ( 50% joint-space narrowing), moderate osteoarthritis; and a grade of 4 ( 50% join- space narrowing), severe osteoarthritis. Univariate and multivariate analysis On univariate analysis, age had a significant inverse relation to pain score, with younger sub- jects having significantly higher scores (p = 0.03). Compared with white patients, Hispanic or Native American patients had significantly higher pain scores (p < 0.001 for both compari- sons; (Fig 1), but there were no other significant differences among racial or ethnic groups. Patients with higher BMIs also had higher pain scores (p < 0.001). With respect to binary vari- ables (Table 2), patients with a current opioid prescription, depression, fibromyalgia, illicit drug use, current antidepressant or gabapentinoid prescription, uninsured status, smoking, previous knee injection, or recommendation for operative treatment of OA had significantly 4 / 11 PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 Psychosocial and demographic determinants of pain in knee osteoarthritis Fig 1. Ethnicity and pain scores. Mean knee-pain scores in 355 patients with knee osteoarthritis, according to patients’ ethnic or racial group. The error bars represent 95% confidence intervals adjusted by using the Tukey procedure for multiple comparisons. The difference between Hispanic and white patients and between Native American and white patients was significant (p < 0.001 for both comparisons). https://doi.org/10.1371/journal.pone.0195075.g001 Fig 1. Ethnicity and pain scores. Mean knee-pain scores in 355 patients with knee osteoarthritis, according to patients’ ethnic or racial group. The error bars represent 95% confidence intervals adjusted by using the Tukey procedure for multiple comparisons. The difference between Hispanic and white patients and between Native American and white patients was significant (p < 0.001 for both comparisons). https://doi.org/10.1371/journal.pone.0195075.g001 https://doi.org/10.1371/journal.pone.0195075.g001 higher pain scores than those without these characteristics. Patients who said that they drank alcohol had significantly lower pain scores than those who said that they did not. The only variables that were not related to reported pain levels were the sex of the patient, a diagnosis of PTSD, and KL grade (p = 0.2; Fig 2). The mean pain scores (SD) according to KL grade were 4.33 (1.89) for grade 0; 4.93 (0.84) for grade 1; 4.70 (0.63) for grade 2; 4.89 (0.54) for grade 3; and 5.74 (0.65) for grade 4. On multivariate analysis, current opioid prescription, depression, and Native American or Hispanic ethnicity remained in the model, indicating a significant relation between those vari- ables and pain scores (p < 0.001, p = 0.002, and p < 0.001, respectively). PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 Discussion In a univariate analysis of a large, diverse cohort of patients with knee OA, we found that sev- eral psychosocial, sociodemographic, disease, and treatment variables were each significantly associated with higher reported knee-pain scores. A subsequent multivariate analysis showed a significant association between higher pain scores and Hispanic or Native American ethnicity, opioid agent prescription, and depression. There was no relation between pain score and KL grade, a discordance also observed in several earlier studies [4–6]. Many of the factors analyzed have not, to our knowledge, previously been correlated with pain score: Hispanic or Native American ethnicity, age, opioid prescription, fibromyalgia, illicit drug use, antidepressant pre- scription, alcohol use (inverse relation), gabapentinoid prescription, health-insurance status, smoking, previous knee injection, and recommendation for surgical versus nonoperative treat- ment of knee OA. It is important to note that although the results yielded statistical signifi- cance across multiple factors analyzed in this study, the factors with clinical significance, 5 / 11 PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 Psychosocial and demographic determinants of pain in knee osteoarthritis Table 2. Association of binary variables and pain score on univariate analysis. Variable Mean ± standard deviation score (95% CI) P Value Sex Female 5.00 ± 2.91 (4.61–5.39) Male 5.05 ± 2.98 (4.55–5.55) 0.86 Opioid-agent prescription Yes 6.84 ± 2.94 (6.36–7.32) No 4.48 ± 2.08 (4.13–4.83) <0.001 Depression Yes 6.19 ± 2.52 (5.67–6.71) No 4.58 ± 2.94 (4.22–4.95) <0.001 Post-traumatic Stress Disorder Yes 6.14 ± 2.48 (3.85–8.44) No 4.96 ± 2.93 (4.64–5.27) 0.29 Fibromyalgia Yes 6.77 ± 2.07 (5.86–7.69) No 4.86 ± 2.93 (4.54–5.18) 0.003 Illicit drug use Yes 6.89 ± 2.58 (5.60–8.18) No 4.88 ± 2.91 (4.55–5.20) 0.004 Antidepressant prescription Yes 5.76 ± 2.77 (5.15–6.36) No 4.73 ± 2.96 (4.37–5.09) 0.005 Gabapentinoid prescription Yes 6.14 ± 2.53 (5.36–6.92) No 4.82 ± 2.96 (4.48–5.15) 0.006 Alcohol use Yes 4.53 ± 2.73 (4.10–4.96) No 5.36 ± 3.04 (4.91–5.81) 0.009 Health insurance Yes 4.90 ± 2.93 (4.57–5.22) No 6.27 ± 2.66 (5.27–7.26) 0.01 Smoking Yes 5.82 ± 2.39 (5.18–6.46) No 4.85 ± 2.99 (4.50–5.19) 0.02 Previous knee injection Yes 5.47 ± 2.94 (4.92–6.02) No 4.75 ± 2.92 (4.37–5.13) 0.03 Surgery recommended by clinician Yes 5.69 ± 2.94 (4.92–6.02) No 4.87 ± 2.92 (4.53–5.20) 0.04 Abbreviation: CI, confidence interval. https://doi org/10 1371/journal pone 0195075 t002 https://doi.org/10.1371/journal.pone.0195075.t002 according to our stated criteria of a difference in pain score of at least two, are opioid prescrip- tions and illicit drug use. PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 Discussion Other research has examined a possible link between knee OA-associated pain and race/ ethnicity. A study by Creamer et al. [4] included only black and white patients and noted that the former reported higher pain scores than the latter. Additionally, one of the few studies to PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 6 / 11 Psychosocial and demographic determinants of pain in knee osteoarthritis Fig 2. Kellgren-Lawrence grade and pain scores. Mean knee-pain scores in 355 patients with knee osteoarthritis, according to Kellgren-Lawrence Grade. The error bars represent 95% confidence intervals adjusted by using the Tukey procedure for multiple comparisons. There were no significant differences between groups. https://doi.org/10.1371/journal.pone.0195075.g002 PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 Previous studies found that depression affects reported pain in patients with rheu- matoid arthritis [18], and an influence of depression on pain severity in patients with OA has been suggested [6,19]. The higher reported pain scores might reflect pain catastrophizing, which is a tendency to focus on negative pain sensations, thereby exaggerating the pain experi- ence and enhancing feelings of helplessness [20]. It was previously shown that depression can lead to increased pain through pain catastrophizing in patients with musculoskeletal condi- tions [21], and pain catastrophizing has been linked to higher pain scores in patients with OA [22]. Cognitive behavioral therapy or psychotherapy aimed at reducing catastrophizing in these patients and similar patients with fibromyalgia may produce better outcomes than phar- macologic therapy [23]. In agreement with earlier findings [7,8], the sex of the patient was not correlated with pain scores. Interestingly, although the prevalence of knee OA increases with age, pain scores were significantly higher in our study’s younger patients. Psychological factors such as depression may have influenced pain reports in these patients. Indeed, younger adults with OA have been shown to have higher rates of depression and depressive symptoms than older adults with the disease [24,25]. In addition, life stress and hypochondriasis, which are associated with higher pain scores, are more prevalent in younger people [10]. Obesity, a known risk factor for development of OA, has been linked to an increased sever- ity of OA-related pain, although Somers et al. [26] found that BMI was not correlated with pain scale scores. Our results show that patients with higher BMIs were significantly more likely to have higher knee-pain scores. Clinicians counseling obese patients with OA-related knee pain should consider suggesting weight loss as possible treatment. The association between higher pain scores and uninsured status is probably related to socioeconomic status because OA pain and symptomatic knee OA have been correlated with lower income levels [7] and the poverty rate in a community [27], respectively. In addition, “feeling helpless” has been found to constitute an important determinant of higher pain scores [8], and patients without health insurance may have an increased tendency to experience this feeling, perhaps accompanied by a feeling of disenfranchisement. Our study had the usual limitations of a retrospective, observational investigation. In addi- tion, there may have been interactions among the variables assessed that our study design could not identify. Fig 2. Kellgren-Lawrence grade and pain scores. Mean knee-pain scores in 355 patients with knee osteoarthritis, according to Kellgren-Lawrence Grade. The error bars represent 95% confidence intervals adjusted by using the Tukey procedure for multiple comparisons. There were no significant differences between groups. https://doi.org/10.1371/journal.pone.0195075.g002 include Hispanic patients with rheumatic disease found significantly poorer global estimates of participant health, compared with those made by physicians [13]. The diversity of our study cohort provided an opportunity to compare pain scores among four different racial/ethnic groups. Both our univariate and multivariate analyses found that Native American or Hispanic patients had significantly higher pain scores than white patients. Studies of arthritis-associated pain in Native Americans are scarce; however, in an investi- gation by Kramer et al. [14] in which face-to-face interviews were conducted with fifty-six Native Americans living in an urban area, most interviewees commented on their cultural practice of minimizing pain complaints and reported that American Indians do not readily discuss pain. The Native Americans studied by Kramer et al. lived in an urban area, whereas many of our Native American patients live in rural areas and are of different tribes. There may be cultural variations among tribes regarding the pain experience or the reporting of pain in a setting in which few, if any, of the clinicians are Native American. Nevertheless, these studies do show the importance of clinician awareness of possible ethnic-group differences in report- ing OA-pain experience. The association between current opioid agent prescription and higher pain scores possibly stems from neuromodulation of pain sensitivity resulting from chronic opioid use [15,16]. It is possible that some of our patients with an opioid prescription reported greater pain as a mani- festation of drug-seeking behavior whereas other patients may simply have higher baseline pain levels. However, the differences in pharmacodynamics, opioid-receptor interactions, and PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 7 / 11 Psychosocial and demographic determinants of pain in knee osteoarthritis bioavailability of a particular opioid dose likely play an important role in sufficient pain relief. The use of a pain-assessment instrument that evaluates patients’ functional abilities, such as the Western Ontario and McMaster Universities Osteoarthritis Index [17], might be useful when pain assessment is unclear. Our multivariate analysis also found a significant relation between higher pain scores and depression. PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 Supporting information S1 Dataset. Knee osteoarthritis patient raw data. Raw data extracted from the chart review for all patients who met study inclusion criteria. (XLS) S1 Dataset. Knee osteoarthritis patient raw data. Raw data extracted from the chart review for all patients who met study inclusion criteria. (XLS) Some factors not examined, such as whether the patient was currently involved in a workers’ compensation evaluation, might have influenced pain-score reports. Additionally, our study had a small number of black patients, with a large percentage of His- panic and Native American patients. This unique ethnic distribution, however, is a reflection of our state’s composite population. Finally, the pain-evaluation method that was used was a simple numeric rating scale. This assessment can be administered quickly, but it may not suffi- ciently characterize a patient’s pain experience. Despite these limitations, our results indicate that psychosocial and sociodemographic fac- tors significantly affect patients’ reports of their level of OA-related knee pain and that patients with OA do not constitute a homogenous group for which the same management approaches will suffice. Management should be tailored to the individual and characterized by more exten- sive patient education regarding the specific factors contributing to their pain, as well as dis- cussion of appropriate expectations for pain-level. The usefulness of a pain assessment based primarily on a numeric pain score is of variable utility. Extensive reliance on such an PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 8 / 11 Psychosocial and demographic determinants of pain in knee osteoarthritis assessment in clinical decision-making may be inappropriate, especially when invasive proce- dures are contemplated. With the limitations that come with the visual analog scale, clinicians could consider employing various self-reported and physical measures recommended by the Osteoarthritis Research Society International (OARSI) such as the Dallas Pain Questionnaire or the WOMAC Questionnaire. Ultimately, pain-treatment decisions should not be based pri- marily on KL grade but the entire patient presentation. Although clinicians cannot modify some of the factors associated with higher pain scores in our study, they can increase their insight into the possible causes of an observed discrep- ancy between a patient’s clinical appearance with respect to comfort and his or her pain score and perhaps mitigate the influence of factors that can be changed. Clinicians who are aware of the psychosocial and sociodemographic characteristics that affect their patients’ reports of OA-related knee pain will thereby be able to provide more effective, patient- centered care. Acknowledgments We thank William D. Tandberg, MD for performing the statistical analysis and Renee´ Robil- lard and Sahar Freedman for editorial assistance. This study received the Vernon L. Nickel Award of 2015 by the Orthopaedic Rehabilitation Association and was presented at the 68th Annual Meeting of the Association of Bone and Joint Surgeons. Author Contributions Conceptualization: Lauren Eberly, Daniel Wascher. Data curation: Lauren Eberly, Dustin Richter, Daniel Wascher. Data curation: Lauren Eberly, Dustin Richter, Daniel Wascher. Data curation: Lauren Eberly, Dustin Richter, Daniel Wascher. Formal analysis: Lauren Eberly, George Comerci, Justin Ocksrider, Robert Schenck. Investigation: Lauren Eberly, Daniel Wascher. Methodology: Lauren Eberly, Dustin Richter, George Comerci, Justin Ocksrider, Deana Mer- cer, Gary Mlady, Daniel Wascher, Robert Schenck. Project administration: Robert Schenck. References 1. Lawrence RC, Felson DT, Helmick CG, Arnold LM, Choi H, Deyo RA, et al; National Arthritis Data Work- group. Estimates of the prevalence of arthritis and other rheumatic conditions in the United States. Part II. Arthritis Rheum. 2008 Jan; 58(1):26–35. https://doi.org/10.1002/art.23176 PMID: 18163497 2. McAlindon TE, Cooper C, Kirwan JR, Dieppe PA. Knee pain and disability in the community. Br J Rheu- matol. 1992 Mar; 31(3):189–92. PMID: 1540789 3. Sokka T. Assessment of pain in patients with rheumatic diseases. Best Pract Res Clin Rheumatol. 2003 Jun; 17(3):427–49. PMID: 12787511 4. Creamer P, Lethbridge-Cejku M, Hochberg MC. Determinants of pain severity in knee osteoarthritis: effect of demographic and psychosocial variables using 3 pain measures. J Rheumatol. 1999 Aug; 26 (8):1785–92. PMID: 10451078 5. Hannan MT, Felson DT, Pincus T. Analysis of the discordance between radiographic changes and knee pain in osteoarthritis of the knee. J Rheumatol. 2000 Jun; 27(6):1513–7. PMID: 10852280 6. Summers MN, Haley WE, Reveille JD, Alarco´n GS. Radiographic assessment and psychologic vari- ables as predictors of pain and functional impairment in osteoarthritis of the knee or hip. Arthritis Rheum. 1988 Feb; 31(2):204–9. PMID: 3348824 7. Carman WJ. Factors associated with pain and osteoarthritis in the Tecumseh Community Health Study Semin Arthritis Rheum. 1989 May; 18(4 Suppl 2):10–3. PMID: 2786252 8. Davis MA, Ettinger WH, Neuhaus JM, Barclay JD, Segal MR. Correlates of knee pain among US adults with and without radiographic knee osteoarthritis. J Rheumatol. 1992 Dec; 19(12):1943–9. PMID: 1294744 9. Lachance L, Sowers M, Jamadar D, Jannausch M, Hochberg M, Crutchfield M. The experience of pain and emergent osteoarthritis of the knee. Osteoarthritis Cartilage. 2001 Aug; 9(6):527–32. https://doi. org/10.1053/joca.2000.0429 PMID: 11520166 10. Lichtenberg PA, Swensen CH, Skehan MW. Further investigation of the role of personality, lifestyle and arthritic severity in predicting pain. J Psychosom Res. 1986; 30(3):327–37. PMID: 3735177 11. Kellgren JH, Lawrence JS. Radiological assessment of rheumatoid arthritis. Ann Rheum Dis. 1957 Dec; 16(4):485–93. PMID: 13498603 12. Hawker GA, Mian S, Kendzerska T, French M. Measures of adult pain: Visual Analog Scale for Pain (VAS Pain), Numeric Rating Scale for Pain (NRS Pain), McGill Pain Questionnaire (MPQ), Short-Form McGill Pain Questionnaire (SF-MPQ), Chronic Pain Grade Scale (CPGS), Short Form-36 Bodily Pain Scale (SF-36 BPS), and Measure of Intermittent and Constant Osteoarthritis Pain (ICOAP). Arthritis Care Res (Hoboken). 2011 Nov; 63 Suppl 11:S240–52. 13. Castrejo´n I, Yazici Y, Samuels J, Luta G, Pincus T. Project administration: Robert Schenck. Resources: Gary Mlady. Supervision: George Comerci, Daniel Wascher, Robert Schenck. Supervision: George Comerci, Daniel Wascher, Robert Schenck. Validation: Dustin Richter, George Comerci, Robert Schenck. Validation: Dustin Richter, George Comerci, Robert Schenck. Visualization: Lauren Eberly, Dustin Richter, George Comerci, Justin Ocksrider, Deana Mer- cer, Gary Mlady, Robert Schenck. Writing – original draft: Lauren Eberly, Daniel Wascher. Writing – review & editing: Lauren Eberly, Dustin Richter, George Comerci, Justin Ocksri- der, Deana Mercer, Gary Mlady, Robert Schenck. PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 9 / 11 Psychosocial and demographic determinants of pain in knee osteoarthritis PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 References Discordance of global estimates by patients and their physicians in usual care of many rheumatic diseases: association with 5 scores on a Multidimen- sional Health Assessment Questionnaire (MDHAQ) that are not found on the Health Assessment Ques- tionnaire (HAQ). Arthritis Care Res (Hoboken). 2014 Jun; 66(6):934–42. 14. Kramer BJ, Harker JO, Wong AL. Arthritis beliefs and self-care in an urban American Indian population. Arthritis Rheum. 2002 Dec 15; 47(6):588–94. https://doi.org/10.1002/art.10795 PMID: 12522831 15. Doehring A, Oertel BG, Sittl R, Lo¨tsch J. Chronic opioid use is associated with increased DNA methyla- tion correlating with increased clinical pain. Pain. 2013 Jan; 154(1):15–23. https://doi.org/10.1016/j. pain.2012.06.011 PMID: 23273101 16. Pud D, Cohen D, Lawental E, Eisenberg E. Opioids and abnormal pain perception: New evidence from a study of chronic opioid addicts and healthy subjects. Drug Alcohol Depend. 2006 May 20; 82(3):218– 23. https://doi.org/10.1016/j.drugalcdep.2005.09.007 PMID: 16229972 17. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol. 1988 Dec; 15(12):1833–40. PMID: 3068365 18. Parker JC, Wright GE. The implications of depression for pain and disability in rheumatoid arthritis. Arthritis Care Res. 1995 Dec; 8(4):279–83. PMID: 8605267 19. Salaffi F, Cavalieri F, Nolli M, Ferraccioli G. Analysis of disability in knee osteoarthritis. Relationship with age and psychological variables but not with radiographic score. J Rheumatol. 1991 Oct; 18 (10):1581–6. PMID: 1765985 20. Sullivan MJ, Thorn B, Haythornthwaite JA, Keefe F, Martin M, Bradley LA, et al. Theoretical perspec- tives on the relation between catastrophizing and pain. Clin J Pain. 2001 Mar; 17(1):52–64. PMID: 11289089 10 / 11 PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 Psychosocial and demographic determinants of pain in knee osteoarthritis 21. Linton SJ, Nicholas MK, MacDonald S, Boersma K, Bergbom S, Maher C, et al. The role of depression and catastrophizing in musculoskeletal pain. Eur J Pain. 2011 Apr; 15(4):416–22. https://doi.org/10. 1016/j.ejpain.2010.08.009 PMID: 20884261 22. Edwards RR, Bingham CO 3rd, Bathon J, Haythornthwaite JA. Catastrophizing and pain in arthritis, fibromyalgia, and other rheumatic diseases. Arthritis Rheum. 2006 Apr 15; 55(2):325–32. https://doi. org/10.1002/art.21865 PMID: 16583384 23. Dexter P, Brandt K. Distribution and predictors of depressive symptoms in osteoarthritis. J Rheumatol. 1994 Feb; 21(2):279–86. PMID: 8182638 24. Shih M, Hootman JM, Strine TW, Chapman DP, Brady TJ. Serious psychological distress in U.S. adults with arthritis. PLOS ONE \| https://doi.org/10.1371/journal.pone.0195075 April 9, 2018 References J Gen Intern Med. 2006 Nov; 21(11):1160–6. https://doi.org/10.1111/j.1525-1497.2006. 00573.x PMID: 16879706 25. Somers TJ, Keefe FJ, Pells JJ, Dixon KE, Waters SJ, Riordan PA, et al. Pain catastrophizing and pain- related fear in osteoarthritis patients: relationships to pain and disability. J Pain Symptom Manage. 2009 May; 37(5):863–72. https://doi.org/10.1016/j.jpainsymman.2008.05.009 PMID: 19041218 26. Bernardy K, Klose P, Busch AJ, Choy EH, Ha¨user W. Cognitive behavioural therapies for fibromyalgia. Cochrane Database Syst Rev. 2013 Sep 10; 9:CD009796. 27. Callahan LF, Cleveland RJ, Shreffler J, Schwartz TA, Schoster B, Randolph R, et al. Associations of educational attainment, occupation and community poverty with knee osteoarthritis in the Johnston County (North Carolina) osteoarthritis project. Arthritis Res Ther. 2011; 13(5):R169. https://doi.org/10. 1186/ar3492 PMID: 22011570 11 / 11
https://openalex.org/W2087583005	https://europepmc.org/articles/pmc2770517?pdf=render	English	null	Surveillance indicators for potential reduced exposure products (PREPs): developing survey items to measure awareness	Harm reduction journal	2,009	cc-by	7,220	BioMed Central Harm Reduction Journal Open Access Research Surveillance indicators for potential reduced exposure products (PREPs): developing survey items to measure awareness Karen Bogen1, Lois Biener2, Catherine A Garrett2, Jane Allen3, K Michael Cummings4, Anne Hartman5, Stephen Marcus5, Ann McNeill6, Richard J O'Connor4, Mark Parascandola5 and Linda Pederson7 Address: 1Mathematica Policy Research, Cambridge, MA, USA, 2Center for Survey Research, University of Massachusetts Boston, Boston, MA, USA, 3American Legacy Foundation, Washington, DC, USA, 4Department of Health Behavior, Roswell Park Cancer Institute, Buffalo, NY, USA, 5National Cancer Institute, National Institutes of Health, Washington, DC, USA, 6Division of Epidemiology and Public Health, University of Nottingham, Nottingham, UK and 7Beth-El College of Nursing & Health Sciences, Univ of Colorado, Colorado Springs, CO, USA Email: Karen Bogen - kbogen@mathematica-mpr.com; Lois Biener - lois.biener@umb.edu; Catherine A Garrett - catherine.garrett@comcast.net; Jane Allen - JAllen@americanlegacy.org; K Michael Cummings - Michael.Cummings@roswellpark.org; Anne Hartman - Anne_Hartman@nih.gov; Stephen Marcus - marcusst@mail.nih.gov; Ann McNeill - Ann.Mcneill@nottingham.ac.uk; Richard J O'Connor - Richard.Oconnor@roswellpark.org; Mark Parascandola - mark.parascandola@nih.hhs.gov; Linda Pederson - lindap@mindspring.com * Corresponding author Published: 19 October 2009 Harm Reduction Journal 2009, 6:27 doi:10.1186/1477-7517-6-27 Received: 6 January 2009 Accepted: 19 October 2009 This article is available from: http://www.harmreductionjournal.com/content/6/1/27 © 2009 Bogen et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. BioMed Central Harm Reduction Journal 2009, 6:27 doi:10.1186/1477-7517-6-27 Harm Reduction Journal 2009, 6:27 doi:10.1186/1477-7517-6-27 This article is available from: http://www.harmreductionjournal.com/content/6/1/27 This article is available from: http://www.harmreductionjour © 2009 Bogen et al; licensee BioMed Central Ltd. g This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Harm Reduction Journal Open Research Surveillance indicators for potential reduced exposure products (PREPs): developing survey items to measure awareness Karen Bogen1, Lois Biener2, Catherine A Garrett2, Jane Allen3, K Michael Cummings4, Anne Hartman5, Stephen Marcus5, Ann McNeill6, Richard J O'Connor4, Mark Parascandola5 and Linda Pederson7 Research Address: 1Mathematica Policy Research, Cambridge, MA, USA, 2Center for Survey Research, University of Massachusetts Boston, Boston, MA, USA, 3American Legacy Foundation, Washington, DC, USA, 4Department of Health Behavior, Roswell Park Cancer Institute, Buffalo, NY, USA, 5National Cancer Institute, National Institutes of Health, Washington, DC, USA, 6Division of Epidemiology and Public Health, University of Nottingham, Nottingham, UK and 7Beth-El College of Nursing & Health Sciences, Univ of Colorado, Colorado Springs, CO, USA Email: Karen Bogen - kbogen@mathematica-mpr.com; Lois Biener - lois.biener@umb.edu; Catherine A Garrett - catherine.garrett@comcast.net; Jane Allen - JAllen@americanlegacy.org; K Michael Cummings - Michael.Cummings@roswellpark.org; Anne Hartman - Anne_Hartman@nih.gov; Stephen Marcus - marcusst@mail.nih.gov; Ann McNeill - Ann.Mcneill@nottingham.ac.uk; Email: Karen Bogen* - kbogen@mathematica-mpr.com; Lois Biener - lois.biener@umb.edu; Catherine A Garrett - catherine.garrett@comcast.net; Jane Allen - JAllen@americanlegacy.org; K Michael Cummings - Michael.Cummings@roswellpark.org; Anne Hartman - Anne_Hartman@nih.gov; Stephen Marcus - marcusst@mail.nih.gov; Ann McNeill - Ann.Mcneill@nottingham.ac.uk; Richard J O'Connor - Richard.Oconnor@roswellpark.org; Mark Parascandola - mark.parascandola@nih.hhs.gov; Linda Pederson - lindap@mindspring.com Michael Cummings - Michael.Cummings@roswellpark.org; Anne Hartman - Anne_Hartman@nih.gov; Stephen Marcus - marcusst@mail.nih.gov; Ann McNeill - Ann.Mcneill@nottingham.ac.uk; Richard J O'Connor - Richard.Oconnor@roswellpark.org; Mark Parascandola - mark.parascandola@nih.hhs.gov; Linda Pederson - lindap@mindspring.com * Corresponding author Received: 6 January 2009 Accepted: 19 October 2009 Received: 6 January 2009 Accepted: 19 October 2009 Published: 19 October 2009 Open Research Surveillance indicators for potential reduced exposure products (PREPs): developing survey items to measure awareness Karen Bogen1, Lois Biener2, Catherine A Garrett2, Jane Allen3, K Michael Cummings4, Anne Hartman5, Stephen Marcus5, Ann McNeill6, Richard J O'Connor4, Mark Parascandola5 and Linda Pederson7 Open Access Background the fact that without a valid measure of awareness - one that allows the researcher to be confident that the respondent is thinking of the specific products being stud- ied - measures of risk perceptions, current use, and inten- tions regarding future use are of little value. g Over the past decade, tobacco companies have introduced cigarettes and smokeless tobacco products with purport- edly lower levels of some toxins than conventional ciga- rettes and smokeless products. These new products, named by the Institute of Medicine as Potential Reduced Exposure Products (PREPs)[1], have typically been intro- duced into regional test markets in the U.S. rather than nationwide, which means they are not widely recognized by name or description by most consumers[2,3]. Although some in the public health community welcome the introduction of PREPS, as they may offer a harm reduction opportunity to current tobacco users, other public health advocates do not believe that enough research - particularly long term research - has been done to know whether PREPs, even if proven to have reduced toxins, actually present a reduced health risk to smokers [4]. The concern in the public health community is that tobacco users who might have been motivated to quit may abandon those quit plans if they believe that an alter- native, less hazardous option exists with the PREPs. Like- wise, former tobacco users could be tempted back to use and non-users could be tempted to initiate use if they too believe that PREPs present lower health risks than conven- tional tobacco products [5,6]. Having faced a similar situ- ation with the introduction of "light" cigarettes - i.e. new product, insufficient research about its health implica- tions, effective tobacco industry marketing, and subse- quent documentation of no public health benefit and possible public health harm - public health officials are cautious about PREPs. Preliminary studies suggest that smokers are interested in trying PREPs, particularly com- bustible ones, and perceive them to have lower health risks than cigarettes [2,7-9]. On the other hand, despite well-documented evidence that smokeless tobacco is sub- stantially less harmful than cigarettes [10-16], studies demonstrate that most consumers rate them as being as harmful as cigarettes, if not more harmful [17,18]. Thus, public health experts have called for the development of a science to evaluate both the products and the public's response to them, as well as ongoing population surveil- lance [1,19,20]. Abstract Background: Over the past decade, tobacco companies have introduced cigarettes and smokeless tobacco products (known as Potential Reduced Exposure Products, PREPs) with purportedly lower levels of some toxins than conventional cigarettes and smokeless products. It is essential that public health agencies monitor awareness, interest, use, and perceptions of these products so that their impact on population health can be detected at the earliest stages. Methods: This paper reviews and critiques existing strategies for measuring awareness of PREPs from 16 published and unpublished studies. From these measures, we developed new surveillance items and subjected them to two rounds of cognitive testing, a common and accepted method for evaluating questionnaire wording. Results: Our review suggests that high levels of awareness of PREPs reported in some studies are likely to be inaccurate. Two likely sources of inaccuracy in awareness measures were identified: 1) the tendency of respondents to misclassify "no additive" and "natural" cigarettes as PREPs and 2) the tendency of respondents to mistakenly report awareness as a result of confusion between PREPs brands and similarly named familiar products, for example, Eclipse chewing gum and Accord automobiles. Conclusion: After evaluating new measures with cognitive interviews, we conclude that as of winter 2006, awareness of reduced exposure products among U.S. smokers was likely to be between 1% and 8%, with the higher estimates for some products occurring in test markets. Recommended measurement strategies for future surveys are presented. Page 1 of 8 (page number not for citation purposes) (page number not for citation purposes) http://www.harmreductionjournal.com/content/6/1/27 Harm Reduction Journal 2009, 6:27 http://www.harmreductionjournal.com/content/6/1/27 http://www.harmreductionjournal.com/content/6/1/27 Background There are three major challenges in developing surveil- lance measures. The first challenge is the fact that aware- ness is currently very low. Many PREPs are available only in limited test markets. Another challenge is that the PREPs themselves and their marketing often change with- out notice, making comparisons in awareness, use and risk perceptions difficult to make over time. A third chal- lenge is the fact that there is no single, agreed-upon defi- nition of a PREP across the public health community. Regardless of these hurdles, it is important to develop sur- vey measures now, to be prepared for the large scale intro- duction and marketing of PREPs that is coming, and it is vital to have stable measures that can be used across sur- veys so that comparisons can be made with confidence that observed changes over time and across surveys are due to true changes and not to differences in survey meas- ures. Page 2 of 8 (page number not for citation purposes) Methods Methods Currently available survey measures, as of 2006, were identified by reviewing published literature on consumer reactions to PREPs (using a number of search terms, including PREPs, new tobacco products, and reduced exposure) and by soliciting input about ongoing and unpublished studies from members of the project advi- sory group, members of the Tobacco Harm Reduction Network, and other members of the tobacco research community. Additional file 1 is a list of the 16 studies that were identified, seven in the published literature (A1 thru A7) and nine of which were unpublished (A8 through A16). We reviewed all of the PREPs-related questions as well as data on responses to the questions where they were available. A full listing of PREPs-related questions from the included studies is available from the first author. This review of measures and results included analysis of the question features and creation of a taxonomy of question structure. From this we developed new surveillance items using many of the previously identified item features. We subjected the new questions to two rounds of cognitive testing, a common and accepted method for evaluating questionnaire wording [23]. Cognitive testing is done with intensive, one-on-one interviews in which partici- pants are usually asked to summarize their understanding of the meaning of questions and to provide a narrative explanation of how they arrived at an answer to a ques- tion by concurrently thinking out loud as they answer the question. This allows an evaluation of the question-and- answer process, question comprehension, how answers This paper describes the development of survey items that could aid in the ongoing surveillance of PREPs in U.S. markets. The first phase in the development effort was a review and critique of current survey measures. The sec- ond phase included developing a surveillance instrument and evaluating it by means of two rounds of in-depth cog- nitive interviews. This paper addresses measurement of product awareness. Other domains, including risk percep- tion, current use, interest in use, and trial, are covered in the full technical reports [21,22]. The decision to focus first on awareness, and not the other domains, is based on Page 2 of 8 (page number not for citation purposes) Page 2 of 8 (page number not for citation purposes) http://www.harmreductionjournal.com/content/6/1/27 http://www.harmreductionjournal.com/content/6/1/27 Harm Reduction Journal 2009, 6:27 The Structure of PREPs Awareness Questions Figure 1 The Structure of PREPs Awareness Questions. The Struc Figure 1 Q g The Structure of PREPs Awareness Questions. This figure shows that the initial distinction between question types is that they ask about PREPs in the abstract or they ask about specific PREPs brands. Within each of those groups, there are further divisions, as shown in the figure. g The Structure of PREPs Awareness Questions. This figure shows that the initial distinction between question types is that they ask about PREPs in the abstract or they ask about specific PREPs brands. Within each of those groups, there are further divisions, as shown in the figure. of such products. There have been two ways of asking about this type of conceptual awareness: Description of PREPs as "less harmful" Five of the reviewed surveys [18, 24, A8, A10, A15] had a question that asks whether a respondent is aware of a new cigarette described generally as designed to be "less harm- ful". Items using this strategy yield rather high awareness estimates, with reports of over one-third of smokers hav- ing heard of such "less harmful" cigarettes [18, A8, A15]. Results and Discussion The following sections address surveillance of awareness of PREPs and are organized to first summarize current measures, second to critique those measures and use them as a way to develop a new measure of awareness, third to present the results of the cognitive interviewing of a new measure or awareness, and fourth to recommend a sur- veillance measure. There are measures relating to both combustible and non-combustible PREPs, and while the following summary focuses on combustibles, the results for non-combustibles are presented where they are differ- ent or important. When referring to the unpublished stud- ies or survey items, we note their sequence number from Additional file 1. Description of the mechanism for harm reduction A second way to measure general awareness of PREPs is to start the question with a description of the mechanism by which the harm reduction is supposed to be achieved, such as the following example: Tobacco companies have recently introduced products that look like cigarettes, but which heat the tobacco, instead of burning it. They claim that these products contain less tar and produce less environmental tobacco smoke. They also say they that they give a lower concentration of cancer-causing chemicals in the smoke. (A9) Methods This figure shows that the initial distinction between question types is that they ask about PREPs in the abstract or they ask about specific PREPs brands. Within each of those groups, there are further divisions, as shown in the figure. AWARENESS PREPs in the abstract Specific brands Describe as less harmful Describe the mechanism Recognition Recall Read brand name Show ad AWARENESS PREPs in the abstract Specific brands Describe as less harmful Describe the mechanism Recognition Recall Read brand name Show ad were formed, and what the answers mean, towards the goal of making sure that there is consistency of interpreta- tion across respondents [23]. In January 2007, seven cognitive interviews were con- ducted in the Boston, MA area, where at least one combus- tible PREP had been available, and one month later, eight interviews were done in Austin, TX, a test market for Camel Snus, a smokeless PREP. The respondents, 11 smokers and four former smokers, were recruited from newspaper advertisements and campus postings. The interviews were conducted by two senior level researchers and one professional interviewer. Respondents were paid $50 for their participation. The interviews, which used concurrent think-aloud techniques, took approximately 45 minutes to complete. All interviews were audio-taped with participants' permission. Variations in wording were tested in these cognitive interviews and probing questions were used to understand respondents' comprehension of the questions. For example, if a respondent said she was aware of a PREP, she would be asked to describe what she knew about it to see if she was really talking about the right product. We probed sufficiently to determine whether or not we believed she answered the tested aware- ness item correctly and, thus, made a judgment as to whether or not the tested item yielded an accurate or inac- curate indicator of awareness. Recognition Show ad Presenting names of one or more specific PREPs Presenting names of one or more specific PREPs Presenting names of one or more specific PREPs Critique of Existing Measures q f g The two ways of asking about PREPs awareness in the abstract - describing them as "less harmful" and describ- ing the mechanism - each has limitations. The description of PREPs as less harmful followed by a question asking if they've heard of such a product results in artificially high estimates of awareness because respondents may under- stand the question to include products marketed as "nat- ural" or "additive free". This is apparent when there is a follow-up question (unaided recall) asking the respond- ent to name a "less harmful" product, and they name brands such as American Spirit and Winston No Bull, which are not PREPs. It is not surprising that consumers view products such as American Spirit and Winston No Bull as less harmful, since they are marketed as "natural" or "additive-free", messages that are associated with being healthier. There is also a history of consumers' misunder- standing the relative health risk of different tobacco prod- ucts, including smokeless [18] and light cigarettes [27]. The abstract question about PREPs awareness by describ- ing the mechanism of the PREP is a difficult approach for ongoing surveillance. The difficulty is that the mechanism used in the different products are highly variable and may not be known or understandable when described to the consumer. Further, it is not a good choice for ongoing sur- veillance because of the variability. A good surveillance item relies on stable measures that can be compared across years and across markets as the new products emerge. g p Use of a recognition question, in which the brand name is provided and respondents indicate whether they have heard of it, often yields high reports of knowledge of spe- cific brands, as high as 53% [25]. Show a product advertisement In two studies [26, A14], respondents were shown adver- tisements for specific PREPs brands and were then asked if they had seen the ads for these brands before. Among smokers in the one study with available data, awareness measured this way was just below 8% for the PREP shown [26]. http://www.harmreductionjournal.com/content/6/1/27 Harm Reduction Journal 2009, 6:27 Summary of Existing Measures In our review of current measures, we identified the struc- ture of PREPs awareness questions. Figure 1 is a summary of the structure, showing that the initial distinction between questions is whether they ask about PREPs in the abstract or about specific PREPs brands. Within each of those groups, there are further divisions. When asked if they had heard of these products, only about 13% of smokers said Yes. As shown on the right side of Figure 1, two general approaches have been used to measure awareness of spe- cific PREPs products: 1) recognition of specific products by either presenting brand names or showing a product advertisement and 2) unaided recall of products. As shown on the left side of the figure, a common strategy for measuring awareness of a product is to describe the general concept and ask whether the respondent is aware Page 3 of 8 (page number not for citation purposes) Page 3 of 8 (page number not for citation purposes) http://www.harmreductionjournal.com/content/6/1/27 Unaided recall of brand names Generally, use of recall strategies, which require respond- ents to come up with brand names on their own, yields very low reports, and many of the brands that respondents name are not PREPs at all. In two of the studies, none of the brands of "less harmful" tobacco products recalled were PREPs [24, A10]. Some of the named products were so-called "natural" cigarettes (such as American Spirit) or nicotine replacement products (such as Nicorette and Nicoderm). In the other four studies that used unaided recall [18, A8, A9, A15], the reports of actual PREPs brands ranged from about 1% to about 8% of smokers. There were many reports of non-PREPs, including American Spirit and Quest, a nicotine step-down brand. Table 1 summarizes the current measures of awareness and the estimates they yield. The two ways of asking about awareness of specific PREPs brands - through recognition or recall of brand names - also have limitations. As shown in summary Table 1, the rates of awareness resulting from recognition measures are often very high, ranging from 11% of smokers (A11) to 53% of smokers [25]. We suspected that these rates were over-reports because a number of PREPs brands share names with other products. For example, Eclipse, Accord and Omni are all combustible PREPS that share their At the time of this review, only two surveys [25, A13] had attempted to measure awareness of non-combustible PREPs by naming them (Ariva, Revel, Exalt, and Stone- wall) along with other types of PREPs and asking the respondent if he/she had heard of them. Among the four smokeless products listed, rates of awareness among smokers ranged from 3% to 6% [25]. Table 1: Current Measures of Awareness of Combustible PREPs Table 1: Current Measures of Awareness of Combustible PREPs Table 1: Current Measures of Awareness of Combustible PREPs HOW SURVEYS ASK ABOUT AWARENESS Percent of smokers affirming awareness Studies with data (citation numbers shown) Describe PREPs as "less harmful" 19-40% [18]; (A8. A10, A15) Describe mechanism of PREP 13% (A9) Recognition - read names 11-53% [24,25]; (A11) Recognition - show a product advertisement 8% [26] Unaided recall 0-8% [18,24]; (A8, A9, A10, A15) Citations A8 through A15 refer to unpublished studies listed in Additional file 1. Citations A8 through A15 refer to unpublished studies listed in Additional file 1. *Citations A8 through A15 refer to unpublished studies listed in Additional file 1. http://www.harmreductionjournal.com/content/6/1/27 http://www.harmreductionjournal.com/content/6/1/27 Harm Reduction Journal 2009, 6:27 name with, respectively, a chewing gum, a car, and a hotel. When respondents say they recognize those names, we cannot be certain that they are thinking of the cigarette brand or if they simply recognize the name more generally and, therefore, report having heard of it. An alternative recognition approach, showing a product advertisement, has the same limitation as describing the mechanism, which is that it is an impractical approach for ongoing sur- veillance. priate descriptions of the brands from the two respond- ents that confirmed their legitimate recollection. A third person who said Yes to the conceptual awareness question could not recall any brand names but described PREPs in a way that suggested that she really was aware of them (i.e. she reported that they were lower in carcinogens and that she had seen a TV documentary about them). The three other participants who said Yes to the combustible PREPs awareness question went on to name non-PREPs like American Spirit, Winston No Bull, nicotine products, and even Newport Lights. Use of an unaided recall awareness question typically fol- lows a question that asks about PREPs in the abstract. As described earlier, we know that what respondents name in these unaided recall questions is often not a PREP at all, but other products that they consider less harmful, such as products advertised as more natural, additive free, or quit aids. Unaided recall (If yes to conceptual awareness) Can you recall any brand names of these products? Considering only the recognition responses (ignoring recall for the moment, even though recognition would be different if it did not follow recall), we would conclude that 4 out of 15 participants were aware of combustible PREPs, based only on those who said they recognized a real PREP and in probing, convinced us that they truly were aware. Having the distracter names on the lists works as a benchmark as to how much name recognition is attributable to a familiar name, and not the specific PREP. (If yes to above) Please tell me the names of any you recall. Conceptual awareness New types of cigarettes are now being developed that are supposed to be less harmful than ordinary ciga- rettes. Have you heard of such products? Follow-on recognition task I'm going to read you the names of some (other) rela- tively new cigarettes. For each one, please tell me whether or not you have ever heard of it. Since the new smokeless PREPs were not all being adver- tised as being potentially less harmful, the introductory question required a somewhat different approach to max- imize the likelihood that someone who was aware of any of the products would understand what was being referred to. The unaided recall and follow-up recognition ques- Table 1: Current Measures of Awareness of Combustible PREPs The tradeoff, then, is between a small under-estimate of awareness if a respondent is required to name a PREP in order to be considered aware (one of the three who was truly aware of PREPs was not counted because she failed to name a PREP) or an even larger over-estimate if only the yes/no conceptual awareness question is used to esti- mate awareness (six who said they were aware instead of the three who said they were aware and really were). We used these critiques of existing measures to develop a new awareness measure to test in cognitive interviews. False recognition reports were common among partici- pants. We included distracter brand names - two out of seven brands listed in the Boston interviews and three out of eight brands listed in Austin. Ten of the 15 participants interviewed recognized 12 of the listed brands, but half of those mentions were for our distracter products - three said they "recognized" Kool Silver and three "recognized" Westin. As a result of this high level of false recognition, we are doubtful of the reports of actual PREPs because two mentions were for Eclipse and two were for Marlboro Ultrasmooth, both of which could be confused with other products and be "recognized" by virtue of their familiar- sounding names. Probes into what was known about these brands (legitimate and bogus) yielded no convinc- ing information that would lead us to believe that the respondent was truly aware of the PREPs. For the two rounds of cognitive interviews, we asked first about conceptual awareness of products described as less harmful, followed by unaided recall to identify over- reports in the conceptual awareness question, and then recognition items, including fictitious items to gauge false reports of similar sounding names. Following are the items used to measure awareness of combustible PREPs: Conceptual awareness Page 5 of 8 (page number not for citation purposes) Unresolved Issue - Definition of a PREP Unresolved Issue Definition of a PREP In order to monitor awareness of PREPs, researchers need some agreement about what constitutes a PREP. Although the definition originated by the Institute of Medicine includes pharmaceutical agents, such as nicotine replace- ment products [1], they were not the focus of our study and, thus, were excluded from our definition. Likewise, our definition excludes quit aids, such as Quest, and is modeled after the one provided in Hatsukami and Hecht, focusing on "tobacco products that have been modified or designed in some way to reduce users' exposure to tobacco toxins" [19]. Arguably, the essential characteristic of most definitions is that they are talking about products designed to yield reduced exposure to toxins. The prob- lem is that many of these products are not currently being advertised as entailing reduced exposure. To an important extent, the design of the surveillance items being recom- mended in this report is determined by current PREPs marketing practices. Specifically, because the new smoke- less products are not being advertised as less harmful than cigarettes, the recommended awareness questions do not include a statement that they are "supposed to be" or "claim to be" less harmful. Most of the combustible PREPs for which awareness is being assessed were advertised as less harmful during the time that they were being actively marketed, so the awareness question recommended here includes language about reduced harm. However, market- ing strategies are likely to change. Philip Morris did not make any health claims while test-marketing Marlboro Ultrasmooth, and the Eclipse and Advance web sites, which originally featured extensive health claims, are no longer readily accessible. This issue complicates assess- ment of awareness of PREPs because although the public health community may be aware that the product design suggests an effort to reduce exposure to particular toxins, the general public may be unaware of that effort while still being aware of the new product. New types of smokeless tobacco products are now available that are put in the mouth but don't involve chewing or spitting. Some come in teabag-like pouches and some come in the form of a lozenge or tablet. Have you heard of any products like this? Since Austin is a test market for one of the new smokeless PREPs, the results there were very different from results in Boston, where awareness was minimal. Cognitive Interview Results In cognitive interviews, six of 15 participants said Yes to the conceptual awareness question about combustible PREPs. Of those, only two recalled the name of legitimate PREPs. Cognitive probes (e.g., What do you know about them? What have you heard about them?) yielded appro- Page 5 of 8 (page number not for citation purposes) Page 5 of 8 (page number not for citation purposes) http://www.harmreductionjournal.com/content/6/1/27 Harm Reduction Journal 2009, 6:27 tions were the same as was used for the combustible PREPs. viewer-administered or web surveys where skips are possi- ble and open-ended responses are feasible. This strategy is not feasible for self-administered surveys. Recommendation for Measuring Awareness of PREPs Recommendation for Measuring Awareness of PREPs We recommend that awareness of PREPS, both combusti- ble and non-combustible, be measured using the concep- tual awareness question, followed by unaided recall of a brand so that those who name non-PREPs can be identi- fied as unaware. These are the items that would be used at present to estimate PREPs awareness. We also recommend asking the recognition series, including distracter names, as a way to monitor changes in both real and likely false recognition of PREPs. However, we do not recommend using the recognition data, alone or in conjunction with the other data, to estimate current awareness of PREPs at this time. As awareness increases, and the false reports drop to a small proportion of recognized brands, it will be time to consider how to combine the recall and recogni- tion data to estimate awareness. Additional file 2 shows the recommended series of items to measure awareness of PREPs. It is important to consider that this recommended strategy for measuring awareness is feasible for inter- Page 6 of 8 (page number not for citation purposes) Unresolved Issue - Definition of a PREP Essentially no one in the Boston study group was aware of any smokeless PREPs based on the conceptual description and unaided recall. With regard to recognition, two participants thought they had each heard of one of the named prod- ucts, but there is no way to know whether these recogni- tions were legitimate or whether they were simply recognitions of a familiar brand name. In Austin, however, all eight respondents said "yes" to the conceptual awareness question. Five of these participants named either Camel Snus or a Skoal product in the fol- lowup unaided recall question. If we accept as awareness "yes" to the conceptual awareness questions followed by a correct brand name recalled, we would conclude that five of the eight participants in Austin were aware of smokeless PREPs. The actual estimate for Austin should be six of eight respondents, because one individual could not name a brand but clearly described the Stonewall product. Hence, this strategy, as with the comparable combustible measure, slightly underestimated awareness in a test mar- ket. Surveillance in other important areas This paper covers only one of the domains of surveillance, but there are a number of others that are important for ongoing consideration. Specifically, interest in use, risk perception and trial will be important areas to monitor as PREPs become more common. Once one is reasonably confident of having an accurate indicator of product awareness, it will be important to have good measures in these areas and to know whether consumers are over-esti- mating or under-estimating the risks or potential health gains of various products. This will be particularly impor- tant as marketing messages change. Page 6 of 8 (page number not for citation purposes) Page 6 of 8 (page number not for citation purposes) http://www.harmreductionjournal.com/content/6/1/27 http://www.harmreductionjournal.com/content/6/1/27 http://www.harmreductionjournal.com/content/6/1/27 Harm Reduction Journal 2009, 6:27 Because all indications are that population levels of awareness of PREPs are quite low, national population monitoring of patterns of current use and impact on changes in other tobacco use behavior seems unnecessary at this time. As levels of PREPs awareness and trial increase, as is likely to be the case with the new smokeless products being introduced by the major cigarette manu- facturers, we would advocate introduction of current use measures in the surveillance program. Given the long lead time required to introduce questions on Federal-level sur- veys, it is, perhaps, appropriate to consider introducing those questions now, so as to be prepared for the increase in use when it happens. Likewise, communications research that examines advertising, packaging, and health claims for specific products are critical and should be con- ducted. In order to be prepared to respond should these products become more widely available, it is important to learn how best to construct public health messages and policy to both maximize potential public health benefits of less harmful tobacco products and minimize negative health consequences. studies suggest. In order to monitor changes in awareness, we propose using a set of questions that appear to result in reasonably accurate estimates of awareness of a partic- ular group of products; the proposed items are presented in Additional file 2. Given the currently very low levels of awareness and the decision of the tobacco companies to introduce most of their new products into test markets, it may fall on states to use their rapid surveillance systems to take the lead in measuring awareness. Surveillance in other important areas States in which there are regional test markets could add questions to the Behavioral Risk Factor Surveillance System (BRFSS) and the Adult Tobacco Survey (ATS) to start the PREPs surveillance proc- ess. At this stage, state surveys would be more successful than national studies that will not have enough respond- ents within test market regions to detect small movements in awareness. These states could compare test market areas to non-test market areas within their states, as well as make comparisons across states, if surveys adopt the same wording. Once awareness increases, it will be time for sur- veys to include surveillance measures in other important domains. However, Camel Snus, a smokeless PREP, was released nationally in 2009 [28], so it is certainly the right time to consider adding an awareness surveillance meas- ure at the national level as well. In addition to the PREPs covered in this report, there are a variety of other products that may be believed to be less harmful than cigarettes, and which may be used by con- sumers in an effort to reduce their tobacco-related health risks. These include conventional smokeless tobacco, medicinal nicotine, e-cigarettes, other non-tobacco nico- tine products, and tobacco products that claim to be "nat- ural" (i.e. have no additives), etc. A thorough understanding of population perceptions about ways to reduce the risks of smoking (aside from quitting), along with the prevalence of behaviors believed by consumers to reduce risk, would be a useful endeavor. It must be acknowledged that the evaluation of items in this paper was limited by the low number of respondents and by the fact that those included were of relatively high socio-economic status. Prior to any large scale use of these survey items, we recommend further pre-tests be done using other survey modes (telephone and self-adminis- tered), and that an effort be made to include respondents with lower levels of education and with greater cultural diversity. It is also important to recognize that population surveil- lance is just one way to collect data about PREPs. In the development of a science to evaluate PREPs, other strate- gies, including lab research, qualitative studies, and review of tobacco industry information, are needed to get a full picture. This paper focuses on population surveil- lance measures from surveys and does not cover the full array of monitoring strategies. Authors' contributions KB and LB worked on the review of previous studies, designed the research project, completed the cognitive interviews of draft questions, and worked on the two reports on which this manuscript is based. KB drafted this manuscript; LB contributed extensively to its completion. CAG provided all manner of research support on the ear- lier reports for this project, including literature reviews, and on this manuscript. SM and MP were the research officers on the original project from which this manu- script derives. AM, RO, MP and LP were on the project advisory group, which helped design the interview ques- tions and to interpret the results. All co-authors partici- Surveillance in other important areas However, surveys are par- ticularly useful for estimating prevalence of awareness, interest and use, as well as perceptions, and it is important to develop accurate surveillance measures to track changes in these areas over time, as the number of PREPs on the market increases. Competing interests p g The authors declare that they have no competing interests. http://www.harmreductionjournal.com/content/6/1/27 Harm Reduction Journal 2009, 6:27 12. pated in discussions of the findings, provided numerous reviews of this manuscript, and read and approved the manuscript. 12. Gartner CE, Hall WD, Vos T, Bertram MY, Wallace AL, Lim SS: Assessment of Swedish snus for tobacco harm reduction: An epidemiological modelling study. The Lancet 2007. 13. Luo J, Ye W, Zendehdel K, Adami J, Adami H-O, Boffetta P, Nyrén O: Oral use of Swedish moist snuff (snus) and risk for cancer of the mouth, lung, and pancreas in male construction workers: A restrospective cohort study. The Lancet 2007, 369(9578):2015-2020. Additional file 2 ( ) 18. O'Connor RJ, Hyland A, Giovino GA, Fong GT, Cummings KM: Smoker Awareness of and Beliefs About Supposedly Less- Harmful Tobacco Products. American Journal of Preventive Medi- cine 2005, 29(2):85-90. [http://www.biomedcentral.com/content/supplementary/1477- 7517-6-27-S2.DOC] 19. Hatsukami D, Hecht SS: Hope or Hazard? What Research Tells Us About "Potentially Reduced-Exposure" Tobacco Prod- ucts. 2005 [http://www.tturc.umn.edu/documents/hope_or_hazard- 3.pdf]. Accessed January 5, 2009 p ] J y 20. Hatsukami DK, Lemmonds C, Tomar SL: Smokeless tobacco use: harm reduction or induction approach? Prev Med 2004, 38(3):309-317. Additional material Additional file 1 Studies included in the summary. This is a list of the published and unpublished studies that were included in this study summary. They rep- resent the currently available survey measures as of 2006. Click here for file [http://www.biomedcentral.com/content/supplementary/1477- 7517-6-27-S1.DOC] Additional file 2 Recommended items. This is a list of the recommended series of items to measure awareness of PREPs. Click here for file [http://www.biomedcentral.com/content/supplementary/1477- 7517-6-27-S2.DOC] ( ) 14. Critchley J, Unal B: Health effects associated with smokeless tobacco: A systematic review. Thorax 2003, 58(5):435-443. Acknowledgements This research was supported by contract # NCI-60036-NG from the National Cancer Institute, and by Grant # CA086257. The authors would like to thank the other members of the project advisory group, who pro- vided valuable ideas at all stages of this project: Jack Fowler, Gary Giovino, Lynn Kozlowski, and Gordon Willis. ( ) 21. Biener L, Bogen K, Garrett CA: PREPs Surveillance: Summary of Existing Survey Measures and Recommendations for Ongo- ing Efforts. 2006 [http://www.csr.umb.edu/pdf/tech_report/ PREPs%20Surv%20Report%201.pdf]. Accessed 7/1/08 p p ] 22. Biener L, Bogen K: PREPs Surveillance: Summary of Cognitive Interviews and Recommendations for PREPs Surveillance Measures. 2007 [http://www.csr.umb.edu/pdf/tech_report/ PREPs%20Surv%20Report%202.pdf]. Accessed 7/1/08 Additional file 1 15. Levy DT, Mumford EA, Cummings KM, Gilpin EA, Giovino G, Hyland A, Sweanor D, Warner KE: The relative risks of a low-nitro- samine smokeless tobacco product compared with smoking cigarettes: estimates of a panel of experts. Cancer Epidemiology, Biomarkers, and Prevention 2004, 13(12):2035-2042. 16. Savitz DA, Meyer RE, Tanzer JM, Mirvish SS, Lewin F: Public health implications of smokeless tobacco use as a harm reduction strategy. American Journal of Public Health 2006, 96(11):1934-1939. ' 17. O'Connor RJ, McNeill A, Borland R, Hammond D, King B, Boudreau C, Cummings KM: Smokers' beliefs about the relative safety of other tobacco products: findings from the ITC collaboration. Nicotine & Tob Research 2007, 9(10):1033-1042. References 1. Stratton K, Shetty P, Wallace R, Bondurant S, eds: Clearing the Smoke: Assessing the Science Base for Tobacco Harm Reduction Washington, D.C.: National Academy Press; 2001. 1. Stratton K, Shetty P, Wallace R, Bondurant S, eds: Clearing the Smoke: Assessing the Science Base for Tobacco Harm Reduction Washington, D.C.: National Academy Press; 2001. 1. Stratton K, Shetty P, Wallace R, Bondurant S, eds: Clearing the Smoke: Assessing the Science Base for Tobacco Harm Reduction Washington, D.C.: National Academy Press; 2001. p p ] 23. Willis GB: Cognitive Interviewing: A Tool for Improving Questionnaire Design Thousand Oaks, CA: Sage Publications, Inc; 2005. 24. Hund LM, Farrelly MC, Allen JA, Chou RH, St. Claire AW, Vallone DM, Healton CG: Findings and implications from a national study on potential reduced exposure products (PREPs). Nic- otine & Tobacco Research 2006, 8(6):791-797. y 2. Pederson LL, Nelson DE: Literature review and summary of perceptions, attitudes, beliefs, and marketing of potentially reduced exposure products: communication implications. Nicotine & Tobacco Research 2007, 9(5):525-534. ( ) 25. Hamilton WL, diStefano Norton G, Ouellette TK, Rhodes WL, Kling R, Connolly GN: Smokers' responses to advertisements for regular and light cigarettes and potential reduced-exposure tobacco products. Nicotine and Tobacco Research 2004, 6(Supple- ment 3):S353-S362. ( ) 3. Hickman N, Klonoff EA, Landrine H, Kashima K, Parekh B, Fernandez S, Thomas K, Brouillard C: Preliminary Investigation of the Advertising and Availability of PREPs, the New "Safe" Tobacco Products. Journal of Behavioral Medicine 2004, 27(4):413-424. ) 26. Parascandola M, Hurd A, Augustson E: Consumer Awareness and Attitudes Related to New Potential Reduced-Exposure Tobacco Products. Am J Health Behav 2008, 32(4):431-437. ( ) 4. Hatsukami DK, Ebbert JO, Feuer RM, Stepanov I, Hecht SS: Chang- ing smokeless tobacco products: New tobacco-delivery sys- temts. American Journal of Preventive Medicine 2007, 33(6S):s368-s378. J ( ) 27. Kozlowski LT, Goldberg ME, Yost BA, White EL, Sweeney CT, Pillit- teri JL: Smokers' misperceptions of light and ultra-light ciga- rettes may keep them smoking. American Journal of Preventive Medicine 1998, 15(1):9-16. ( ) 5. Joseph AM, Hennrikus D, Thoele MJ, Krueger R, Hatsukami D: Com- munity tobacco control leaders' perceptions of harm reduc- tion. Tobacco Control 2004, 13(2):108-113. ( ) 28. RJ Reynolds' Snus Goes National: Convenience Store News 2008 [http:/ /www.csnews.com/csn/cat_management/tobacco/ article_display.jsp?vnu_content_id=1003875892]. ( ) 6. Martin EG, Warner KE, Lantz PM: Tobacco harm reduction: what do the experts think? Tobacco Control 2004, 13(2):123-128. 6. Conclusion A review of existing research on population awareness of PREPs as well as results of cognitive interviews suggest that awareness of PREPs is quite low at this time, except in active test-markets, and is lower than some published Page 7 of 8 (page number not for citation purposes) Page 7 of 8 (page number not for citation purposes) http://www.harmreductionjournal.com/content/6/1/27 28. RJ Reynolds' Snus Goes National: Convenience Store News 2008 [http:/ /www.csnews.com/csn/cat_management/tobacco/ article_display.jsp?vnu_content_id=1003875892]. References Martin EG, Warner KE, Lantz PM: Tobacco harm reduction: what do the experts think? Tobacco Control 2004, 13(2):123-128. 7. Parascandola M, Augustson E, O'Connell ME, Marcus S: Consumer awareness and attitudes related to new potential reduced- exposure tobacco product brands. Nicotine and Tobacco Research 2009, 11(7):886-895. p , ( ) 7. Parascandola M, Augustson E, O'Connell ME, Marcus S: Consumer awareness and attitudes related to new potential reduced- exposure tobacco product brands. Nicotine and Tobacco Research 2009, 11(7):886-895. , ( ) 8. Caraballo RS, Pederson LL, Gupta N: New tobacco products: do smokers like them? Tobacco Control 2006, 15(1):39-44. , ( ) 9. Shiffman S, Pillitteri JL, Burton SL, Di Marino ME: Smoker and ex- smoker reactions to cigarettes claiming reduced risk. Tobacco Control 2004, 13:78-84. 10. Foulds J, Kozlowski L: Snus-what should the public-health response be? The Lancet 2007. p 11. Foulds J, Ramström LM, Burke M, Fagerstrom K: Effect of smoke- less tobacco (snus) on smoking and public health in Sweden. Tobacco Control 2003, 12(4):349-359. Page 8 of 8 (page number not for citation purposes) Page 8 of 8 (page number not for citation purposes)
https://openalex.org/W2914955048	https://www.nature.com/articles/s41467-019-11271-0.pdf	English	null	Sharing information between related diseases using Bayesian joint fine mapping increases accuracy and identifies novel associations in six immune mediated diseases	bioRxiv (Cold Spring Harbor Laboratory)	2,019	cc-by	21,052	ARTICLE ARTICLE Stochastic search and joint ﬁne-mapping increases accuracy and identiﬁes previously unreported associations in immune-mediated diseases T he underlying genetic contribution to many complex diseases and traits has been investigated with great success by genome-wide association studies (GWAS). Various approaches have identiﬁed thousands of variants associated with a spectrum of diseases. In particular, much progress has been made in the genetics of immune-mediated diseases (IMD), revealing a complex pattern of shared and overlapping genetic etiology1,2. gy Fine-mapping—the process of distinguishing causal genetic variants from their neighbours—is an essential step to enable the design of functional assays required to understand the mechan- ism by which the region impacts disease risk, but it is complicated by linkage disequilibrium (LD)3. The problem is often approa- ched through stepwise regression4,5, which assumes that statis- tical inference of the best joint model (i.e. a model with multiple causal SNPs) can be derived by starting with the most signiﬁcant SNP, then conditioning on this and adding the next most signiﬁcant, continuing this conditioning until no con- ditionally signiﬁcant SNPs remain. It has been noted that the SNP with the smallest p-value need not be causal, especially if it is in LD with two causal SNPs6. Alternative Bayesian ﬁne-mapping methods have been developed, which use a stochastic search instead of stepwise search7–9. Stepwise and stochastic search results may disagree9 and although stochastic search generally demonstrates improved accuracy10 these techniques have not yet been widely adopted. While one of the strengths of Bayesian methods is that multiple competing models can be identiﬁed with posterior support for each, for the purposes of comparing stochastic search and stepwise search results, we chose to focus on discrepancies between the best models chosen for each. In all regions, the model preferred by stochastic search either had equal or better Bayesian Information Criterion (BIC) and equal or larger number of variants compared to the model chosen by stepwise search (Supplementary Fig. 6). For 16 regions (18 disease-region pairs) the stepwise model was nested in that of stochastic search (treating SNPs in the same SNP group as equivalent; Supple- mentary Table 2). In six regions (6 disease-region pairs) there appeared to be two separate signals, both weak (2 × 10–10 < p < 4 × 10−6 by single SNP logistic regression) with stochastic search posterior support fairly evenly shared between the two SNP groups, and the SNP selected by stepwise search falls in the group with slightly less posterior support (Table 1). Stochastic search and joint ﬁne-mapping increases accuracy and identiﬁes previously unreported associations in immune-mediated diseases Jennifer L. Asimit 1, Daniel B. Rainbow 2, Mary D. Fortune 1, Nastasiya F. Gri Linda S. Wicker 2 & Chris Wallace 1,3 Jennifer L. Asimit 1, Daniel B. Rainbow 2, Mary D. Fortune 1, Nastasiya F. Grinberg 3, Linda S. Wicker 2 & Chris Wallace 1,3 Thousands of genetic variants are associated with human disease risk, but linkage dis- equilibrium (LD) hinders ﬁne-mapping the causal variants. Both lack of power, and joint tagging of two or more distinct causal variants by a single non-causal SNP, lead to inac- curacies in ﬁne-mapping, with stochastic search more robust than stepwise. We develop a computationally efﬁcient multinomial ﬁne-mapping (MFM) approach that borrows informa- tion between diseases in a Bayesian framework. We show that MFM has greater accuracy than single disease analysis when shared causal variants exist, and negligible loss of precision otherwise. MFM analysis of six immune-mediated diseases reveals causal variants unde- tected in individual disease analysis, including in IL2RA where we conﬁrm functional effects of multiple causal variants using allele-speciﬁc expression in sorted CD4+ T cells from genotype-selected individuals. MFM has the potential to increase ﬁne-mapping resolution in related diseases enabling the identiﬁcation of associated cellular and molecular phenotypes. 1 MRC Biostatistics Unit, University of Cambridge, Cambridge Institute of Public Health, Forvie Site, Robinson Way, Cambridge Biomedical Campus, Cambridge CB2 0SR, UK. 2 JDRF/Wellcome Diabetes and Inﬂammation Laboratory, Wellcome Trust Center for Human Genetics, Nufﬁeld Department of Medicine, University of Oxford, Oxford OX3 7BN, UK. 3 Department of Medicine, Cambridge Biomedical Campus, University of Cambridge, Box 157, Level 4, Cambridge CB2 0QQ, UK. Correspondence and requests for materials should be addressed to J.L.A. (email: jennifer.asimit@mrc-bsu.cam.ac.uk) or to C.W. (email: cew54@cam.ac.uk) 1 NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.103 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 T statistically distinguishable within the group. When we discuss a SNP group model, e.g. model A + B, we mean the collection of models that include exactly one SNP from group A and exactly one SNP from group B, and no others. We consider posterior support for each grouped model (GPP) as the sum of posterior probabilities overall SNP models in that group when interpreting the stochastic search results. SNP group membership is shown in Supplementary Data 3. Stochastic search and joint ﬁne-mapping increases accuracy and identiﬁes previously unreported associations in immune-mediated diseases In a further four regions (ﬁve disease-region pairs) we found non-nested stochastic/stepwise mismatches, which could not be explained simply. y p Here, we systematically compare stepwise and stochastic approaches by application to dense genotype data for six IMD, aiming to address the frequency and causes of disagreement between results. We ﬁnd that stochastic search solutions are more likely to be correct than stepwise search results when sample sizes are large, but that they can face similar issues to stepwise searches when sample sizes are small. We also observe a striking sharing of causal variants between different IMD, consistent with previous reports1,2, which motivates us to propose a Bayesian multinomial stochastic search method, in which multiple related diseases can be simultaneously ﬁne-mapped. This allows us to borrow infor- mation between diseases and achieve correct ﬁne-mapping solu- tions at smaller sample sizes than when considering individual diseases alone. We show that posterior probabilities under our proposed model can be decomposed into quantities available from single disease analyses, allowing it to be applied without excessive additional computational overhead. Joint tagging of stochastic search models by stepwise SNPs. We investigated these ﬁve mismatch cases further, both mathemati- cally and using simulation, hypothesising that they may reﬂect cases where the SNP with smallest p-value acts to tag both of two distinct causal variants17. We walk through these results using the example of ATD in a chromosome 10p region. Haplotype ana- lysis, which estimates effects for all observed combinations of alleles across these three SNPs, illustrates how the minor allele of stepwise search-selected SNP rs706779 (a member of group J) tends to be carried together with the minor alleles of stochastic search-selected SNPs rs61839660 (group A) and rs11594656 (group C) (Fig. 1a). Considering the haplotypes formed from rs61839660/A, rs706779/J and rs11594656/C, we see that while haplotypes carrying the rs706779:C allele in the presence of either rs61839660:T or rs11594656:A (haplotypes TCT or CCA) are protective for ATD, a haplotype carrying rs706779:C in combi- nation with rs61839660:C and rs11594656:T (CCT, frequency 13%) is indistinguishable from the common (susceptible) haplo- type CTT (Fig. 1a, p = 0.24, Wald test). NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications Results h Doing so, we found that 20–40% of potential common causal variant pairs (MAF > 5%) had a potential joint tag, though this was highly variable across regions (Fig. 2b, c, Supplementary Data 4) and should be considered an upper limit because our assumption of equal effect sizes may not be justiﬁed. Together, these results better characterise and quantify the potential frequency of joint tagging, in which a non-causal SNP carried on population haplotypes together with distinct causal SNPs with similar effects may have a smaller single SNP p-value than either causal variant itself. This can cause stepwise search to get stuck on the tag, whereas stochastic search will ﬁnd both causal variants, if the sample sizes are large enough. With smaller sample sizes, stochastic search may also choose the tag, because such samples may not contain enough information to overcome the strong penalty needed by more complex models to avoid over-ﬁtting. Thus, joint tagging may potentially affect many more cases than identiﬁed above by the simple comparison of stepwise and stochastic search results from ﬁxed sample sizes. Proposed method for ﬁne-mapping multiple diseases. We noticed a striking overlap between the ﬁne-mapping results for different diseases in these regions, with 20 of 30 regions with two or more associated diseases showing evidence of overlap (Sup- plementary Fig. 8), consistent with previous reports of shared genetic etiology between the diseases2, which inspired the crea- tion of the ImmunoChip. This motivated us to exploit the sharing between diseases, extending the stochastic search approach to jointly analyse multiple diseases, borrowing information between them, to help overcome sample size limitations. We use a mul- tinomial logistic regression framework, the natural extension of the binomial logistic model, where each individual is assumed to belong to exactly one disease group or a pooled group of controls shared between diseases. This formally accounts for the sharing of controls between diseases in different studies. Each row summarises results for a single region, deﬁned by chromosome, start and end coordinates (hg19), with neighbouring or which SNP group it belongs to, by a letter in front of the SNP rs ID; the SNP group size, p-value of the SNP, and stochastic search p-values from these models are given. Results h Table 1 Regions having conﬂicting models selected by stepwise and stochastic search Region Disease SW model SW P-value SW model GPP SW SNP group size SS model P-value(s) SS GPP SNP group size(s) LD 2q-100544954-101038647 (AFF3) iRA A/rs10209110 3.79 × 10−9 0.427 32 C/rs13415465 1.96 × 10−8 0.514 93 0.46 2q-231076289-231235886 (SP110, SP140, SP140L) iCEL C/rs62192167 8.49 × 10−7 0.374 4 B/rs12694846 3.62 × 10−6 0.44 18 0.12 7p-50246236-50365063 (IKZF1) T1D A/rs2168587 8.10 × 10−7 0.123 1 D/rs17552787 2.46 × 10−6 0.457 20 0.094 7p-50366637-50694384 (DDC, FIGNL1, GRB10, IKZF1) T1D B/rs34046423 2.08 × 10−10 0.226 35 A/rs10264390 2.89 × 10−9 0.657 37 0.4 15q-67414055-67469568 (SMAD3) iCEL A/rs2289261 2.48 × 10−7 0.324 15 B/rs8024330 1.74 × 10−6 0.342 20 0.24 20p-1497197-1689461 (SIRPD/SIRPB1) T1D B/rs202535 6.79 × 10−9 0.221 9 C/rs202536 1.25 × 10−8 0.391 21 0.49 2q-204446380-204816382 (CTLA4) T1D G/rs3087243 3.89 × 10−17 0.00281 32 H/rs231779 + E/rs370078940 2.10 × 10−21 1.36 × 10−8 0.765 52 31 0.50 0.07 iRA G/rs3087243 1.54 × 10−7 0.00787 32 H/rs34029700 + E/rs7422494 7.04 × 10−9 4.39 × 10−9 0.753 52 31 0.26 0.05 4q-122973062-123565302 (IL2/IL21) T1D D/rs77516441 3.97 × 10−14 0.0193 13 F/rs13122213 + A/rs6837165 1.05 × 10−9 3.43 × 10−16 0.85 53 106 0.20 0.23 10p-6030000-6220000 (IL2RA) ATD J/rs706779 4.63 × 10−8 0.011 2 C/rs2476491 + A/rs61839660 2.89 × 10−9 1.96 × 10−8 0.954 8 31 0.34 0.13 14q-101290463-101328739 (MEG3) T1D C/rs34552516 9.69 × 10−10 0.0814 5 B/rs1054000 + A/rs11160606 1.13 × 10−11 2.36 × 10−6 0.777 5 16 0.29 0.30 NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.or g pp y g Finally, we showed that the pattern of LD between three SNPs (two causal and a third tag), together with MAF (minor allele frequency) and effect sizes, determine whether a tag SNP has the smallest expected p-value (Fig. 2a, Supplementary Note 1). At the extremes of this pattern, there is a non-zero probability that the tag model will be erroneously selected, even by a criterion such as BIC which penalises the larger model (Supplementary Note 2). While we cannot identify how many cases of joint tagging may exist in our GWAS data because the causal variants are unknown, we can quantify what proportion of 3 SNP LD matrices match this pattern under an assumption of equal odds ratios at the causal variants. Results h Stochastic and stepwise search differences in 10% of regions. We systematically applied stepwise and stochastic search ﬁne- mapping to dense genotyping data from ImmunoChip studies of six IMD: type 1 diabetes (T1D)11, multiple sclerosis (MS)12, autoimmune thyroid disease (ATD)13, celiac disease (CEL)14, juvenile idiopathic arthritis (JIA)15 and rheumatoid arthritis (RA) 16 (sample sizes given in Supplementary Table 1) in 90 densely mapped regions with at least one associated disease (Supple- mentary Data 1), 204 disease-region combinations in total. Results are given in Supplementary Data 2–3. For RA and CEL, we performed parallel analyses in UK-only and UK + interna- tional samples (iRA and iCEL, respectively). Simulations showed that if the J model (any model with exactly one SNP from group J) were true, both stepwise and stochastic search would correctly identify it (Fig. 1b, Supplementary Table 3). In contrast, if the A + C model (2-SNP model with a SNP from each of groups A and C) were true, stepwise got stuck on J, while stochastic search moved from selecting J at lower sample sizes, to A + C at higher sample sizes (Fig. 1b, Supplementary Table 4, further examples in other regions/ diseases in Supplementary Tables 5–8, SNP group membership in Supplementary Data 3). A small perturbation on the simulated effect sizes for A + C led both methods to select C or A + C directly, indicating that the potential for joint tagging was dependent on the combined effect sizes. Unlike stepwise search which produces a single best model, stochastic search results are a posterior probability distribution across typically thousands of potential causal variant models. To make these more interpretable, SNPs in high LD which meet the criteria of substitutability (see Methods) were grouped. The identiﬁcation of SNP groups is a feature of stochastic search— generally, SNPs in a group have high LD and similar evidence for association, such that a single candidate causal variant is not We explored a broader range of combined effect sizes mathematically, ﬁnding that there was a high probability of J having the smallest p-value when A and C were causal only when A and C had similar odds ratios; and that our observed data fell 2 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 within this region (Fig. 1c). A similar pattern was seen at all other mismatch regions (Supplementary Fig. 7). NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 ARTICLE C C C T C C C T T T T A rs61839660 rs706779 rs11594656 SNP Haplotypes 0 10 20 30 40 50 Fq (%) Frequency −0.4 −0.3 −0.2 −0.1 0.0 Log OR Effects A + C causal a c Null J AC Other Null J AC A+C Other Null C A + C J Other C A + C nullJ Other Null A A+CC J Other Null A + C CJ Other J A < C A > C Stepwise Stochastic 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations b 1.0 Pr J top 1.00 < > 0.75 0.50 0.25 0.00 0.5 0.0 log(OR) SNP A –0.5 –1.0 –1.0 –0.5 0.0 log(OR) SNP C 0.5 1.0 Fig. 1 Evidence for joint tagging. a Haplotype analysis of SNPs selected by stepwise search and GUESSFM for ATD in region 10p-6030000-6220000. A representative SNP from each SNP group is shown. Each column represents one SNP, with possible alleles colour coded according to major or minor. Each row is a haplotype—a speciﬁc combination of alleles across all SNPs—with frequency in UK controls and effect on disease risk (log OR + 95% CI). There are four common haplotypes. Three carry the minor allele at the J SNP rs706799, but only those that also carry minor allele at A or C show a signiﬁcant effect on disease risk. b Comparison of stepwise and stochastic search applied to simulated data. Causal variants were simulated as follows: J: single causal variant J, OR = 0.8; A < C causal variants A + C, odds ratios A:0.81, C:0.74; A > C: causal variants A + C, odds ratios A:0.74, C:0:8. Potential models include J (green), C (blue), A + C (purple), A (red) and null (black); any other models are grouped together as grey. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 Potential models include J (green), C (blue), A + C (purple), A (red) and null (black); any other models are grouped together as grey. The y-axis shows the proportion of simulations in which the stepwise approach chose the indicated model (adding SNPs while p < 10−6) or the average posterior probabilities for each model for the stochastic search approach. Sample size (x-axis) is the number of cases and controls. c Assuming A and C are causal, this plot shows the probability that J has the smallest p-value as a function of the effect sizes (log odds ratios) at A and C. The estimated effects for A and C from real data are shown by a point, and the simulations from b by < and > for A < C and A > C conditions, respectively. Source data for b are provided in Supplementary Tables 3–4 GUESSFM), takes only 15–83 s. We can deal with multiple populations, with not all populations represented for all diseases, by noting that when controls are not shared, the joint log Bayes factor is a simple sum of logistic log Bayes factors, allowing us to ﬁt a multinomial to the samples from common populations with shared controls, and add disease-speciﬁc log Bayes factor terms from logistic models ﬁtted to the distinct populations. computational time and memory will be required to evaluate all conﬁgurations, and to store these results. We provide solutions for both challenges. First, we show the log Bayes factor for a multinomial model that simultaneously considers all diseases can be approximated by a quantity that can be rapidly calculated—the sum of the log Bayes factors for the corresponding logistic models for each individual disease and an offset term determined by sample and model sizes (Supplementary Note 3). Second, we show that the marginal (single disease) model posteriors from the multinomial model can be calculated without needing to store the individual conﬁguration Bayes factors (Supplementary Note 3). Results h The LD column lists the r2 between the stepwise SNP and the SNP(s) from the stocha We introduce the concept of conﬁgurations—sets of causal variant models for each disease, and we borrow information between the diseases by means of a prior, which upweights conﬁgurations that share one or more causal variants between diseases by a factor κ (Fig. 3). Such a parameter is also used in colocalisation analysis, with values ranging from 1001,18 to 100019. In the case of MFM, it may be easier to elicit a prior on the chance of any sharing in causal variants between a pair of diseases, and we show in Supplementary Note 3 how this value can be used to derive κ for two or more diseases. In all our simulations and analyses, we chose κ so that the prior on any pair of diseases sharing at least one causal variant in a region where they are both associated is 0.5, compatible with conclusions of previous IMD studies of IMDs1. One obvious challenge for dealing with conﬁgurations, is that the number of models that needs to be considered for each disease is already large, and the number of possible conﬁgurations is the product of these. This implies that exponentially increased 3 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 Potential models include J (green), C (blue), A + C (purple), A (red) and null (black); any other models are grouped together as grey. The y-axis shows the proportion of simulations in which the stepwise approach chose the indicated model (adding SNPs while p < 10−6) or the average posterior probabilities for each model for the stochastic search approach. Sample size (x-axis) is the number of cases and controls. c Assuming A and C are causal, this plot shows the probability that J has the smallest p-value as a function of the effect sizes (log odds ratios) at A and C. The estimated effects for A and C from real data are shown by a point, and the simulations from b by < and > for A < C and A > C conditions, respectively. Source data for b are provided in Supplementary Tables 3–4 Null J AC Other Null J AC A+C Other Null C A + C J Other C A + C nullJ Other Null A A+CC J Other Null A + C CJ Other J A < C A > C Stepwise Stochastic 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations b b Stochastic Sample size (1000 s) Fig. 1 Evidence for joint tagging. a Haplotype analysis of SNPs selected by stepwise search and GUESSFM for ATD in region 10p-6030000-6220000. A representative SNP from each SNP group is shown. Each column represents one SNP, with possible alleles colour coded according to major or minor. Each row is a haplotype—a speciﬁc combination of alleles across all SNPs—with frequency in UK controls and effect on disease risk (log OR + 95% CI). There are four common haplotypes. Three carry the minor allele at the J SNP rs706799, but only those that also carry minor allele at A or C show a signiﬁcant effect on disease risk. b Comparison of stepwise and stochastic search applied to simulated data. Causal variants were simulated as follows: J: single causal variant J, OR = 0.8; A < C causal variants A + C, odds ratios A:0.81, C:0.74; A > C: causal variants A + C, odds ratios A:0.74, C:0:8. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 The y-axis shows the proportion of simulations in which the stepwise approach chose the indicated model (adding SNPs while p < 10−6) or the average posterior probabilities for each model C C C T C C C T T T T A rs61839660 rs706779 rs11594656 SNP Haplotypes 0 10 20 30 40 50 Fq (%) Frequency −0.4 −0.3 −0.2 −0.1 0.0 Log OR Effects A + C causal a c 1.0 Pr J top 1.00 < > 0.75 0.50 0.25 0.00 0.5 0.0 log(OR) SNP A –0.5 –1.0 –1.0 –0.5 0.0 log(OR) SNP C 0.5 1.0 A + C causal c 1.0 Pr J top 1.00 < > 0.75 0.50 0.25 0.00 0.5 0.0 log(OR) SNP A –0.5 –1.0 –1.0 –0.5 0.0 log(OR) SNP C 0.5 1.0 c a log(OR) SNP C rs61839660 rs706779 rs11594656 SNP 0 10 20 30 40 50 Fq (%) −0.4 −0.3 −0.2 −0.1 0.0 Log OR Null J AC Other Null J AC A+C Other Null C A + C J Other C A + C nullJ Other Null A A+CC J Other Null A + C CJ Other J A < C A > C Stepwise Stochastic 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations b 1.0 0.5 0.0 log(OR) SNP A 0.5 1.0 Fig. 1 Evidence for joint tagging. a Haplotype analysis of SNPs selected by stepwise search and GUESSFM for ATD in region 10p-6030000-6220000. A representative SNP from each SNP group is shown. Each column represents one SNP, with possible alleles colour coded according to major or minor. Each row is a haplotype—a speciﬁc combination of alleles across all SNPs—with frequency in UK controls and effect on disease risk (log OR + 95% CI). There are four common haplotypes. Three carry the minor allele at the J SNP rs706799, but only those that also carry minor allele at A or C show a signiﬁcant effect on disease risk. b Comparison of stepwise and stochastic search applied to simulated data. Causal variants were simulated as follows: J: single causal variant J, OR = 0.8; A < C causal variants A + C, odds ratios A:0.81, C:0.74; A > C: causal variants A + C, odds ratios A:0.74, C:0:8. NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 We consider the patterns of three-way LD between each possible trio of SNPs, nominating the ﬁrst two as causal, and the third as a potential tag. a For each pair of potential causal SNPs, we can predict whether the third SNP is a tag according to the pairwise correlation between that SNP and the two potentially causal SNPs (r1, r2). Red (yellow) areas indicate settings where the third SNP is (is not) a potential tag for SNPs 1 and 2. In this example, the potentially causal SNPs have equal MAF, equal effect on disease risk (equal odds ratios, OR) and are uncorrelated. Then, if the third SNP is (A) uncorrelated or weakly correlated with either SNP 1 or 2, or (B, C) negatively correlated with one and positively with the other, we would not expect it to act as a tag. On the other hand, if it were (D, E) strongly positively or negatively correlated with both causal variants, we would expect it to act as a joint tag. b shows the result of searching all possible SNP trios in UK ImmunoChip control data, and quantifying the proportion of trios that correspond to joint tagging in each region, assuming the causal variants have equal OR; the pattern is individually rare, consistently <5%. c shows the proportion of SNP pairs for which at least one potential tag exists, which can be substantial—about 40% overall. The blue curves show the loess ﬁts to the points. Source data are provided in Supplementary Data 4 nominal posterior probability, which are in high LD and rarely selected together in any model (Methods). nominal posterior probability, which are in high LD and rarely selected together in any model (Methods). One of the multi-SNP regions that showed differences across multiple diseases was on chromosome 2q, harbouring the candidate gene CTLA4. In stepwise analysis, iRA, T1D, ATD and CEL all converge on a single SNP model, in the group labelled G in the stochastic search results, while for iCEL a single SNP is selected in group I (Table 3, Fig. 5a). For single disease stochastic search, we ﬁnd CEL (UK-only) and ATD have a single signal in the group labelled G, matching the stepwise results, while RA and T1D both have two signals, in groups labelled E and H, represented by causal variant conﬁguration E + H. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 The iCEL result is more uncertain, with the posterior spread between I + K, I or E + G. Note that K is also the second selected SNP for iCEL stepwise regression (p = 4 × 10–6), although it doesn’t reach our adopted signiﬁcance threshold. Simulations show that G may tag an E + H model (Fig. 5b-c, Supplementary Tables 5–6, SNP group membership in Supplementary Data 3). MFM increases chance of selecting the correct model. We examined the performance of MFM by simulation. We found that when causal variants overlapped between diseases, MFM was able to recover the correct models at smaller sample sizes than indi- vidual disease analysis (Fig. 4a, b, Supplementary Data 5, 6), i.e. sharing information between diseases contributed to a gain in accuracy similar to increasing sample size for each disease. When no causal variants were shared, multinomial and independent approaches gave similar results (Fig. 4c, Supplementary Data 7), i.e. sharing information did not tend to mislead as long as there were strong signals in each disease. When one disease had no causal variants, multinomial and independent results were again similar (Fig. 4d, Supplementary Data 8); i.e. no information is gained but there is also no noticeable loss in accuracy in doing so. g p p pp y ) MFM ﬁnds increased support for E + H for RA and T1D while the CEL and iCEL results become more concentrated with support for G or E + G (Table 3). While we suggested G may tag E + H, MFM maintains strongest support for G in ATD, although there is also posterior support for H in combination with other groups (group marginal posterior probability of inclusion, gMPPI = 0.60). A previous attempt to ﬁne-map autoimmune disease association, by colocalisation analysis of T1D, RA and CEL (using the same UK data as here) came to similar conclusions, ﬁnding strong support for E + H models for iRA and T1D and either G or E + G for CEL1. However, a more recent analysis of T1D and RA, also in largely the same samples, identiﬁed a different pair of variants, rs3087243 (G) and rs117701653 (C)20 for both diseases using an exhaustive search of all one and two SNP models. MFM analysis of up to six IMD. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 a E 0.5 –0.5 0 –0.5 0.5 r2 r1 A C B D –1 E –1 a 0 E 0.5 –0.5 0 –0.5 0.5 r2 r1 A C B D –1 E 5 10 15 0 500 1000 1500 # SNPs % Tagging Trios that are tags b 0 20 40 60 0 500 1000 1500 # SNPs % Tagging At least one tag c –1 Fig. 2 Potential frequency of joint tagging. We consider the patterns of three-way LD between each possible trio of SNPs, nominating the ﬁrst two as causal, and the third as a potential tag. a For each pair of potential causal SNPs, we can predict whether the third SNP is a tag according to the pairwise correlation between that SNP and the two potentially causal SNPs (r1, r2). Red (yellow) areas indicate settings where the third SNP is (is not) a potential tag for SNPs 1 and 2. In this example, the potentially causal SNPs have equal MAF, equal effect on disease risk (equal odds ratios, OR) and are uncorrelated. Then, if the third SNP is (A) uncorrelated or weakly correlated with either SNP 1 or 2, or (B, C) negatively correlated with one and positively with the other, we would not expect it to act as a tag. On the other hand, if it were (D, E) strongly positively or negatively correlated with both causal variants, we would expect it to act as a joint tag. b shows the result of searching all possible SNP trios in UK ImmunoChip control data, and quantifying the proportion of trios that correspond to joint tagging in each region, assuming the causal variants have equal OR; the pattern is individually rare, consistently <5%. c shows the proportion of SNP pairs for which at least one potential tag exists, which can be substantial—about 40% overall. The blue curves show the loess ﬁts to the points. Source data are provided in Supplementary Data 4 0 5 10 15 0 500 1000 1500 # SNPs % Tagging Trios that are tags b 0 20 40 60 0 500 1000 1500 # SNPs % Tagging At least one tag c b b a c % Tagging Fig. 2 Potential frequency of joint tagging. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 These insights solve both computational time and memory challenges: joint analysis of 2–6 diseases in the IL2RA region, (after individual stochastic search results were generated with Finally, to enable interpretation of the posterior probability of thousands of models for each disease, which typically contain many models differing only by the exchange of one SNP for another in high LD, we formalise the method for grouping SNPs across multiple diseases by hierarchical clustering of SNPs according to their LD (r2) and the probability of being jointly required to explain disease, grouping SNPs selected with some 4 ARTICLE NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 MFM analysis Conﬁguration CB, B + D = models B for disease 1 and B + D for disease 2 Prior for CB, B + D depends on , and overlap between models B, B + D Multinomial BFB, B + D approximated from binomial BFB, BFB + D and sample size Calculate prosterior over all conﬁgurations, store only marginals for each disease Model Marginal Posterior Disease 1 Disease 2 sparsity sharing sparsity Fig. 3 Schematic diagram for Multinomial Fine-Mapping (MFM) method. MFM is used for multiple diseases with shared controls and, for simplicity, only two diseases are shown. After selection of a sparsity prior parameterised by π, stochastic search is applied individually to each disease, as in standard analyses. Marginal results are combined in MFM, using an approximation to the multinomial Bayes factor (BF) and with information shared between diseases using a sharing prior, which upweights joint models with shared causal variants by a factor κ. Memory constraints are dealt with by storing only the marginal posterior distributions for each disease in groups E (CEL, RA, T1D), G (CEL and ATD) and H (RA and T1D, and possibly ATD) and possibly two low frequency causal variants in groups C and K (RA, CEL). identiﬁed by stochastic search for CEL and iCEL as having similar protective effects across all diseases and low minor allele frequencies (<0.05). The two SNPs are unlinked (r2 < 0.01) and in low LD with other genotyped or imputed SNPs outside their groups (r2 < 0.2). The 2-SNP models E + H identiﬁed here, and G + C20 have similar BIC in our data for iRA and iCEL (Supplementary Data 10), but the greater number of SNPs in the E and H groups mean that E + H encompasses many more possible causal variant pairs and so has greater grouped posterior support. Additionally, individual E + H models have a clearly better ﬁt than G + C for T1D (Supplementary Data 10). In total, results in this region exemplify the difﬁculty with ﬁne-mapping multiple causal variants in the presence of complex LD, and suggest the region likely contains three common causal variants, Our previous report of stochastic-stepwise mismatch focused on MS and T1D in the IL2RA region9. We identiﬁed four groups of SNPs corresponding to four causal variants for T1D, with results agreeing between stepwise and stochastic search9. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 iﬁed by stochastic search for CEL and iCEL as having similar ctive effects across all diseases and low minor allele uencies (<0.05). The two SNPs are unlinked (r2 < 0.01) and w LD with other genotyped or imputed SNPs outside their in groups E (CEL, RA, T1D), G (CEL and ATD) and H (RA and T1D, and possibly ATD) and possibly two low frequency causal variants in groups C and K (RA, CEL). Our previous report of stochastic stepwise mismatch focused Model Data Model Data Model Data Posterior Model Data Posterior A BFA PA B BFB PB D BFD PD B + D BFB + D PB + D ... A B D B + D ... A B D B + D ... A B D B + D ... ... BFA BFB BFD BFB + D ... BFA BFB BFD BFB + D ... ... PA PB PD PB + D ... PA PB PD PB + D ... A BFA PA B BFB PB D BFD PD B + D BFB + D PB + D ... ... ... sparsity Genotype data disease2 Genotype data disease1 Genotype data controls Stochastic search (GUESSFM) Stochastic search (GUESSFM) Standard analysis MFM analysis Conﬁguration CB, B + D = models B for disease 1 and B + D for disease 2 Prior for CB, B + D depends on , and overlap between models B, B + D Multinomial BFB, B + D approximated from binomial BFB, BFB + D and sample size Calculate prosterior over all conﬁgurations, store only marginals for each disease Model Marginal Posterior Disease 1 Disease 2 sparsity sharing Schematic diagram for Multinomial Fine-Mapping (MFM) method. MFM is used for multiple diseases with shared controls and, for simplicity, only iseases are shown. After selection of a sparsity prior parameterised by π, stochastic search is applied individually to each disease, as in standard ses. Marginal results are combined in MFM, using an approximation to the multinomial Bayes factor (BF) and with information shared between ses using a sharing prior, which upweights joint models with shared causal variants by a factor κ. Memory constraints are dealt with by storing only arginal posterior distributions for each disease Model Data Posterior Model Data Posterior A BFA PA B BFB PB D BFD PD B + D BFB + D PB + D ... NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 We applied MFM to all 30 ImmunoChip regions with at least two associated diseases (Sup- plementary Data 9, visualised at https://chr1swallace.github.io/ MFM-output/index.html). We identiﬁed seven regions for which the top model by independent stochastic search and MFM dif- fered (Table 2). Four of these were single SNP models under independent analysis, which moved to an alternative single SNP in MFM. For three of these four, the difference was seen in analysis of a UK-only subset, so that we could consider inde- pendent analysis of the UK + international data, which included more samples but used the more conventional analysis method as an adjudicator. In all three cases, this adjudicator matched the MFM analysis of the UK-only data, suggesting that UK inde- pendent analysis was limited by power, and that UK MFM ana- lysis increased power, allowing conclusions to be drawn that were consistent with those seen in a larger single disease analysis. We compared the models suggested by all these studies across all diseases by BIC (Supplementary Data 10) and using haplotype analysis (Fig. 5d). This visually highlighted rs117701653/C identiﬁed for iRA by exhaustive search20 and rs76676160/K 5 5 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 ... ... A BFA PA B BFB PB D BFD PD B + D BFB + D PB + D ... ... ... sparsity Genotype data disease2 Genotype data disease1 Genotype data controls Stochastic search (GUESSFM) Stochastic search (GUESSFM) Standard analysis Model Data Model Data ... A B D B + D ... A B D B + D ... A B D B + D ... ... BFA BFB BFD BFB + D ... BFA BFB BFD BFB + D ... ... PA PB PD PB + D ... PA PB PD PB + D ... ... ... ... Stan MFM analysis Conﬁguration CB, B + D = models B for disease 1 and B + D for disease 2 Prior for CB, B + D depends on , and overlap between models B, B + D Multinomial BFB, B + D approximated from binomial BFB, BFB + D and sample size Calculate prosterior over all conﬁgurations, store only marginals for each disease Model Marginal Posterior Disease 1 Disease 2 sparsity sharing Fig. 3 Schematic diagram for Multinomial Fine-Mapping (MFM) method. MFM is used for multiple diseases with shared controls and, for simplicity, on two diseases are shown. After selection of a sparsity prior parameterised by π, stochastic search is applied individually to each disease, as in standar analyses. Marginal results are combined in MFM, using an approximation to the multinomial Bayes factor (BF) and with information shared between diseases using a sharing prior, which upweights joint models with shared causal variants by a factor κ. Memory constraints are dealt with by storing on the marginal posterior distributions for each disease Model Data Model Data A B D B + D ... A B D B + D ... A B D B + D ... BFA BFB BFD BFB + D ... BFA BFB BFD BFB + D ... PA PB PD PB + D ... PA PB PD PB + D ... NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 Other Other Null D AC B A.D Null D A C B A.D Other Other Null DA C A.C Null C DA A.C Disease 1 Disease 2 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations a Other Other Null C A.C C Null A.C Other Other B D Null A.D Null B D A.D A < C A < D 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 Sample size (1000 s) Propn simulations b Other Other Null C D B A.D Null C D B A.D Other Other Null C A.C Null C A.C A = D C 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations c other Other Null D A B A.D Null D A B A.D Other Other DB Null D B Null A = D NULL 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations d Method a a Independent Shared Fig. 4 Comparison of MFM analysis and single disease analysis. Causal variants were simulated for two diseases with models deﬁned by SNP groups from the IL2RA region. MFM is shown by solid lines and independent analyses by dashed lines. Throughout, disease 1 has causal variants A + D, while causal variants for disease 2 vary. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 How- ever, while stepwise search identiﬁed a single SNP for MS, rs2104286 (group B), stochastic search identiﬁed two distinct variants in groups A and D (posterior probability 55%), and suggested that rs2104286/B was a joint tag for these groups (r2 = 0.334 and 0.301, respectively)9, a conclusion supported by haplotype analysis and simulations here (Fig. 6, Supplementary Tables 7, 8). NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications 6 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 a, b Disease 2 has causal variants A + C and the odds ratio of A, ORA, is the same for both diseases; a A has a stronger effect than C and D; ORA = 1.4 (both), ORD = 1.25 (disease 1), ORC = 1.25 (disease 2). b A has a weaker effect than C and D; ORA = 1.25 (both), ORD = 1.4 (disease 1), ORC = 1.4 (disease 2). c Disease 2 has only C causal; ORA = ORD = 1.25 (disease 1), ORC = 1.25 (disease 2). d Disease 2 has no causal variants (no association). Potential models include A (red), B (green), C (blue), D (yellow), A + D (orange), A + C (purple) and null (black); any other models are grouped together as grey. The y-axis shows the average posterior probabilities for each model. a, b MFM can identify the true two causal variant model at smaller sample sizes than independent analysis in simulated data when there is sharing between diseases. c, d When there is no sharing (c) or one disease has no true associations (d), no information is gained by using MFM but there is only minimal loss in accuracy in doing so. Source data are provided in Supplementary Data 5–8 Sample size (1000 s) other Other Null D A B A.D Null D A B A.D Other Other DB Null D B Null A = D NULL 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations d Other Other Null C D B A.D Null C D B A.D Other Other Null C A.C Null C A.C A = D C 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations c d d c C Method a a Independent Shared Fig. 4 Comparison of MFM analysis and single disease analysis. Causal variants were simulated for two diseases with models deﬁned by SNP groups from the IL2RA region. MFM is shown by solid lines and independent analyses by dashed lines. Throughout, disease 1 has causal variants A + D, while causal variants for disease 2 vary. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 a, b Disease 2 has causal variants A + C and the odds ratio of A, ORA, is the same for both diseases; a A has a stronger effect than C and D; ORA = 1.4 (both), ORD = 1.25 (disease 1), ORC = 1.25 (disease 2). b A has a weaker effect than C and D; ORA = 1.25 (both), ORD = 1.4 (disease 1), ORC = 1.4 (disease 2). c Disease 2 has only C causal; ORA = ORD = 1.25 (disease 1), ORC = 1.25 (disease 2). d Disease 2 has no causal variants (no association). Potential models include A (red), B (green), C (blue), D (yellow), A + D (orange), A + C (purple) and null (black); any other models are grouped together as grey. The y-axis shows the average posterior probabilities for each model. a, b MFM can identify the true two causal variant model at smaller sample sizes than independent analysis in simulated data when there is sharing between diseases. c, d When there is no sharing (c) or one disease has no true associations (d), no information is gained by using MFM but there is only minimal loss in accuracy in doing so. Source data are provided in Supplementary Data 5–8 Fig. 4 Comparison of MFM analysis and single disease analysis. Causal variants were simulated for two diseases with models deﬁned by SNP groups from the IL2RA region. MFM is shown by solid lines and independent analyses by dashed lines. Throughout, disease 1 has causal variants A + D, while causal variants for disease 2 vary. a, b Disease 2 has causal variants A + C and the odds ratio of A, ORA, is the same for both diseases; a A has a stronger effect than C and D; ORA = 1.4 (both), ORD = 1.25 (disease 1), ORC = 1.25 (disease 2). b A has a weaker effect than C and D; ORA = 1.25 (both), ORD = 1.4 (disease 1), ORC = 1.4 (disease 2). c Disease 2 has only C causal; ORA = ORD = 1.25 (disease 1), ORC = 1.25 (disease 2). d Disease 2 has no causal variants (no association). Potential models include A (red), B (green), C (blue), D (yellow), A + D (orange), A + C (purple) and null (black); any other models are grouped together as grey. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 a, b Disease 2 has causal variants A + C and the odds ratio of A, ORA, is the same for both diseases; a A has a stronger effect th C d D OR 1 4 (b th) OR 1 25 (di 1) OR 1 25 (di 2) b A h k ff t th C d D OR 1 25 (b th) OR 1 4 Other Other Null D AC B A.D Null D A C B A.D Other Other Null DA C A.C Null C DA A.C Disease 1 Disease 2 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations a Other Other Null C A.C C Null A.C Other Other B D Null A.D Null B D A.D A < C A < D 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 Sample size (1000 s) Propn simulations b b a Propn simulations Other Other Null D AC B Null D A C B Other Other Null DA C Null C DA 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 Sample size (1000 s) Propn simulatio Other Other Null Null Other Other B Null A.D Null B 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 Sample size (1000 s) Propn simulatio Other Other Null C D B A.D Null C D B A.D Other Other Null C A.C Null C A.C A = D C 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations c other Other Null D A B A.D Null D A B A.D Other Other DB Null D B Null A = D NULL 1 2 3 4 5 1 2 3 4 5 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations d Method a a Independent Shared Fig. 4 Comparison of MFM analysis and single disease analysis. Causal variants were simulated for two diseases with models deﬁned by SNP groups from the IL2RA region. MFM is shown by solid lines and independent analyses by dashed lines. Throughout, disease 1 has causal variants A + D, while causal variants for disease 2 vary. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 The last column gives the mean r2 between the SNP group(s) of independent analyses and those of MFM. The other diseases that were used in MFM are listed under Other Diseases Table 3 Summary results for ﬁne-mapping in CTLA4 and IL2RA Region Disease SW model SW P Indep. model Indep. PP MFM (UK) model MFM (UK) PP MFM (Int.) model MFM (Int.) PP CTLA4 ATD G/rs11571297 1.22 × 10−24 G 0.842 G 0.593 G 0.374 G + H 0.349 H + I 0.273 G + H 0.236 E + H 0.102 CEL G/rs3087243 1.48 × 10−12 G 0.641 G 0.517 G + K 0.136 E + G 0.281 iCEL I/rs2162610 3.74 × 10−14 I + K 0.351 E + G 0.829 I 0.14 E + G 0.115 iRA G/rs3087243 1.54 × 10−7 E + H 0.753 E + H 0.805 A + E + H 0.142 T1D G/rs3087243 3.89 × 10−17 E + H 0.765 E + H 0.687 E + H 0.904 G 0.135 IL2RA ATD J/rs706779 4.63 × 10−8 A + C 0.954 A + C 0.985 A + C 0.986 MS B/rs2104286 1.13 × 10−13 B 0.632 A + D 0.883 A + D 0.901 A + D 0.188 H 0.114 iRA I/rs706778 7.55 × 10−8 I 0.966 I 0.695 A 0.201 T1D A/rs61839660 3.60 × 10−34 A + C + E + F 0.622 A + C + E + F 0.684 A + C + E + F 0.674 C/rs11594656 5.85 × 10−12 A + E + F + H + I 0.201 A + E + F + H + I 0.178 A + E + F + H + I 0.191 E/rs12220852 8.79 × 10−10 For each disease, the following are provided: selected stepwise (SW) model and conditional SNP p-values, high PP models (and PP) for each of independent analyses (Indep.), MFM (UK samples only) and MFM with international samples. CTLA4 and IL2RA are the regions 2q-204446380-2048163 and 10p-6030000-6220000, respectively Table 2 Regions with conﬂicting models chosen by independent disease analysis and MFM Table 3 Summary results for ﬁne-mapping in CTLA4 and IL2RA Region Disease SW model SW P Indep. model Indep. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 PP MFM (UK) model MFM (UK) PP MFM (Int.) model MFM (Int.) PP CTLA4 ATD G/rs11571297 1.22 × 10−24 G 0.842 G 0.593 G 0.374 G + H 0.349 H + I 0.273 G + H 0.236 E + H 0.102 CEL G/rs3087243 1.48 × 10−12 G 0.641 G 0.517 G + K 0.136 E + G 0.281 iCEL I/rs2162610 3.74 × 10−14 I + K 0.351 E + G 0.829 I 0.14 E + G 0.115 iRA G/rs3087243 1.54 × 10−7 E + H 0.753 E + H 0.805 A + E + H 0.142 T1D G/rs3087243 3.89 × 10−17 E + H 0.765 E + H 0.687 E + H 0.904 G 0.135 IL2RA ATD J/rs706779 4.63 × 10−8 A + C 0.954 A + C 0.985 A + C 0.986 MS B/rs2104286 1.13 × 10−13 B 0.632 A + D 0.883 A + D 0.901 A + D 0.188 H 0.114 iRA I/rs706778 7.55 × 10−8 I 0.966 I 0.695 A 0.201 T1D A/rs61839660 3.60 × 10−34 A + C + E + F 0.622 A + C + E + F 0.684 A + C + E + F 0.674 C/rs11594656 5.85 × 10−12 A + E + F + H + I 0.201 A + E + F + H + I 0.178 A + E + F + H + I 0.191 E/rs12220852 8.79 × 10−10 For each disease, the following are provided: selected stepwise (SW) model and conditional SNP p-values, high PP models (and PP) for each of independent analyses (Indep.), MFM (UK samples only) and MFM with international samples. CTLA4 and IL2RA are the regions 2q-204446380-2048163 and 10p-6030000-6220000, respectively Table 3 Summary results for ﬁne-mapping in CTLA4 and IL2RA For each disease, the following are provided: selected stepwise (SW) model and conditional SNP p-values, high PP models (and PP) for each of independent analyses (Indep.), MFM (UK samples only) and MFM with international samples. CTLA4 and IL2RA are the regions 2q-204446380-2048163 and 10p-6030000-6220000, respectively For each disease, the following are provided: selected stepwise (SW) model and conditional SNP p-values, high PP models (and PP) for each of and MFM with international samples. CTLA4 and IL2RA are the regions 2q-204446380-2048163 and 10p-6030000-6220000, respectively rs706778 and rs11256557 in the haplotype analysis, Supplemen- tary Fig. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 Table 2 Regions with conﬂicting models chosen by independent disease analysis and MFM Region Disease Other diseases Independent MFM Mean r2 between groups 1p-2406887-2785671 (MMEL1, TNFRSF14) RA CEL, MS D/rs4648662 C/rs10752749 0.36 1p-2406887-2785671 (MMEL1, TNFRSF14) iRA iCEL, MS C/rs141426426 C/rs10797431 1 6q-90806835-91039808 (BACH2) RA ATD, T1D G/rs56258221 C/rs72928038 0.33 6q-90806835-91039808 (BACH2) iRA ATD, T1D C/rs72928038 C/rs72928038 1 18p-12738413-12924117 (PTPN2) CEL T1D F/rs34799913 C/rs12967678 0.4 18p-12738413-12924117 (PTPN2) iCEL iRA, T1D C/rs67878610 C/rs12967678 1 7p-37363978-37440453 (ELMO1) MS CEL A/ rs1962401 C/rs77801025 0.47 2q-204446380-204816382 (CTLA4) iCEL ATD, iRA, T1D I/rs2162610 + K/rs76676160 G/rs3087243 + E/rs3116499 (I,G): 0.14 (I,E): 0.17 (K,G): 0.031 (K,E): 0.004 10p-6030000-6220000 (IL2RA) MS ATD, iRA, T1D B/rs2104286/ A/rs12722496 + D/rs7089861 0.2 0.3 16p-11017058-11307024 (DEXI) MS T1D A/rs11643622 B/rs12708716 + D/rs4780346 0.3 0.3 Each row summarises results for a single region, deﬁned by chromosome, start and end coordinates (hg19), with a previously reported candidate gene name shown for orientation. The best model for each method is selected by group posterior probability (GPP) and for each method the best SNP models for each group(s) are given as representatives of the group models. The last column gives the mean r2 between the SNP group(s) of independent analyses and those of MFM. The other diseases that were used in MFM are listed under Other Diseases Table 2 Regions with conﬂicting models chosen by independent disease analysis and MFM Region Disease Other diseases Independent MFM Mean r2 between groups 1p-2406887-2785671 (MMEL1, TNFRSF14) RA CEL, MS D/rs4648662 C/rs10752749 0.36 1p-2406887-2785671 (MMEL1, TNFRSF14) iRA iCEL, MS C/rs141426426 C/rs10797431 1 6q-90806835-91039808 (BACH2) RA ATD, T1D G/rs56258221 C/rs72928038 0.33 6q-90806835-91039808 (BACH2) iRA ATD, T1D C/rs72928038 C/rs72928038 1 18p-12738413-12924117 (PTPN2) CEL T1D F/rs34799913 C/rs12967678 0.4 18p-12738413-12924117 (PTPN2) iCEL iRA, T1D C/rs67878610 C/rs12967678 1 7p-37363978-37440453 (ELMO1) MS CEL A/ rs1962401 C/rs77801025 0.47 2q-204446380-204816382 (CTLA4) iCEL ATD, iRA, T1D I/rs2162610 + K/rs76676160 G/rs3087243 + E/rs3116499 (I,G): 0.14 (I,E): 0.17 (K,G): 0.031 (K,E): 0.004 10p-6030000-6220000 (IL2RA) MS ATD, iRA, T1D B/rs2104286/ A/rs12722496 + D/rs7089861 0.2 0.3 16p-11017058-11307024 (DEXI) MS T1D A/rs11643622 B/rs12708716 + D/rs4780346 0.3 0.3 Each row summarises results for a single region, deﬁned by chromosome, start and end coordinates (hg19), with a previously reported candidate gene name shown for orientation. The best model for each method is selected by group posterior probability (GPP) and for each method the best SNP models for each group(s) are given as representatives of the group models. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 The y-axis shows the average posterior probabilities for each model. a, b MFM can identify the true two causal variant model at smaller sample sizes than independent analysis in simulated data when there is sharing between diseases. c, d When there is no sharing (c) or one disease has no true associations (d), no information is gained by using MFM but there is only minimal loss in accuracy in doing so. Source data are provided in Supplementary Data 5–8 While our previous analysis included UK and non-UK (international) cases and controls for MS, here we used only the UK subset, and both stepwise and stochastic search identiﬁed B (group posterior probability, GPP = 0.632), with the A + D model having only GPP = 0.188, consistent with results that stochastic selection of a joint tag depends on sample size (Table 3). A more recent stepwise analysis of a larger, international sample has identiﬁed two SNPs, rs11256593 and rs1272255921. rs12722559 (r2 0.323 with rs2104286/B) is in our group H (GPP = 0.114, third strongest stochastic search model) while rs12722559 (r2 0.482 with rs2104286/B, MPPI = 1.20 × 10−5) was not in our SNP groups. In our UK data, we found the best ﬁtting models were A + D (BIC 19299.46) and B (19302.88), both signiﬁcantly better ﬁts than rs11256593 + rs12722559 (BIC 19320.06). For ATD, stepwise search identiﬁed a 1-SNP model, rs706779/ J, consistent with previous analyses of ATD13,22, and matching the top reported SNP for another IMD, Vitiligo23, while stochastic search selected a two SNP model, A + C (Table 1, Supplementary Data 11). MFM maintained support for the A + C model for ATD, and preferred the 2-SNP A + D model for MS (Table 3), agreeing with our previous stochastic search results for a larger UK + international MS dataset9. Limited power may also 7 NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 The y-axis shows the proportion of simulations in which the stepwise approach chose the indicated model (adding SNPs while p < 10-6) or the average posterior probabilities for each model for the stochastic search approach. Sample size (x-axis) is the number of cases and controls. c Assuming E and H are causal, this plot shows the probability that G has the smallest p-value as a function of the effect sizes (log odds ratios) at E and H. The estimated effects for E and H from T1D data are shown by a point, and the simulations from b by < and > for E < H and E > H conditions, respectively. d Haplotype analysis of SNP groups with support in any analysis. Each row represents one SNP, with possible alleles colour coded according to major or minor. Each column is a haplotype—a speciﬁc combination of alleles across all SNPs—with frequency in UK controls and effect on disease risk (log OR + 95% CI). MAF is shown as a percentage on a log scale to allow frequencies of rarer haplotypes to be distinguished. Source data for b are provided in Supplementary Tables 5, 6 Fig. 5 Analysis of chromosome 2q region containing CTLA4. a Map showing positions of SNPs (GRCh37) colour coded by SNP group. SNPs in the same group are in high LD. b Comparison of stepwise and stochastic search applied to simulated data. Causal variants were simulated as follows: G: causal variant G, OR = 1.25; E < H and E > H causal variants E + H, E < H:ORE = 1.19, ORH = 1.24 (observed in T1D data); E > H: ORE = 1.24, ORH = 1.19. Possible models include E (red), G (yellow), H (blue), E + H (purple) and null (black); any other models are grouped together as grey. The y-axis shows the proportion of simulations in which the stepwise approach chose the indicated model (adding SNPs while p < 10-6) or the average posterior probabilities for each model for the stochastic search approach. Sample size (x-axis) is the number of cases and controls. c Assuming E and H are causal, this plot shows the probability that G has the smallest p-value as a function of the effect sizes (log odds ratios) at E and H. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 9) selected for RA-international are carried along with the minor protective alleles of groups A, C and D and it is possible that the group I SNPs are tagging three IL2RA SNP groups. affect RA-international in this region, for which individual analysis picked group I (97%) and MFM support was split between groups A (20%) and I (70%). g Our results emphasise the importance of the A group, which is selected for three of the four diseases (T1D, MS and ATD). This group of SNPs have been previously associated with variation in the expression of IL2RA mRNA and of its encoded protein, CD25, in CD4+ memory T cells24,25, and a recent allele-speciﬁc expression (ASE) study has pinpointed the causal variant affecting mRNA expression among the set as rs6183966026— notably the same variant identiﬁed in an IBD GWAS of 67,852 individuals27 and an eczema/dermatitis phenotype in a GWAS of thousands of phenotypes for 337,000 samples in the UK Biobank28, supporting the notion that this SNP has a common effect across multiple diseases that is missed in stepwise analysis of MS and ATD. We note that the direction of effect for rs61839660 is opposite in IBD and eczema/dermatitis (risk allele T) compared to T1D, MS, ATD and JIA15 (risk allele C). We note also that the minor alleles of group I SNPs (represented by Allele-speciﬁc expression conﬁrms functional effects. In addi- tion to linking group A SNPs to IL2RA expression, we have shown that SNPs in group D decrease the percentage of CD25 expressing naive T cells9,24. Here, we extend our analysis of IL2RA mRNA expression to examine any effects of rs2104286/B in the context of groups A and D. ASE assays compare relative expression between paternally and maternally inherited chro- mosomes in individuals heterozygous for a putative functional SNP according to the allele each chromosome carries at the SNP. It is a powerful design, because the within-individual comparison controls for between individual biological variation resulting from other genetic and environmental differences. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 We quantiﬁed ASE of IL2RA mRNA in memory and naive CD4+ T cells NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications 8 CD28 CTLA4 ICOS rs11681201/E rs1980421/E rs1980422/E rs7422494/E rs7426056/E rs7588874/E rs67895606/E rs117701653/C rs4673266/E rs231389/E rs13429019/E rs5837875/E rs3116499/E rs62184035/E rs370078940/E rs6745050/G rs144115812/G rs34636506/G rs1968351/G rs13030124/G rs60306514/G rs61272514/G rs34050244/G rs34461055/G rs35702010/G rs12990970/G rs1863800/G rs112783914/G rs2882971/G rs116604011/G rs114019067/G rs2882972/G rs13033315/G rs231804/G rs11571304/G rs231806/G rs11393378/G rs231811/G rs2162610/I rs77040243/K rs231775/H rs231779/H rs3087243/G rs231723/H rs231724/H rs2307982/H rs231725/H rs231726/H rs1427676/H rs231727/H rs11571302/G rs7565213/G rs231729/H rs11571297/G rs137967632/H rs2882973/H rs79570815/K rs75309573/K rs34029700/H rs1365965/H rs6737395/H rs1991416/H rs2352546/H rs34302643/H rs73993041/H rs3096849/H rs3096851/H rs3096852/H rs3116523/H rs3116504/H rs2162608/H rs3116518/H rs76676160/K rs6752680/I rs3116505/H rs3096857/H rs6714468/I rs3096865/H rs12464033/H rs76210392/K rs138585195/K a Null G Null G H Null G H Null H E + H G Null G H Null E + H G H G E < H E > H Stepwise Stochastic 1 2 3 4 5 2 4 6 2 4 6 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations b c −0.5 0.0 0.5 Log OR Disease ATD CEL iCEL iRA T1D 1 3 10 30 MAF (log scale) G T T T T T T T T G G G A A A A A A G G G G G A A A A A A A A A A C C C A A A A A A A A C T T T T T T C C C rs76676160/K rs231775/H rs3087243/G rs2162610/I rs117701653/C rs1980422/E SNP Reference haplotype allele a a False True d 1.0 H + E causal Pr G top 1.00 0.75 0.50 0.25 0.00 0.5 0.0 log(OR) SNP E –0.5 –1.0 –1.0 –0.5 0.0 log(OR) SNP H 0.5 1.0 Fig. 5 Analysis of chromosome 2q region containing CTLA4. a Map showing positions of SNPs (GRCh37) colour coded by SNP group. SNPs in the same group are in high LD. b Comparison of stepwise and stochastic search applied to simulated data. Causal variants were simulated as follows: G: causal variant G, OR = 1.25; E < H and E > H causal variants E + H, E < H:ORE = 1.19, ORH = 1.24 (observed in T1D data); E > H: ORE = 1.24, ORH = 1.19. Possible models include E (red), G (yellow), H (blue), E + H (purple) and null (black); any other models are grouped together as grey. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 CD28 CTLA4 ICOS rs11681201/E rs1980421/E rs1980422/E rs7422494/E rs7426056/E rs7588874/E rs67895606/E rs117701653/C rs4673266/E rs231389/E rs13429019/E rs5837875/E rs3116499/E rs62184035/E rs370078940/E rs6745050/G rs144115812/G rs34636506/G rs1968351/G rs13030124/G rs60306514/G rs61272514/G rs34050244/G rs34461055/G rs35702010/G rs12990970/G rs1863800/G rs112783914/G rs2882971/G rs116604011/G rs114019067/G rs2882972/G rs13033315/G rs231804/G rs11571304/G rs231806/G rs11393378/G rs231811/G rs2162610/I rs77040243/K rs231775/H rs231779/H rs3087243/G rs231723/H rs231724/H rs2307982/H rs231725/H rs231726/H rs1427676/H rs231727/H rs11571302/G rs7565213/G rs231729/H rs11571297/G rs137967632/H rs2882973/H rs79570815/K rs75309573/K rs34029700/H rs1365965/H rs6737395/H rs1991416/H rs2352546/H rs34302643/H rs73993041/H rs3096849/H rs3096851/H rs3096852/H rs3116523/H rs3116504/H rs2162608/H rs3116518/H rs76676160/K rs6752680/I rs3116505/H rs3096857/H rs6714468/I rs3096865/H rs12464033/H rs76210392/K rs138585195/K a a c 1.0 H + E causal Pr G top 1.00 0.75 0.50 0.25 0.00 0.5 0.0 log(OR) SNP E –0.5 –1.0 –1.0 –0.5 0.0 log(OR) SNP H 0.5 1.0 Null G Null G H Null G H Null H E + H G Null G H Null E + H G H G E < H E > H Stepwise Stochastic 1 2 3 4 5 2 4 6 2 4 6 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations b b c log(OR) SNP E p ( ) −0.5 0.0 0.5 Log OR Disease ATD CEL iCEL iRA T1D 1 3 10 30 MAF (log scale) G T T T T T T T T G G G A A A A A A G G G G G A A A A A A A A A A C C C A A A A A A A A C T T T T T T C C C rs76676160/K rs231775/H rs3087243/G rs2162610/I rs117701653/C rs1980422/E SNP Reference haplotype allele a a False True d g( ) d Reference haplotype allele a a False True Fig. 5 Analysis of chromosome 2q region containing CTLA4. a Map showing positions of SNPs (GRCh37) colour coded by SNP group. SNPs in the same group are in high LD. b Comparison of stepwise and stochastic search applied to simulated data. Causal variants were simulated as follows: G: causal variant G, OR = 1.25; E < H and E > H causal variants E + H, E < H:ORE = 1.19, ORH = 1.24 (observed in T1D data); E > H: ORE = 1.24, ORH = 1.19. Possible models include E (red), G (yellow), H (blue), E + H (purple) and null (black); any other models are grouped together as grey. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 The y-axis shows the proportion of simulations in which the stepwise approach chose the indicated model (adding SNPs while p < 10-6) or the average posterior probabilities for each model for the stochastic search approach. Sample size (x-axis) is the number of cases and controls. c Assuming E and H are causal, this plot shows the probability that G has the smallest p-value as a function of the effect sizes (log odds ratios) at E and H. The estimated effects for E and H from T1D data are shown by a point, and the simulations from b by < and > for E < H and E > H conditions, respectively. d Haplotype analysis of SNP groups with support in any analysis. Each row represents one SNP, with possible alleles colour coded according to major or minor. Each column is a haplotype—a speciﬁc combination of alleles across all SNPs—with frequency in UK controls and effect on disease risk (log OR + 95% CI). MAF is shown as a percentage on a log scale to allow frequencies of rarer haplotypes to be distinguished. Source data for b are provided in Supplementary Tables 5, 6 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 6 Analysis of chromosome 10p region containing IL2RA a Map showing positions of SNPs (GRCh37) in groups A, B and D. SNPs in the same group are in high LD, with colour used to indicate group membership. b Haplotype analysis of SNPs selected by stepwise search and GUESSFM for MS. Each row represents one SNP, with possible alleles colour coded according to major or minor. Each column is a haplotype—a speciﬁc combination of alleles across all SNPs—with frequency in UK controls and effect on disease risk (log OR + 95% CI). There are four common haplotypes. Three appear protective, carrying the minor allele at either A or D, but only two carry the minor allele at B. c Comparison of stepwise and stochastic search applied to simulated data. Causal variants were simulated as follows: B: single causal variant B, OR = 0.8; A < D causal variants A + D, odds ratios A:0.84, D:0.77; A~D: causal variants A + D, odds ratios A:0.81, D:0:8 (observed in MS data); A > D: causal variants A + D, odds ratios A:0.77, D:0:84. Potential models include A (red), B (green), D (blue), A + D (purple) and null (black); any other models are grouped together as grey. The y-axis shows the proportion of simulations in which the stepwise approach chose the indicated model (adding SNPs while p < 10-6) or the average posterior probabilities for each model for the stochastic search approach. Sample size (x-axis) is the number of cases and controls. d Assuming A and D are causal, this plot shows the probability that B has the smallest p-value as a function of the effect sizes (log odds ratios) at A and D. The estimated effects for A and D from MS data are shown by a point, and the simulations from c by < and > for A < D and A > D conditions respectively. Source data for c are provided in Supplementary Tables 7, 8 isolated from 36 donors selected by genotype from a bioresource (www.cambridgebioresource.org.uk) to be heterozygous at SNPs in group A (A-het), D (D-het) or both (A + D-het). To control for other potential effects, we chose donors homozygous for SNPs in groups C and F. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 A A IL2RA RBM17 rs41294681/A rs28360493/A rs12722563/A rs12722558/A rs12722552/A rs12722522/A rs12722508/A rs7909519/A rs61839660/A rs12722496/A rs12722495/A rs12722504/B rs2104286/B rs12722488/B rs3118475/D rs7089861/D rs62626317/D rs41295055/D rs56382813/D 1 0.6 0.6 0.6 0.5 0.5 0.5 0.5 0.5 0.5 0.1 0.1 0.1 0.6 0.6 0.6 0.5 0.5 0.5 0.5 0.5 0.5 0.1 0.1 0.1 1 1 0.9 0.9 0.9 0.7 0.8 0.8 0.3 0.3 0.3 1 0.9 0.9 0.9 0.7 0.8 0.8 0.3 0.2 0.3 0.9 0.9 0.9 0.7 0.8 0.8 0.3 0.3 0.3 1 1 0.8 0.9 0.9 0.3 0.3 0.3 1 0.8 0.9 0.9 0.3 0.3 0.3 0.8 1 1 0.3 0.3 0.3 0.9 0.9 0.3 0.3 0.3 1 0.3 0.3 0.3 0.3 0.3 0.3 1 0.9 0.4 0.2 0.4 0.4 0.3 1 0.4 0.2 0.4 0.4 0.3 0.4 0.2 0.4 0.4 0.3 0.6 1 1 0.8 0.6 0.6 0.7 1 0.8 0.8 a −0.4 −0.3 −0.2 −0.1 0.0 Log OR 0 20 40 60 Fq (%) C C C T C C T T C C T T rs61839660/A rs2104286/B rs41295055/D SNP Major allele a a FALSE TRUE b Null B A + D Null B A + D Null D B Null A A + D B Null B D Null B A + D Null B Null B A + D D D B A < D A~D A > D Stepwise Stochastic 1 2 3 4 5 2 4 6 2 4 6 2 4 6 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations c d 1.0 A + D causal 0.5 0.0 log(OR) SNP D log(OR) SNP A –0.5 –1.0 –1.0 –0.5 0.0 0.5 Pr B top 1.00 0.75 0.50 0.25 0.00 1.0 D Fig. 6 Analysis of chromosome 10p region containing IL2RA a Map showing positions of SNPs (GRCh37) in groups A, B and D. SNPs in the same group are in high LD, with colour used to indicate group membership. b Haplotype analysis of SNPs selected by stepwise search and GUESSFM for MS. Each row represents one SNP, with possible alleles colour coded according to major or minor. Each column is a haplotype—a speciﬁc combination of alleles across all SNPs—with frequency in UK controls and effect on disease risk (log OR + 95% CI). There are four common haplotypes. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 Source data for c are provided in Supplementary Tables 7, 8 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467 019 11271 0 IL2RA RBM17 rs41294681/A rs28360493/A rs12722563/A rs12722558/A rs12722552/A rs12722522/A rs12722508/A rs7909519/A rs61839660/A rs12722496/A rs12722495/A rs12722504/B rs2104286/B rs12722488/B rs3118475/D rs7089861/D rs62626317/D rs41295055/D rs56382813/D 1 0.6 0.6 0.6 0.5 0.5 0.5 0.5 0.5 0.5 0.1 0.1 0.1 0.6 0.6 0.6 0.5 0.5 0.5 0.5 0.5 0.5 0.1 0.1 0.1 1 1 0.9 0.9 0.9 0.7 0.8 0.8 0.3 0.3 0.3 1 0.9 0.9 0.9 0.7 0.8 0.8 0.3 0.2 0.3 0.9 0.9 0.9 0.7 0.8 0.8 0.3 0.3 0.3 1 1 0.8 0.9 0.9 0.3 0.3 0.3 1 0.8 0.9 0.9 0.3 0.3 0.3 0.8 1 1 0.3 0.3 0.3 0.9 0.9 0.3 0.3 0.3 1 0.3 0.3 0.3 0.3 0.3 0.3 1 0.9 0.4 0.2 0.4 0.4 0.3 1 0.4 0.2 0.4 0.4 0.3 0.4 0.2 0.4 0.4 0.3 0.6 1 1 0.8 0.6 0.6 0.7 1 0.8 0.8 a −0.4 −0.3 −0.2 −0.1 0.0 Log OR 0 20 40 60 Fq (%) C C C T C C T T C C T T rs61839660/A rs2104286/B rs41295055/D SNP Major allele a a FALSE TRUE b IL RB rs41294681/A rs28360493/A rs12722563/A rs12722558/A rs12722552/A rs12722522/A rs12722508/A rs7909519/A rs61839660/A rs12722496/A rs12722495/A rs12722504/B rs2104286/B rs12722488/B rs3118475/D rs7089861/D rs62626317/D rs41295055/D rs56382813/D 1 0.6 0.6 0.6 0.5 0.5 0.5 0.5 0.5 0.5 0.1 0.1 0.1 0.6 0.6 0.6 0.5 0.5 0.5 0.5 0.5 0.5 0.1 0.1 0.1 1 1 0.9 0.9 0.9 0.7 0.8 0.8 0.3 0.3 0.3 1 0.9 0.9 0.9 0.7 0.8 0.8 0.3 0.2 0.3 0.9 0.9 0.9 0.7 0.8 0.8 0.3 0.3 0.3 1 1 0.8 0.9 0.9 0.3 0.3 0.3 1 0.8 0.9 0.9 0.3 0.3 0.3 0.8 1 1 0.3 0.3 0.3 0.9 0.9 0.3 0.3 0.3 1 0.3 0.3 0.3 0.3 0.3 0.3 1 0.9 0.4 0.2 0.4 0.4 0.3 1 0.4 0.2 0.4 0.4 0.3 0.4 0.2 0.4 0.4 0.3 0.6 1 1 0.8 0.6 0.6 0.7 1 0.8 0.8 a 7 −0.4 −0.3 −0.2 −0.1 0.0 Log OR 0 20 40 60 Fq (%) C C C T C C T T C C T T rs61839660/A rs2104286/B rs41295055/D SNP Major allele a a FALSE TRUE b b a Major allele a a FALSE TRUE d 1.0 A + D causal 0.5 0.0 log(OR) SNP D log(OR) SNP A –0.5 –1.0 –1.0 –0.5 0.0 0.5 Pr B top 1.00 0.75 0.50 0.25 0.00 1.0 A A Null B A + D Null B A + D Null D B Null A A + D B Null B D Null B A + D Null B Null B A + D D D B A < D A~D A > D Stepwise Stochastic 1 2 3 4 5 2 4 6 2 4 6 2 4 6 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Sample size (1000 s) Propn simulations c D d c Fig. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 The estimated effects for E and H from T1D data are shown by a point, and the simulations from b by < and > for E < H and E > H conditions, respectively. d Haplotype analysis of SNP groups with support in any analysis. Each row represents one SNP, with possible alleles colour coded according to major or minor. Each column is a haplotype—a speciﬁc combination of alleles across all SNPs—with frequency in UK controls and effect on disease risk (log OR + 95% CI). MAF is shown as a percentage on a log scale to allow frequencies of rarer haplotypes to be distinguished. Source data for b are provided in Supplementary Tables 5, 6 9 NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 Three appear protective, carrying the minor allele at either A or D, but only two carry the minor allele at B. c Comparison of stepwise and stochastic search applied to simulated data. Causal variants were simulated as follows: B: single causal variant B, OR = 0.8; A < D causal variants A + D, odds ratios A:0.84, D:0.77; A~D: causal variants A + D, odds ratios A:0.81, D:0:8 (observed in MS data); A > D: causal variants A + D, odds ratios A:0.77, D:0:84. Potential models include A (red), B (green), D (blue), A + D (purple) and null (black); any other models are grouped together as grey. The y-axis shows the proportion of simulations in which the stepwise approach chose the indicated model (adding SNPs while p < 10-6) or the average posterior probabilities for each model for the stochastic search approach. Sample size (x-axis) is the number of cases and controls. d Assuming A and D are causal, this plot shows the probability that B has the smallest p-value as a function of the effect sizes (log odds ratios) at A and D. The estimated effects for A and D from MS data are shown by a point, and the simulations from c by < and > for A < D and A > D conditions respectively. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 ASE results from donors 3 and 4 were consistent with those of D-hets and A + D hets, respectively, shown in Fig. 7b, even though the status of the B SNP was different. These rare donors are consistent with our conclusions that differences seen in IL2RA mRNA expression are controlled by the A and D SNPs, in memory and naive CD4+ T cells, respectively, and argue that the B SNP tags two functionally distinct groups of SNPs, A and D. One key result from our analysis is that sample sizes in the low tens of thousands may still not be large enough to robustly ﬁne- map multiple causal variants. This motivates continued collection of GWAS samples for diseases too infrequent to be found in large numbers in the Biobank style datasets, and greater sharing of data between researchers working on related diseases to better map the most likely genetic causal variants. A particular note of caution is raised by the genomic locations where we ﬁnd discrepancies between stochastic and stepwise results. These are almost entirely those with the strongest biological prior for involvement in these diseases, and also those with typically the strongest effects, and thus greatest power. We question whether these regions are most likely to give rise to discrepancies because they harbour the largest numbers of potential effects or whether, if we had access to much larger datasets, we would see similar discrepancies genome-wide. Our analysis of six diseases reveals several cases where there appear to be multiple functional haplotypes—i.e. more than one IMD causal variant in a region—that affect different diseases differently. Thus, these functional haplotypic maps are essential for designing biological follow-up experiments, for which we need to decide not just what variants to test, but also what variants to hold constant to avoid confounding the effect of the variant of immediate interest. Note that in our ASE work, testing of B heterozygotes, which are in fact a 2:1 mix of D heterozygotes and A heterozygotes, would have resulted in bimodal results in both the memory and naive CD4+ T cells subsets. The ability of sto- chastic search to suggest alternative models provides us the knowledge to compare such models biologically, thereby allowing homogeneous phenotypic groups to emerge that were differently associated with the A and D SNP groups. Discussion Fi i Fine-mapping is a general problem in statistical genetics, important in its own right and for informing integrative down- stream analyses19,29. We have shown that there are candidate causal SNP models for which stepwise regression does not con- verge to the correct solution, even with increasing sample size, and described the constraints on LD that give rise to this joint tagging phenomenon. In principle, exhaustive search could overcome the problem, but scalability is a substantial problem: in a 1000-SNP region: there are a manageable 0.5 million 2-way models but over 166 million 3-way models (and 41.5 billion 4-way models), which cannot be ﬁt in reasonable time for a logistic model that requires optimisation for each model. This exponential growth of model numbers has motivated different approaches to scalable search strategies using linear models and exhaustive search6,30 or specialised search strategies7,8,31. We show that a logistic model stochastic search9 is feasible and does tend to the correct solution as sample sizes increase. However, even stochastic search methods are limited by existing sample sizes when there are multiple causal variants in proximity, and may produce similar results to stepwise methods when sample sizes are insufﬁcient. MFM could be easily adapted as a wrapper around any of the linear model methods above, provided that the linear model is considered an acceptable approximation to a logistic model and that controls are either shared completely or not at all. MFM borrows information across diseases and is thus related to, but distinct from, methods that compare two19 or more32 traits, which integrate over the ﬁne-mapping posteriors of indi- vidual traits, upweighting models that share causal variants, to determine whether there is evidence for sharing. Here, we exploit a prior belief that traits studied are enriched for colocalisation to determine the marginal ﬁne-mapping posterior for each trait, and remove the common colocalisation assumptions of independent datasets and a single causal variant per trait in any region. We also avoid enforcing identity of causal SNPs or their effect sizes between different diseases, as in analysis of an overarching disease phenotype (e.g. autoimmune disease20). It is clear from our results that, causal variants may differ between diseases in the same region and that, even when causal variants are shared, magnitude and direction of effects may differ between diseases. MFM could be applied to other collections of diseases where causal variants may tend to be shared, such as psychiatric NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 Our approach can be expanded in a haplotype-directed manner to other accessible immune cell types to determine cell-speciﬁc and activation- speciﬁc inﬂuences of each disease-associated SNP group (A, C, D, E and F) on IL2RA mRNA expression, enabling a more complete picture of how particular haplotypes mediate protection or sus- ceptibility to disease. The association of the minor alleles of the A haplotype with disease protection for T1D, MS and ATD, but with disease susceptibility for eczema and IBD, could be caused by A-mediated regulation of IL2RA expression in two different cell types: one critical for T1D, MS and ATD disease pathogen- esis, the other type pivotal for eczema and IBD. Alternatively, the genetically-determined level of CD25 on memory CD4 T cells could inﬂuence their likelihood of differentiating into particular types of cytokine-producing effector cells, a phenotype beneﬁcial for some diseases but not others. We propose that, rather than attempting to colocalise eQTL signals and disease associations that are both determined by stepwise analysis35, disease haplotype-directed searches for allele-speciﬁc expression exem- pliﬁed in this study will lead to greater clarity when unraveling cellular mechanisms in immune-based diseases. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 Neither donor 1 or 2 demonstrated an ASE in either the memory or naive T cells, an expected result for donor 1 who does not carry a minor allele at A, D or B, and a result from donor 2 showing that the minor allele at B is not associated with either phenotype. ASE results from donors 3 and 4 were consistent with those of D-hets and A + D hets, respectively, shown in Fig. 7b, even though the status of the B SNP was different. These rare donors are consistent with our conclusions that differences seen in IL2RA mRNA expression are controlled by the A and D SNPs, in memory and naive CD4+ T cells, respectively, and argue that the B SNP tags two functionally distinct groups of SNPs, A and D. diseases33 or metabolic-related traits34 if appropriate priors can be elicited for each collection. This might be possible from prior work as here, but if not we recommend that a range of plausible values be considered, with robust results identiﬁed as those which remain similar under different priors. expression in A haplotype donors. However, it is not signiﬁcant, and we did not observe an increase in CD25+ naive T cells associated with the MS-protective A allele in a previous study24. p p yp , g , and we did not observe an increase in CD25+ naive T cells associated with the MS-protective A allele in a previous study24. Additionally, we identiﬁed four individuals, three of whom carry rare IL2RA haplotypes (Fig. 7c): donor 1 carries a common haplotype combination that is homozygous across A, B, D; donor 2 carries the minor allele at B in the absence of a minor allele at either A or D, donor 3 carries a minor allele at D but not B, and donor 4 also carries a minor allele at D but not at B on one haplotype and minor alleles at A and B on the other haplotype (Fig. 7a). Neither donor 1 or 2 demonstrated an ASE in either the memory or naive T cells, an expected result for donor 1 who does not carry a minor allele at A, D or B, and a result from donor 2 showing that the minor allele at B is not associated with either phenotype. NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunication NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 The pattern of LD in the region means that the large majority of A-het and D-het individuals are also hetero- zygous at the B SNP and A + D-het individuals are homozygous at the B SNP (Fig. 7a, Supplementary Data 12), allowing us to directly compare the effects of SNPs in groups A, B and D. the lack of allelic imbalance in memory T cells from D-het individuals who are heterozygous at B. In naive CD4+ T cells, D-het as well as A + D-het heterozygotes had an allelic imbalance with the protective D haplotype producing less IL2RA mRNA than the susceptible or protective A haplotypes, conﬁrming our previous observations of decreased CD25+ naive CD4+ T cells associated with donors having the protective D haplotype9. Again, this is inconsistent with B causality, since only D-het and not A + D-het individuals are heterozygous at B. In A-het donors there appears to be an allelic imbalance in naive CD4+ T cells favouring the MS-protective versus susceptible haplotype, which is the opposite direction to that observed with protection at D and could reﬂect an anticipatory differentiation of naive T cells toward the memory lineage and its phenotype of increased CD25 In memory CD4+ T cells, A-het and A + D-het individuals showed an allelic imbalance with the MS-protective A haplotype producing more IL2RA mRNA, inconsistent with B causing the imbalanced expression since A + D-het individuals tested are homozygous for B (Fig. 7b). Also inconsistent with B causality is 10 ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 expression in A haplotype donors. However, it is not signiﬁcant, and we did not observe an increase in CD25+ naive T cells associated with the MS-protective A allele in a previous study24. Additionally, we identiﬁed four individuals, three of whom carry rare IL2RA haplotypes (Fig. 7c): donor 1 carries a common haplotype combination that is homozygous across A, B, D; donor 2 carries the minor allele at B in the absence of a minor allele at either A or D, donor 3 carries a minor allele at D but not B, and donor 4 also carries a minor allele at D but not at B on one haplotype and minor alleles at A and B on the other haplotype (Fig. 7a). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 Source data are provided in Supplementary Data 12 Fig. 7 Allele-speciﬁc expression analysis of IL2RA. This shows that there are two phenotypes that map to the A and D SNP groups and not the B group, providing functional evidence that the stochastic search better explains the genetic association than stepwise. a Schematic of donor IL2RA genotypes used in allele-speciﬁc expression studies. As the minor alleles for both A and D each usually co-occur with the minor B allele, in A-het and D-het individuals, the B SNP is heterozygous but in A + D-het individuals, the B SNP is homozygous. There are rare exceptions as seen in donors 3 and 4. b Allele-speciﬁc expression of IL2RA in CD4+ central memory T cells and CD4+ naive T cells in A-het, D-het and A + D het donors. The allelic ratios (top:bottom haplotypes shown in panel a) are calculated from 3–4 replicates per individual. P-values were calculated by t-test of allelic ratio in each group compared to gDNA. Donors that contributed to panel c are indicated by a triangle. c Donors with rare IL2RA haplotypes conﬁrm that the B SNP does not explain genotype-phenotype expression of IL2RA. As there is only one donor per genotype we cannot perform statistical testing. We show each of the four replicate allelic ratios (ratio order as in panel b) to indicate the variability of the assay. The bounds of each box correspond to the ﬁrst and third quartiles, and the whiskers extend to the smallest/largest value at most 1.5IQR from the bounds; IQR is the distance between the ﬁrst and third quartiles. Data beyond the whisker ends are plotted individually. Source data are provided in Supplementary Data 12 The SNPs belonging to the above-mentioned groups, as well as the lead SNPs for autoimmune thyroid disease (ATD; rs706799), alopecia areata (AA; rs3118470), rheumatoid arthritis (RA; rs10795791) and ulcerative colitis (UC; rs4147359) were extracted from the generated data for analyses via stepwise regression and stochastic search; the lead SNP for multiple sclerosis forms group B. This extraction was done for computational efﬁciency, and is based on the previous analysis of MS and T1D that identiﬁed these SNP groups as contributing the majority of the posterior probability9. All other SNPs contribute negligible posterior probability and we assume this in the simulations. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 A het A B C D E F D het A B C D E F A + D het A B C D E F Donor 1 A B C D E F Donor 2 A B C D E F Donor 3 A B C D E F Donor 4 A B C D E F Legend Non-reference allele Either allele Reference allele a p = 2.4 × 10−7 gDNA Central memory T cells Naive T cells A het D het A + D het A het D het A + D het 0.5 1.0 1.5 2.0 Genotype group Allelic ratio b gDNA Central memory T cells Naive T cells Dnr 1 Dnr 2 Dnr 3 Dnr 4 Dnr 1 Dnr 2 Dnr 3 Dnr 4 0.5 1.0 1.5 2.0 Genotype group Allelic ratio c p = 8.8 × 10−5 p = 1 × 10−4 p = 0.0099 p = 0.037 p = 0.036 Fig. 7 Allele-speciﬁc expression analysis of IL2RA. This shows that there are two phenotypes that map to the A and D SNP groups and not the B grou providing functional evidence that the stochastic search better explains the genetic association than stepwise. a Schematic of donor IL2RA genotypes use n allele-speciﬁc expression studies. As the minor alleles for both A and D each usually co-occur with the minor B allele, in A-het and D-het individuals, th B SNP is heterozygous but in A + D-het individuals, the B SNP is homozygous. There are rare exceptions as seen in donors 3 and 4. b Allele-speciﬁc expression of IL2RA in CD4+ central memory T cells and CD4+ naive T cells in A-het, D-het and A + D het donors. The allelic ratios (top:bottom haplotypes shown in panel a) are calculated from 3–4 replicates per individual. Methods Si l ti Simulations—single trait. Simulations were carried out under a realistic scenario that mimics the MAF and r2 in the IL2RA region. We simulated haplotypes for 345 SNPs in chromosome 10p-6030000-6220000 (GRCh37/hg19), based on the CEU 1000 Genomes Phase 3 data36 (ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/ 20130502/) using HapGen237. Code to perform the simulations can be found in https://github.com/jennasimit/MFMextra. Causal variants were selected within SNP groups for each disease model (see Supplementary Data 11) with various OR relating the odds of disease in heterozygote carriers of the non-reference allele compared to the homozygote reference allele. We assumed a multiplicative model throughout. Simulations—single trait. Simulations were carried out under a realistic scenario that mimics the MAF and r2 in the IL2RA region. We simulated haplotypes for 345 SNPs in chromosome 10p-6030000-6220000 (GRCh37/hg19), based on the CEU 1000 Genomes Phase 3 data36 (ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/ 20130502/) using HapGen237. Code to perform the simulations can be found in https://github.com/jennasimit/MFMextra. Causal variants were selected within SNP groups for each disease model (see Supplementary Data 11) with various OR relating the odds of disease in heterozygote carriers of the non-reference allele compared to the homozygote reference allele. We assumed a multiplicative model throughout. TURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications 11 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 P-values were calculated by t-test of allelic ratio in each group compared et 2 p = 2.4 × 10−7 gDNA Central memory T cells Naive T cells A het D het A + D het A het D het A + D het 0.5 1.0 1.5 2.0 Genotype group Allelic ratio b p = 8.8 × 10−5 p = 1 × 10−4 p = 0.0099 p = 0.037 p = 0.036 b G yp g p gDNA Central memory T cells Naive T cells Dnr 1 Dnr 2 Dnr 3 Dnr 4 Dnr 1 Dnr 2 Dnr 3 Dnr 4 0.5 1.0 1.5 2.0 Genotype group Allelic ratio c c Dnr 1 Dnr 2 Dnr 3 Dnr 4 Dnr 1 Dnr 2 Dnr 3 Dnr 4 Genotype group Fig. 7 Allele-speciﬁc expression analysis of IL2RA. This shows that there are two phenotypes that map to the A and D SNP groups and not the B group, providing functional evidence that the stochastic search better explains the genetic association than stepwise. a Schematic of donor IL2RA genotypes used in allele-speciﬁc expression studies. As the minor alleles for both A and D each usually co-occur with the minor B allele, in A-het and D-het individuals, the B SNP is heterozygous but in A + D-het individuals, the B SNP is homozygous. There are rare exceptions as seen in donors 3 and 4. b Allele-speciﬁc expression of IL2RA in CD4+ central memory T cells and CD4+ naive T cells in A-het, D-het and A + D het donors. The allelic ratios (top:bottom haplotypes shown in panel a) are calculated from 3–4 replicates per individual. P-values were calculated by t-test of allelic ratio in each group compared to gDNA. Donors that contributed to panel c are indicated by a triangle. c Donors with rare IL2RA haplotypes conﬁrm that the B SNP does not explain genotype-phenotype expression of IL2RA. As there is only one donor per genotype we cannot perform statistical testing. We show each of the four replicate allelic ratios (ratio order as in panel b) to indicate the variability of the assay. The bounds of each box correspond to the ﬁrst and third quartiles, and the whiskers extend to the smallest/largest value at most 1.5IQR from the bounds; IQR is the distance between the ﬁrst and third quartiles. Data beyond the whisker ends are plotted individually. 3. Keep the ﬁrst Nk rows from Gk, k = 0, 1, 2. We focused on a target odds (TO) of 1, such that there is an equal probability of sharing to non-sharing. Results for a range of TO from 9 (no sharing more likely than sharing of causal variants) to 0.35 (sharing more likely than distinct causal variants) are in Supplementary Data 5–8. Mathematical predictions of minimum univariate p-value. We used sunbeam plots to characterise how changing the odds ratio of two causal SNPs in a model can change the probability that a third variant will have the minimum p-value (and hence be selected ﬁrst in any stepwise ﬁne-mapping algorithm). We utilised components of the simGWAS package (http://github.com/chr1swallace/ simGWAS) to calculate expected GWAS Z scores for any given set of causal variants and their effect sizes, across those causal variants and their neighbouring SNPs38. We considered the behaviour of Z scores at each of two nominated causal variants (following Fig. 1, let us refer to these variants as A and C) with a third SNP, not itself causal, but potentially correlated with both A and C (in Fig. 1, this is SNP J). For each of a range of possible odds ratios, we computed which of the three SNPs had the smallest expected p-value, and coloured that square of the grid correspondingly. When the log odds ratios of both A and C were close to 0, then no SNP had a low p-value and it was not possible to ﬁnd signiﬁcant evidence of disease association in the region. This section of the grid was coloured white. Superimposed upon the grid is a point corresponding to the odds ratio we computed for A and C from the real dataset. Code to produce these plots is at https://github.com/chr1swallace/MFM-paper/tree/master/sunbeams. Haplotype analyses. Haplotype analyses were performed by ﬁrst phasing the genotypes across selected SNPs using an E–M algorithm and selecting 10 multiply imputed samples from the posterior (snphap, https://github.com/chr1swallace/ snphap). These samples were analysed in parallel and results combined using standard multiple imputation functions in the R package MICE41. Code to implement these steps is available at https://github.com/chr1swallace/snpHaps. All analyses included the ﬁrst four PCs, and country as an additional covariate for iCEL and iRA to account for population structure. Allele-speciﬁc expression. Samples were obtained from the Cambridge BioR- esource (www.cambridgebioresource.org.uk) as part of the ‘Genes and Mechanisms of Type 1 Diabetes’ study and were of self-reported white ethnicity. 3. Keep the ﬁrst Nk rows from Gk, k = 0, 1, 2. We simulated either shared conﬁgurations where each disease was under the inﬂuence of two causal variants, one shared between diseases (A) and one unique to each disease (one from C, one from D); or independent conﬁgurations, where the two diseases were under the inﬂuence of distinct causal variants (one from each of A and D for one disease and one from C for the other disease) or one disease had no associations in the region (one from each of A and D for one disease and none for the other disease). All causal variants were assigned an odds ratio of 1.25 or 1.4. For both diseases, equal-sized case-control samples consisting of N cases and N controls were considered for N ranging from 1000 to 5000; each simulation setting had 100 replications. Code to perform these steps is available at https://github.com/chr1swallace/ MFM-analysis. Code to perform these steps is available at https://github.com/chr1swallace/ MFM-analysis. SNP grouping. SNPs with marginal posterior probability of inclusion >0.001 were grouped according to criteria of substitutability—that is, any single SNP in a group can be used in place of any other to give similar model ﬁts (assessed by Bayes factors) and no more than 1 SNP in the same group is needed to deﬁne any models with even modest posterior support. We reasoned that this meant SNPs would need to be in LD—high r2—and rarely selected together in models— i.e. model selection correlation (rmodel) should be negative; both rmodel and r2 are used so that our SNP grouping is informed by both model posteriors and LD. We hierarchically cluster SNPs within each disease according to r2 × sign(rmodel) using complete linkage, and group SNPs by cutting the tree such that all SNPs within a group must have pairwise r2 > 0.5, pairwise rmodel < 0, and marginal posterior probability that both are included in a model was <0.01. We then identify overlapping groups deﬁned in different diseases, and merge or split groups when they meet this criteria. The speciﬁc algorithm is deﬁned in the group.multi function in https://github.com/chr1swallace/GUESSFM/blob/ master/R/groups.R. We compared the independent stochastic search analyses of each disease with the multinomial approach with upweighted sharing based on a range of target odds (i.e. prior odds of no sharing of causal variants between one disease and any other disease). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 (RA)16 and autoimmune thyroid disease (ATD)13 (Supplementary Table 1). All genome coordinates are from build GRCh37. (RA)16 and autoimmune thyroid disease (ATD)13 (Supplementary Table 1). All genome coordinates are from build GRCh37. https://github.com/jennasimit/MFMextra. First we used HapGen2 to generate a population of 100,000 individuals based on the CEU 1000 Genomes Phase 3 data. Causal variants for each trait were randomly selected within particular SNP groups for a certain disease model (see Supplementary Data 11); when the same SNP group contained a causal variant for both diseases, one variant was selected from the group and set as causal for both diseases. This reﬂects our belief that if causal variants for two diseases are known to belong to the same small SNP group, it is likely that the same SNP is causal for both diseases rather than different SNPs in the same high-LD group. https://github.com/jennasimit/MFMextra. First we used HapGen2 to generate a population of 100,000 individuals based on the CEU 1000 Genomes Phase 3 data. Causal variants for each trait were randomly selected within particular SNP groups To ensure controls could be combined across datasets, we restricted analysis for the multinomial model to UK samples, and used principal component analysis including 1000 Genomes data to exclude two individuals who fell outside individual country clusters. Genotypes were compared between datasets to ensure exclusion of duplicate samples. Data were split into subsets according to the densely genotyped regions targeted by the ImmunoChip (Supplementary Data 1) and imputed to 1000 Genomes phase 336 using SHAPEIT39 and IMPUTE240. Phased reference data were downloaded from https://mathgen.stats.ox.ac.uk/ impute/1000GP_Phase3.html. Country and the ﬁrst four principal components were included as covariates in all regressions to account for population structure. SNPs were excluded if they had info scores < 0.3, certainty < 0.98, \|Z\| for HWE > 4 in UK controls, MAF < 0.5% in UK controls, call rate < 0.99 in any case or control group, or an absolute difference in certain genotype call rates between controls and any case group of >5%. for a certain disease model (see Supplementary Data 11); when the same SNP group contained a causal variant for both diseases, one variant was selected from the group and set as causal for both diseases. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 Set j = 1 and repeat the following steps while n0 < N0 or n1 < N1 or n2 < N2. k 2. Set j = 1 and repeat the following steps while n0 < N0 or n1 < N1 or n2 < N2. Stochastic search ﬁne-mapping of single diseases was performed using GUESSFM (http://github.com/chr1swallace/GUESSFM). Initial searches were performed after tagging at r2 < 0.99 with an optimistic binomial prior for the number of causal variants per region with expectation set at 3 to allow good mixing of the chains. Reanalysis of the expanded tag sets for SNPs in models included in the model set with total posterior probability 0.99 was performed using approximate Bayes factors and the more conservative prior expectation of two causal variants per region using GUESSFM. GUESSFM results were combined using the methods proposed in this paper (details in Supplementary Notes 3), as implemented in the R package MFM (http://github.com/jennasimit/MFM). We set the prior odds that two diseases shared any causal variants to 1 (i.e. a 50% probability that they share none). For a number of diseases, d > 2, we set the prior that the diseases share no causal variants to 0:5 ﬃﬃﬃﬃﬃﬃ d1 p , where the exponent is the geometric mean of the exponents in the (nonsensical) extremes 0.5d-1, which assumes all diseases are independent and 0.5 which assumes all diseases are completely dependent. a. For k = 1, 2 determine pjk ¼ logit1ðβ0 þ P 1 i¼1 βikxijÞ and generate uniform random variables u1, u2. b. If u1 > p1 and u1 > p2, then n0 = n0 + 1, append gj to G0, j = j+ 1, and go to beginning of step 2. c. Else, if u1 ≤p1 and u1 > p2, then n1 = n1 + 1, append gj to G1, j=j=+ 1, and go to beginning of step 2. d. Else, if u1 ≤p2 and u1 > p1, then n2 = n2 + 1, append gj to G2, j = j + 1, and go to beginning of step 2. g g g p e. Else, if u2 < 0.5 and n1 < N1, then n1 = n1 + 1and append gj to G1, j = j + 1. Otherwise, n2 = n2 + 1and append gj to G2, j = j + 1. Go to beginning of step 2. ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 This reﬂects our belief that if causal variants for two diseases are known to belong to the same small SNP group, it is likely that the same SNP is causal for both diseases rather than different SNPs in the same high-LD group. g g p Logistic regression models with the selected causal variants and odds ratios (OR) were then used to assign each individual as either a member of the controls, disease 1 cases, or disease 2 cases until the desired number of individuals in each group was attained; let ORjk be the odds ratio for causal variant j and disease k. The prevalence for both diseases was set to 0.1, as our purpose is to generate cases and controls for method comparison. In particular, the following steps were used to ascertain control/disease 1/disease 2 status, where xij is the number of non- reference alleles of variant i for individual j (i.e. genotype score), gj is the vector of genotype scores for individual j, β0 = log(0.1), βik = log(ORik) is the effect of causal variant i for disease k, and m1 is the number of causal variants. Forward stepwise regression was performed using univariate logistic regressions across all SNPs in the region. The SNP with the strongest association (smallest p-value) was selected, then all two-SNP models containing the selected SNP and any other SNP were considered, and the process repeated until no SNP could be added with a marginal p < 10−6, a less stringent threshold (than the conventional 5 × 10−8), chosen so that less than genome-wide signiﬁcant associations can be considered when other diseases are also associated, in which case borrowing power may be informative. 1. Let nk be the number of individuals ascertained to group k (controls are group 0, groups 1 and 2 consist of members with disease 1 and 2, respectively) and Gk be the matrix of genotype scores for individuals in group k. Initialise nk = 0 and Gk as a null vector. 1. Let nk be the number of individuals ascertained to group k (controls are group 0, groups 1 and 2 consist of members with disease 1 and 2, respectively) and Gk be the matrix of genotype scores for individuals in group k. Initialise nk = 0 and Gk as a null vector. k 2. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 The total number of SNPs in the region is not disregarded and is used in the prior probability calculations for the SNPs that are analysed. For each replication a stepwise regression model was ﬁt, adding SNPs to the model using a p-value threshold of 1 × 10−6. To generate stochastic search results, we used GUESSFM9, setting a prior of three causal variants for the region to encourage good mixing of the chains in the initial Bayesian variable selection, and setting the prior to a more conservative two causal variants per region to obtain ﬁnal model posterior probabilities (PP). Model ﬁts were summarised by the proportion of times each model was selected via stepwise regression or the mean of the GUESSFM posterior probabilities for each model. Simulations—multiple traits. We adapted the HapGen2 simulation outlined above to simulate datasets for two case and one control set; code is available in 12 NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 CD45RA+ and CD27+, whereas CD4+ central memory T cells were sorted as CD3+ CD4+ CD8−CD127med/high CD25low-med CD45RA−and CD27+. CD45RA+ and CD27+, whereas CD4+ central memory T cells were sorted as CD3+ CD4+ CD8−CD127med/high CD25low-med CD45RA−and CD27+. 6. Chen, W. et al. Fine mapping causal variants with an approximate bayesian method using marginal test statistics. Genetics 200, 719–736 (2015). To phase the direction of effect from the four donors carrying rare IL2RA haplotypes (Fig. 7a, c), their haplotypes were compared to those found in the 1000 Genome Project CEU data to assess the allele frequency of the ASE readout SNP (rs12244380, A or G), to predict which allele is most likely to be carried. For donor 1, the E haplotype carries the G allele with frequency 73% whereas the susceptible haplotype carries the A allele 60% of the time. For donor 2, it is most likely the B and E alleles are on the same haplotype (20 examples where they are together vs four examples where they are on different chromosomes), and here the B + E haplotype carries the A allele of rs12244380 (100%). For donor 3, all examples of the D haplotype lacking the B allele carry the A allele of rs12244380 (14/14), whereas the E haplotype carries the G allele of rs12244380 73% of the time. Lastly, for donor 4, the A haplotype carries the G allele of rs12244380 88% of the time, and for all examples of the D haplotype lacking B carries the A allele of rs12244380 (7/7). Where multiple assays were performed on the same donor, we retained those with the smallest standard deviation of allelic ratios but show both results in Supplementary Data 12 g g 7. Benner, C. et al. FINEMAP: efﬁcient variable selection using summary data from genome-wide association studies. Bioinformatics 32, 1493–1501 (2016). 8. Newcombe, P. J., Conti, D. V. & Richardson, S. JAM: a scalable bayesian framework for joint analysis of marginal SNP effects. Genet. Epidemiol. 40, 188–201 (2016). 9. Wallace, C. et al. Dissection of a complex disease susceptibility region using a bayesian stochastic search approach to ﬁne mapping. PLoS Genet. 11, e1005272 (2015). 10. Pasaniuc, B. & Price, A. L. Dissecting the genetics of complex traits using summary association statistics. Nat. Rev. Genet. 18, 117–127 (2017). 11. Onengut-Gumuscu, S. et al. Data availability 18. Guo, H. et al. Integration of disease association and eQTL data using a Bayesian colocalisation approach highlights six candidate causal genes in immune-mediated diseases. Hum. Mol. Genet. 24, 3305–3313 (2015). y Complete results from our analyses are available at https://chr1swallace.github.io/MFM- output/index.html and have been deposited at ﬁgshare under https://doi.org/10.6084/m9. ﬁgshare.8289677. Data were obtained from the study authors for each of the six autoimmune diseases that we analysed. Original genotype data may be requested from the original study authors: ATD ImmunoChip, Cooper et al. (https://www.ncbi.nlm.nih.gov/ pubmed/22922229); RA ImmunoChip, Eyre et al. (https://www.ncbi.nlm.nih.gov/pmc/ articles/PMC3882906/); JIA ImmunoChip, Hinks et al. (https://www.ncbi.nlm.nih.gov/ pubmed/23603761). MS ImmunoChip data were accessed through application to the International Multiple Sclerosis Genetic Consortium (IMSGC; http://www.imsgenetics.org/). Primary analysis of the MS data is presented by IMSGC (https://www.ncbi.nlm.nih.gov/ pmc/articles/PMC3832895/). The primary analysis of the Celiac ImmunoChip is by Trynka et al. (https://www.nature.com/articles/ng.998) and the genotype data are hosted by the European Bioinformatics Institute, under accession number EGAS00000000053. T1D ImmunoChip data are available from dbGaP (Study Accession: phs000180.v3.p2) and 2000 T1D samples were genotyped as part of the WTCCC (and controls) - data access is described at https://www.wtccc.org.uk/info/access_to_data_samples.html. 19. Giambartolomei, C. et al. Bayesian test for colocalisation between pairs of genetic association studies using summary statistics. PLoS Genet. 10, e1004383 (2014). 20. Westra, H.-J. et al. Fine-mapping identiﬁes causal variants for RA and T1D in DNASE1L3, SIRPG, MEG3, TNFAIP3 and CD28/CTLA4 loci. bioRxiv 151423 (2017). 21. International Multiple Sclerosis Genetics Consorti et al. The Multiple Sclerosis Genomic Map: Role of peripheral immune cells and resident microglia in susceptibility. bioRxiv 143933. https://doi.org/10.1101/143933 (2017). p y p g 22. Farh, K. K.-H. et al. Genetic and epigenetic ﬁne mapping of autoimmune disease variants. Nature 518, 337–343 (2015). 23. Jin, Y. et al. Genome-wide association analyses identify 13 new susceptibility loci for generalized vitiligo. Nat. Genet. 44, 676–680 (2012). d l ll ﬁ h f h loci for generalized vitiligo. Nat. Genet. 44, 676–680 (2012). 24. Dendrou, C. A. et al. Cell-speciﬁc protein phenotypes for the autoimmune locus IL2RA using a genotype-selectable human bioresource. Nat. Genet. 41, 1011–1015 (2009). URLs 13. Cooper, J. D. et al. Seven newly identiﬁed loci for autoimmune thyroid disease. Hum. Mol. Genet 21, 5202–5208 (2012). For Global Biobank Engine, Stanford, CA, see http://gbe.stanford. edu/ [accessed January 2018]. For Global Biobank Engine, Stanford, CA, see http://gbe.stanford. edu/ [accessed January 2018]. 14. Trynka, G. et al. Dense genotyping identiﬁes and localizes multiple common and rare variant association signals in celiac disease. Nat. Genet. 43, 1193–1201 (2011). For Extended information in searchable format, see https:// chr1swallace.github.io/MFM-output. 15. Hinks, A. et al. Dense genotyping of immune-related disease regions identiﬁes 14 new susceptibility loci for juvenile idiopathic arthritis. Nat. Publ. Group 45, 664–669 (2013). Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article. 16. Eyre, S. et al. High-density genetic mapping identiﬁes new susceptibility loci for rheumatoid arthritis. Nat. Genet. 44, 1336–1340 (2012). 17. Hormozdiari, F., Kostem, E., Kang, E. Y., Pasaniuc, B. & Eskin, E. Identifying causal variants at loci with multiple signals of association. Genetics 198, 497–508 (2014). ARTICLE Fine mapping of type 1 diabetes susceptibility loci and evidence for colocalization of causal variants with lymphoid gene enhancers. Nat. Genet. 47, 381–386 (2015). 12. International Multiple Sclerosis Genetics Consortium (IMSGC). et al. Analysis of immune-related loci identiﬁes 48 new susceptibility variants for multiple sclerosis. Nat. Genet. 45, 1353–1360 (2013). Code availability 25. Garg, G. et al. Type 1 diabetes-associated IL2RA variation lowers IL-2 signaling and contributes to diminished CD4+ CD25+ regulatory T cell function. J. Immunol. 188, 4644–4653 (2012). All code used is freely available as R libraries and R scripts. Multinomial Fine-mapping (MFM) software is available at https://jennasimit.github.io/MFM. Custom code for our analyses is available at https://github.com/chr1swallace/MFM-analysis and https://github. com/chr1swallace/MFM-paper. Software for simulations to evaluate MFM is available at https://jennasimit.github.io/MFMextra/. 26. Rainbow, D. B. et al. A rare IL2RA haplotype identiﬁes SNP rs61839660 as causal for autoimmunity. bioRxiv 108126 (2017). 27. Huang, H. et al. Fine-mapping inﬂammatory bowel disease loci to single- variant resolution. Nature 547, 173–178 (2017). Received: 26 February 2019 Accepted: 2 July 2019 Received: 26 February 2019 Accepted: 2 July 2019 Received: 26 February 2019 Accepted: 2 July 2019 28. Ferreira, M. A. et al. Shared genetic origin of asthma, hay fever and eczema elucidates allergic disease biology. Nat. Genet. 49, 1752–1757 (2017). 29. Burren, O. S. et al. Chromosome contacts in activated T cells identify autoimmune disease candidate genes. Genome Biol. 18, 165 (2017). g 30. Guan, Y. & Stephens, M. Practical issues in imputation-based association mapping. PLoS Genet. 4, e1000279 (2008). 30. Guan, Y. & Stephens, M. Practical issues in mapping. PLoS Genet. 4, e1000279 (2008). 3. Keep the ﬁrst Nk rows from Gk, k = 0, 1, 2. Informed consent was obtained from all volunteers for the collection and use of the per- ipheral blood samples. The NHS Cambridgeshire Research Ethics committee approved this work involving human participants for allele-speciﬁc expression assays. Data and samples were treated anonymously and conﬁdentially. Allele-speciﬁc expression analysis was performed as described in Burren et al., 201729 but modiﬁed to start with sorted CD4+ naive and central memory T cells. CD4+ naive T cells were sorted as CD3+ CD4+ CD8−CD127med/high CD25low-med Fine-mapping analyses of ImmunoChip-genotyped diseases. We collated individual genotype data generated using the ImmunoChip for a total of 61,641 individuals, formed of controls and six disease cohorts: MS (UK subset)12, T1D11, juvenile idopathic arthritis (JIA, UK subset)15, celiac disease14, rheumatoid arthritis 13 NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunication TURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications Received: 26 February 2019 Accepted: 2 July 2019 Author contributions Sparks Childhood Arthritis Response to Medication Study was funded by Sparks, UK (08ICH09) and the Big Lottery Fund, UK (RG/1/010135231). MS—US National Institutes of Health, the Wellcome Trust, the UK MS Society, the UK Medical Research Council, the US National MS Society, the Cambridge National Institute for Health Research (NIHR) Biomedical Research Centre, DeNDRon, the Bibbi and Niels Jensens Foundation, the Swedish Brain Foundation, the Swedish Research Council, the Knut and Alice Wallenberg Foundation, the Swedish Heart–Lung Foundation, the Foundation for Strategic Research, the J.A. developed the MFM method, performed statistical analyses and interpreted results, wrote the paper. D.R. performed ASE assays and analysis and interpreted results, wrote the paper. M.F. performed mathematical analyses and conceived the sunbeam plots. N.G. performed mathematical analyses and veriﬁed the analytical methods. L.W. supervised ASE work, interpreted the results of ASE analyses and all the ﬁne-mapping analyses, wrote the paper. C.W. conceived the study, developed the MFM method, performed statistical analyses and interpreted results, wrote the paper. All authors read and agreed the manuscript. Howard Martin, Fay Rodger and Ruth Littleboy for running the Illumina MiSeq in the Stratiﬁed Medicine Core Laboratory NGS hub, Department of Medical Genetics, Cam- bridge University, Addenbrooke’s Hospital, Cambridge. This research was supported by the Cambridge NIHR BRC Cell Phenotyping Hub. In particular, we wish to thank Anna Petrunkina Harrison, Simon McCallum, Christopher Bowman, Natalia Savinykh, Esther Perez and Jelena Markovic Djuric for their advice and support in cell sorting. We also thank Helen Stevens, Pamela Clarke, Gillian Coleman, Sarah Dawson, Jennifer Denesha, Simon Duley, Meeta Maisuria-Armer and Trupti Mistry for acquisition and preparation of samples. We thank Stephen Eyre for helpful comments and disease investigators ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-11271-0 Stockholm County Council, Karolinska Institutet, INSERM, Fondation d’Aide pour la Recherche sur la Sclrose en Plaques, Association Franaise contre les Myopathies, Infra- strutures en Biologie Sant et Agronomie (GIS–IBISA), the German Ministry for Edu- cation and Research, the German Competence Network MS, Deutsche Forschungsgemeinschaft, Munich Biotec Cluster M4, the Fidelity Biosciences Research Initiative, Research Foundation Flanders, Research Fund KU Leuven, the Belgian Charcot Foundation, Gemeinntzige Hertie Stiftung, University Zurich, the Danish MS Society, the Danish Council for Strategic Research, the Academy of Finland, the Sigrid Juselius Foundation, Helsinki University, the Italian MS Foundation, Fondazione Car- iplo, the Italian Ministry of University and Research, the Torino Savings Bank Foun- dation, the Italian Ministry of Health, the Italian Institute of Experimental Neurology, the MS Association of Oslo, the Norwegian Research Council, the South–Eastern Norwegian Health Authorities, the Australian National Health and Medical Research Council, the Dutch MS Foundation and Kaiser Permanente. RA—Arthritis Foundation, the US National Institutes of Health, and Arthritis Research UK (21754). T1D—National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), National Institute of Allergy and Infectious Diseases (NIAID), National Human Genome Research Institute (NHGRI), National Institute of Child Health and Human Development (NICHD), and the JDRF and supported by U01 DK062418. 37. Su, Z., Marchini, J. & Donnelly, P. HAPGEN2: simulation of multiple disease SNPs. Bioinformatics 27, 2304–2305 (2011). 38. Fortune, M. & Wallace, C. simGWAS: a fast method for simulation of large scale case-control GWAS summary statistics. Bioinformatics https://doi.org/ 10.1093/bioinformatics/bty898 (2018). y 39. Delaneau, O., Zagury, J.-F. & Marchini, J. Improved whole-chromosome phasing for disease and population genetic studies. Nat. Methods 10, 5–6 (2013). 40. Howie, B. N., Donnelly, P. & Marchini, J. A ﬂexible and accurate genotype imputation method for the next generation of genome-wide association studies. PLoS Genet. 5, e1000529 (2009). 41. van Buuren, S. & Groothuis-Oudshoorn, K. mice: multivariate imputation by chained equations in R. J. Stat. Softw., Artic. 45, 1–67 (2011). Acknowledgements The authors would like to thank the Rivas lab for making the Global Biobank Engine resource available. We gratefully acknowledge the participation of all NIHR Cambridge BioResource volunteers. We thank the Cambridge BioResource staff for their help with volunteer recruitment. We thank members of the Cambridge BioResource SAB and Management Committee for their support of our study and the National Institute for Health Research Cambridge Biomedical Research Centre for funding. Access to Cam- bridge BioResource volunteers and to their data and samples are governed by the Cambridge BioResource SAB. Documents describing access arrangements and contact details are available at www.cambridgebioresource.org.uk. We thank Graeme Clark, Additional information Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- 019-11271-0. Competing interests: The authors declare no competing interests. Author contributions Health Research Cambridge Biomedical Research Centre for funding. Access to Cam- bridge BioResource volunteers and to their data and samples are governed by the Cambridge BioResource SAB. Documents describing access arrangements and contact details are available at www.cambridgebioresource.org.uk. We thank Graeme Clark, Howard Martin, Fay Rodger and Ruth Littleboy for running the Illumina MiSeq in the Stratiﬁed Medicine Core Laboratory NGS hub, Department of Medical Genetics, Cam- bridge University, Addenbrooke’s Hospital, Cambridge. This research was supported by the Cambridge NIHR BRC Cell Phenotyping Hub. In particular, we wish to thank Anna Petrunkina Harrison, Simon McCallum, Christopher Bowman, Natalia Savinykh, Esther Perez and Jelena Markovic Djuric for their advice and support in cell sorting. We also thank Helen Stevens, Pamela Clarke, Gillian Coleman, Sarah Dawson, Jennifer Denesha, Simon Duley, Meeta Maisuria-Armer and Trupti Mistry for acquisition and preparation of samples. We thank Stephen Eyre for helpful comments and disease investigators Stephen Eyre, Wendy Thomson, Gosia Trynka, David van Heel, Steve Rich and the Type 1 Diabetes Genetics Consortium, Stephen Sawcer and the International Multiple Sclerosis Genetics Consortium, Matthew Simmonds, Stephen Gough, Jayne Franklyn, and Oliver Brand for sharing genotype data. This work was supported by the MRC (MC UU 00002/ 4, MR/R021368/1), the Wellcome Trust (096388, 099772, 107212, 107881), the JDRF (9- 2011-253 and 5-SRA-2015-130-A-N), the National Institute for Health Research Cam- bridge Biomedical Research Centre (BRC), and Dementia Platforms UK. Funding of collection and genotyping of samples: Controls—British 1958 Birth Cohort (MRC grant G0000934, Wellcome Trust (grant 068545/Z/02). DNA control samples were prepared and provided by S. Ring, R. Jones, M. Pembrey, W. McArdle, D. Strachan and P. Burton. UK Blood Services collection of common controls (UKBS-CC collection)—funded by Wellcome Trust (076113/C/04/Z) and by a National Institute for Health Research pro- gram grant to National Health Service Blood and Transplant (RP-PG-0310-1002). Genotyping of control samples was supported, in part, by grants from Juvenile Diabetes Research Foundation International (JDRF) and the US NIH (U01 DK062418). ATD— Wellcome Trust Celiac disease—Wellcome Trust, by grants from the Celiac Disease Consortium and an Innovative Cluster approved by the Netherlands Genomics Initiative, by the Dutch Government (BSIK03009 to C. Wijmenga) and the Netherlands Organi- sation for Scientiﬁc Research (NWO, grant 918.66.620) and by the US National Institutes of Health grant 1R01CA141743 and Fondo de Investigacin Sanitaria grants FIS08/1676 and FIS07/0353. JIA—Arthritis Research UK (grant 17552). References 31. Bottolo, L. et al. GUESS-ing polygenic associations with multiple phenotypes using a GPU-based evolutionary stochastic search algorithm. PLoS Genet. 9, e1003657 (2013). 1. Fortune, M. D. et al. Statistical colocalization of genetic risk variants for related autoimmune diseases in the context of common controls. Nat. Genet. https://doi.org/10.1038/ng.3330 (2015). 1. Fortune, M. D. et al. Statistical colocalization of genetic risk variants for related autoimmune diseases in the context of common controls. Nat. Genet. https://doi.org/10.1038/ng.3330 (2015). 32. Giambartolomei, C. et al. A Bayesian framework for multiple trait colocalization from summary association statistics. Bioinformatics 34, 2538–2545 (2018). p g g 2. Cotsapas, C. et al. Pervasive sharing of genetic effects in autoimmune disease. PLoS Genet. 7, e1002254 (2011). 2. Cotsapas, C. et al. Pervasive sharing of genetic effects in autoimmune disease. PLoS Genet. 7, e1002254 (2011). 33. Cross-Disorder, Group of the Psychiatric Genomics Consortium. Identiﬁcation of risk loci with shared effects on ﬁve major psychiatric disorders: a genome-wide analysis. Lancet 381, 1371–1379 (2013). 3. Spain, S. L. & Barrett, J. C. Strategies for ﬁne-mapping complex traits. Hum. Mol. Genet. 24, R111–R119 (2015). p y Identiﬁcation of risk loci with shared effects on ﬁve major psychiatric disorders: a genome-wide analysis. Lancet 381, 1371–1379 (2013). 3. Spain, S. L. & Barrett, J. C. Strategies for ﬁne-mapping complex traits. Hum. Mol. Genet. 24, R111–R119 (2015). 4. Yang, J. et al. Conditional and joint multiple-SNP analysis of GWAS summary statistics identiﬁes additional variants inﬂuencing complex traits. Nat. Genet. 44, 369–75, S1–3 (2012). 4. Yang, J. et al. Conditional and joint multiple-SNP analysis of GWAS summary statistics identiﬁes additional variants inﬂuencing complex traits. Nat. Genet. 44, 369–75, S1–3 (2012). g y 34. Vattikuti, S., Guo, J. & Chow, C. C. Heritability and genetic correlations explained by common SNPs for metabolic syndrome traits. PLoS Genet. 8, e1002637 (2012). 5. Cordell, H. J. & Clayton, D. G. A uniﬁed stepwise regression procedure for evaluating the relative effects of polymorphisms within a gene using case/ control or family data: application to HLA in type 1 diabetes. Am. J. Hum. Genet. 70, 124–141 (2002). 35. Schmiedel, B. J. et al. Impact of genetic polymorphisms on human immune cell gene expression. Cell 175, 1701–1715.e16 (2018). 36. 1000 Genomes Project Consortium. et al. A global reference for human genetic variation. Nature 526, 68–74 (2015). 14 NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications © The Author(s) 2019 Reprints and permission information is available online at http://npg.nature.com/ reprintsandpermissions/ Peer review information: Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. Peer reviewer reports are available. Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. © The Author(s) 2019 15 NATURE COMMUNICATIONS \| (2019) 10:3216 \| https://doi.org/10.1038/s41467-019-11271-0 \| www.nature.com/naturecommunications
https://openalex.org/W2159332780	https://europepmc.org/articles/pmc5349079?pdf=render	English	null	Management of a parotid sialocelein a young patient: case report and literature review	Journal of Applied Oral Science	2,010	cc-by	3,801	abstract S ialocele is a subcutaneous cavity containing saliva, caused by trauma or infection in the parotid gland parenchyma, laceration of the parotid duct or ductal stenosis with subsequent dilatation. It is characterized by an asymptomatic soft and mobile swelling on the parotid region. Imaging studies are useful and help establishing the diagnosis, such as sialography, ultrasonography, computed tomography and magnetic resonance imaging. This paper describes a recurrent case of a parotid sialocele in a young female patient. She presented a 6 cm x 5 cm swelling on the left parotid region. The ultrasonographic scan of the area revealed a hypoechoic ovoid well defined image suggesting a cyst. A sialography of the left parotid showed a cavitary sialectasia in a panoramic and anteroposterior view. A conservative management was adopted by percutaneous needle aspiration of the swelling, which was useful to provide material for analysis and helped healing. Dentists should be aware of this pathology and the importance in adopting a conservative treatment whenever it is possible. Key words: Parotid gland, diagnosis. Sialography. Sialocele. sialocele15. Temporomandibular joint surgery13, parotidectomy18, mastoidectomy12, mandibular osteotomies11 and facial abscess drainage25 have been mentioned as potential causes of sialocele, and in all these cases the duct and/or the gland are damaged. Patient’s history usually includes facial trauma or surgery, days or weeks before the onset of the swelling5. www.scielo.br/jaos www.scielo.br/jaos Melissa Rodrigues de ARAUJO1, Bruna Stuchi CENTURION2, Danielle Frota de ALBUQUERQUE3, Luiz Henrique MARCHESANO4, José Humberto DAMANTE5 Melissa Rodrigues de ARAUJO1, Bruna Stuchi CENTURION2, Danielle Frota de ALBUQUERQUE3, Luiz Henrique MARCHESANO4, José Humberto DAMANTE5 Melissa Rodrigues de ARAUJO1, Bruna Stuchi CENTURION2, Danielle Frota de ALBUQUERQUE3, Luiz Henrique MARCHESANO4, José Humberto DAMANTE5 1- DDS, MSc, PhD. Department of Stomatology, Bauru Dental School, University of São Paulo, Bauru, SP, Brazil. 1- DDS, MSc, PhD. Department of Stomatology, Bauru Dental School, University of São Paulo, Bauru, SP, Brazil. 2- DDS, Department of Stomatology, Bauru Dental School, University of São Paulo, Bauru, SP, Brazil. 3- DDS, MSc. Department of Stomatology, Bauru Dental School, University of São Paulo, Bauru, SP, Brazil. 4 MSc PhD Clinical Analysis Laboratory Craniofacial Anomalies Rehabilitation Hospital University of São Paulo Bauru SP Brazil 2- DDS, Department of Stomatology, Bauru Dental School, University of São Paulo, Bauru, SP, Brazil. 2- DDS, Department of Stomatology, Bauru Dental School, University of São Paulo, Bauru, SP, Brazil. 3 DDS MSc Department of Stomatology Bauru Dental School University of São Paulo Bauru SP Brazil 2 DDS, Department of Stomatology, Bauru Dental School, University of São Paulo, Bauru, SP, Brazil. 3- DDS, MSc. Department of Stomatology, Bauru Dental School, University of São Paulo, Bauru, SP, Brazil. 3- DDS, MSc. Department of Stomatology, Bauru Dental School, University of São Paulo, Bauru, SP, Brazil. 4- MSc, PhD. Clinical Analysis Laboratory, Craniofacial Anomalies Rehabilitation Hospital, University of São Paulo, Bauru, SP, Brazil. 4- MSc, PhD. Clinical Analysis Laboratory, Craniofacial Anomalies Rehabilitation Hospital, University of São Paulo, Bauru, SP, Braz 5- DDS, MSc, PhD, Full Professor. Department of Stomatology, Bauru Dental School, University of São Paulo, Bauru, SP, Brazil. y y p y 5- DDS, MSc, PhD, Full Professor. Department of Stomatology, Bauru Dental School, University of São Paulo, Bauru, SP, Brazil. Received: March 13, 2009 - Modification: September 5, 2009 - Accepted: October 9, 2009 J Appl Oral Sci. 2010;18(4):432-6 Received: March 13, 2009 - Modification: September 5, 2009 - Accepted: October 9, 2009 Corresponding address: Melissa Rodrigues de Araujo - Rua Pedro Romildo Dall Stella, 100 - casa 5 - 82.115-470 - Phone: 41 3023 3357 or 8417 0800 - melissararaujo@hotmail.com rresponding address: Melissa Rodrigues de Araujo - Rua Pedro Romildo Dall Stella, 100 - casa 5 - 82.115-470 - Phone: 41 3023 3357 lissararaujo@hotmail.com ress: Melissa Rodrigues de Araujo - Rua Pedro Romildo Dall Stella, 100 - casa 5 - 82.115-470 - Phone: 41 3023 3357 or 8417 0800 - Introduction Sialocele is a subcutaneous cavity containing saliva, usually results from trauma or infection to the parotid gland parenchyma, laceration of the parotid duct or ductal stenosis with subsequent dilatation5. A post-traumatic sialocele is an acquired lesion that arises from extravasation of saliva into the glandular or periglandular tissues secondary to disruption of the parotid duct or parenchyma1,7. Traumatic causes include sharp penetrating wounds in the oral cavity or in the face6,17 and blunt trauma, such as zygomatic and mandible fractures15,19. Extrinsic infections from mandibular teeth3 mumps, actinomycosis, tuberculosis and syphilis have been recognized as causes of parotid fistulae in the past14.i Sialocele is characterized by a swelling on the parotid region. On palpation, the lesion may be soft and mobile and unless secondary infected, the patient has no pain, fever, chills, or erythema of the skin7. The diagnosis is complex and involves a combination of thorough history and clinical assessment of the patient, fine needle aspiration and image analysis. Fine needle aspiration is a standard technique which permits sampling and the fluid may be submitted to further Congenital fistulae and the ones secondary to invasive malignant tumors of the parotid gland or the oral cavity can also be associated to 432 J Appl Oral Sci. Management of a parotid sialocelein a young patient: case report and literature review investigation. Imaging studies include sialography, ultrasonography, computed tomography (CT) and magnetic resonance imaging5,7. The sialography is a technique that may increase sialocele’s pressure causing rupture and fistulae, although this is not a commonly observed complication18. CT may show details of the area, such as a single or multiloculated cystic lesion with regular margins and lower density of the surrounding tissues. The CT differential diagnosis would include retention cyst, sialodochitis, branchial cleft cyst and lymphoepithelial cyst7. the affected side was normal and salivary flow had normal physical aspects without debris or purulent discharge. The presumptive clinical diagnosis was an abscess associated to mild inflammation, considering the patient’s history and the previous episodes. Conventional panoramic radiography, ultrasonography and sialography were performed. The panoramic radiograph showed teeth and bones preserved without any evidence of abnormalities. The ultrasonographic scan revealed a hypoechoic ovoid well defined image suggesting a cyst (Figure 2). Figure 1- Anteroposterior facial view illustrating the 6 x 5 cm swelling on patient’s left cheek Many treatment modalities have been mentioned in the literature. They basically consist of a conservative or a surgical approach. Introduction A conservative modality is based on regular aspiration of the content, compression dressing, and administration of an antisialogogue8,18. Radiotherapy and toxins are other treatment modalities. Administration of botulinum toxin causes temporary chemical denervation of the cholinergic nerve fibers and has been used successfully. It is a highly effective, safe and non invasive method21,22,28. Drugs act by blocking acethylcoline release, thereby inhibiting neurotransmission at the secretomotor parasympathetic autonomic nerve ending responsible for salivation9. When conservative management fails, or when the overlying skin become so thin that there is imminent risk of rupture, surgical treatment is indicated11. Surgery may be performed by repair or reconstruction of the duct, creation of a controlled internal fistula, superficial or total parotidectomy, parasympathetic denervation, and ductal ligation. If the sialocele is left untreated, it may develop into a significantly large facial swelling, fistula formation and may drain extra orally8,29 This paper reports a case of a parotid sialocele in a young patient managed with a conservative approach. J Appl Oral Sci. Case Report Figure 1- Anteroposterior facial view illustrating the 6 x 5 cm swelling on patient’s left cheek A white17-year-old female patient presented with a 4-month history of swelling over her left cheek anteriorly to the ear. She referred 3 previous episodes, the last one having started 1 month before the appointment. She denied trauma to the region, had not have episodes of fever lately and her medical history was not contributory. There was no associated pain or alteration of facial function as well as no motor or sensory deficits were observed. A white17-year-old female patient presented with a 4-month history of swelling over her left cheek anteriorly to the ear. She referred 3 previous episodes, the last one having started 1 month before the appointment. She denied trauma to the region, had not have episodes of fever lately and her medical history was not contributory. There was no associated pain or alteration of facial function as well as no motor or sensory deficits were observed. The swelling measured about 6 cm in length and 5 cm in width. On bimanual palpation an ill-defined and resilient mass was noticed (Figure 1). This mass was evident extra orally with a considerable bulging of the skin in the left parotid region. The lesion was normal and no lymphadenopathy was detected. On intraoral examination, oral mucosa and teeth were healthy. The parotid duct in Figure 2- The longitudinal section in the ultrasonographic scan of the left parotid gland demonstrates a hypoechoic ovoid well-defined image suggesting a cyst The swelling measured about 6 cm in length and 5 cm in width. On bimanual palpation an ill-defined and resilient mass was noticed (Figure 1). This mass was evident extra orally with a considerable bulging of the skin in the left parotid region. The lesion was normal and no lymphadenopathy was detected. On intraoral examination, oral mucosa and teeth were healthy. The parotid duct in Figure 2- The longitudinal section in the ultrasonographic scan of the left parotid gland demonstrates a hypoechoic ovoid well-defined image suggesting a cyst 433 433 2010;18(4):432-6 433 2010;18(4):432-6 J Appl Oral Sci. J Appl Oral Sci. 433 2010;18(4):432-6 2010;18(4):432-6 ARAUJO MR, CENTURION BS, ALBUQUERQUE DF, MARCHESANO LH, DAMANTE JH ARAUJO MR, CENTURION BS, ALBUQUERQUE DF, MARCHESANO LH, DAMANTE JH Sialography of the left parotid was performed using Lipiodol® (Lipiodol; Guerbet, Jacarepaguá, RJ, Brazil) as a substance of contrast. Case Report A partial filling of the gland was enough to show cavitary sialectasia in a panoramic (Figure 3a) and anteroposterior view (Figure 3b), and the contrast was retained in the gland for at least 24 hours (Figure 3c). A percutaneous needle aspiration of the swelling was performed 2 weeks after the sialography. It yielded 4 mL of a clear viscous fluid. The material was submitted to microbiological analysis and showed numerous polymorphonuclear leukocytes. The fluid did not show any bacterial growth and Figure 3- A panoramic (A) and an anteroposterior (B) radiograph taken after sialography of the left parotid gland. A partial filling of the gland was sufficient to show cavitary sialectasia. The contrast was retained in the gland for at least 24 h, as shown in the panoramic radiograph (C) Figure 3- A panoramic (A) and an anteroposterior (B) radiograph taken after sialography of the left parotid gland. A partial filling of the gland was sufficient to show cavitary sialectasia. The contrast was retained in the gland for at least 24 h, as shown in the panoramic radiograph (C) Figure 5- Complete recovery of the skin and remission of the swelling after 30 days Figure 4- Erythema on the overlying skin after the percutaneous needle aspiration Figure 4- Erythema on the overlying skin after the percutaneous needle aspiration Figure 5- Complete recovery of the skin and remission of the swelling after 30 days J Appl Oral Sci. J Appl Oral Sci. 434 2010;18(4):432-6 434 J Appl Oral Sci. 434 2010;18(4):432-6 . 434 2010;18(4):432-6 2010;18(4):432-6 Management of a parotid sialocelein a young patient: case report and literature review may be very helpful2. Sialography of the parotid gland was mandatory in revealing the sialectasia in our case and is also useful to distend the duct when it is involved4. Lipiodol® is a solution containing iodine that might have acted as an antibacterial agent and helped reducing the facial swelling combined with the aspiration. In our case report the sialography served as a diagnostic tool and helped healing the involved gland. There was a clear relief of the signs after sialography. presented high amylase levels (7,810 units/L). The aspiration procedure was sufficient to empty the cavity. The patient was advertised to compression dressing twice a day. Seven days after the percutaneous aspiration a great decrease of the swelling was observed. Case Report The overlying skin showed a discreet erythema on the surface of the mass, without signs of infection, and the patient did not refer pain (Figure 4). At the 30-day follow-up visit, complete recovery was noticed, and the residual swelling had totally disappeared (Figure 5). The patient was followed up during two years and a half and no recurrence was detected. Another key to the diagnosis and treatment was fine-needle aspiration, which provided material for analysis and helped empting the gland. Parotid secretion has a high amylase content that is usually around 10,000 units/L19. In the present case, the amylase content was 7,810 units/L, confirming the diagnosis of parotid saliva extravasation. Discussion Primary parotid gland cysts are very rare, representing five per cent of all parotid tumors10,23,25,. Mucoceles are round and well defined lesions that contain mucus, when they occur in the parotid gland are called sialoceles. For practical purposes they may be regarded as being of either extravasation or retention type. The term mucous extravasation phenomenon (cyst) is used when mucus is extruded into the connective tissues and is surrounded by a granulation tissue, while mucous retention cyst is used to describe a cyst with retained mucin that is lined with ductal epithelium which may have undergone squamous or oncocytic metaplasia26,27. The factors that determine a mucocele are the rate of mucus production and the speed of phagocytosis of the extravasated mucus. The majority of the mucoceles previously reported in the parotid gland are of retention type (ductal cyst)26. A variety of treatments have been proposed for parotid sialoceles20. These include multiple aspirations and compression dressings; late primary repair or reconstruction of the duct; creation of a controlled internal fistula; superficial or total parotidectomy; parasympathetic denervation (sectioning of the auriculotemporal nerve); antisialogogues; radiation therapy and ductal ligation. Most of these procedures are invasive with risks of injury of the facial nerve, with variable and often poor success rates7. The anticholinergic drugs have many undesired side effects such as xerostomia, constipation, photophobia, tachycardia and urinary retention16. Atropine and glycopyrrolate are antisialagogue drugs that may be used to treat sialoceles4. The present case reported did not require any invasive treatment or administration of a drug. A conservative treatment such as suggested by Landau and Stewart17 (1985), the injection of an antibacterial solution to perform a sialography, the aspiration of the content and the compression dressing were capable to solve the case. Parotid sialoceles are lesions that occur after trauma or injury in the face causing accumulation of saliva in the area1,4,20. It has not been described yet a case where the patient could not associate a history of trauma, injury or surgery. It is possible that our patient suffered a trauma of low intensity and could not remember it, but it should be something else to explain the 3 recurrent episodes of tumefaction. Dentists should be aware of this pathology and the importance in adopting a conservative treatment whenever it is possible. J Appl Oral Sci. p g g ; ( ) 40. 3- Bailey BM. A persistent parotid fistula following the extraction of mandibular teeth. J Laryngol Otol. 1984;98(10):1051-4. 4- Barron R, Margulis A, Icekson M, Zeltser R, Eldad A, Nahlieli O. Iatrogenic parotid sialocele following rhytidectomy: diagnosis and treatment. Plast Reconstr Surg. 2001;108(6). 5- Bater MC. An unusual case of preauricular swelling: a giant parotid sialocele. Int J Oral Maxillofac Surg. 1998;27(2):125-6. 6- Burch RJ. Spontaneous closure of a parotid gland fistula with the aid of banthine; report of a case. Oral Surg Oral Med Oral Pathol. 1953;6(10):1191-4. 1- Akinbami BO. Traumatic diseases of parotid gland and sequalae. Review of literature and case reports. Niger J Clin Pract. 2009;12(2):212-5. 2- Antoniadis K, Karakasis D, Tzarou V, Skordalaki A. Benign cysts of the parotid gland. Int J Oral Maxillofac Surg. 1990;19(3):139- 40. 3- Bailey BM. A persistent parotid fistula following the extraction of mandibular teeth. J Laryngol Otol. 1984;98(10):1051-4. 4- Barron R, Margulis A, Icekson M, Zeltser R, Eldad A, Nahlieli O. Iatrogenic parotid sialocele following rhytidectomy: diagnosis and treatment. Plast Reconstr Surg. 2001;108(6). 5- Bater MC. An unusual case of preauricular swelling: a giant parotid sialocele. Int J Oral Maxillofac Surg. 1998;27(2):125-6. 6- Burch RJ. Spontaneous closure of a parotid gland fistula with the aid of banthine; report of a case. Oral Surg Oral Med Oral Pathol. 1953;6(10):1191-4. 2- Antoniadis K, Karakasis D, Tzarou V, Skordalaki A. Benign cysts of the parotid gland. Int J Oral Maxillofac Surg. 1990;19(3):139- 40. 3- Bailey BM. A persistent parotid fistula following the extraction of mandibular teeth. J Laryngol Otol. 1984;98(10):1051-4. References Initially, our patient had a resilient, ill-defined mass, which was difficult to palpate, probably due to the position of the cystic diffuse inflammation, under the dense parotid fascia, which makes physical examination unreliable. Clinical assessment may be inaccurate in these cases4,24. However, soft and mobile lesions had also been described when they are more superficial7. 1- Akinbami BO. Traumatic diseases of parotid gland and sequalae. Review of literature and case reports. Niger J Clin Pract. 2009;12(2):212-5. 1- Akinbami BO. Traumatic diseases of parotid gland and sequalae. Review of literature and case reports. Niger J Clin Pract. 2009;12(2):212-5. 2- Antoniadis K, Karakasis D, Tzarou V, Skordalaki A. Benign cysts of the parotid gland. Int J Oral Maxillofac Surg. 1990;19(3):139- 40. 2- Antoniadis K, Karakasis D, Tzarou V, Skordalaki A. Benign cysts of the parotid gland. Int J Oral Maxillofac Surg. 1990;19(3):139- 40. 2- Antoniadis K, Karakasis D, Tzarou V, Skordalaki A. Benign cysts of the parotid gland. Int J Oral Maxillofac Surg. 1990;19(3):139- 40. 3- Bailey BM. A persistent parotid fistula following the extraction of mandibular teeth. J Laryngol Otol. 1984;98(10):1051-4. Our patient did not present fever or other compromising signs in any episode of swelling20. The management of a patient with a swelling in the parotid region requires careful clinical evaluation. Fine-needle aspiration or biopsy is necessary for a definite diagnosis. Sialography, computed tomography and ultrasonographic scans 435 435 J Appl Oral Sci. 2010;18(4):432-6 ARAUJO MR, CENTURION BS, ALBUQUERQUE DF, MARCHESANO LH, DAMANTE JH ARAUJO MR, CENTURION BS, ALBUQUERQUE DF, MARCHESANO LH, DAMANTE JH 19- Meyer RA, Gordon RC. Method for repair of traumatic pseudocyst of parotid duct: report of case. J Oral Surg. 1969;27(4):281-3. 19- Meyer RA, Gordon RC. Method for repair of traumatic pseudocyst of parotid duct: report of case. J Oral Surg. 1969;27(4):281-3. 7- Canosa A, Cohen MA. Post-traumatic parotid sialocele: report of two cases. J Oral Maxillofac Surg. 1999;57(6):742-5. 19- Meyer RA, Gordon RC. Method for repair of traumatic pseudocyst of parotid duct: report of case. J Oral Surg. 1969;27(4):281-3. 8- Cholankeril JV, Scioscia PA. Post-traumatic sialoceles and mucoceles of the salivary glands. Clin Imaging. 1993;17(1):41-5. 8- Cholankeril JV, Scioscia PA. Post-traumatic sialoceles and mucoceles of the salivary glands. Clin Imaging. 1993;17(1):41-5. 9- Chow TL, Kwok SP. Use of botulinum toxin type A in a case of persistent parotid sialocele. Hong Kong Med J. 2003;9(4):293-4. 10- Cohen MN, Rao U, Shedd DP. Benign cysts of the parotid gland. J Surg Oncol. 1984;27(2):85-8. 20- Monfared A, Ortiz J, Roller C. Distal parotid duct pseudocyst as a result of blunt facial trauma. Ear Nose Throat J. 2009;88(8):15-7. 21- Parekh D, Glezerson G, Stewart M, Esser J, Lawson HH. Post-traumatic parotid fistulae and sialoceles. A prospective study of conservative management in 51 cases. Ann Surg. 1989;209(1):105-11. 11- Demetriades D, Rabinowitz B. Management of parotid sialoceles: a simple surgical technique. Br J Surg. 1987;74(4):309. 11- Demetriades D, Rabinowitz B. Management of parotid sialoceles: a simple surgical technique. Br J Surg. 1987;74(4):309. 12- Dierks EJ, Granite EL. Parotid sialocele and fistula after mandibular osteotomy. J Oral Surg. 1977;35(4):299-300 22- Pereira KD, Smith SL, Mitchell RB. Parotid sialocele in a 10-year-old girl. Ear Nose Throat J. 2007;86(1):27-8. 12- Dierks EJ, Granite EL. Parotid sialocele and fistula after mandibular osteotomy. J Oral Surg. 1977;35(4):299-300 10-year-old girl. Ear Nose Throat J. 2007;86(1):27-8. 23- Pieterse AS, Seymour AE. Parotid cysts. An analysis of 16 cases and suggested classification. Pathology. 1981;13(2):225-34. 24- Rejali D, Simo R, Small M. Mucocele mimicking a Warthin's tumour recurrence. J Laryngol Otol. 1998;112(11):1092-4. 25- Richardson GS, Clairmont AA, Erickson ER. Cystic lesions of the parotid gland. Plast Reconstr Surg. 1978;61(3):364-70. 26- Seifert, G., Miehlke, A., Haubrich, J., Chilla, R. Diseases of the salivary gland. Stuttgart: Georg Thieme Verlag; 1986. p. 171-80. 27- Tal H, Altini M, Lemmer J. Multiple mucous retention cysts of the oral mucosa. Oral Surg Oral Med Oral Pathol. 1984;58(6):692- 5. 23- Pieterse AS, Seymour AE. Parotid cysts. 7- Canosa A, Cohen MA. Post-traumatic parotid sialocele: report of two cases. J Oral Maxillofac Surg. 1999;57(6):742-5. 8- Cholankeril JV, Scioscia PA. Post-traumatic sialoceles and mucoceles of the salivary glands. Clin Imaging. 1993;17(1):41-5. 9- Chow TL, Kwok SP. Use of botulinum toxin type A in a case of persistent parotid sialocele. Hong Kong Med J. 2003;9(4):293-4. 10- Cohen MN, Rao U, Shedd DP. Benign cysts of the parotid gland. J Surg Oncol. 1984;27(2):85-8. 11- Demetriades D, Rabinowitz B. Management of parotid sialoceles: a simple surgical technique. Br J Surg. 1987;74(4):309. 12- Dierks EJ, Granite EL. Parotid sialocele and fistula after mandibular osteotomy. J Oral Surg. 1977;35(4):299-300 13- Dolwick MF, Kretzschmar DP. Morbidity associated with the preauricular and perimeatal approaches to the temporomandibular joint. J Oral Maxillofac Surg. 1982;40(11):699-700. 14- Hemenway WG, Bergstrom L. Parotid duct fistula: a review. South Med J. 1971;64(8):912-8. 15- Hutchison IL, Ryan D. A parotid fistula and sialocele complicating temporomandibular joint surgery. Br J Oral Maxillofac Surg. 1989;27(3):203-8. 16- Krausen AS, Ogura JH. Sialoceles: medical treatment first. Trans Sect Otolaryngol Am Acad Ophtalmol Otolaryngol. 1977;84(5):ORL890-5. 17- Landau R, Stewart M. Conservative management of post- traumatic parotid fistulae and sialoceles: a prospective study. Br J Surg. 1985;72(1):42-4. 18- Langdon JD. Complications of parotid gland surgery. J Maxillofac Surg. 1984;12(5):225-9. ARAUJO MR, CENTURION BS, ALBUQUERQUE DF, MARCHESANO LH, DAMANTE JH An analysis of 16 cases and suggested classification. Pathology. 1981;13(2):225-34. 13- Dolwick MF, Kretzschmar DP. Morbidity associated with the preauricular and perimeatal approaches to the temporomandibular joint. J Oral Maxillofac Surg. 1982;40(11):699-700. 14- Hemenway WG, Bergstrom L. Parotid duct fistula: a review. South Med J. 1971;64(8):912-8. 15- Hutchison IL, Ryan D. A parotid fistula and sialocele complicating temporomandibular joint surgery. Br J Oral Maxillofac Surg. 1989;27(3):203-8. 16- Krausen AS, Ogura JH. Sialoceles: medical treatment first. Trans Sect Otolaryngol Am Acad Ophtalmol Otolaryngol. 1977;84(5):ORL890-5. 28- Vargas H, Galati LT, Parnes SM. A pilot study evaluating the treatment of postparotidectomy sialoceles with botulinum toxin type A. Arch Otolaryngol Head Neck Surg. 2000;126: 421-4. 29- Witt RL. The incidence and management of siaolocele after parotidectomy. Otolaryngol Head Neck Surg. 2009;140(6):871-4. 17- Landau R, Stewart M. Conservative management of post- traumatic parotid fistulae and sialoceles: a prospective study. Br J Surg. 1985;72(1):42-4. 18- Langdon JD. Complications of parotid gland surgery. J Maxillofac Surg. 1984;12(5):225-9. J Appl Oral Sci. 436 2010;18(4):432-6 J Appl Oral Sci. 436 2010;18(4):432-6 436 20 436 J Appl Oral Sci.
https://openalex.org/W2338615199	https://europepmc.org/articles/pmc4489166?pdf=render	English	null	Thermochromic phantom for therapeutic ultrasound daily quality assurance	Journal of therapeutic ultrasound	2,015	cc-by	1,323	© 2015 Qureshi et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver (http:// creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Thermochromic phantom for therapeutic ultrasound daily quality assurance Farhan Qureshi1*, Zachary Larrabee1, Chris Roth1, Arik Hananel1, Matt Eames1, David Moore1, John Snell1, Neal Kassell1, Jean-Francois Aubry2 From Current and Future Applications of Focused Ultrasound 2014. 4th International Symposium Washington, D.C, USA. 12-16 October 2014 Figure 1 Phantom after Sonication with HIFU Transducer Background/introduction Errors in power output ranging from −100% to +210% have been reported in a multitude of physiotherapy transducers.[1] Differences in power output can arise even after careful calibration on an annual or bi-annual schedule, which can either result in harm to the patient or non-effective treatment. Therefore, easy implementa- tion of daily quality assurance is of great importance. We propose a simple, easy to use DQA phantom which allows the user to assess the power output of the focused ultrasound transducer, and determine if it has changed significantly after calibration. The basis of this phantom is the use of a highly attenuating ultrasound absorber with a surface layer of thermochromic liquid crystals (TLC). The use of thermochromic materials as a technique for use in ultrasound phantoms has been an ongoing area of interest. Many of these techniques require complicated set-ups, with large water tanks and imaging systems to produce very accurate data.[2,3] Simpler phantoms have been developed for use in benchtop settings, but still with a large emphasis on complicated image analysis.[4] Our proposed phantom will be a product where the end-user can visually assess the size of the lesion formed as a function of power out- put, without resorting to complicated image analysis. Figure 1 Phantom after Sonication with HIFU Transducer a visual check as to whether or not the transducer’s power output has changed. A Matlab code has been developed to numerically solve the bio-heat equation in finite differences time domain with a generation term dependent on the induced pressure field. The code gener- ates plots with a color map corresponding to the thermo- chromic material’s color changing properties, and the acoustic absorber’s acoustic attenuation and thermal diffu- sivity properties. The attenuation coefficient of the mate- rial, provided by Precision Acoustics Ltd, was 28.8 ±2.6 dB∙cm-1∙MHz-1. The thermal diffusivity was tested by Decagon Devices, Inc. using the transient line heat source method, and was determined to be 0.09 ±0.01 mm2/s. The peak positive pressure is adjusted in the simulation so that the lesion size generated by the simulation corresponds to the actual lesion size measured after sonication, giving an estimate of the pressure at focus. Qureshi et al. Journal of Therapeutic Ultrasound 2015, 3(Suppl 1):P72 http://www.jtultrasound.com/content/3/S1/P72 Qureshi et al. Journal of Therapeutic Ultrasound 2015, 3(Suppl 1):P72 http://www.jtultrasound.com/content/3/S1/P72 1Focused Ultrasound Foundation, Charlottesville, Virginia, United States Full list of author information is available at the end of the article POSTER PRESENTATION Open Access Methods In the phantom, differences in lesion size due to differ- ences in ambient temperature are revealed by visual inspection. A simple temperature scale will correspond to concentric circles of varying radii printed on the phan- tom. The generated lesion size matches the radius of the corresponding circle at each ambient temperature, giving 1Focused Ultrasound Foundation, Charlottesville, Virginia, United States Full list of author information is available at the end of the article Page 2 of 2 Page 2 of 2 Qureshi et al. Journal of Therapeutic Ultrasound 2015, 3(Suppl 1):P72 http://www.jtultrasound.com/content/3/S1/P72 Qureshi et al. Journal of Therapeutic Ultrasound 2015, 3(Suppl 1):P72 http://www.jtultrasound.com/content/3/S1/P72 Figure 2 Simulated Temperature Elevation Figure 2 Simulated Temperature Elevation Acknowledgements (Funding) Th k f d d b h F Acknowledgements (Funding) This work was funded by the Focused Ultrasound Foundation. Acknowledgements (Funding) This work was funded by the Focused Ultrasound Foundation. Acknowledgements (Funding) This work was funded by the Focused Ultrasound Foundation. References 1. Schabrun S, Walker H, et al: How Accurate are Therapeutic Ultrasound Machines? Hong Kong Physiotherapy Journal 2008, 26:39-44. 2. Martin K, Fernandez R: A thermal beam-shape phantom for ultrasound physiotherapy transducers. Ultrasound Med Biol 1997, 23(8):1267-1274. 3. Munoz A. Lopez Gerardo, Orozc A. Valentino Gerardo: Three Dimensional Temperature Distribution Analysis of Ultrasound Therapy Equipments Using Thermochromic Liquid Crystal Films, New Developments in Liquid Crystals.Georgiy V Tkachenko 2009, ISBN: 978-953-307-015-5, InTech, Available from: http://www.intechopen.com/books/new-developments-in- liquid-crystals/three-dimensional-temperature-distribution-analysis-of- ultrasound-therapy-equipments-using-thermochr. 4. Butterworth I, Barrie J, et al: Exploiting thermochromic materials for the rapid quality assurance of physiotherapy ultrasound treatment heads. Ultrasound Med Biol 2012, 38(5):767-776. doi:10.1186/2050-5736-3-S1-P72 Cite this article as: Qureshi et al.: Thermochromic phantom for therapeutic ultrasound daily quality assurance. Journal of Therapeutic Ultrasound 2015 3(Suppl 1):P72. 1. Schabrun S, Walker H, et al: How Accurate are Therapeutic Ultrasound Machines? Hong Kong Physiotherapy Journal 2008, 26:39-44. 2. Martin K, Fernandez R: A thermal beam-shape phantom for ultrasound physiotherapy transducers. Ultrasound Med Biol 1997, 23(8):1267-1274. 3. Munoz A. Lopez Gerardo, Orozc A. Valentino Gerardo: Three Dimensional Temperature Distribution Analysis of Ultrasound Therapy Equipments Using Thermochromic Liquid Crystal Films, New Developments in Liquid Crystals.Georgiy V Tkachenko 2009, ISBN: 978-953-307-015-5, InTech, Available from: http://www.intechopen.com/books/new-developments-in- liquid-crystals/three-dimensional-temperature-distribution-analysis-of- ultrasound-therapy-equipments-using-thermochr. 4 Butterworth I Barrie J et al Exploiting thermochromic materials for the 1. Schabrun S, Walker H, et al: How Accurate are Therapeutic Ultrasound Machines? Hong Kong Physiotherapy Journal 2008, 26:39-44. 2. Martin K, Fernandez R: A thermal beam-shape phantom for ultrasound physiotherapy transducers. Ultrasound Med Biol 1997, 23(8):1267-1274. 3. Munoz A. Lopez Gerardo, Orozc A. Valentino Gerardo: Three Dimensional Temperature Distribution Analysis of Ultrasound Therapy Equipments Using Thermochromic Liquid Crystal Films, New Developments in Liquid Crystals.Georgiy V Tkachenko 2009, ISBN: 978-953-307-015-5, InTech, Available from: http://www.intechopen.com/books/new-developments-in- liquid-crystals/three-dimensional-temperature-distribution-analysis-of- ultrasound-therapy-equipments-using-thermochr. 4. Butterworth I, Barrie J, et al: Exploiting thermochromic materials for the rapid quality assurance of physiotherapy ultrasound treatment heads. Ultrasound Med Biol 2012, 38(5):767-776. doi:10.1186/2050-5736-3-S1-P72 Cite this article as: Qureshi et al.: Thermochromic phantom for therapeutic ultrasound daily quality assurance. Journal of Therapeutic Ultrasound 2015 3(Suppl 1):P72. 1. Schabrun S, Walker H, et al: How Accurate are Therapeutic Ultrasound Machines? Hong Kong Physiotherapy Journal 2008, 26:39-44. Results and conclusions A method for creating a simple DQA phantom has been devised. The main goal of the phantom is to verify that the output power of the transducer does not change compared to initial calibration. The end user places the phantom in their transducer system, and sonicates for a set power and time. The power output of the transducer is then checked visually using a system of concentric circles printed on the phantom. This system is a cheap and effective way to pro- duce DQA phantoms, both because of the ease of use, and because the use of TLC makes visual changes in the phan- tom reversible and reproducible over a long period of time. Authors’ details 1 1Focused Ultrasound Foundation, Charlottesville, Virginia, United States. 2Institut Langevin, Paris, France. Published: 30 June 2015 References Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission y 4. Butterworth I, Barrie J, et al: Exploiting thermochromic materials for the rapid quality assurance of physiotherapy ultrasound treatment heads. Ultrasound Med Biol 2012, 38(5):767-776. doi:10.1186/2050-5736-3-S1-P72 Cite this article as: Qureshi et al.: Thermochromic phantom for therapeutic ultrasound daily quality assurance. Journal of Therapeutic Ultrasound 2015 3(Suppl 1):P72. doi:10.1186/2050-5736-3-S1-P72 Cite this article as: Qureshi et al.: Thermochromic phantom for therapeutic ultrasound daily quality assurance. Journal of Therapeutic Ultrasound 2015 3(Suppl 1):P72.
https://openalex.org/W4288682761	https://www.biochemia-medica.com/assets/images/upload/xml_tif/bm-32-3-030704.pdf	English	null	Designing and validating an autoverification system of biochemical test results in Hatay Mustafa Kemal University, clinical laboratory	Biochemia medica	2,022	cc-by	6,869	Abstract Introduction: Autoverification (AV) is a postanalytical tool that uses algorithms to validate test results according to specified criteria. The Clinical and Laboratory Standard Institute (CLSI) document for AV of clinical laboratory test result (AUTO-10A) includes recommendations for laboratories needing guidance on implementation of AV algorithms. The aim was to design and validate the AV algorithm for biochemical tests. Materials and methods: Criteria were defined according to AUTO-10A. Three different approaches for algorithm were used as result limit checks, which are reference range, reference range ± total allowable error, and 2nd and 98th percentile values. To validate the algorithm, 720 cases in middleware were tested. For actual cases, 3,188,095 results and 194,520 reports in laboratory information system (LIS) were evaluated using the AV system. Cohen’s kappa (κ) was calculated to determine the degree of agreement between seven independent reviewers and the AV system. Results: The AV passing rate was found between 77% and 85%. The highest rates of AV were in alanine transaminase (ALT), direct bilirubin (DBIL), and magnesium (Mg), which all had AV rates exceeding 85%. The most common reason for non-validated results was the result limit check (41%). A total of 328 reports evaluated by reviewers were compared to AV system. The statistical analysis resulted in a κ value between 0.39 and 0.63 (P < 0.001) and an agreement rate between 79% and 88%. Conclusions: Our improved model can help laboratories design, build, and validate AV systems and be used as starting point for different test gro- ups. Keywords: automation; biochemistry; laboratory organization and management; validation/evaluation Keywords: automation; biochemistry; laboratory organization and management; validation/evaluation Submitted: February 07, 2022 Accepted: June 24, 2022 Designing and validating an autoverification system of biochemical test results in Hatay Mustafa Kemal University, clinical laboratory Bahar Ünlü Gül1, Oğuzhan Özcan2, Serdar Doğan2, Abdullah Arpaci2 1Department of Medical Biochemistry, Kars Harakani Public Hospital, Kars, Turkey 2Department of Medical Biochemistry, Hatay Mustafa Kemal University, Hatay, Turkey 1Department of Medical Biochemistry, Kars Harakani Public Hospital, Kars, Turkey 2Department of Medical Biochemistry, Hatay Mustafa Kemal University, Hatay, Turkey Corresponding author: abdullaharpaci@mku.edu.tr 1 ©Copyright by Croatian Society of Medical Biochemistry and Laboratory Medicine. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creative- commons.org/licenses/by/4.0/) which permits users to read, download, copy, distribute, print, search, or link to the full texts of these articles in any medium or format and to remix, transform and build upon the material, provided the original work is properly cited and any changes properly indicated. Biochem Med (Zagreb) 2022;32(3):030704 https://doi.org/10.11613/BM.2022.030704 Introduction The result verification process, which is the most important control step of the postanalytical phase, can be performed in two ways which are manually or by using autoverification (AV) systems. The re- sults are evaluated by the laboratory technician and then by the clinical biochemist (2). The appro- priate results are verified and sent to the relevant clinic via the laboratory information system (LIS). Corrective actions are initiated for the discordant results. These actions are carried out in three phas- Clinical laboratories are centres where millions of tests are analysed that guide the diagnosis and treatment of patients. The workload of clinical lab- oratories is increasing day by day because of the expansion of test panels, increasing numbers of samples analysed, high-quality expectations, and aiming to report test results in a shorter time. This increased workload causes clinical biochemists to allocate a large part of their working hours for the manual verification of test results (1). https://doi.org/10.11613/BM.2022.030704 Designing and validating an autoverification system of biochemical test results Ünlü Gül B. et al. es preanalytical, analytical, and postanalytical ac- cording to the type of laboratory errors. Common preanalytical errors include haemolysed sample, clotted sample, inadequate volume, wrong sam- ple tube, and incorrect identification. At the ana- lytical phase, quality control (QC) or calibration failure, interference, and sample mix-ups errors can occur. Some postanalytical errors include de- lay in reporting, incorrect calculation, and critical results being delayed or not reported. The evalua- tion of each parameter at laboratory phases can be quite time-consuming for laboratories where the number of experts is insufficient compared to the number of tests that are ordered. Additionally, the evaluation process is subjective because it de- pends on the personnel and does not include a standardisation. There is a need for AV systems that use multiple algorithms in accordance with the developing technology for allowing effective time management, preventing possible laboratory errors, and providing more consistent test results (3,4). and validate specific AV rules for laboratory tests (6). However, it is recommended that each labora- tory create the cut-off values required for autoveri- fying test results according to the needs of its pa- tient population (3). Autoverification algorithms use various criteria to determine the reportability of test results. These criteria include instrument flags, QC checks, moving averages, serum indices, critical values, delta checks, and analytical meas- urement ranges (AMRs). Materials and methods Design of AV algorithm Introduction Those algorithms can also examine patient or sample information from elec- tronic medical records, times of sampling, and de- mographic information including age, sex, diag- nosis, and inpatient/outpatient status (8). Although the AUTO 10-A guidelines have been around for nearly 15 years, there remains a lack of standardisation, especially regarding algorithms, validation rules, and verification limits. In recent years, there has been an increasing body of litera- ture on AV, but still, there is a limited number of studies on this topic in Turkey. In this respect, this study aims to contribute to the understanding of the AV process, provide well-designed AV rules that can be used in clinical laboratories, rapid and accurate verification of test results without human intervention, and raise work efficiency. It is the first time that we developed multi-rule algorithms for the verification of biochemical tests in Hatay Mustafa Kemal University (HMKU), Central Labora- tory. We presented a detailed description of how to design an AV system that can be used as a start- ing point and for applying detailed system valida- tion process. Indeed, our study, which was per- formed at a large district hospital with around 1700 inpatients and 24,000 outpatients treated per month, provides a new approach to the recent lit- erature with respect to designing algorithms to detect pre-analytical errors through consistency checks. Autoverification is a postanalytical process by which laboratory test results are released without manual intervention or review. Autoverification, which is an application of artificial intelligence for clinical laboratories, is thought to be an alternative to manual validation of test results (4). The first al- gorithm describing the use of computers to assist laboratory test validation was published over 50 years ago by Lindberg on the identification of “critical” results (5). This research is aimed at the evaluation of correlated analyses (e.g., urea and creatinine (CREA)) with consistency checks and delta checks when test results exceed the defined limits. Currently, AV ensures a well-designed set of rules, and more specific algorithms have been de- veloped (2,6,7). Approved guidelines include important recom- mendations for laboratories needing guidance on the implementation of AV algorithms. The Auto- verification of Clinical Laboratory Test Results: Ap- proved Guidelines (AUTO 10-A and AUTO-15) were issued by the Clinical Laboratory Standards Insti- tute (CLSI) in 2006 and 2019, respectively (3,4). These guidelines provide a basic template to allow each clinical core laboratory to build, implement, Biochem Med (Zagreb) 2022;32(3):030704 Validation of AV ing System, Arbe Software Trading Company Ltd.) platform which was the middleware system. The middleware was used to provide the connection between the biochemistry autoanalysers and LIS. This electronic connection can process the prean- alytical, analytical, and postanalytical data sent from the autoanalysers within the scope of the de- signed AV rules and send them to LIS. Additionally, the definition of AV algorithms and the creation of simulator patient data were completed through the middleware. It should be validated that the developed algo- rithm follows the expected logic and produces the expected results. The validation of the AV system was performed in two phases. In the first phase, all algorithm criteria were applied to the 720 simulat- ed results to follow the logic of the algorithm and verify the performance of the algorithm. The simu- lated results were created to include all algorithm decision rules for each biochemical test. In the sec- ond phase, the validation of the algorithm was performed using actual test results according to the AUTO-10A guideline (3). In this phase, 194,520 patient reports and 2,025,948 test results, which were previously assayed in our laboratory were collected from July 2019 to May 2020. We con- firmed that the algorithms followed the expected logic, and data were recorded on the middleware after checking the correctness of the calculations. While designing the AV algorithms, the evaluation criteria recommended in the approved guidelines were defined, and the master algorithm template was created (Figure 1). The corrective actions that were cause-specific for the laboratory tests that did not pass any algorithm rule are also given in Figure 1. Figure 1. Master algorithm template for biochemical tests. QC – quality control. AMR – analytical measurement range. EDTA – ethylenediaminetetraacetic acid. LIS – laboratory information sys- tem. AV – autoverification. Is result a repeated test? No No No No No Yes Yes Yes Yes Yes Fail Fail Fail Fail Fail Fail Fail Pass Pass Pass Pass Pass Pass Pass Pass Result below or above analitical measurement range? Critical value control Interference indices check? Are there any previous results? Consistency checks? Delta check control? Result limit check? All tests completed? Autoverification Is result a numeric value Moving average control Instrument error flag? Design of AV algorithm The AV process was carried out on ADVIA 1800 au- toanalysers (Siemens Diagnostics, New York, USA) through the LIOS (Laboratory Information Operat- https://doi.org/10.11613/BM.2022.030704 Biochem Med (Zagreb) 2022;32(3):030704 2 Designing and validating an autoverification system of biochemical test results Ünlü Gül B. et al. Criteria for algorithm Delta check: Delta check is performed to compare the present result for a patient to previous results and evaluate the probability of significant change. Delta check has been performed in previous stud- ies using various methods such as delta difference, delta percent change, rate difference, and rate percent difference (12). Internal QC and calibration: Internal QC is routinely performed in our laboratory and transferred from the autoanalyser to the middleware according to the Westgard QC multi-rules (9). If there are no QC results within 24 hours or if any QC rule is violated, the middleware stops the AV for all samples. Cali- bration periods and the last calibration dates of all tests were defined in the middleware calibration screen. In this study, we chose the rate percent difference, which was calculated as the present value sub- tracted from the previous value then divided by the previous value, followed by dividing the re- sults by the time interval. We chose a time interval as three months based on discussions with physi- cians at the Department of Medical Biochemistry. We used the reference change value (RCV), ex- pressed as a percentage or absolute value which is a common systematic approach in the determina- tion of delta thresholds (9). Instrument error codes (Flags): If any instrument flag is sent from the autoanalyser, the results are held for later manual verification. Moving average: We used the ‘moving average’ method as an additional QC method to help pro- vide the quality of the results. The moving averag- es involving the collection of over the last 20 pa- tient results are analysed. Using these results, a mean value is defined, a warning limit which is mean ± 2 standard deviation (SD) is calculated, and then, a warning message is sent to the user via SMS (Short Message Service) (6,10). Rate percent difference Present result-Previous result x 100 Previous result Time interval = Equation (Eq.) 1. Analytical measurement range (AMR): We used the AMRs recommended by the manufacturer for the autoanalyser (Siemens Diagnostics, New York, USA). If any results were outside the range of the AMR, a warning message was generated, and sam- ple dilution was performed automatically or man- ually. Result limit checks: Result limit checks are used to specify if the result exceeds predetermined thresh- olds. Validation of AV Quality control check Starting step to AV Send results to LIS Manual verification Hold for manual review and corrective actions • QC check: Investigate and corrective action (e.g., QC material, instrument error logs, assay methods, reagent) • Instrument error flags: Investigate and corrective action (e.g., sample switching, wrong sample aliquoted/pipetted, inadequate volume, environment status) • Moving average control: Investigate and corrective action (e.g., patient data control, QC control, etc.) • Abnormal result: Investigate and corrective action (dilution, re-run etc.) • Repeated test: Investigate and corrective action • Result below AMR: Rules dictate how result should be reported (e.g., “< 3” or “0”) • Result above AMR: Manual or automated dilution • Critical results: Call clinician • Indices check: Inspect sample (result-comment/recollect) • Delta check: Investigate cause with corrective action • Result limit check: Investigate cause with corrective action • Consistency check: Corrective action depending on likely cause (e.g., EDTA or citrate contamination, delayed sample, slottes sample etc.) Figure 1. Master algorithm template for biochemical tests. QC – quality control. AMR – analytical measurement range. EDTA – ethylenediaminetetraacetic acid. LIS – laboratory information sys- tem. AV – autoverification. Figure 1. Master algorithm template for biochemical tests. QC – quality control. AMR – analytical measurement range. EDTA – ethylenediaminetetraacetic acid. LIS – laboratory information sys- tem. AV – autoverification. QC – quality control. AMR – analytical measurement range. EDTA – ethylenediaminetetraacetic acid. LIS – tem. AV – autoverification. Biochem Med (Zagreb) 2022;32(3):030704 Biochem Med (Zagreb) 2022;32(3):030704 https://doi.org/10.11613/BM.2022.030704 https://doi.org/10.11613/BM.2022.030704 3 Designing and validating an autoverification system of biochemical test results Ünlü Gül B. et al. Biochem Med (Zagreb) 2022;32(3):030704 Criteria for algorithm Three different limits were used to determine the thresholds as reference range, reference range ± total allowable error (TEa), and 2nd and 98th percentile values (8,13). Reference ranges were provided by the manufacturer, and we used them for each test according to the relevant reagent procedures. Total allowable error values that were taken from the College of American Pathologists (CAP) database were used for each test to deter- mine reference range ± TEa values (14). For each test from the historical data of 2019, 2nd and 98th percentile values were calculated, and outliers were excluded by using box-blot analyses. Critical Values: We used the critical values recom- mended by the General Directorate of Health Ser- vices, Department of Investigation, and Diagnostic Services (Ankara, Turkey) (11). If any results exceed- ed the range of the critical value, they were manu- ally verified by a laboratory technician, and a phone call was immediately made to clinicians. Serum Indices: The haemolysis, icterus and lipae- mia (HIL) interferences were detected using multi- ple spectrophotometric readings by the biochem- istry analyser in the serum specimen for the spe- cific tests. If analytes like potassium (K), aspartate aminotransaminase (AST), and lactate dehydroge- nase (LDH) were higher in the haemolytic samples that required examination by a technician, these results were held for manual verification. Consistency checks: Consistency rule checks are cross-checks established based on two or more different correlated tests. The cross-checks includ- ed in the algorithm in the last control step are giv- en in Table 1. https://doi.org/10.11613/BM.2022.030704 Biochem Med (Zagreb) 2022;32(3):030704 4 Designing and validating an autoverification system of biochemical test results Ünlü Gül B. et al. Criteria for algorithm Citrate contamination Calcium decreased by 50% Sodium increased by 5 mmol/L Chloride decreased by 10 mmol/L Clotted sample Sodium < 136 mmol/L Potassium < 3.5 mmol/L Calcium < 8.4 mmol/L Glucose < 3.9 mmol/L Ethylenediaminetetraacetic acid (EDTA) contamination Potassium > 7 mmol/L and calcium < 8 mmol/L, or ALP < 50 U/L, or magnesium < 2 mmol/L Delayed sample Glucose < 2.2 mmol/L Potassium > 6 mmol/L Haemolysis < 50 mg/dL Discordant results ALT/AST ratio < 0.25 or > 4 Albumin / Total Protein ratio < 0.25 or > 1 Direct Bilirubin / Total Bilirubin ratio > 1 HDL cholesterol / Total Cholesterol ratio > 0.75 Intravenous glucose contamination Sodium < 136 Chloride < 98 Potassium > 5 Glucose > 6.1 Intravenous saline contamination Increased chloride normal sodium, low or critical low potassium Monoclonal protein interference High lipemia index (> +2) with low/normal triglyceride Glomerular filtration rate It is evaluated together with the creatinine test Indirect bilirubin It is evaluated together with direct bilirubin and total bilirubin ALP – alkaline phosphatase. ALT – alanine aminotransaminase. AST – aspartate aminotransaminase. HDL cholesterol – high density lipoprotein cholesterol. Table 1. Consistency rule checks Table 1. Consistency rule checks The study was carried out with the approval of the HMKU Ethics Committee, with protocol numbered 2019/91, and resolution numbered 03. This study was supported financially by the HMKU Scientific Research Projects (BAP; project number: 19.U.016) Unit, for which we are thankful. Comparison of manual review to AV Patient reports, which were randomly selected by the middleware to compare manual review and AV results, were manually verified by seven expert re- viewers. The number of the autoverified reports was 255 out of a total of 328 test reports, and 73 test reports were manually verified. The seven ex- perts who performed the manual verification of the patient reports consisted of two medical bio- chemists, two medical biochemistry research as- sistant doctors, and three laboratory technicians with a wide range of professional knowledge. Us- ing the obtained data, the AV and manual review results were compared statistically, and their agreement rate was evaluated. The AV system in- cluded 30 biochemical tests analysed in the Cen- tral Laboratory and most frequently ordered by clinicians (Table 2). Statistical analysis All data of the study were analysed using the IBM SPSS Statistics software, Version 21.0, (SPSS Inc., New York, USA). P < 0.05 was considered signifi- cant. The categorical variables are expressed as counts and percentages. Second and 98th percen- tile values were calculated for each test from the historical data of 2019, and outlier values were eliminated by using box-blot analyses. Our results were statistically analysed according to three dif- ferent result limit checks which were reference Biochem Med (Zagreb) 2022;32(3):030704 https://doi.org/10.11613/BM.2022.030704 5 Designing and validating an autoverification system of biochemical test results Ünlü Gül B. et al. Table 2. Gray – indicate the highest passing rates for biochemical tests based on different result limit checks. TEa – total allowable error. AV – autoverification. Alb – albumin. ALP – alkaline phosphatase. ALT – alanine aminotransaminase. AMY – amylase. AST – aspartate aminotransaminase. BUN – blood Urea Nitrogen. CK – creatine kinase. CK-MB – creatine, direct kinase isoenzyme MB. DBIL – direct bilirubin. Fe – iron. Phos – inorganic phosphate. GGT – gamma-glutamyltransferase. Glc – glucose. HDL – high density lipoprotein cholesterol. Ca – calcium. Cl – chloride. CHOL – cholesterol. CREA – creatinine. LDH – lactate dehydrogenase. LDL – low density lipoprotein cholesterol. Lip – lipase. Mg – magnesium. K – potassium. Na – sodium. TBIL – total bilirubin. TP – total Protein. TRSF – transferrine. TG – triglyceride. UIBC – unsaturated Iron-Binding Capacity. UA – uric acid. Statistical analysis Autoverification passing rates for all biochemical tests Biochemical parameters Units Reference range Reference range ± %TEa 2nd and 98th percentile Number of tests Autoverify % Stop AV % Autoverify % Stop AV % Autoverify % Stop AV % Alb g/L 83 17 89 11 93 7 90,541 ALP U/L 81 19 87 13 89 11 63,217 ALT U/L 88 12 92 9 92 8 166,761 AMY U/L 85 15 90 10 92 8 40,663 AST U/L 83 17 87 13 91 10 142,545 BUN mmol/L 81 19 84 16 87 14 148,408 CK U/L 85 15 88 12 89 12 15,722 CK-MB U/L 59 41 69 31 89 11 5379 DBIL µmol/L 87 13 90 10 90 10 62,760 Fe µmol/L 52 48 54 46 90 11 19,902 Phos mmol/L 84 16 88 12 89 12 43,134 GGT U/L 74 26 80 20 87 13 58,915 Glc mmol/L 66 34 74 26 85 15 147,452 HDL mmol/L 56 44 85 15 88 12 11,852 Ca mmol/L 82 18 90 10 90 10 98,106 Cl mmol/L 79 21 87 13 87 13 60,847 CHOL mmol/L 68 32 80 20 91 9 11,672 CREA µmol/L 65 35 79 21 79 21 173,905 LDH U/L 71 29 80 20 91 7 57,363 LDL mmol/L 42 57 63 37 93 7 12,691 Lip U/L 84 16 88 12 90 10 37,945 Mg mmol/L 87 13 87 13 87 13 47,594 K mmol/L 64 36 68 32 68 32 140,852 Na mmol/L 83 17 89 11 90 11 144,863 TBIL µmol/L 84 16 85 15 89 12 63,069 TP g/L 70 30 75 25 77 23 54,860 TRSF g/L 23 77 37 63 89 12 130 TG mmol/L 65 35 77 23 90 10 21,004 UIBC µmol/L 55 45 78 22 90 10 19,544 UA mmol/L 81 19 87 13 90 10 64,252 Total 2,025,948 Table 2. Autoverification passing rates for all biochemical tests Gray – indicate the highest passing rates for biochemical tests based on different result limit checks. TEa – total allowable error. AV – autoverification. Alb – albumin. ALP – alkaline phosphatase. ALT – alanine aminotransaminase. AMY – amylase. AST – aspartate aminotransaminase. BUN – blood Urea Nitrogen. CK – creatine kinase. CK-MB – creatine, direct kinase isoenzyme MB. DBIL – direct bilirubin. Fe – iron. Phos – inorganic phosphate. GGT – gamma-glutamyltransferase. Glc – glucose. HDL – high density lipoprotein cholesterol. Statistical analysis Ca – calcium. Cl – chloride. CHOL – cholesterol. CREA – creatinine. LDH – lactate dehydrogenase. LDL – low density lipoprotein cholesterol. Lip – lipase. Mg – magnesium. K – potassium. Na – sodium. TBIL – total bilirubin. TP – total Protein. TRSF – transferrine. TG – triglyceride. UIBC – unsaturated Iron-Binding Capacity. UA – uric acid. Biochem Med (Zagreb) 2022;32(3):030704 https://doi.org/10.11613/BM.2022.030704 6 Designing and validating an autoverification system of biochemical test results Ünlü Gül B. et al. range, reference range ± TEa, and 2nd and 98th percentile values. Cohen’s κ value was calculated to determine the degree of agreement between seven independent reviewers’ assessments and the AV system. The κ coefficient ranges from - 1 to + 1. A κ value of + 1 indicates the highest agree- ment, while a κ value of - 1 indicates the highest disagreement between the users (15). Using the 2nd and 98th percentiles as a limit range increased the AV pass rates for TRSF, LDL, and Fe by 65%, 51%, and 37%, respectively, and these were also the tests that showed the highest in- crease in their AV pass rates. In the AV process, the most common reasons for manually verified test results are given in Figure 3. A total of 2,025,948 tests were included, of which 83% (1,672,430) were autoverified when the refer- ence range ± TEa criterion was used as the limit range (Figure 2). Result limit check was observed to be the most common reason for stopping AV (41%) (Figure 3). Result limit checks, which were the most common reason for non-validated re- sults, were the same for all three different decision limits, and the most common reasons were result limit checks, moving averages, and serum indices, respectively. Results The test-based AV rate for all biochemical tests was 85% over a period of one year when the 2nd and 98th percentile values were used as result lim- it checks. When the reference ranges were used, the AV pass rate was found to be the lowest at 77% (Figure 2). The test-based AV pass rate according to three dif- ferent limit ranges are detailed in Table 2. The highest rates of AV were found for the alanine ami- notransaminase (ALT), direct bilirubin (DBIL), and magnesium (Mg) parameters, which all had AV rates exceeding 85%. The least validated tests were transferrin (TRSF), low-density lipoprotein cholesterol (LDL), and iron (Fe) according to the reference range, and the AV pass rates of these pa- rameters were 23%, 42%, and 52%, respectively. A total of 328 actual patient reports were included, of which 255 (78%) were autoverified, and 73 (22%) were manually verified, according to the reference range ± TEa criterion. The degree of agreement between the users and the middleware was ana- lysed using the κ statistic, and the results are given in Table 3. Accordingly, the statistical analysis re- sulted in a κ value between 0.39 and 0.63 (from minimal to moderate agreement, P < 0.001). The https://doi.org/10.11613/BM.2022.030704 Biochem Med (Zagreb) 2022;32(3):030704 7 Figure 2. Autoverification passing rate on test-based according to different verify limits. Validated – It refers to the percentage of tests that were passed according to the algorithm criteria in Figure 1. Rejected – It re- fers to the percentage of tests that were not passed at least one of the algorithm criteria in Figure 1. TEa – total allowable error. Figure 3. Cause analysis of manually verified test results. The reference range ± TEa was used as the decision limit. TEa – total allowable error. Results Reference range ± TEa Reference range Total number of tests 2,000,000 1,750,000 1,500,000 1,250,000 1,000,000 750,000 500,000 250,000 0 77% 83% 85% 23% 18% 15% Validated Rejected 1,562,304 463,644 1,672,430 353,518 1,722,725 303,223 2nd and 98th percentile 100 90 80 70 60 50 40 30 20 10 0 41% 28% 18% 9% 2% % 1.8% 0.09% Result limit check Moving average Interference indices Delta check Critical value Repeat testing Analytical measurement range Reference range ± TEa Reference range Total number of tests 2,000,000 1,750,000 1,500,000 1,250,000 1,000,000 750,000 500,000 250,000 0 77% 83% 85% 23% 18% 15% Validated Rejected 1,562,304 463,644 1,672,430 353,518 1,722,725 303,223 2nd and 98th percentile 100 90 80 70 60 50 40 30 20 10 0 41% 28% 18% 9% 2% % 1.8% 0.09% Result limit check Moving average Interference indices Delta check Critical value Repeat testing Analytical measurement range Figure 3. Cause analysis of manually verified test results. The reference range ± TEa was used as the decision limit. TEa – total allowable error. Figure 2. Autoverification passing rate on test-based according to different verify limits. Figure 2. Autoverification passing rate on test-based according to different verify limits. Validated – It refers to the percentage of tests that were passed according to the algorithm criteria in Figure 1. Rejected – It re- fers to the percentage of tests that were not passed at least one of the algorithm criteria in Figure 1. TEa – total allowable error. Validated – It refers to the percentage of tests that were passed according to the algorithm criteria in Figure 1. Rejected – It re- fers to the percentage of tests that were not passed at least one of the algorithm criteria in Figure 1. TEa – total allowable error. Biochem Med (Zagreb) 2022;32(3):030704 https://doi.org/10.11613/BM.2022.030704 Designing and validating an autoverification system of biochemical test results Ünlü Gül B. et al. Results Reviewers Autoverification Rules (LIOS) Agreement Sensitivity Specificity κ value Kappa approximate significance V R Reviewer 1 V 243 27§ 88 90 79 0.63 P < 0.001 R 12‡ 46† Reviewer 2 V 249* 42§ 85 86 84 0.49 P < 0.001 R 6‡ 31† Reviewer 3 V 234* 26§ 86 90 69 0.58 P < 0.001 R 21‡ 47† Reviewer 4 V 234* 31§ 84 88 67 0.52 P < 0.001 R 21‡ 42† Reviewer 5 V 210* 23§ 79 90 53 0.46 P < 0.001 R 45‡ 50† Reviewer 6 V 240* 43§ 82 85 67 0.41 P < 0.001 R 15‡ 30† Reviewer 7 V 247* 48§ 83 84 76 0.39 P < 0.001 R 8‡ 25† True positive. †True negative. ‡False positive. §False negative. Agreement – True Positive + (True Negative / Total number) x 100. Sensitivity – True Positive / (True Positive + False Negative) x 100. Specificity – True Negative / (True Negative + False Positive) x 100. κ value – 0-0.20 none; 0.21-0.39 minimal; 0.40-0.59 weak; 0.60-0.79 moderate; 0.80-0.90 strong; > 0.90 almost perfect (16). P < 0.001 – Highly statistically significant degree of agreement. V – validated. R – rejected. LIOS – laboratory information operating system. Table 3. Degree of agreement between the LIOS and each expert reviewer Table 3. Degree of agreement between the LIOS and each expert reviewer True positive. †True negative. ‡False positive. §False negative. Agreement – True Positive + (True Negative / Total number) x 100. Sensitivity – True Positive / (True Positive + False Negative) x 100. Specificity – True Negative / (True Negative + False Positive) x 100. κ value – 0-0.20 none; 0.21-0.39 minimal; 0.40-0.59 weak; 0.60-0.79 moderate; 0.80-0.90 strong; > 0.90 almost perfect (16). P < 0.001 – Highly statistically significant degree of agreement. V – validated. R – rejected. LIOS – laboratory information operating system. It is necessary to design AV rules for laboratories like our facility where approximately 1100 samples are analysed for routine biochemical tests. Bio- chemical tests constitute a large part of the ana- lysed patient samples. When a large number of re- sults in the queue are examined, fatigue can de- velop, and this is admittedly a potential risk factor for laboratory errors (13). Results In this regard, in our study, we developed algorithms to detect and minimise common pre-analytical errors including ethylenediaminetetraacetic acid (EDTA) or citrate contamination, clotted sample, and delayed sam- ple through consistency rule checks (Table 1). In previous research, it has been reported that the most commonly used criteria in AV algorithms are AMRs, critical values, instrument error codes, se- rum indices, and delta check values (18,19). In our study, we also included moving averages, refer- ence ranges, result limit checks, and consistency checks in the AV rules to create a more specific multi-rule algorithm (Figure 1). Therefore, the find- ings of this study suggested that test results were strongest agreement between user 1 and the mid- dleware was found to be a statistically significant and moderate agreement (κ = 0.63; P < 0.001). Biochem Med (Zagreb) 2022;32(3):030704 Discussion Autoverification is a powerful tool that uses rule- based systems to evaluate and validate test results without manual intervention. Currently, laborato- ries use AV in different groups of tests, including routine tests which are biochemistry, immunoas- says, haematology, coagulation, blood gas, and urinalysis (16). It has apparent benefits in improv- ing test quality, reducing error rates, decreasing turnaround time, and enhancing the efficiency of laboratory verification. In contrast, manual verifi- cation is a time-consuming activity with inherent subjectivity, and thereby, it cannot provide suffi- ciently accurate verification of test results (17). To overcome these limitations, we designed and im- plemented a middleware-based multi-rule system for AV in biochemical tests (Figure 1). https://doi.org/10.11613/BM.2022.030704 Biochem Med (Zagreb) 2022;32(3):030704 8 Designing and validating an autoverification system of biochemical test results Ünlü Gül B. et al. reviewed in more detail with a multi-rule algo- rithm. study by Gruenberg et al. revealed a higher AV passing rate, greater than 90%, using a ± 60% del- ta check for 23,410 CREA results (25). In a previous study in which delta check limits were evaluated as < 20%, AV rates between 50% and 75% were obtained. Understandably, the AV passing rates re- ported in multiple studies have shown differences (7,26,27). These differences were probably due to using different result limit checks and delta check limits and developments in AV rules over time. The AV system in this study was designed accord- ing to three different decision limits that have been recommended in the literature (6,13,20). The AV passing rate was found between 77% and 85% based on the algorithms we developed. The high- est AV passing rate was in the parameters of ALT, DBIL, and Mg, which all had rates exceeding 85%. Torke et al. used the midpoint between the medi- an of the reference range and AMR as a limit range, and the AV passing rate they found increased to 62% (19). Our passing rates were higher than their system, at a verification rate of approximately 77%. A higher percentage of test results (83%) was au- toverified based on the reference range ± TEa cri- terion. Previous studies have shown that by using the reference range ± TEa criterion, the AV passing rate varies between 73% and 85% (8, 20). A study conducted by Shih et al. Discussion concluded that the rate of AV was 92% for 25,526 patient reports and 42 bio- chemical tests using the 2nd and 98th percentiles of cumulative patient data (21). We found that the AV rate obtained using historical patient data was lower (85%). These differences could be explained by differences in the limit ranges, the groups of tests that are studied, and the laboratory equip- ment that is used. Additionally, critical values and AMRs were not autoverified but held for later manual verification in our study. The least validat- ed tests were TRFS, LDL, and Fe when the refer- ence range criterion was used. One reason for this could be that the Fe test shows higher diurnal vari- ation than the other tests (22). The reason for the variability in the LDL test was thought to be the differences in genetics and dietary habits in the population in which the study was conducted. The dramatic increase in the passing rate of the trans- ferrin test was probably related to the lower num- ber of test orders compared to other tests. l d h d d h f d l In the AV process, one of the important issues is the cause analysis of manually verified test results. In our study, the most common reasons for non- validated results were result limit checks, moving averages, and serum indices, respectively. Rimac et al. reported that among 31 different biochemical tests, the least common reason for non-validated results was the critical value (2%) (27). Similarly, the same rate of critical value was 2% in our study, and this result was consistent with data reported in the literature. The degree of agreement between AV and the seven expert reviewers’ assessments was found significant (agreement rates between 79% and 88%, P < 0.001), which pointed out that our system was valid (Table 3). Mohy-Sediq et al. compared AV system results to results provided by 4 reviewers, and the agreement rates were between 73% and 77%, which were lower than those in our study (28). The validation of AV rules is crucial in ensuring that the AV system operates as intended and re- quires high attention to detail. Additionally, there is a need for developing middleware that allows well-designed algorithms and validation (29). This study had some limitations. References net]. 2020 Sept. [Cited 2021 Jul 9]. Available from: https:// dosyamerkez.saglik.gov.tr/Eklenti/15143,karar-siniri-esik- deger-kriek42009846pdf.pdf?0. (in Turkish) net]. 2020 Sept. [Cited 2021 Jul 9]. Available from: https:// dosyamerkez.saglik.gov.tr/Eklenti/15143,karar-siniri-esik- deger-kriek42009846pdf.pdf?0. (in Turkish) 1. T.C. Sağlık Bakanlığı. Tıbbi laboratuvarda onay destek siste- mi [Internet]. 2018 Mar [Cited 2021 Jul 9]. Available from: https://shgmtetkikdb.saglik.gov.tr/Eklenti/15045/0/tibbi- laboratuvarda-onay-desteek41916132pdf.pdf (in Turkish) 12. Randell EW, Yenice S. Delta Checks in the clinical laboratory. Crit Rev Clin Lab Sci. 2019;56:75-97. https://doi.org/10.1080 /10408363.2018.1540536 2. Fernandez DC, Avinash SS, Malathi M, Shivashanka- ra AR, Kumar A, Fernandez PA. Establishing the referen- ce change values (RCVs) and validating the delta check au- to-verification in a clinical biochemistry laboratory. Muller J Med Sci Res. 2017;8:42-6. https://doi.org/10.4103/0975- 9727.199363 13. Randell EW, Yenice S, Khine Wamono AA, Orth M. Autove- rification of test results in the core clinical laboratory. Clin Biochem. 2019;73:11-25. https://doi.org/10.1016/j.clinbio- chem.2019.08.002 14. Collage of American Pathologist (CAP). Participant summary. Available from: https://www.cap.org/member- resources/councils-committees/chemistry-participant-re- ports. Accessed Dec 9th 2021. 3. Clinical and Laboratory Standard Institute (CLSI). AUTO 10-A Autoverification of clinical laboratory test results - First Edition. CLSI document; Wayne (PA): 2006. 4. Clinical and Laboratory Standard Institute (CLSI). AUTO 15 Autoverification of Medical Laboratory Results for Specific Disciplines - Second Edition. CLSI document; Wayne (PA): 2019. 15. McHugh ML. Interrater reliability: the kappa statistic. Bi- ochem Med (Zagreb). 2012;22:276-82. https://doi. org/10.11613/BM.2012.031 16. Topcu DI, Gulbahar O. A model to establish autoverificati- on in the clinical laboratory. Clin Biochem. 2021;93:90-8. https://doi.org/10.1016/j.clinbiochem.2021.03.018 5. Lindberg DA. Collection, evaluation, and transmission of hospital laboratory data. Methods Inf Med. 1967;6:97-107. https://doi.org/10.1055/s-0038-1636364 17. Jin D, Wang Q, Peng D, Wang J, Li B, Cheng Y, et al. Deve- lopment and implementation of an LIS-based validation system for autoverification toward zero defects in the auto- mated reporting of laboratory test results. BMC Med Inform Decis Mak. 2021;21:174. https://doi.org/10.1186/s12911- 021-01545-3 6. Li J, Cheng B, Ouyang H, Xiao T, Hu J, Cai Y. Designing and evaluating autoverification rules for thyroid function profi- les and sex hormone tests. Ann Clin Biochem. 2018;55:254- 63. https://doi.org/10.1177/0004563217712291 7. Wu J, Pan M, Ouyang H, Yang Z, Zhang Q, Cai Y. Establis- hing and evaluating autoverification rules with intelli- gent guidelines for arterial blood gas analysis in a clini- cal laboratory. SLAS Technol. 2018;23:631-40. https://doi. org/10.1177/2472630318775311 18. Rimac V, Jokic A, Podolar S, Vlasic Tanaskovic J, Honovic L, Lenicek Krleza J. Potential conflict of interest None declared. sistent test results. Since starting to use the AV sys- tem, our laboratory accelerated verification so we can save more time and focus on verifying the ab- normal test results. Our improved model can help design, build, and validate an AV system and be used as a starting point for different test groups. None declared. Discussion As the compari- son of AV and user results was evaluated on previ- ously analysed and validated patient results, the discordant results could not be reanalysed. Pro- spective studies are needed for such an analysis. Our experience suggests that designing and using a comprehensive AV algorithm requires carefully created rules and the performance of a well-de- signed validation process. The AV system based on a middleware enabled more rapid and routine evaluation of test results, minimised the require- ments for manual work and provided more con- Several studies have recommended the use of del- ta checks in AV algorithms (12,23,24). In this study, the delta check limits were determined by using the RCV values obtained from the biological varia- tion database (2). In our study, the delta check limit used for the CREA test was ± 13%, and the AV rates were between 65% and 79% (Table 2). A recent Biochem Med (Zagreb) 2022;32(3):030704 https://doi.org/10.11613/BM.2022.030704 9 Designing and validating an autoverification system of biochemical test results Ünlü Gül B. et al. References General position of Croatian medical bi- ochemistry laboratories on autovalidation: survey of the Working Group for Post-analytics of the Croatian Society of Medical Biochemistry and Laboratory Medicine. Biochem Med (Zagreb). 2020;30:020702. https://doi.org/10.11613/ BM.2020.020702 8. Demirci F, Akan P, Kume T, Sisman AR, Erbayraktar Z, Se- vinc S. Artificial neural network approach in laboratory test reporting: Learning Algorithms. Am J Clin Pathol. 2016;146:227-37. https://doi.org/10.1093/ajcp/aqw104 19. Torke N, Boral L, Nguyen T, Perri A, Chakrin A. Process im- provement and operational efficiency through test result autoverification. Clin Chem. 2005;51:2406-8. https://doi. org/10.1373/clinchem.2005.054395 9. Ricós C, Alvares V, Cava F, García-Lario JV, Hernández A, Ji- menez CV, et al. Desirable biological variation databa- se specifications. The 2014 update. Available from: https:// www.westgard.com/biodatabase1.htm. Accessed Dec 9th 2021. 20. Froom P, Saffuri-Elias E, Barak M. Autovalidation rates in an outpatient coagulation laboratory. Int J Lab Hematol. 2015;37:680-685. 10. Lukić V, Ignjatović S. Optimizing moving average con- trol procedures for small-volume laboratories. Biochem Med (Zagreb). 2019;29:030710. https://doi.org/10.11613/ BM.2019.030710 21. Shih MC, Chang HM, Tien N, Hsiao CT, Peng CT. Building and validating an autoverification system in the clinical che- mistry laboratory. Lab Med. 2011;42:668-73. https://doi. org/10.1309/LM5AM4IIXC4OIETD 11. T.C. Sağlık Bakanlığı. Karar sınırı (eşik değer), kritik değer (panik değer) ve ölçüm birimlerinin harmonizasyonu [Inter- Biochem Med (Zagreb) 2022;32(3):030704 https://doi.org/10.11613/BM.2022.030704 10 Designing and validating an autoverification system of biochemical test results Ünlü Gül B. et al. 22. Nguyen LT, Buse JD, Baskin L, Hossein Sadrzadeh SM, Nau- gler C. Influence of diurnal variation and fasting on serum iron concentrations in a community-based population. Clin Biochem. 2017;50:1237-42. https://doi.org/10.1016/j.clinbi- ochem.2017.09.018 26. Wang Z, Peng C, Kang H, Fan X, Mu R, Zhou L, et al. Design and evaluation of a LIS-based autoverification system for coagulation assays in a core clinical laboratory. BMC Med Inform Decis Mak. 2019;19:123. https://doi.org/10.1186/ s12911-019-0848-2 27. Rimac V, Lapic I, Kules K, Rogic D, Miler M. Implementa- tion of the autovalidation algorithm for clinical chemi- stry testing in the laboratory information system. Lab Med. 2018;49:284-91. https://doi.org/10.1093/labmed/lmx089 23. Xia LY, Cheng XQ, Liu Q, Liu L, Qin XZ, Zhang L, et al. [Deve- loping and application of an autoverification system for cli- nical chemistry and immunology test results]. Zhonghua Yi Xue Za Zhi. 2017;97:616-21. (in Chinese) 28. Sediq AME, Abdel-Azeez AGH. Designing an autoverifica- tion system in Zagazig University Hospitals Laboratori- es: preliminary evaluation on thyroid function profile. Ann Saudi Med. 2014;34:427-32. https://doi.org/10.5144/0256- 4947.2014.427 24. References Fernández-Grande E, Valera-Rodriguez C, Sáenz-Mateos L, Sastre-Gómez A, García-Chico P, Palomino-Muñoz TJ. Im- pact of reference change value (RCV) based autoverificati- on on turnaround time and physician satisfaction. Biochem Med (Zagreb). 2017;27:342-9. https://doi.org/10.11613/ BM.2017.037 29. Marquardt B. A step-by-step process to 95% autoverificati- on. CAP Today [Internet]. 2015 Dec [cited 2021 Jun 19]. Avai- lable from: https://www.captodayonline.com/step-by-step- autoverification/ 25. Gruenberg JM, Stein TA, Karger AB. Determining the utility of creatinine delta checks: A large retrospective analysis. Clin Biochem. 2018;53:139-42. https://doi.org/10.1016/j. clinbiochem.2018.01.023 https://doi.org/10.11613/BM.2022.030704 Biochem Med (Zagreb) 2022;32(3):030704 11 11 11
W3193006812.txt	http://uu.diva-portal.org/smash/get/diva2:1627318/FULLTEXT01	en		Multiplex protein screening of biomarkers associated with major bleeding in patients with atrial fibrillation treated with oral anticoagulation	Journal of thrombosis and haemostasis	2,021	cc-by	8,419	Received: 27 May 2021 \| Accepted: 12 August 2021 DOI: 10.1111/jth.15498 ORIGINAL ARTICLE Multiplex protein screening of biomarkers associated with major bleeding in patients with atrial fibrillation treated with oral anticoagulation Agneta Siegbahn1,2 \| Johan Lindbäck1 \| Ziad Hijazi1,3 \| Mikael Åberg2 \| 4 5 4 John H. Alexander \| John W. Eikelboom \| Renato D. Lopes \| Tymon Pol3 \| Jonas Oldgren1,3 \| Christopher B. Granger4 \| Salim Yusuf5 \| Lars Wallentin1,3 1 Uppsala Clinical Research Center, Uppsala University, Uppsala, Sweden 2 Department of Medical Sciences, Clinical Chemistry and Science for Life Laboratory, Uppsala University, Uppsala, Sweden 3 Department of Medical Sciences, Cardiology, Uppsala University, Uppsala, Sweden 4 Duke Clinical Research Institute, Duke Health, Durham, North Carolina, USA 5 Abstract Background: Oral anticoagulants (OAC) in patients with atrial fibrillation (AF) prevent thromboembolic events, but are associated with significant risk of bleeding. Objectives: To explore associations between a wide range of biomarkers and bleeding risk in patients with AF on OAC. Method: Biomarkers were analyzed in a random sample of 4200 patients, 204 cases with major bleedings, from ARISTOTLE. The replication cohort included 344 cases Population Health Research Institute, Hamilton, Ontario, Canada with major bleeding and 1024 random controls from RE-LY. Plasma samples obtained Correspondence Agneta Siegbahn, Department of Medical Sciences, Clinical Chemistry, Uppsala University, SE-751 85 Uppsala, Sweden. Email: agneta.siegbahn@medsci.uu.se cular and inflammation panels and conventional immunoassays. The associations Funding information The ARISTOTLE trial was funded by Bristol-Myers Squibb Co, Princeton, NJ, USA and Pfizer Inc., New York, NY, USA, and coordinated by the Duke Clinical Research Institute, Durham, NC, USA and Uppsala Clinical Research Center, Uppsala, Sweden. The RE-LY trial was funded by Boehringer Ingelheim, Ingelheim, Germany. The analyses were supported by The Swedish Foundation for Strategic Research [grant number RB13-0197], the Swedish Heart-Lung Foundation [grant number 20090183] and Science for Life Laboratory, Uppsala University, Uppsala, Sweden. Roche Diagnostics, Rotkreuz, Results: Out of 268 proteins, nine biomarkers were independently associated with at randomization were analyzed by the Olink Proximity Extension Assay cardiovasbetween biomarker levels and major bleeding over 1 to 3 years of follow-up were evaluated by random survival forest/Boruta analyses and Cox regression analyses to assess linear associations and hazard ratios for identified biomarkers. bleeding in both cohorts. In the replication cohort the linear hazard ratios (95% confidence intervals) per interquartile range were for these biomarkers: TNF-R1 1.748 (1.456, 2.098), GDF-15 1.653 (1.377, 1.985), EphB4 1.575 (1.320, 1.880), suPAR 1.548 (1.294, 1.851), OPN 1.476 (1.240, 1.757), OPG 1.397 (1.156, 1.688), TNF-R 2 1.360 (1.144,1.616), cTnT-hs 1.232 (1.067, 1.423), and TRAIL-R 2 1.202 (1.069, 1.351). Conclusions: In patients with AF on OAC, GDF-15, cTnT-hs, and seven novel biomarkers were independently associated with major bleedings and reflect pathophysiologic processes of inflammation, apoptosis, oxidative stress, vascular calcification, coagulation, and fibrinolysis. Investigations of the utility of these markers to refine Manuscript Handled by: Ton Lisman Final decision: Ton Lisman, 12 August 2021 ClinicalTrials.gov identifier: NCT00412984 and NCT00262600 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. © 2021 The Authors. Journal of Thrombosis and Haemostasis published by Wiley Periodicals LLC on behalf of International Society on Thrombosis and Haemostasis. 2726 \| wileyonlinelibrary.com/journal/jth J Thromb Haemost. 2021;19:2726–2737. \| SIEGBAHN et al. Switzerland, provided the pre-commercial assay of GDF-15. 2727 risk stratification and guide the management of patients at high risk of bleeding are warranted. KEYWORDS anticoagulants, atrial fibrillation, biomarkers, hemorrhages, proteomics 1 \| I NTRO D U C TI O N ESSENTIALS Atrial fibrillation (AF) is associated with a five-fold increased risk of • Biomarkers associated with bleeding can help with risk thromboembolisms, mainly stroke, independently of other risk fac1,2 tors. prediction in patients with atrial fibrillation (AF) on oral Oral anticoagulants (OAC) significantly reduce the risk of anticoagulants (OACs). thromboembolic events but, unfortunately, are also associated with • New analytical technologies allow measurements of a significant risk of major bleeding complications.3,4 Accurate assess- hundreds of biomarkers in small plasma volumes. ment to balance the risk of stroke and systemic embolic events (S/ • In a targeted proteomic approach we screened plasma SEE), against the risk of major bleeding is therefore an important from 5568 patients with AF on OAC. therapeutic goal in the clinical management of these patients.5 • Nine out of 268 investigated biomarkers were signifi- Age, prior hemorrhage, severe renal disease, and anemia (hemo- cantly and independently associated with the risk of globin) have been independently associated with an increased risk of bleeding. 6 major bleeding in patients with AF. Using different combinations of these mainly clinical variables has resulted in at least five validated risk scores for better prediction of bleeding events in patients with AF.7 However, in recent years circulating protein biomarkers have been shown to add substantial incremental information about bleeding risk in patients with AF on OAC treatment. These biomarkers 2 \| M E TH O D S 2.1 \| Patient population encompass markers of oxidative stress, myocardial injury, renal impairment, anemia, inflammation, and coagulation and fibrinolytic ac- 2.1.1 \| Identification cohort tivity.8 To improve the prognostication of major bleeding in patients with AF we recently identified the strongest clinical variables and The Apixaban for Reduction in Stroke and Other Thromboembolic biomarkers associated with the outcome—that is, growth differenti- Events in Atrial Fibrillation (ARISTOTLE) was a double-blind, double- ation factor 15 (GDF-15), cardiac Troponin T (cTnT-hs), and hemoglo- dummy, randomized clinical trial that enrolled 18,201 patients with bin analyzed by conventional methods. These biomarkers together AF and at least one CHADS2 risk factor for stroke or systemic em- with age and history of prior bleeding were included in the novel bolism between December 2006 and April 2010. Exclusion criteria risk score, the ABC bleeding score.9 This bleeding score improved included conditions other than AF that required anticoagulation the discrimination for prediction of major bleeding compared to risk (e.g., prosthetic heart valve) and severe renal insufficiency (serum scores based on clinical variables and performed correct reclassifi- creatinine >2.5 mg/dl [221 µmol/L] or calculated creatinine clear- 9,10 cation of bleeding risk. ance <25 ml/min). Patients were randomized to warfarin (n = 9081) Proximity Extension Assay (PEA; Target 96, Olink Proteomics) or apixaban (n = 9120). The patients were followed for 1 to 3 years is an example of a new analytical high-throughput technology that (median 1.74). Details of the ARISTOTLE trial have previously been allows simultaneous measurements of hundreds of biomarkers in published.4,12 1 μl plasma. This technology is useful for screening multiple protein biomarkers for associations with cardiovascular disease (CVD) and responses to treatment.11 In the present study, we used three Target 2.1.2 \| Replication cohort 96 multiplex immunoassays—C VD II, CVD III, and Inflammation— including 276 proteins, all with proposed involvement in cardiovas- The Randomized Evaluation of Long-Term Anticoagulation Therapy cular disease, metabolism, inflammation, and immunity. Our main (RE-LY) trial was a prospective, multicenter, randomized trial com- purpose was to identify novel prognostic biomarkers and potential paring two blinded doses of dabigatran with open label warfarin pathophysiological processes associated with major bleeding events that enrolled 18,113 patients with AF between December 2005 and in patients with AF treated with OACs. Identification of biomarkers March 2009. Exclusion criteria included severe heart valve disorder, associated with bleeding can help with risk prediction and the devel- recent stroke, creatinine clearance less than 30 ml/min, or active opment of novel therapies that might mitigate the risk of bleeding in liver disease.3,13 Patients were followed for 1 to 3 years (median patients with AF taking OACs. 1.96). 2728 \| In both trials, patients at certain centers participated in the bio- SIEGBAHN et al. 2.2.1 \| Proteomic profiling marker substudies and provided, at randomization, venous blood samples. EDTA plasma was immediately centrifuged and frozen in The proteomic analyses were performed at the Clinical Biomarkers aliquots and stored at –70°C until analyzed centrally at the Uppsala Facility, Science for Life Laboratory, Uppsala University, with- Clinical Research Center (UCR) Laboratory, Uppsala University, out information on any other data. The determinations were per- Uppsala, Sweden. In both trials ethics committee approval was ob- formed with the high-throughput PEA technique using the Target tained for all investigational sites, and all patients provided written 96 Multiplex CVD II96x96, CVD III96x96, and Inflammation96x96 panels informed consent. (Olink Proteomics, www.olink.com/products), which together simultaneously measured 276 selected proteins in plasma potentially related to CVD and inflammation. The PEA technology uses pairs of 2.1.3 \| Outcome assessment antibodies equipped with DNA reporter molecules.11 When binding to their correct targets, antibody pairs give rise to new DNA am- The primary safety endpoint in both the ARISTOTLE and RE-LY plicons each ID-barcoding their respective antigens. The amplicons trials was major bleeding, adapted from the International Society are subsequently quantified using the Fluidigm BioMark™ HD real- on Thrombosis and Haemostasis (ISTH) definition.12–14 A blinded time PCR platform. Inter-plate variability was adjusted by intensity Clinical Events Committee reviewed and centrally adjudicated all normalization with the plate median as the normalization factor. For suspected bleeding events in both studies. Major bleeding was de- data analysis Olink NPX Manager software was used. The results fined as acute or subacute clinically overt bleeding accompanied provide relative values, normalized protein expression (NPX) data, by at least one of the following: (1) decrease in hemoglobin level of which are log2 transformed and one-unit-higher NPX represents a ≥2 g/L; (2) a transfusion of at least two units of packed red blood doubling of the measured protein concentration. The PEA assays cells; and/or (3) fatality or occurrence in a critical area or organ: in- have high reproducibility and repeatability with mean intra-assay tracranial, intraspinal, intraocular, pericardial, intraarticular, intra- and inter-assay coefficients of variation around 8% and 12%, respec- muscular with compartment syndrome, or retroperitoneal. tively, and average inter-site variation at 15%.11 A good correlation between biomarkers analyzed with the PEA technique and analyzed by conventional immunoassays has previ- 2.1.4 \| Multimarker screening study design ously been shown.16 The protein markers in the identification cohort included three panels, CVD II, CVD III, and Inflammation, and are The identification of biomarkers associated with major bleedings was detailed in Tables S1 and S2. Of the 276 PEA proteins, 10 were avail- based on a random sample of 4200 patients from the ARISTOTLE able on more than one panel, resulting in 266 unique markers. As biomarker substudy (N = 14,780 patients with available measure- initial results in the identification cohort identified biomarkers from ments on cTnT-hs, NT-proBNP [N-terminal pro-B-t ype natriuretic the CVD II and CVD III panels as more strongly associated with the peptide], GDF-15, and cystatin C), the identification cohort, includ- outcome major bleeding, the inflammation panel was omitted in the ing 204 patients with and 3996 patients without major bleedings replication subset. during follow-up. Replication was performed by a case-cohort design and included all 344 cases with major bleedings and a random sample of 1024 patients without major bleeding in the RE-LY biomarker substudy (N = 8549 patients with available measurements on cTnT-hs, NT-proBNP, and GDF-15). 2.3 \| Statistical analyses A random survival forest algorithm17 was used to evaluate the simultaneous association between variables and major bleeding. The 2.2 \| Biochemical analyses evaluation included levels of 263 PEA markers, four conventional markers (NT-proBNP, cTnT-hs, GDF-15, and IL-6), renal function (cystatin C in the ARISTOTLE and CKD-EPI in the RE-LY studies, The plasma concentrations of high-sensitivity cTnT-hs, NT-proBNP, respectively), and 13 clinical characteristics (randomized treat- and GDF-15 were determined by Roche immunoassays using a Cobas ment, age, gender, body mass index [BMI], smoking, hypertension, Analytics e601 (Roche Diagnostics). Interleukin 6 (IL-6) was ana- diabetes, hemoglobin, previous myocardial infarction, stroke/tran- lyzed using the high-sensitivity sandwich ELISA immunoassay (R&D sient ischemic attack [TIA], peripheral artery disease, heart failure, Systems Inc.) and Cystatin C with the ARCHITECT system ci8200 and previous bleeding). The total number of biomarkers analyzed (Abbott Laboratories) using the particle-enhanced turbidimetric im- with the random survival forest algorithm was therefore 268 (5 munoassay (PETIA) from Gentian. Estimated glomerular filtration analyzed by conventional analyses + 266 biomarkers analyzed with rate (eGFR) was calculated based on centrally determined creatinine PEA excluding 3 PEA duplicate biomarkers that were analyzed by levels using the CKD-EPI (Chronic Kidney Disease Epidemiology conventional analyses). The number of trees was 5000, splits were Collaboration) equation.15 done according to a maximally selected statistic criterion, and the \| SIEGBAHN et al. 2729 variables were ranked according to their scaled permutation variable using the R environment for statistical computing, version 3.3.119 importance, measured as the change in Harrell’s C-index before and using the ranger package. 20 after permuting the variable of interest. The scaling is done over the trees and not over the variables and therefore does not yield values between 0 and 1. Subjects with all PEA markers missing were excluded, in ARISTOTLE 316 patients. In RE-LY there was none. There were only a few partially missing values and these were singly im- 3 \| R E S U LT S 3.1 \| Baseline characteristics puted using multivariate imputations by chained equations. An identical approach was used in the RE-LY evaluation, with a total of 184 Baseline characteristics of the identification and validation cohorts PEA markers. The largest proportion of missing values for clinical are shown in Tables 1 and 2, respectively. In both studies, patients variables was 0.5% in ARISTOTLE and 2% in RE-LY; for the PEA bio- who experienced major bleeding events during follow-up were at markers 6% in ARISTOTLE and 0.1% in RE-LY. A Boruta analysis was baseline older and had more previous cardiovascular events includ- used to confirm which of the variables in the random survival forest ing prior bleedings. At baseline, these patients also had higher con- 18 In centrations of the cardiovascular biomarkers, NT-proBNP, cTnT-hs, short, the Boruta analysis performs multiple runs of random sur- analysis had a larger than random association with outcomes. and GDF-15. The inflammation marker IL-6 was also higher in this vival forest comparing all variables to random variables, which are group of patients with AF. Slightly lower baseline levels of hemo- shuffled copies of the original variables. Variables performing bet- globin were found in patients with bleeding events during follow-up. ter than the maximum random variable importance are classified as Complete lists of all 266 biomarkers and their relative concentra- confirmed, variables performing worse are rejected, and variables tions (NPX values) in patients with or without bleeding events and that cannot be confirmed or rejected are classified as tentative. In limit of detection (LOD) are presented in Tables S1 and S2 for the the Boruta analysis the number of trees was lowered to 2000 due to identification and replication cohorts, respectively. performance issues and a maximum of 100 random survival forests were run. Biomarkers with a top ranking in the random survival forest analysis and confirmed in the Boruta analysis in both cohorts, 3.2 \| Identification substudy were considered to have confirmed association with the risk of major bleeding. In the ARISTOTLE cohort, 13 clinical variables and 268 biomarkers The pairwise correlation between PEA biomarkers and estab- were analyzed in the random survival forest analysis and the 50 vari- lished conventional biomarkers was assessed by the Spearman ables with highest variable importance are presented in Figure S1A correlation. in supporting information. In this cohort osteopontin (OPN), age, Cox regression analyses were performed including each of the and GDF-15 were identified as having the strongest association with established standard immunoassays (naturally log-transformed) and major bleeding. According to the corresponding Boruta analysis, the PEA biomarkers, one at a time, assuming a linear association with age and the two top biomarkers were followed by 14 biomarkers the log hazard rate. In the validation cohort, weighted Cox regres- which had confirmed or tentatively confirmed importance for pre- sion analyses were done, in which each subject was given a weight dicting bleeding events: cTnT-hs, interleukin-17 receptor A (IL-17RA), corresponding to the reciprocal of the sampling probability. Thus, tumor necrosis factor receptor 1 (TNF-R1), ephrin type-B receptor 4 the cases were given a weight of 1.0 and the randomly sampled con- (EphB4), cystatin C, trefoil factor 3 (TFF3), TNF-R 2, TNF-related ap- trols were given a weight of 1/0.1716528. optosis inducing ligand receptor 2 (TRAIL-R 2), CD 40, soluble uroki- On the inflammation panel, 16 of the proteins had more than nase plasminogen activator receptor (suPAR). spondin-2 (SPON2), 80% of the measurements below the limit of detection and these C-C motif chemokine 17 (CCL17), lymphotoxin-beta receptor (LTBR), were not candidates for inclusion in the Cox regression models. The and osteoprotegerin (OPG) (Figure 1). The linear association with Cox regression analyses were performed in an unadjusted model bleeding events for each of these individual biomarkers were inves- and a model adjusted for baseline clinical characteristics (age, gen- tigated by unadjusted (Figure S2A) and adjusted for clinical charac- der, BMI, smoking, hypertension, diabetes, prior myocardial infarc- teristics (Table S3) Cox regression analyses. tion, prior stroke/TIA, peripheral artery disease, heart failure, prior bleeding, and randomized treatment) and adjusting also for renal function (cystatin-C) and the established markers of bleeding risk 3.3 \| Replication substudy (GDF-15 and cTnT-hs). Results were presented as the relative hazard for an interquartile difference of each marker with corresponding The variable importance for the 50 highest ranked variables in the 95% confidence intervals and P-values. Thus, the hazard ratio can be random survival forest including 12 clinical variables and 186 bio- interpreted as the relative hazard comparing the two biomarker val- markers in the RE-LY cohort are shown in Figure S1B. The random ues defining the inner 50% of the distribution, that is, the third ver- survival forest and Boruta analyses identified age and 28 biomark- sus the first quartile. The incremental discriminative value for each ers with confirmed importance for bleeding, of which 9 biomarkers biomarker was illustrated by the C-index. All analyses were done were also identified in the ARISTOTLE cohort (Figure 2 and Table 3). 2730 \| SIEGBAHN et al. Baseline characteristics Randomly selected patients N = 3996 Cases with major bleeding N = 204 Age 69.0 (62.0–75.0) 74.0 (68.0–79.0) Gender/female 1465 (36.7%) 69 (33.8%) Body mass index (BMI) 28.6 (25.4–32.8) 28.2 (24.5–31.6) Current smoker 362 (9.1%) 11 (5.4%) Hypertension 3505 (87.7%) 172 (84.3%) Diabetes 1001 (25.1%) 57 (27.9%) Prior myocardial infarction 513 (12.8%) 33 (16.2%) Previous PCI/CABG 536 (13.4%) 41 (20.1%) Prior stroke 715 (17.9%) 55 (27.0%) Peripheral arterial disease 185 (4.6%) 17 (8.3%) Heart failure 1254 (31.4%) 53 (26.0%) Prior bleeding 665 (16.6%) 48 (23.5%) Warfarin 1964 (49.1%) 124 (60.8%) NT-proBNP (ng/L) 689.0 (363.0, 1257.0) 827.0 (499.5, 1311.0) cTnT-hs (ng/L) 10.7 (7.4, 16.4) 14.3 (9.2, 20.2) GDF−15 (ng/L) 1353.0 (969.8, 2041.2) 1856.5 (1195.0, 2550.2) Cystatin C (mg/L) 1.0 (0.8, 1.2) 1.1 (0.9, 1.3) eGFR (ml/min) 74.8 (57.5, 96.4) 65.1 (48.8, 82.8) IL−6-hs (ng/L) 2.3 (1.5, 4.0) 2.7 (1.6, 4.6) CRP (mg/L) 2.2 (1.0, 4.7) 2.0 (0.9, 5.5) Hemoglobin (g/dl) 14.2 (13.2, 15.3) 13.7 (12.6, 14.7) TA B L E 1 Baseline characteristics and concentrations of established biomarkers in the identification cohort, the ARISTOTLE trial Note: Continuous variables presented as median (Q1-Q3). Categorial variables are presented as numbers (percentage). Abbreviations: CABG, coronary artery bypass grafting; CRP, C-reactive protein; cTnT-hs, cardiac troponin T measured with a high-sensitivity assay; eGDFR, estimated glomerular filtration rate; GDF-15, growth differentiation factor-15; IL-6, interleukin 6; NT-proBNP, N-terminal pro-B-t ype natriuretic peptide; PCI, percutaneuos coronary intervention. Variables with strongest association to major bleeding events were characteristics—age, gender, BMI, smoking, hypertension, diabetes, age, TNF-R1, TNF-R 2, GDF-15, hemoglobin, interleukin 18 (IL-18) prior myocardial infarction, prior stroke/TIA, peripheral artery dis- binding protein, fibroblast growth factor 23 (FGF-23), TRAIL-R 2, and ease, heart failure, prior bleeding, and randomized treatment. The troponin (cTnT-hs). associations of the nine prognostic biomarkers with bleeding by The linear associations between these biomarkers and major splines are shown in Figure S3A,B. bleeding by unadjusted (Figure S2B) and adjusted for clinical variables Cox regression analyses are shown in Table S4. The correlation of the nine identified biomarkers and established cardiovascular biomarkers, NT-proBNP, cTnT-hs, GDF-15, and renal function (cystatin C) are shown in Table 4. TNF-R1, TNF-R 2 and 3.4 \| Biomarkers associated with bleeding in both cohorts According to the random survival forest and Boruta analyses the following nine biomarkers were confirmed in both cohorts: GDF- TRAIL-R 2, GDF-15, EphB4, and suPAR were moderately correlated with renal function (cystatin C) in both studies, rho ≥0.5. TNF-R1, TRAIL-R 2, and suPAR also correlated with GDF-15 in both studies. 4 \| DISCUSSION 15, cTnT-hs, OPN, EphB4, TNF-R1, TNF-R 2, suPAR, TRAIL-R 2, and OPG (Table 3). The linear associations between the identified nine In this study we used a targeted proteomic approach to identify and biomarkers and major bleeding by unadjusted and adjusted Cox re- confirm new biomarkers indicating risk of bleeding based on one gression analyses in both cohorts are shown in Table 3. The C-index random sample and one case-cohort in patients with AF treated varied in the ARISTOTLE cohort between 0.685 and 0.715 and in the with oral anticoagulation. Within each cohort, the strength of the RE-LY cohort between 0.668 and 0.694 by adding one of the nine biomarkers’ association with the risk of bleeding were ranked ac- identified biomarkers to the baseline model including the clinical cording to two different principles, each capturing different aspects \| SIEGBAHN et al. TA B L E 2 Baseline characteristics and concentrations of established biomarkers in the validation cohort, the RE-LY trial Baseline characteristics Randomly selected patients N = 1024 Cases with major bleeding N = 344 Age 72.0 (66.0–77.0) 76.0 (71.0–8 0.0) Gender/female 383 (37.4%) 115 (33.4%) Body mass index (BMI) 27.9 (25.0–31.2) 27.3 (24.3–3 0.9) Current smoker 80 (7.8%) 22 (6.4%) Hypertension 814 (79.5%) 279 (81.1%) Diabetes 212 (20.7%) 100 (29.1%) Prior myocardial infarction 172 (16.8%) 74 (21.5%) Prior stroke 204 (19.9%) 71 (20.6%) Peripheral arterial disease 36 (3.5%) 20 (5.8%) Heart failure 298 (29.1%) 99 (28.8%) Warfarin 343 (33.5%) 117 (34.0%) 2731 NT-proBNP (ng/L) 816.0 (382.2, 1452.5) 935.0 (468.8, 1705.8) cTnT-hs (ng/L) 11.9 (7.6, 19.0) 16.9 (10.7, 28.3) GDF−15 (ng/L) 1455.0 (1083, 2122.0) 2056.5 (1434.2, 3055.5) Cystatin C (mg/L) 1.0 (0.8, 1.2) 1.1 (0.9, 1.4) eGFR (ml/min) 65.2 (54.3, 74.2) 59.7 (49.0, 70.2) IL−6-hs (ng/L) 2.3 (1.4, 3.8) 2.8 (1.8, 5.1) CRP (mg/L) 2.5 (1.2, 5.4) 3.2 (1.4, 7.5) Hemoglobin (g/dl) 14.3 (13.2, 15.3) 13.4 (12.4, 14.7) Note: Continuous variables presented as median (Q1-Q3). Categorial variables are presented as numbers (percentage). Abbreviations: CABG, coronary artery bypass grafting; CRP, C-reactive protein; cTnT-hs, cardiac troponin T measured with a high-sensitivity assay; eGDFR, estimated glomerular filtration rate; GDF-15, growth differentiation factor-15; IL-6, interleukin 6; NT-proBNP, N-terminal pro-B-t ype natriuretic peptide; PCI, percutaneuos coronary intervention. of the association. The random survival forest and Boruta algorithms cardiovascular disease. 23 It is a marker of cellular aging and inflam- evaluated the biomarkers simultaneously while the Cox regression matory activity, and a major risk indicator of hemorrhages in patients analyses evaluated the biomarkers individually. We confirmed that with AF treated with OAC and in patients with acute coronary syn- the biomarkers GDF-15 and cTnT are strongly and independently drome (ACS). 21,22,24 The underlying mechanism of the association of related to the risk of major bleeding in these patients. 8,9,21,22 In ad- GDF-15 and risk of major bleeding is not fully understood and has dition, we were also able to identify seven novel biomarkers—OPN, yet to be revealed. Beyond its inflammatory activity, one possible OPG, TNF-R1, TNF-R 2, TRAIL-R 2, EphB4, and suPAR—out of 268 process leading to enhanced bleeding risk might be that upon cellu- investigated proteins to be significantly and independently associ- lar stress and tissue damage, such as vascular vulnerability, GDF-15 ated with the risk of bleeding. These nine novel biomarkers repre- is secreted. sent different pathophysiological processes and warrant further In the present study OPN and TNF-R1 were strong predictors of investigations concerning their utility for assessment of bleeding major bleeding in both cohorts. OPN is a secreted multifunctional risk and identification of new therapeutic targets that might prevent glucophosphoprotein that plays major roles in physiological as well as bleeding in patients with AF treated with OAC. in pathophysiological processes. 25 Several vascular cell types, such as monocytes/macrophages, fibroblasts, endothelial cells, vascular 4.1 \| Identified biomarkers and pathophysiological implications smooth muscle cells, and myocytes, upregulate and secrete OPN. 26 OPN has also been suggested to be involved in and serve as a biomarker for vascular calcification. 27,28 Similar to OPN, OPG, another member of the TNF receptor superfamily, is a modulator of vascular The identified biomarkers represent a broad array of pathophysi- calcification and correlated with coronary calcium scores in patients ologic processes. We attempt, therefore, to put our findings into with acute coronary disease, which is a risk factor for non–coronary perspective of these pathways. artery bypass grafting--related major bleeding in ACS. 29,30 In pa- The cytokine GDF-15 is secreted by a broad range of cells upon hypoxia and oxidative stress and is strongly associated with tients with ACS treated with dual antiplatelet therapy, OPG was recently identified as an independent biomarker of bleeding events.31 2732 \| SIEGBAHN et al. F I G U R E 1 Variable importance in the ARISTOTLE cohort. Boruta analysis in the identification, ARISTOTLE, cohort of the significance of variable importance for ISTH major bleeding in the random survival forest analysis including clinical variables, biomarkers assayed by conventional methods and by Proximity Extension Assay. Color coding according to the Boruta analysis: green = confirmed, yellow = tentative, and red = rejected Not much attention has been focused on the role of OPN in levels are independently associated with higher risk of renal disease AF. However, very recently, OPN was found to induce atrial fibro- progression, cardiovascular events, and mortality in patients with sis32 and to be significantly associated with incident AF.33 It is also diabetes.38 In a large population-based cohort, individuals with high strongly associated with future ischemic stroke in patients with AF levels of sTNF-R1 had increased risk of myocardial infarction, cardio- during anticoagulant treatment. 34 Recently, it was shown that hu- vascular death, and stroke, and the plasma levels were suggested as mans express multiple OPN isoforms that have different functional surrogate markers of arterial injury.39 The TNF alpha/TNF-R1 sys- effects. 35 However, the importance of these isoforms in vascular pathophysiology, CVD, and AF is still elusive. tem also induces release of endothelial cell-derived extracellular vesicles, which participate in vascular damage by contributing to The strong association of TNF-R1 and bleeding risk assessment apoptosis and promoting inflammation of the endothelial cells.40 in patients with AF treated with OAC identified in the present study Another so-called death receptor, TRAIL-R 2 has been attributed to has not been previously reported. TNF-alpha binds mainly to TNF- be the main apoptosis-inducing receptor.41 It induces apoptosis in R1, which belongs to the TNF receptor superfamily and is expressed macrophages, vascular smooth muscle cells, and endothelial cells on all tissues in the body including cardiomyocytes. 36 Upon activa- tion of TNF-R1 the induced signaling pathways mainly result in apoptosis and necrosis. 37 and inflammatory activity of atherosclerotic plaques, and thereby contributes to the plaque vulnerability phenotype.42 Protease cleavage by ADAM17 of the receptor The association between OPN and bleeding risk found in our releases the soluble form (sTNF-R1) and increased sTNF-R1 plasma two large cohorts of patients with AF may reflect effects of calcified \| SIEGBAHN et al. 2733 F I G U R E 2 Variable importance in the RE-LY cohort. Boruta analysis in the replication, RE-LY, cohort of the significance of variable importance for ISTH major bleeding in the random survival forest analysis including clinical variables, biomarkers assayed by conventional methods and by Proximity Extension Assay. Color coding according to the Boruta analysis: green = confirmed, yellow = tentative, and red = rejected arteries on bleeding rather than the arrhythmia. Infiltration of calcium Matrix metalloproteinases (MMPs) and members of the disin- into the vessel wall may lead to weakening of the wall by reducing tegrin and metalloproteinase (ADAM) family, predictors of incident compensatory compliance and the biological and mechanical hemo- AF, participate in the regulation of Eph functions and signaling upon static capacity of the vessel. Thus, this may limit vascular ability to Eph–ephrin interaction often results in cell detachment and repul- appropriately respond to trauma and thereby predispose the vessel to sion.34,47,48 We have previously described that members of the Eph 43 bleeding. Activation of the TNF alpha/TNF-R1 system and TRAIL-R2 RTK family, including EphB4, can be proteolytically cleaved in their can induce endothelial cell damage and apoptosis of the vessel wall, ectodomain by tissue factor/factor VIIa, the main initiator of blood which provides a possible mechanism of bleeding propensity. coagulation, leading to cell repulsion.49 Thus, the Eph family of RTKs EphB4 is a member of the Eph tyrosine kinase receptor, family are novel co-receptors and proteolytical substrates of the coagula- which constitutes the largest receptor tyrosine kinase (RTK) family tion system with consequences on cellular functions beyond blood in the genome.44 The Eph–ephrin signaling acts as a global cell posi- coagulation.49 The activation of blood coagulation in AF might lead tioning system and the EphB4/EphrinB2 system plays an important to proteolytic cleavage of EphB4 and a soluble form measured in role in vascular development and angiogenesis. 45 The Eph receptor– plasma. To our knowledge, the present study is the first to show an ephrin signaling is also involved in inflammatory processes such as association between EphB4 and AF including bleeding events in AF monocyte adhesion, transmigration through vascular endothelium, patients. Whether EphB4 is merely a specific biomarker of bleeding and atherosclerotic plaque development.46 propensity reflecting increased expression on vascular cells in AF or 2734 \| SIEGBAHN et al. TA B L E 3 Associations between ISTH major bleeding and biomarkers with confirmed importance in the random survival forest analyses in both studies and results from adjusted Cox regression analyses ARISTOTLE Hazard ratio (95% CI) RE-LY Hazard ratio (95% CI) Biomarker Unadjusted Adjusted Unadjusted Adjusted OPN 1.988 (1.667, 2.372) 1.518 (1.221, 1.867) 1.777 (1.525, 2.071) 1.265 (1.046, 1.529) GDF−15 1.731 (1.496, 2.001) 1.195 (0.925, 1.544) 2.056 (1.772, 2.385) 1.574 (1.293, 1.915) EphB4 1.378 (1.220, 1.557) 1.153 (1.014, 1.310) 1.968 (1.683, 2.302) 1.380 (1.122, 1.698) TNF-R1 1.894 (1.605, 2.236) 1.372 (1.042, 1.806) 2.152 (1.838, 2.521) 1.614 (1.272, 2.049) cTnT-hs 1.596 (1.404, 1.814) 1.137 (0.914, 1.415) 1.457 (1.284, 1.654) 1.128 (0.966, 1.317) TNF-R 2 1.765 (1.490, 2.090) 1.231 (0.968, 1.567) 1.494 (1.255, 1.778) 1.231 (1.017, 1.492) suPAR 1.777 (1.493, 2.116) 1.317 (1.023, 1.700) 1.830 (1.566, 2.137) 1.288 (1.039, 1.598) TRAIL-R 2 1.307 (1.191, 1.436) 1.085 (0.949, 1.240) 1.290 (1.163, 1.432) 1.143 (1.009, 1.294) OPG 1.650 (1.388, 1.963) 1.161 (0.937, 1.440) 1.793 (1.535, 2.093) 1.205 (0.986, 1.473) Notes: The hazard ratios correspond to a comparison of the third and the first sample quartiles or, since the association is assumed linear, an interquartile difference. The Cox regression model is adjusted for clinical characteristics—age, gender, BMI, smoking, hypertension, diabetes, prior MI, prior stroke/TIA, prior PAD, prior HF, prior bleeding and randomized treatment, renal function, cystatin C, and the biomarkers GDF-15 and cTnT-hs. Abbreviations: BMI, body mass index; CI, confidence interval; cTnT-hs, cardiac troponin T measured with a high-sensitivity assay; EphB4, ephrin type-B receptor 4; GDF-15, growth differentiation factor-15; HF, heart failure; MI, myocardial infarction; OPG, osteoprotegerin; OPN, osteopontin; PAD, peripheral artery disease; RF, random survival forest; suPAR, soluble urokinase plasminogen activator receptor; TIA, transient ischemic attack; TNF-R , tumor necrosis factor receptor; TRAIL-R 2, TNF-related apoptosis inducing ligand receptor 2. GDF−15 NT-proBNP cTnT-hs Cystatin C Biomarker ARI RE-LY ARI RE-LY ARI RE-LY ARI RE-LY GDF−15 1 1 0.35 0.37 0.49 0.44 0.52 0.49 cTnT-hs 0.49 0.44 0.38 0.40 1 1 0.48 0.36 TNFR−1 0.55 0.55 0.29 0.28 0.45 0.40 0.64 0.61 OPN 0.40 0.35 0.28 0.26 0.34 0.32 0.37 0.33 suPAR 0.53 0.50 0.33 0.32 0.40 0.33 0.55 0.48 EphB4 0.42 0.41 0.22 0.19 0.38 0.29 0.51 0.49 TNF-R 2 0.49 0.48 0.22 0.23 0.39 0.31 0.59 0.56 TRAIL-R 2 0.63 0.60 0.39 0.41 0.49 0.42 0.64 0.60 OPG 0.44 0.41 0.25 0.28 0.33 0.34 0.29 0.16 TA B L E 4 Spearman correlations between the significant PEA biomarkers and established biomarkers presented from in the identification, ARISTOTLE, and replications RE-LY cohorts Abbreviations: cTnT-hs, cardiac troponin T measured with a high-sensitivity assay; EphB4, ephrin type-B receptor 4; GDF-15, growth differentiation factor-15; NT-proBNP, N-terminal pro-B-t ype natriuretic peptide; PEA, Proximity Extension Assay; OPG, osteoprotegerin; OPN, osteopontin; suPAR, soluble urokinase plasminogen activator receptor; TNF-R , tumor necrosis factor receptor; TRAIL-R 2, TNF-related apoptosis inducing ligand receptor 2. whether it, considering its complex biology, plays a causative role in clots and degradation of blood clotting factors and extracellular AF remains to be elucidated. matrix.53,54 Inflammatory cytokines cleave uPAR from the cell sur- The inflammatory biomarker suPAR is involved in the devel- face into a soluble form that is engaged in innate immune responses opment of atherosclerosis and is associated with the presence through regulation of cell adhesion and migration.54 An elevated and severity of coronary artery disease, cardiovascular death, and plasma level of suPAR has also been suggested as a biomarker of myocardial infarction.50,51 uPAR is expressed on a variety of cells, heart failure and of incident AF.55,56 In the present study suPAR including leukocytes and endothelial cells, and localizes active uPA was found to have a strong prognostic value in prediction of future to the cell surface, which forms an extracellular proteolytic enzyme bleeding events in patients with AF on OAC, confirming the modu- system.52 The uPA system cleaves plasminogen into plasmin and latory role of fibrinolysis on the hemostatic balance. These results thereby plays a mandatory role in fibrinolysis by dissolution of fibrin are in accordance with previous studies, which clearly demonstrated \| SIEGBAHN et al. 2735 that elevated D-dimer levels, reflected as a marker of fibrinolysis, of bleeding events, with 204 in the identification cohort and 344 in 57 the replication cohort. The PEA inflammation panel was not applied GDF-15, suPAR, EphB4, and the death receptors were moderately in the validation cohort. These differences may have influenced the correlated with cystatin C. Renal impairment is a well-known risk results to some degree. are associated with increased bleeding events in AF patients. factor for CVD and thromboembolism but also bleeding propensity, The results in this study were obtained in individuals with AF en- which has been attributed to an imbalance in the coagulation sys- rolled in clinical trials evaluating antithrombotic therapy, which may tem.58,59 suPAR, TNF-R1, and TRAIL-R 2 also correlated moderately limit external generalizability. The biomarker assays used in the PEA technique result in val- with GDF-15, thus further indicating that inflammatory as well as coagulation processes and endothelial dysfunction are involved. ues expressed in relative units, NPX values. The PEA technique is This is the first study to report a protein multimarker approach excellent for screening purposes but for clinical applications abso- in screening of new biomarkers of importance for bleeding events lute concentrations will be required. Prior validation studies have, in patients with AF and OAC. The identification of nine biomarkers however, clearly demonstrated that biomarkers analyzed with the independently associated with bleeding confirms the hypothesis that PEA technique have good further investigation concordance with molecular biomarkers can be potentially useful to indicate the under- established immunoassays.16 lying biology and pathways of importance for adverse events in cardiovascular disease, including AF. A multimarker screening may also identify future targets for therapies to reduce bleeding or prevent 5 \| CO N C LU S I O N S thromboembolism without increasing bleeding. Important molecular mechanisms by which these biomarkers may confer bleeding risk are In two well-characterized clinical trial cohorts of patients with AF on cellular aging, inflammation and innate immune responses, vascular OAC we confirmed, out of 268 screened biomarkers, the importance remodeling and calcification, endothelial cell damage, apoptosis, fi- of GDF-15 and cTnT-hs and identified seven novel biomarkers that brosis, coagulation activity, and fibrinolysis. The exact mechanisms contribute to prognostication of the risk of major bleedings. These remain, however, elusive. At least OPN and suPAR are not only mark- biomarkers showed association with inflammatory processes (GDF- ers of bleeding, but also markers of specific pathological processes, 15, OPN, OPG, TNF-R1, TNF-R 2, TRAIL-R 2, EphB4, and suPAR), vas- such as fibrosis in the heart, underlying incident AF. Impaired vessel cular remodeling and calcification (GDF-15, OPN, OPG), endothelial walls in combination with decreased coagulation propensity, due to cell damage and cell survival (TNF-R1, TNF-R 2, TRAIL-R 2), coagula- OAC, and activation of fibrinolysis in patients with AF at older age tion (EphB4), and fibrinolysis (suPAR). Further in-depth studies are might be one plausible explanation for these biomarkers as indicators required to elucidate the underlying biological processes for the risk of future bleeding. In the clinical setting GDF-15 has already clearly of major bleeding in AF. Whether any individual or a combination of improved the risk prediction beyond clinical risk factors. Whether the these novel biomarkers might improve prognostication of bleeding newly identified biomarkers may add clinical value to the assessment in the clinical setting and guide future therapies that may reduce of major bleeding risks in patients with AF needs to be validated. bleeding warrant further investigations using quantitative assays in these and other cohorts. 4.1.1 \| Strengths and limitations C O N FL I C T S O F I N T E R E S T Dr. Siegbahn: institutional research grants from AstraZeneca, The strength of this study is that it is based on cross-validation of Boehringer Ingelheim, Bristol-Myers Squibb/Pfizer, GlaxoSmithKline, proteomic findings in two large materials from two clinical cohorts Roche Diagostics and consultancy fees from Olink Proteomics. Mr. with complete follow-up and independent ascertainments of all Lindbäck: institutional research grants from Boehringer Ingelheim, outcome events. Considering the problem of multiplicity and mass- Bristol-Myers Squibb/Pfizer. Dr. Hijazi: lecture fees from Boehringer significance when evaluating the 268 biomarkers, it was alleviated Ingelheim, Roche, Bristol-Myers Squibb, and Pfizer; consulting by using a random survival forest and Boruta algorithm, as well as fees from Merck Sharp & Dohme, Roche, Bristol-Myers Squibb, comparing the results in two independent cohorts. We based our and Pfizer. Dr. Åberg: none. Dr. Alexander: institutional research selection of important biomarkers on a non-linear machine learning grants from Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, approach using the same criteria as we have previously published.60 Cryolife, CSL Behring, Ferring, Glaxosmithkline, and XaTek and The random survival forest and Boruta algorithms allow for non- consulting fees/honoraria from AbbVie, Bristol-Myers Squibb, linear associations and complex interactions among the variables. In Cryolife, Glaxosmithkline, Pfizer, and Portola. Dr. Eikelboom: hon- the Cox regression analyses, on the other hand, we assumed a linear oraria and institutional research grants from AstraZeneca, Bayer, association between the log relative hazard of major bleedings and Boehringer Ingelheim, Bristol-Myers Squibb/Pfizer, Daiichi Sankyo, each marker one at a time, making it possible to estimate average GlaxoSmithKline, Janssen, Sanofi Aventis, and Eli Lilly, as well as a adjusted hazard ratios in a more conventional way. personnel award from the Heart and Stroke Foundation. Dr. Lopes: There were differences between the two cohorts in regard to institutional research grant and consulting fees from Bristol-Myers study design, random sample versus case cohort, and to the number Squibb; institutional research grant from GlaxoSmithKline; consulting 2736 \| SIEGBAHN et al. fees from Bayer, Boehringer Ingleheim, Pfizer, Merck, Portola. Dr. Pol: none. Dr. Oldgren: consulting and lecture fees from Boehringer Ingelheim, Bayer, Bristol-Myers Squibb, Pfizer. Dr. Granger: grants 11. and personal fees from GlaxoSmithKline, Boehringer Ingelheim, Bristol-Myers Squibb, Pfizer, Sanofi-Aventis, Takeda, The Medicines 12. Company, Janssen, Bayer, Hoffmann-L a Roche; grants from Medtronics Foundation, Merck & Co., Armetheon; personal fees from Lilly, AstraZeneca, Daiichi Sankyo, Ross Medical Corporation, Salix Pharmaceuticals, Gilead. Dr. Yusuf: grants, speaker fees, and 13. travel support from Boehringer Ingelheim. Dr. Wallentin: institutional research grants, consultancy fees, lecture fees, and travel support from Bristol-Myers Squibb/Pfizer, AstraZeneca, GlaxoSmithKline, 14. Boehringer Ingelheim; institutional research grants from Merck & Co, Roche; consultancy fees from Abbott; holds two patents involv- 15. ing GDF-15. AU T H O R C O N T R I B U T I O N S 16. AS and LW conceived the research and planned the design of the study. AS, JL, ZH, TP, JO, and LW performed formal analysis of the study results. JL performed the statistical analyses. AS wrote the manuscript. All coauthors critically revised the manuscript. All coauthors have approved the final manuscript for submission. 17. 18. 19. ORCID Mikael Åberg https://orcid.org/0000-0002-7858-8233 REFERENCES 1. Kannel WB, Wolf PA, Benjamin EJ, Levy D. Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based estimates. Am J Cardiol. 1998;82:2n-9n. 2. Kirchhof P, Breithardt G, Aliot E, et al. Personalized management of atrial fibrillation: proceedings from the fourth atrial fibrillation competence NETwork/European heart rhythm association consensus conference. Europace. 2013;15:1540-1556. 3. Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361:1139-1151. 4. Granger CB, Alexander JH, McMurray JJ, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2011;365:981-992. 5. Berg DD, Ruff CT, Morrow DA. Biomarkers for risk assessment in atrial fibrillation. Clin Chem. 2021;67:87-95. 6. Hylek EM, Held C, Alexander JH, et al. Major bleeding in patients with atrial fibrillation receiving apixaban or warfarin: the ARISTOTLE trial (apixaban for reduction in stroke and other thromboembolic events in atrial fibrillation): predictors, characteristics, and clinical outcomes. J Am Coll Cardiol. 2014;63:2141-2147. 7. Lane DA, Lip GYH. Stroke and bleeding risk stratification in atrial fibrillation: a critical appraisal. Eur Heart J Suppl. 2020;22:O14-O27. 8. Hijazi Z, Oldgren J, Siegbahn A, Wallentin L. Application of biomarkers for risk stratification in patients with atrial fibrillation. Clin Chem. 2017;63:152-164. 9. Hijazi Z, Oldgren J, Lindbäck J, et al. The novel biomarker-based ABC (age, biomarkers, clinical history)-bleeding risk score for patients with atrial fibrillation: a derivation and validation study. Lancet. 2016;387:2302-2311. 10. Berg DD, Ruff CT, Jarolim P, et al. Performance of the ABC scores for assessing the risk of stroke or systemic embolism and bleeding in 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. patients with atrial fibrillation in ENGAGE AF-TIMI 48. Circulation. 2019;139:760-771. Assarsson E, Lundberg M, Holmquist G, et al. Homogenous 96-plex PEA immunoassay exhibiting high sensitivity, specificity, and excellent scalability. PLoS One. 2014;9:e95192. Lopes RD, Alexander JH, Al-Khatib SM, et al. Apixaban for reduction in stroke and other thromboemboLic events in atrial fibrillation (ARISTOTLE) trial: design and rationale. Am Heart J. 2010;159:331-339. Ezekowitz MD, Connolly S, Parekh A, et al. Rationale and design of RE-LY: randomized evaluation of long-term anticoagulant therapy, warfarin, compared with dabigatran. Am Heart J. 2009;157:805- 810, 10 e1–2. Schulman S, Kearon C. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost. 2005;3:692-694. Hijazi Z, Hohnloser SH, Andersson U, et al. Efficacy and safety of apixaban compared with warfarin in patients with atrial fibrillation in relation to renal function over time: insights from the ARISTOTLE randomized clinical trial. JAMA Cardiol. 2016;1:451-460. Siegbahn A, Eriksson S, Lindbäck J, Wallentin L. A comparison of the proximity extension assay with established immunoassays. Science. 2017:22-25. Sponsored supplement. Ishwaran H, Kogalur UB, Blackstone EH, Lauer MS. Random survival forests. Ann App Stat. 2008;2:841-860. Kursa MB. Robustness of random forest-based gene selection methods. BMC Bioinformatics. 2014;15:8. R_Core_Team. A language and environment for statistical computing. R Foundation for Statistical Computing. 2017. Wright MN, Ziegler A. Ranger: a fast implementation of random forests for high dimensional data in C++ and R. J Stat Softw. 2017;77:1-17. Hijazi Z, Oldgren J, Andersson U, et al. Growth-differentiation factor 15 and risk of major bleeding in atrial fibrillation: insights from the randomized evaluation of long-term anticoagulation therapy (RE-LY) trial. Am Heart J. 2017;190:94-103. Wallentin L, Hijazi Z, Andersson U, et al. Growth differentiation factor 15, a marker of oxidative stress and inflammation, for risk assessment in patients with atrial fibrillation: insights from the apixaban for reduction in stroke and other thromboembolic events in atrial fibrillation (ARISTOTLE) trial. Circulation. 2014;130:1847-1858. Wollert KC, Kempf T, Wallentin L. Growth differentiation factor 15 as a biomarker in cardiovascular disease. Clin Chem. 2017;63:140-151. Hagström E, James SK, Bertilsson M, et al. Growth differentiation factor-15 level predicts major bleeding and cardiovascular events in patients with acute coronary syndromes: results from the PLATO study. Eur Heart J. 2016;37:1325-1333. Scatena M, Liaw L, Giachelli CM. Osteopontin: a multifunctional molecule regulating chronic inflammation and vascular disease. Arterioscler Thromb Vasc Biol. 2007;27:2302-2309. O'Brien ER, Garvin MR, Stewart DK, et al. Osteopontin is synthesized by macrophage, smooth muscle, and endothelial cells in primary and restenotic human coronary atherosclerotic plaques. Arterioscler Thromb. 1994;14:1648-1656. Berezin AE, Kremzer AA. Circulating osteopontin as a marker of early coronary vascular calcification in type two diabetes mellitus patients with known asymptomatic coronary artery disease. Atherosclerosis. 2013;229:475-481. Schoenhagen P. Osteopontin, coronary calcification, and cardiovascular events: future diagnostic and therapeutic targets for disease prevention? Eur Heart J. 2006;27:766-767. Tousoulis D, Siasos G, Maniatis K, et al. Novel biomarkers assessing the calcium deposition in coronary artery disease. Curr Med Chem. 2012;19:901-920. \| SIEGBAHN et al. 30. Généreux P, Madhavan MV, Mintz GS, et al. Relation between coronary calcium and major bleeding after percutaneous coronary intervention in acute coronary syndromes (from the acute catheterization and urgent intervention triage strategy and harmonizing outcomes with revascularization and stents in acute myocardial infarction trials). Am J Cardiol. 2014;113:930-935. 31. Ueland T, Åkerblom A, Ghukasyan T, et al. Osteoprotegerin Is associated with major bleeding but not with cardiovascular outcomes in patients with acute coronary syndromes: insights from the PLATO (platelet inhibition and patient outcomes) trial. J Am Heart Assoc. 2018;7:e007009. 32. Lin R, Wu S, Zhu D, Qin M, Liu X. Osteopontin induces atrial fibrosis by activating Akt/GSK-3beta/beta-c atenin pathway and suppressing autophagy. Life Sci. 2020;245:117328. 33. Molvin J, Jujic A, Melander O, et al. Exploration of pathophysiological pathways for incident atrial fibrillation using a multiplex proteomic chip. Open Heart. 2020;7:e001190. 34. Hijazi Z, Wallentin L, Lindbäck J, et al. Screening of multiple biomarkers associated with ischemic stroke in atrial fibrillation. J Am Heart Assoc. 2020;9:e018984. 35. Cabiati M, Svezia B, Matteucci M, et al. Myocardial expression analysis of osteopontin and its splice variants in patients affected by end-stage idiopathic or ischemic dilated cardiomyopathy. PLoS One. 2016;11:e0160110. 36. Idriss HT, Naismith JH. TNF alpha and the TNF receptor superfamily: structure-function relationship(s). Microsc Res Tech. 2000;50:184-195. 37. Holbrook J, Lara-Reyna S, Jarosz-Griffiths H, McDermott M. Tumour necrosis factor signalling in health and disease. F1000Res. 2019;8:F1000. Faculty Rev-111. 38. Carlsson AC, Ruge T, Kjøller E, et al. 10-year associations between tumor necrosis factor receptors 1 and 2 and cardiovascular events in patients with stable coronary heart disease: a CLARICOR (effect of clarithromycin on mortality and morbidity in patients with ischemic heart disease) trial substudy. J Am Heart Assoc. 2018;7:e008299. 39. Mattisson IY, Björkbacka H, Wigren M, et al. Elevated markers of death receptor-activated apoptosis are associated with increased risk for development of diabetes and cardiovascular disease. EBioMedicine. 2017;26:187-197. 40. Lee SK, Yang SH, Kwon I, Lee OH, Heo JH. Role of tumour necrosis factor receptor-1 and nuclear factor-κB in production of TNF- α-induced pro-inflammatory microparticles in endothelial cells. Thromb Haemost. 2014;112:580-588. 41. Wang S, El-Deiry WS. TRAIL and apoptosis induction by TNF- family death receptors. Oncogene. 2003;22:8628-8633. 42. Gonçalves I, Singh P, Tengryd C, et al. sTRAIL-R 2 (soluble TNF [tumor necrosis factor]-related apoptosis-inducing ligand receptor 2) a marker of plaque cell apoptosis and cardiovascular events. Stroke. 2019;50:1989-1996. 43. Johnson RC, Leopold JA, Loscalzo J. Vascular calcification: pathobiological mechanisms and clinical implications. Circ Res. 2006;99:1044-1059. 44. Pasquale EB. Eph-ephrin bidirectional signaling in physiology and disease. Cell. 2008;133:38-52. 45. Salvucci O, Tosato G. Essential roles of EphB receptors and EphrinB ligands in endothelial cell function and angiogenesis. Adv Cancer Res. 2012;114:21-57. 46. Funk SD, Orr AW. Ephs and ephrins resurface in inflammation, immunity, and atherosclerosis. Pharmacol Res. 2013;67:42-52. 47. Atapattu L, Lackmann M, Janes PW. The role of proteases in regulating Eph/ephrin signaling. Cell Adh Migr. 2014;8:294-3 07. 2737 48. Ravassa S, Ballesteros G, López B, et al. Combination of circulating type I collagen-related biomarkers is associated with atrial fibrillation. J Am Coll Cardiol. 2019;73:1398-1410. 49. Eriksson O, Ramström M, Hörnaeus K, Bergquist J, Mokhtari D, Siegbahn A. The Eph tyrosine kinase receptors EphB2 and EphA2 are novel proteolytic substrates of tissue factor/coagulation factor VIIa. J Biol Chem. 2014;289:32379-32391. 50. Fuhrman B. The urokinase system in the pathogenesis of atherosclerosis. Atherosclerosis. 2012;222:8-14. 51. Nikorowitsch J, Borchardt T, Appelbaum S, et al. Cardio-renal biomarker soluble urokinase-t ype plasminogen activator receptor is associated with cardiovascular death and myocardial infarction in patients with coronary artery disease independent of troponin, C-reactive protein, and renal function. J Am Heart Assoc. 2020;9:e015452. 52. Mahmood N, Mihalcioiu C, Rabbani SA. Multifaceted role of the urokinase-t ype plasminogen activator (uPA) and its receptor (uPAR): diagnostic, prognostic, and therapeutic applications. Front Oncol. 2018;8:24. 53. Plesner T, Behrendt N, Ploug M. Structure, function and expression on blood and bone marrow cells of the urokinase-t ype plasminogen activator receptor, uPAR. Stem Cells. 1997;15:398-4 08. 54. Mondino A, Blasi F. uPA and uPAR in fibrinolysis, immunity and pathology. Trends Immunol. 2004;25:450-455. 55. Borné Y, Persson M, Melander O, Smith JG, Engström G. Increased plasma level of soluble urokinase plasminogen activator receptor is associated with incidence of heart failure but not atrial fibrillation. Eur J Heart Fail. 2014;16:377-383. 56. Westin O, Rasmussen LJH, Andersen O, Buch E, Olsen JE, Friberg J. Soluble urokinase plasminogen activator receptor (suPAR) as a predictor of incident atrial fibrillation. J Atr Fibrillation. 2018;10:1801. 57. Christersson C, Wallentin L, Andersson U, et al. D-dimer and risk of thromboembolic and bleeding events in patients with atrial fibrillation–observations from the ARISTOTLE trial. J Thromb Haemost. 2014;12:1401-1412. 58. Olesen JB, Lip GY, Kamper AL, et al. Stroke and bleeding in atrial fibrillation with chronic kidney disease. N Engl J Med. 2012;367:625-635. 59. Tran L, Pannier B, Lacolley P, et al. A case-control study indicates that coagulation imbalance is associated with arteriosclerosis and markers of endothelial dysfunction in kidney failure. Kidney Int. 2020;S0085–2538(20):31530-31531. 60. Wallentin L, Eriksson N, Olszowka M, et al. Plasma proteins associated with cardiovascular death in patients with chronic coronary heart disease: a retrospective study. PLoS Med. 2021;18:e1003513. S U P P O R T I N G I N FO R M AT I O N Additional supporting information may be found online in the Supporting Information section. How to cite this article: Siegbahn A, Lindbäck J, Hijazi Z, et al. Multiplex protein screening of biomarkers associated with major bleeding in patients with atrial fibrillation treated with oral anticoagulation. J Thromb Haemost. 2021;19:2726– 2737. https://doi.org/10.1111/jth.15498
https://openalex.org/W2725740837	http://rbep.inep.gov.br/ojs3/index.php/emaberto/article/download/3182/2917	Portuguese	null	Políticas públicas para formação de professores: Pibid, Mestrados Profissionais e PNEM	Em Aberto	2,019	cc-by	7,952	Políticas públicas para formação de professores: Pibid, mestrados profissionais e PNEM Marinalva Vieira Barbosa Natália A. Morato Fernandes 23 Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 Abstract Abstract Public policies for teacher training: Pibid, professional master´s degree and PNEM This article aims to analyze the conceptions, the teacher profile, the dialogues and the contradictions that guide the current formative policies; based on the documents which created the Programa Institucional de Bolsa de Iniciação à Docência (Pibid), professional master’s degrees and the Pacto Nacional pelo Fortalecimento do Ensino Médio (PNEM). Specifically, it is interesting to uncover what teacher’s profile these policies seek to build and whether they are related to each other and to the goals of the Plano Nacional de Educação (PNE 2015-2024). Keywords: policies; education; teaching; conceptions. Resumo O objetivo deste artigo é analisar as concepções, o perfil docente, os diálogos e as contradições que orientam as políticas formativas atuais, com base nos documentos de criação do Programa Institucional de Bolsa de Iniciação à Docência (Pibid), dos mestrados profissionais (MPs) e do Pacto Nacional pelo Fortalecimento do Ensino Médio (PNEM). Especificamente, interessa saber qual perfil docente tais políticas buscam construir e se mantêm relação entre si e entre as metas e objetivos do Plano Nacional de Educação (PNE 2014-2024). Palavras-chave: políticas; formação; docência; concepções. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 Abstract Public policies for teacher training: Pibid, professional master´s degree and PNEM Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 Introdução c) possibilitam, entre a formação inicial e continuada que oferecem, uma espécie de continum, uma vez que o bolsista do Pibid pode vir a ser o aluno dos mestrados profissionais e/ou poderá estar entre aqueles que atuam no ensino médio, que são também público da formação oferecida pelo PNEM. No que diz respeito à estrutura deste artigo, inicialmente apresentamos uma breve discussão sobre a formação de professores no Brasil, enfocando as políticas, as contradições e as influências que a organização das políticas para formação sofre em razão das demandas sociais e econômicas. Nas análises, após um resumo de apresentação de cada programa, o foco recai sobre as concepções presentes nos seus respectivos objetivos e, consequentemente, sobre o professor que cada um visa formar. A título de conclusão, buscamos delinear os diálogos entre as propostas desses programas, suas contradições e de que maneira eles contribuem para a efetivação das metas 15, 16, 17 e 18 do PNE 2014-2024. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 Introdução As políticas públicas para formação inicial e continuada de professores, dentro do escopo de preocupações relacionadas à melhoria da educação, no Brasil das últimas décadas, ganharam relevância nas pesquisas acadêmicas e também nos investimentos realizados pelos governos das esferas municipal, estadual e federal com vistas à melhoria da educação básica. Na mesma direção, houve investimento considerável em busca da melhoria da formação de professores, entretanto, esse tema continua presente nos debates em razão dos problemas que apresenta. A formação de professores é não só uma demanda não esgotada, mas também segue como um grande desafio, principalmente quando se focalizam os avanços ou não das políticas públicas implantadas (Azevedo; Aguiar, 2001; Gatti, 2008; Gatti; Barretto; André, 2011; Montandon, 2012). 24 Diante dessa atualidade, considerando os objetivos formativos presentes nos documentos legais de criação (portarias, decretos e documentos de apresentação) do Programa Institucional de Bolsa de Iniciação à Docência (Pibid), Programas de mestrados profissionais (MPs) e Pacto Nacional pelo Fortalecimento do Ensino Médio (PNEM), procuramos responder a três perguntas: 1) Como a formação inicial e continuada são concebidas? Especificamente, qual perfil do profissional é delineado nos objetivos desses programas? 1) Como a formação inicial e continuada são concebidas? Especificamente, qual perfil do profissional é delineado nos objetivos desses programas? 2) Há diálogo entre os objetivos desses programas no que diz respeito à formação? 2) Há diálogo entre os objetivos desses programas no que diz respeito à formação? 3) A formação proposta pelos três programas dialoga com o que as metas do Plano Nacional de Educação (PNE 2014-2024) estabelecem? A escolha desses três programas para análise se justifica pelo seguinte: a) são de abrangência nacional, portanto alcançam professores de todas as regiões do Brasil; Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 b) alcançam professores de todos os níveis da educação básica, principalmente a partir do ensino fundamental; b) alcançam professores de todos os níveis da educação básica, principalmente a partir do ensino fundamental; c) possibilitam, entre a formação inicial e continuada que oferecem, uma espécie de continum, uma vez que o bolsista do Pibid pode vir a ser o aluno dos mestrados profissionais e/ou poderá estar entre aqueles que atuam no ensino médio, que são também público da formação oferecida pelo PNEM. Regulamentação e política governamental 25 No quadro das transformações sociais, políticas e econômicas das últimas décadas, especificamente nos anos seguintes ao processo de redemocratização do Brasil, a formação de professores ganhou destaque no âmbito das políticas educacionais, como consequência dos debates que ganharam força sobre a qualidade do ensino oferecido na escola básica. Ou seja, como tais discussões vêm ocorrendo também em contexto de importantes demandas por uma educação básica que atenda às exigências econômicas relacionadas à mão de obra qualificada para o mercado de trabalho contemporâneo, por isso, a escola passou a ser vista como a responsável por criar as condições para que os indivíduos pudessem desenvolver suas capacidades e aprender os conteúdos necessários para compreender e para participar das relações sociais, econômicas e culturais cada vez mais plurais. A qualidade da educação e a formação de professores passaram a ser pauta nas agendas e discussões sobre educação no Brasil e nas ações e políticas propostas pelo Ministério da Educação (MEC). Passaram a ser faces de uma mesma questão e, consequentemente, esta última ganhou ressonância no processo de elaboração e aprovação da Lei de Diretrizes e Bases da Educação Nacional (LDB – Lei nº 9.394/1996). Naquele momento, a meta era fazer com que, em dez anos, todos os professores que atuavam na educação básica, nos seus diferentes níveis, alcançassem a formação superior. Não tendo sido alcançada essa meta, o parágrafo 4º do art. 87, foi revogado. Diante disso, a formação de professores tornou-se um tema crucial, implicado em qualquer discussão que visasse propor políticas públicas para a educação escolar, devido ao reconhecimento de que os desafios apresentados para a escola contemporânea exigiam/exigem um patamar cada vez mais elevado de formação do seu corpo docente. Tal preocupação está registrada no art. 61 da LDB, que define a formação dos profissionais da educação como resposta às necessidades e objetivos dos diferentes níveis de ensino e às características de cada fase de desenvolvimento do educando. No artigo seguinte, define-se que a formação mínima do professor da educação básica deverá ser realizada em nível superior, em instituições de ensino superior (IES): Art. 62. A formação de docentes para atuar na educação básica far-se-á em nível superior, em curso de licenciatura plena, admitida, como formação mínima para o exercício do magistério na educação infantil e nos cinco primeiros anos do ensino fundamental, a oferecida em nível médio, na modalidade normal. (Redação dada pela lei nº 13.415/2017). Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 Regulamentação e política governamental § 5º A União, o Distrito Federal, os Estados e os Municípios incentivarão a formação de profissionais do magistério para atuar na educação básica pública mediante programa institucional de bolsa de iniciação à docência a estudantes matriculados em cursos de licenciatura, de graduação plena, nas instituições de educação superior. (Incluído pela Lei nº 12.796/2013). (Brasil, 1996). No decorrer dos anos 2000, ocorreu a expansão da oferta de programas e projetos de formação (Brasil. Capes, 2013). A parceria entre universidade e escola foi intensificada. Consequentemente, a qualidade dessa formação e a sua efetivação em transformações das práticas de ensino e aprendizagem também passaram a ser um tema de grandes preocupações, uma vez que, na ponta desse processo, estão as exigências sociais e econômicas mais amplas de produção de um indivíduo capaz de compreender, agir e propor mudanças no contexto global e neoliberal em que está inserido (Mazzeu, 2009; Tommasi; Warde; Haddad, 2009). 26 O MEC, de acordo com Weber (2015), ao longo dos anos que precederam a aprovação da LDB, mostrou-se bastante proativo na execução da tarefa de oferecer meios para impulsionar a efetivação de ações de formação inicial e continuada de professores. Dentre elas, a mais expressiva, segundo a autora, foi a instituição da Política Nacional de Formação de Profissionais de Magistério da Educação Básica, que ficou sob a Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (Capes), especificamente na Diretoria de Formação de Professores da Educação Básica. Essa política, no momento da sua instituição, em 2009, fundamentalmente, visou firmar o reconhecimento da importância do professor na realização do processo educativo escolar e, em consequência disso, a necessidade de sua valorização profissional. Entretanto, todas essas mudanças, efetivadas à luz do texto da LDB e em documentos oficiais que a desdobraram, não foram suficientes para produzir a formação requerida para os profissionais da educação básica, de maneira a atender às necessidades atuais da escola básica, principalmente no que diz respeito ao que dela é demandado. Os investimentos feitos também não resultaram no número de profissionais necessários para ocupar a demanda da escola básica. Ainda é muito comum associar o desempenho dos alunos da educação básica em exames nacionais e internacionais à qualidade de formação do professor. Regulamentação e política governamental Nos textos críticos ao ensino realizado pela escola básica, os objetos criticados são muitas vezes os mesmos de décadas anteriores a todo esse movimento em busca da melhoria da educação pela via do investimento na formação inicial e continuada. Consequentemente, nas metas 15, 16, 17 e 18 do PNE 2014-2024, questões como formação específica, formação em nível de pós-graduação, valorização profissional e plano de carreira compõem o eixo central intitulado “Valorização dos Profissionais da Educação: formação, remuneração, carreira e condições de trabalho”. O objetivo da nota técnica da Meta 15 é, na vigência do PNE, garantir a formação específica do professor. Essa nota trata de um problema antigo, bastante presente nas discussões sobre as políticas de formação, qual seja: o desvio de função. Segundo o texto da nota, no Brasil de 2013, devido à escassez de licenciados para determinadas áreas, ainda era comum professores atuarem em disciplinas para as quais não foram formados. O percentual de professores com formação compatível com a disciplina que lecionava era de 50% e, com pequenas variações, esse percentual era similar para todos os níveis de escolarização: educação infantil, ensino fundamental e médio. Na conclusão dos resultados apresentados na nota técnica da Meta 15, afirma-se que é necessário ampliar os esforços e investimentos na formação superior nas áreas específicas em que atuam muitos professores da educação básica. Nesse sentido, a Meta 16 já aponta um direcionamento desse investimento, qual seja: a formação em nível de pós-graduação de 50% dos professores hoje ativos e também a oferta de formação continuada em suas áreas de atuação. 27 A “escassez oculta” ou a falta de profissionais com formação adequada à área em que atuam, de fato, põe em evidência dois problemas: a desvalorização da carreira docente e a falta de um plano de carreira que contemple também as questões relacionadas à remuneração. As metas 17 e 18 voltam-se para a valorização dos profissionais da educação básica, prevendo inclusive a equiparação salarial com as demais profissões com o mesmo nível de formação, e para o estabelecimento de planos de carreiras baseados no piso salarial previsto na Constituição Federal. Regulamentação e política governamental Em síntese, tais metas do PNE representam o reconhecimento de que ainda são necessárias políticas governamentais que visem criar condições efetivas para que esse profissional se sinta valorizado socialmente pelo esforço que precisa depreender na formação das gerações futuras, na perspectiva do que lhe é demandado pela sociedade contemporânea (Gatti; Barretto; André, 2011). Pibid, MPs e PNEM: abrangências e concepções Contextualização e eixos de sustentação Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 Contextualização e eixos de sustentação Os programas analisados neste artigo foram criados com vistas a alcançar professores que atuam em diferentes níveis do magistério da educação básica, enfocando a formação inicial e continuada. Criados em diferentes momentos da implantação e execução da Política Nacional para Formação de Profissionais do Magistério da Educação Básica, têm como eixo comum e orientador a “equalização de oportunidades educacionais e padrão mínimo de qualidade do ensino” (Brasil, 2008). Esse eixo sintetiza as atribuições da União, conforme previsto no parágrafo 1º, do art. 211, da Constituição Federal e que embasou o Decreto nº 8.752/2016 (que retoma o Decreto nº 6.755/2009). Esse decreto instituiu a Política Nacional de Formação de Profissionais do Magistério da Educação Básica, cujos objetivos, conforme o art. 4º, estavam previstos para ser efetivados a partir da “criação dos Fóruns Estaduais Permanentes de Apoio à Formação Docente, em regime de colaboração entre a União, os Estados, o Distrito Federal e os Municípios, e por meio de ações e programas específicos do Ministério da Educação” (Brasil, 2009). No contexto dessas regulamentações, o Pibid, criado em 2007 por meio da Portaria Normativa MEC nº 38/2007, e, instituído no âmbito da Capes, pela Portaria Normativa Capes nº 122/2009, teve dois objetivos: a) incentivar a iniciação à docência durante a formação inicial de professores nos cursos de licenciatura de instituições públicas de ensino superior, valorizando o magistério e aqueles que optaram pela carreira docente; b) elevar a qualidade das ações que, nessas instituições, estavam voltadas para a formação inicial de professores. 28 Consequentemente, para a execução desses objetivos primeiros, estão previstos outros que alcançam a escola ao estimularem a integração da educação superior com a educação básica por meio da parceria entre secretarias estaduais e municipais de educação e universidades públicas, num trabalho colaborativo com vistas à melhoria do ensino nas escolas públicas – consideradas também como parceiras, corresponsáveis pela formação dos futuros professores. Na modalidade mestrados profissionais (MPs), foram criados alguns programas voltados para a formação de professores da educação básica pública. Os MPs ganharam força na empolgação e ebulição que programas como o Pibid geraram nas universidades federais. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 Contextualização e eixos de sustentação Assim, o MEC, por meio da Portaria nº 17/2009, que retoma a Portaria Normativa MEC nº 07/2009, definiu os MPs como cursos de pós-graduação stricto sensu com foco na prática profissional e que, por isso, deveriam habilitar o profissional concluinte a: a) a atuar em atividades técnico-científicas e de inovação, por meio da incorporação do método científico; b) possuir uma formação apropriada que o permita aplicar o conhecimento adquirido durante a formação com rigor metodológico; c) incorporar (mantendo-se atualizado) os avanços científicos e tecnológicos de forma a construir solução para problemas específicos no seu ambiente de trabalho. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 Com o mesmo intuito de oferecer formação continuada para professores em atuação na educação básica, o MEC, também por meio da pactuação entre universidades, estados e municípios e demais entidades engajadas em questões educativas, criou – por meio da Portaria nº 1.140/2013 – e implementou, em 2014, o Pacto Nacional pelo Fortalecimento do Ensino Médio (PNEM), cujos objetivos e ações foram organizados em torno dos seguintes eixos: a) garantir a formação de professores e demais profissionais da escola; a) garantir a formação de professores e demais profissionais da escola; b) ampliar os espaços de formação dos profissionais que atuam nessa etapa de ensino; a) garantir a formação de professores e demais profissionais da escola; b) ampliar os espaços de formação dos profissionais que atuam nessa etapa de ensino; b) ampliar os espaços de formação dos profissionais que atuam nessa etapa de ensino; c) induzir ao redesenho curricular do ensino médio; d) disponibilizar recursos pedagógicos e novas tecnologias. Esses objetivos estão em consonância com a LDB e com as atuais Diretrizes Curriculares Nacionais do Ensino Médio (DCNEM), definidas na Resolução CNE/CEB nº 2/2012. Em síntese, o Pibid alcança a formação inicial de professores em todos os níveis de escolarização – da educação infantil ao ensino médio. Os MPs, nas áreas em que foram instituídos, oferecem formação continuada a professores de diferentes níveis da educação básica – do ensino fundamental ao médio. Por sua vez, o PNEM alcança o professor do ensino médio, portanto, dos últimos anos da educação básica. Nesse sentido, o público do Pibid tem a possibilidade de se fazer presente nos demais programas, sendo que o público dos MP’s certamente pode ser o público do PNEM. Contextualização e eixos de sustentação Aliás, a possibilidade de relação entre os públicos desses programas está assinalada no Regulamento do Pibid (Portaria nº 46/2016, anexo I, art. 7º, inciso II), em que a Capes coloca como uma das características necessárias aos projetos institucionais a articulação, nas propostas das IES, do Pibid com os programas institucionais do MEC e, dentre os enumerados, está o PNEM. Trata-se, portanto, de um conjunto de programas que alcança o professor da educação básica nos seus diferentes momentos e/ou níveis de formação e atuação e, consequentemente, oferecem subsídios para a construção da concepção do que seja o fazer docente nos dias atuais. 29 29 Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 A formação e o perfil do professor nos objetivos do Pibid, MPs e PNEM Para a análise a ser desenvolvida neste tópico, tomamos como base a concepção de que a profissão docente se constitui por meio do diálogo entre elementos teóricos e aqueles advindos das experiências práticas. Nesse sentido, as políticas formativas integram a constituição desse conjunto amplo de conhecimentos e saberes que o professor sustenta sobre a própria profissão, uma vez que, nas suas diferentes proposições, visam inserir o professor em contextos de aprendizagens teórico- práticos. Por isso, é fundamental dar atenção às concepções que norteiam tais políticas. Conforme já dito, o Pibid objetiva alcançar o futuro professor na formação inicial. Dada a sua abrangência e proporção dos investimentos, trata-se de um programa que formará um número importante de uma geração de professores que ministrarão e/ou ministram aulas em todos os níveis da educação básica. No Regulamento do Pibid constam os seguintes objetivos: Art. 4º [...] Art. 4º [...] I incentivar a formação de docentes em nível superior para a educação básica; I incentivar a formação de docentes em nível superior para a educação básica; II contribuir para a valorização do magistério; III elevar a qualidade da formação inicial de professores nos cursos de licenciatura, promovendo a integração entre educação superior e educação básica; IV inserir os licenciandos no cotidiano de escolas da rede pública de educação, proporcionando-lhes oportunidades de criação e participação em experiências metodológicas, tecnológicas e práticas docentes de caráter inovador e interdisciplinar que busquem a superação de problemas identificados no processo de ensino e aprendizagem; IV inserir os licenciandos no cotidiano de escolas da rede pública de educação, proporcionando-lhes oportunidades de criação e participação em experiências metodológicas, tecnológicas e práticas docentes de caráter inovador e interdisciplinar que busquem a superação de problemas identificados no processo de ensino e aprendizagem; V incentivar escolas públicas de educação básica, mobilizando seus professores como coformadores dos futuros docentes e tornando-as protagonistas nos processos de formação inicial para o magistério; VI contribuir para a articulação entre teoria e prática necessárias à formação dos docentes, elevando a qualidade das ações acadêmicas nos cursos de licenciatura; VII contribuir para que os estudantes de licenciatura se insiram na cultura escolar do magistério, por meio da apropriação e da reflexão sobre instrumentos, saberes e peculiaridades do trabalho docente. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 A formação e o perfil do professor nos objetivos do Pibid, MPs e PNEM 30 VIII articular-se com os programas de formação inicial e continuada de professores da educação básica, de forma a contribuir com a criação ou com o fortalecimento de grupos de pesquisa que potencialize a produção de conhecimento sobre ensinar e aprender na Educação Básica; IX comprometer-se com a melhoria da aprendizagem dos estudantes nas escolas onde os projetos institucionais são desenvolvidos. (Brasil. Capes, 2016, anexo 1). O inciso I dialoga com o previsto no art. 62 da LDB: possibilitar que o professor tenha formação em nível superior; enquanto que o inciso II, na mesma linha, pretende a valorização do magistério. Cabe destacar que, ao longo do texto, essa última questão não apresenta maiores definições acerca do que seria e de como se daria essa valorização. Ao leitor fica subentendido que o conjunto de ações formativas propostas nos objetivos levaria a essa valorização. As questões relacionadas às condições de trabalho e sua precariedade e à valorização profissional e salarial – que, inclusive, são as que mais afastam o futuro professor da docência –, não são abordadas nem postas como condição fundamental a ser enfocada durante essa formação. Com isso, cria-se a ideia de que a melhoria do trabalho docente depende somente da qualidade da formação do professor e, consequentemente, das práticas e conhecimentos que poderá manejar. Os incisos III, VI e VII tocam numa questão histórica: a necessidade de fortalecer a relação entre universidade e escola e de superar a dicotomia entre teoria e prática vigente durante a formação na universidade e também durante o exercício docente. De diferentes perspectivas, esses três incisos apontam para aquilo que Imbernón (2000) chamou de um discurso paradoxal, contraditório porque postula que o nível acadêmico é importante para essa formação e, ao mesmo tempo, aponta as suas fragilidades, dentre as quais se destaca a não superação da concepção que dá maior valor aos conhecimentos teóricos, específicos, em detrimentos dos conhecimentos práticos. Nesse sentido, a exigência prevista nos objetivos do programa para que haja equilíbrio e acomodação entre formação teórica e prática instala o dilema da formação culta versus a formação para a prática. A formação e o perfil do professor nos objetivos do Pibid, MPs e PNEM Isso ocorre porque, por meio das ações do Pibid, deve-se, ao mesmo tempo, formar um professor culto – com capacidade para pensar as práticas de ensino considerando conhecimentos filosóficos e científicos –, e um professor prático/técnico – que, a partir da experiência de formação no programa, seja “capaz de entrar numa sala de aula e, aplicando regras relativas à conduta e aos conhecimentos a serem transmitidos, seja capaz de se desempenhar a contento diante dos alunos” (Saviani, 2011, p. 13). Os incisos V e VII entendem a escola como locus da formação e o professor que lá está como coformador dos futuros professores. Especificamente, o inciso V contrapõe-se a uma prática muito denunciada no âmbito da realização dos estágios supervisionados: a transformação da escola em mero figurante dessa formação, uma vez que dela é demandada a oferta de espaço e de supervisor do estágio. No reverso do discurso de defesa da necessidade de fazê-la protagonista, está o que a coloca como não protagonista desse processo. Nessa perspectiva, o documento traz também a afirmação implícita (VII) de que a docência se constitui como prática social a partir do entrecruzamento de conhecimento teórico e de saberes da prática. A ação docente que possibilita ao professor ocupar a posição de um profissional transformador de seus alunos constitui-se com base nas apropriações e reflexões sobre os modos de confecção e funcionamento dos instrumentos didático-pedagógicos; admite-se também que há saberes e peculiaridades nessa profissão que são somente acessíveis no contexto escolar. 31 O inciso VIII aborda duas questões distintas. A primeira diz respeito à ideia de combater a fragmentação, dispersão e descontinuidade existentes nas IES, em decorrência (acredita-se) da diversificação dos modelos de funcionamento dessas instituições, e nas instâncias superiores propositoras das políticas de formação. A segunda parte enfoca a criação de grupos de pesquisa, reafirmando assim a importância da formação acadêmica acrescida de outra demanda: a produção de conhecimento sobre o que significa ensinar e aprender na educação básica. Nessa perspectiva, o professor que está na escola é colocado na posição de quem tem o que dizer sobre a formação que se realiza nas IES. Um sujeito de conhecimento e não somente da prática. Por fim, o inciso IX traz a exigência de que as IES se comprometam também com a melhoria da aprendizagem (e não só do ensino) dos alunos da educação básica. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 A formação e o perfil do professor nos objetivos do Pibid, MPs e PNEM Isso traz, para o âmbito da formação, a proposição de que a aprendizagem do aluno que está na escola é também responsabilidade das IES. Nestas é muito presente a visão de que o aprendizado do aluno da educação básica é responsabilidade da escola, que, inclusive, é muito criticada por não realizar essa tarefa a contento. Nesse sentido, o comprometimento é indireto e precisa ocorrer na efetivação da formação de um professor capaz de mudar a qualidade da aprendizagem na escola. Um professor transformador da educação escolar hoje realizada. Por sua vez, os mestrados profissionais, tendo como foco o professor em atuação, buscam oferecer uma formação mais completa do que os cursos de especialização (e os que ficaram conhecidos nos documentos oficiais como sendo de aperfeiçoamento), ancorada em conhecimentos científicos que permitam ao professor atuar no mercado de trabalho. Assim, o art. 3º da Portaria MEC nº 17/2009, que retoma a Portaria MEC nº 7/2009, define os MPs como cursos de pós-graduação stricto sensu, e o art. 4º explicita os seus objetivos: Art. 3º O mestrado profissional é definido como modalidade de formação pós-graduada stricto sensu [...]. Art. 4º São objetivos do mestrado profissional: I capacitar profissionais para o exercício da prática profissional avançada e transformadora de procedimentos, visando atender demandas sociais, organizacionais ou profissionais e do mercado de trabalho; II transferir conhecimento para a sociedade, atendendo demandas específicas e de arranjos produtivos com vistas ao desenvolvimento nacional, regional ou local; III promover a articulação integrada da formação profissional com entidades demandantes de naturezas diversas, visando melhorar a eficácia e a eficiência das organizações públicas e privadas por meio da solução de problemas e geração e aplicação de processos de inovação apropriados; 32 IV contribuir para agregar competitividade e aumentar a produtividade em empresas, organizações públicas e privadas. (Brasil. MEC, 2009). Cabe destacar que a semântica do termo “mestrado profissional”, em contraponto ao termo “mestrado acadêmico”, estabelece uma diferenciação que, de fato, traz na sua gênese a crítica à formação docente que ocorre nos modelos vigentes nas IES, por ser muito teórica e afastada das demandas da escola e da sociedade. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 A formação e o perfil do professor nos objetivos do Pibid, MPs e PNEM O inciso I dá visibilidade a essa diferenciação ao definir que o objetivo deve ser a capacitação para a prática profissional avançada e transformadora de procedimentos (não de humanos), o que remete às exigências dos organismos internacionais que definem que as políticas públicas educacionais devem ter como meta responder às demandas de um mercado de trabalho competitivo e globalizado (Tommasi; Warde; Haddad, 2009). Isso inclui a exigência de uma formação de professor universalizada/profissionalizada, logo técnica. Se os objetivos do Pibid apresentam a prevalência da concepção de formação de um professor culto, os dos MPs voltam-se para a técnica; daí, a necessidade de ser avançada, transformadora de procedimentos para que atenda às demandas sociais, organizacionais ou profissionais do mercado de trabalho. O professor a ser formado por essa perspectiva – inciso II –, precisa assumir a posição do técnico capaz de transferir (não construir) conhecimento de modo a atender às demandas do campo produtivo relacionadas aos contextos nacionais, regionais e locais. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 Os incisos III e IV trazem exigências que são do campo empresarial e econômico: melhoria da eficácia e eficiência, solução de problemas, geração de processos inovadores, aumento de produtividade e competitividade. Esses termos remetem às altas exigências do mercado econômico contemporâneo e, consequentemente, do trabalho como força produtiva do capital empresarial. Quando traduzidos para o campo da educação e da formação, dadas as especificidades e precariedades que hoje permeiam a prática profissional dos professores, tais exigências tendem a ser tomadas como simples e, rapidamente, são assimiladas como modelos a serem seguidos. Além disso, essas características tecnicistas podem levar o professor a ter dificuldades para compreender os ritmos das escolas e para incorporar essas exigências no contexto escolar, sobretudo na escola pública. Ao serem usadas como eixos para estruturar os mestrados profissionais voltados à formação docente, tais exigências deixam de considerar a realidade da profissão docente no que diz respeito às condições de trabalho. Inovação, competitividade e produtividade pressupõem acesso às tecnologias e conhecimentos atuais, tal como se dá no contexto empresarial, algo ainda não disponível na e para a maioria das escolas básicas brasileiras. Os objetivos postos pela Capes – como diretrizes para concepção, organização e execução dos MPs – são constitutivos e definidores do modo como essa formação tem sido trabalhada nas IES. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 A formação e o perfil do professor nos objetivos do Pibid, MPs e PNEM Ou seja, os MPs se configuram como uma nova trajetória de formação não mais centrada na dimensão acadêmica e sim na dimensão profissional. Com base nessas diretrizes legais, dá-se a criação dos inúmeros programas de MPs (com funcionamento em rede ou não) para formar professores para a educação básica em diferentes áreas do conhecimento. As versões em rede desses programas, que aprofundam o caráter neoliberal e racionalista presentes nas políticas formativas, apresentam as seguintes características organizativas: objetivos e currículo vigentes em âmbito nacional, orientações teórico-metodológicas comuns, modelos para a produção dos trabalhos de conclusão de curso e, em muitos, bibliografia das disciplinas unificada. Esse processo incide, principalmente, na relação que o professor estabelece com o conhecimento e com os próprios saberes da prática, uma vez que provoca a homogeneização dos modos de olhar, questionar e buscar respostas para os problemas que se apresentam como objeto de estudo e orientam a busca de soluções. 33 No que diz respeito ao PNEM, nos documentos que orientam as ações de formação continuada de professores e coordenadores do ensino médio para os anos de 2013 e 2014, afirma-se que, nos fundamentos e objetivos desse programa, está a busca por uma escola que ofereça, no decorrer do ensino médio, uma formação integral que possibilite o desenvolvimento do ser humano na sua totalidade, não cindido pela divisão social advinda da ideia de uma formação para o trabalho. A formação escolar é concebida, portanto, como meio de emancipação com vistas às leituras de mundo e à participação dos bens sociais, culturais e produtivos. Tal concepção, expressa no texto orientador do Pacto, fundamenta-se nas Diretrizes Curriculares para o Ensino Médio (DCNEM) na medida em que concebe o trabalho docente englobando ciência, cultura e novas tecnologias, o que implica considerar, no processo formativo, as dimensões da ação humana, o aspecto educativo do trabalho, a pesquisa como eixo fundamental da ação pedagógica e, principalmente, os direitos humanos como princípio central da prática educativa (Brasil. MEC, 2014). Esses princípios expressos nas DCNEM, e tomados como referenciais pelo PNEM, colocam-se na perspectiva de superar a dualidade que, historicamente, marca o ensino médio no Brasil – de um lado, ensino propedêutico para as classes favorecidas com perspectiva de continuação dos estudos, e, de outro, ensino técnico profissionalizante para os filhos da classe trabalhadora. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 A formação e o perfil do professor nos objetivos do Pibid, MPs e PNEM A concepção norteadora desses documentos é a da formação humana integral, que não se coaduna com a que toma a formação escolar como preparação para o trabalho a partir do uso dos conhecimentos científico-tecnológicos. Para tanto, defende-se que a formação do professor precisa ser pensada a partir de múltiplos espaços e propostas, que podem ser concretizados na educação formal, tais como nos cursos de licenciaturas e nos espaços de formação continuada. A partir disso, o documento orientador das ações de formação continuada de professores e coordenadores pedagógicos do ensino médio em 2014 traz as seguintes afirmações: Objetivo geral: Formar em nível de aperfeiçoamento ou extensão todos os professores e coordenadores pedagógicos que atuam no Ensino Médio com vistas à valorização da formação do professor e dos profissionais da Educação a partir do diálogo entre conhecimentos teóricos e experiências docentes e de gestão pedagógica. Descrição do curso: O curso privilegia a articulação entre a teoria e a prática no processo de formação docente, fundado no domínio de conhecimentos científicos e didáticos. Considera a escola como lócus de formação continuada e (re)construção coletiva do projeto político-pedagógico em suas articulações com as concepções de juventude e direito à qualidade social da educação. 34 Metodologia: A proposta do curso compreende o professor como um sujeito epistêmico, que elabora e produz conhecimentos com base na compreensão da realidade e nas possibilidades de transformação da sociedade [...]. (Brasil. MEC, 2014, p. 5). O objetivo “formar (...) a partir do diálogo entre conhecimentos teóricos e experiências docentes” é detalhado no item seguinte: “o curso privilegia a articulação entre a teoria e a prática no processo de formação docente, fundado no domínio de conhecimentos científicos e didáticos”. Logo, o professor é concebido como um “sujeito epistêmico, que elabora e produz conhecimentos com base na compreensão e nas possibilidades de transformação da sociedade”. Tais fundamentos permitem interpretar que o Pacto pretende uma formação continuada que coloca o professor na posição de sujeito portador de saberes práticos e que, a partir disso, pode promover discussões teórico-práticas com vistas à construção de respostas de caráter científico para os problemas pedagógicos contextuais que se apresentam na e para a escola. Dito de outro modo, por meio das práticas formativas, possibilitará as condições necessárias para refletir sobre as reais necessidades da escola e, à luz dos seus conhecimentos teóricos, poderá construir respostas para os problemas que se apresentam no decorrer do trabalho docente. A formação e o perfil do professor nos objetivos do Pibid, MPs e PNEM Para tanto, os conhecimentos teóricos advindos desse processo de formação precisam se desdobrar de modo a se transformarem e/ou dialogarem com os conhecimentos relacionados com as questões mais amplas, tais como educação, economia, política e sociedade e suas relações com os aspectos sociocognitivos da aprendizagem e do desenvolvimento humano. Como sujeito epistêmico, esse docente é um sujeito ético, com responsabilidades sociais, individuais e investigativas – aqui entra também o papel da pesquisa no avanço na compreensão do seu próprio trabalho. As características do curso de formação do Pacto, conforme apontadas acima, evidenciam a estratégia da formação em serviço tendo, talvez, dois objetivos principais: a) ampliar o alcance do programa ao realizar a formação na escola (usando a hora atividade); b) preparar o maior número possível de professores e coordenadores pedagógicos do ensino médio para a reforma que se pretende para esse nível de ensino – o redesenho curricular a partir de áreas do conhecimento. Nessa perspectiva, “a formação contínua tende a articular-se em primeira linha com os objetivos do sistema, nomeadamente com o desenvolvimento da reforma” (Nóvoa, 1992, p. 9). Desse modo, o desenvolvimento profissional dos professores e o desenvolvimento organizacional da escola seriam consequências do processo desencadeado pelo programa. 35 Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 Considerações finais No que diz respeito às três perguntas que nortearam a escrita deste artigo, é possível responder que a formação inicial e continuada no Pibid e no PNEM são concebidas, com maior ou menor ênfase, objetivando a constituição do professor como um sujeito culto, logo conhecedor e produtor de conhecimentos teóricos considerados fundamentais à compreensão e transformação da sua prática profissional. Nesse sentido, a formação acadêmica é criticada naquilo que não dialoga com a escola, mas é posta como necessária às mudanças que as políticas educacionais visam promover na educação básica. Pleiteia-se nesses dois programas a mudança na forma de efetivação. Já os MPs reivindicam uma formação técnica, voltada para as exigências das economias e, consequentemente, do mercado de trabalho globalizado. Concebidas à luz das exigências de profissionalização das áreas administrativas e exatas, dentre outras, as diretrizes da Capes que orientam a criação de MPs se ancoram num conjunto de discursos que coloca a formação oferecida pelas IES em xeque. A relação entre condições de trabalho, valorização do magistério, carreira e salários e melhoria da qualidade da educação básica não é abordada no documento do Pibid, embora, conforme assinalado nas análises, pode-se partir do princípio que o conjunto de ações demandadas pelo programa criaria as condições para que o professor se engajasse também em políticas que visam a mudanças nesses aspectos. O PNEM, ao trazer nas suas diretrizes a concepção de que as mudanças educacionais prescindem de compreensão e engajamento e, principalmente, que o trabalho docente engloba um empenho na formação e desenvolvimento humano em todos os seus aspectos, coloca de forma contundente a necessidade de que sejam garantidas as condições de trabalho e a valorização do profissional docente. As diretrizes dos MPs, pelo que é próprio dos seus fundamentos, não sinalizam a necessidade de abordar tais questões – ao contrário, apontam para uma perspectiva que tende a reforçar a precarização humana nos contextos profissionais a serviço da produtividade, dos resultados e de seus avanços. No que diz respeito ao diálogo com o atual PNE, os três programas apresentam objetivos que corroboram com as metas 15, 16 e 17 do plano. No caso do PNEM, também com a meta 18. Considerações finais Assim, tendo em vista o proposto no PNE, tais programas, considerando a importância do processo educativo para a formação das novas gerações e as características exigidas desse profissional, representam um passo relevante no sentido da implementação de políticas de formação atreladas também às de valorização da carreira para além do ofício. Porém, tendo em vista as metas do PNE, faz-se necessária a inclusão, nessas políticas formativas, de diretrizes que não só formem os profissionais docentes com uma compreensão/concepção de que a melhoria da educação passa pela melhoria das suas condições de trabalho, salários e valorização do magistério, mas também que as atrelem a ações governamentais que avancem na criação de condições e possibilidades de reconhecimento e respeito condizentes com o papel social atribuído a esse profissional, garantindo o desenvolvimento de recursos humanos para atender às demandas sociais e econômicas (Gatti; Barretto; André, 2011). 36 Referências bibliográficas AZEVEDO, J. M. L.; AGUIAR, M. Â. A produção do conhecimento sobre a política educacional no Brasil: um olhar a partir da Anped. Educação & Sociedade, Campinas, v. 22, n. 77, p. 49-70, dez. 2001. Disponível em: <http://www.scielo. br/scielo.php?script=sci_arttext&pid=S0101-73302001000400004>. Acesso em: 2 mar. 2017. BRASIL. Constituição (1988). Constituição da República Federativa do Brasil. Texto constitucional promulgado em 5 de outubro de 1988, com as alterações adotadas pelas Emendas Constitucionais nº 1/1992 a 56/2007 e pelas Emendas Constitucionais de Revisão nº 1 a 6/1994. Brasília: Senado Federal, 2008. BRASIL. Constituição (1988). Constituição da República Federativa do Brasil. Texto constitucional promulgado em 5 de outubro de 1988, com as alterações adotadas pelas Emendas Constitucionais nº 1/1992 a 56/2007 e pelas Emendas Constitucionais de Revisão nº 1 a 6/1994. Brasília: Senado Federal, 2008. BRASIL. Decreto nº 6.755, de 29 de janeiro de 2009. Institui a Política Nacional de Formação de Profissionais do Magistério da Educação Básica, disciplina a atuação da Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – CAPES no fomento a programas de formação inicial e continuada, e dá outras providências. Diário Oficial da União, Brasília, DF, 30 jan. 2009. Revogado pelo Decreto nº 8.752/2016. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 BRASIL. Decreto nº 8.752, de 9 de maio de 2016. Dispõe sobre a Política Nacional de Formação dos Profissionais da Educação Básica. Diário Oficial da União, Brasília, DF, 10 maio 2016. BRASIL. Lei nº 9.394, de 20 de dezembro de 1996. Estabelece as diretrizes e bases da educação nacional. Diário Oficial da União, Brasília, DF, 23 dez. 1996. Disponível em: http://www.planalto.gov.br/ccivil_03/leis/L9394.htm >. Acesso em: 2 mar. 2017. BRASIL. Conselho Nacional de Educação (CNE). Câmara de Educação Básica (CEB). Parecer CNE/CEB nº 5/2011. Diretrizes curriculares nacionais para o ensino médio. Brasília, 2011. Disponível em: <http://portal.mec.gov.br/index. php?option=com_docman&view=download&alias=13448-diretrizes-curiculares- nacionais-2013-pdf&Itemid=30192>. Acesso em: 4 mar. 2017. BRASIL. Conselho Nacional de Educação (CNE). Câmara de Educação Básica (CEB). Resolução nº 2, de 30 de janeiro de 2012. Define diretrizes curriculares nacionais para o ensino médio. Brasília, 2012. Disponível em: <http://pactoensinomedio. mec.gov.br/images/pdf/resolucao_ceb_002_30012012.pdf>. BRASIL. Conselho Nacional de Educação (CNE). Câmara de Educação Básica (CEB). Resolução nº 2, de 30 de janeiro de 2012. Define diretrizes curriculares nacionais para o ensino médio. Brasília, 2012. Disponível em: <http://pactoensinomedio. mec.gov.br/images/pdf/resolucao_ceb_002_30012012.pdf>. BRASIL. Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (Capes). Portaria Normativa Capes nº 122, de 16 de setembro de 2009. Referências bibliográficas Dispõe sobre o PIBID – Programa Institucional de Bolsa de Iniciação à Docência, no âmbito da CAPES. Disponível em: <http://www.capes.gov.br/images/stories/download/bolsas/ PortariaNormativa122_PIBID.pdf>. Acesso em: 3 mar. 2017. BRASIL. Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (Capes). Portaria Normativa Capes nº 122, de 16 de setembro de 2009. Dispõe sobre o PIBID – Programa Institucional de Bolsa de Iniciação à Docência, no âmbito da CAPES. Disponível em: <http://www.capes.gov.br/images/stories/download/bolsas/ PortariaNormativa122_PIBID.pdf>. Acesso em: 3 mar. 2017. 37 BRASIL. Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (Capes). Diretoria de Formação de Professores da Educação Básica (DEB). Relatório de gestão Prodocência. Brasília, 2013. Disponível em: <http://www.capes.gov.br/ images/stories/download/bolsas/1892014-relatorio-PRODOCENCIA.pdf>. Acesso em: 2 mar. 2017. de gestão Prodocência. Brasília, 2013. Disponível em: <http://www.capes.gov.br/ images/stories/download/bolsas/1892014-relatorio-PRODOCENCIA.pdf>. Acesso em: 2 mar. 2017. BRASIL. Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (Capes). Portaria nº 46, de 11 de abril de 2016. Aprova o regulamento do Programa Institucional de Bolsa de Iniciação à Docência – Pibid. 2016. Disponível em: <https://www.capes.gov.br/images/stories/download/legislacao/15042016- Portaria-46-Regulamento-PIBID-completa.pdf>. BRASIL. Instituto Nacional de Estudos e Pesquisas Educacionais Anísio Teixeira (Inep). Plano Nacional de Educação PNE 2014-2024: linha de base. Brasília: Inep, 2015. Disponível em: <http://portal.inep.gov.br/ documents/186968/485745/Plano+Nacional+de+Educa%C3%A7%C3% A3o+PNE+2014-2024++Linha+de+Base/c2dd0faa-7227-40ee-a520- 12c6fc77700f?version=1.1>. Acesso em: 3 mar. 2017. BRASIL. Ministério da Educação (MEC). Portaria normativa nº 38, de 12 de dezembro de 2007. Dispõe sobre o Programa de Bolsa Institucional de Iniciação à Docência – PIBID. Diário Oficial da União, Brasília, DF, 13 dez. 2007. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 BRASIL. Ministério da Educação (MEC). Portaria normativa nº 7, de 22 de junho de 2009. Dispõe sobre o mestrado profissional no âmbito da Fundação Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Capes. Diário Oficial da União, Brasília, DF, 23 jun. 2009. BRASIL. Ministério da Educação (MEC). Portaria normativa nº 17, de 28 de dezembro de 2009. Dispõe sobre o mestrado profissional no âmbito da Fundação Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – CAPES. Diário Oficial da União, Brasília, DF, 29 dez. 2009. BRASIL. Ministério da Educação (MEC). Portaria normativa nº 17, de 28 de dezembro de 2009. Dispõe sobre o mestrado profissional no âmbito da Fundação Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – CAPES. Diário Oficial da União, Brasília, DF, 29 dez. 2009. BRASIL. Ministério da Educação (MEC). Portaria nº 1.140, de 22 de novembro de 2013. Referências bibliográficas Institui o Pacto pelo Fortalecimento do Ensino Médio e define suas diretrizes gerais, forma, condições e critérios para a concessão de bolsas de estudo e pesquisa no âmbito do ensino médio público, nas redes estaduais e distrital de educação. Diário Oficial da União, Brasília, DF, 9 dez. 2013. BRASIL. Ministério da Educação (MEC). Portaria nº 1.140, de 22 de novembro de 2013. Institui o Pacto pelo Fortalecimento do Ensino Médio e define suas diretrizes gerais, forma, condições e critérios para a concessão de bolsas de estudo e pesquisa no âmbito do ensino médio público, nas redes estaduais e distrital de educação. Diário Oficial da União, Brasília, DF, 9 dez. 2013. BRASIL. Ministério da Educação (MEC). Secretaria de Educação Básica (SEB). Pacto Nacional pelo Fortalecimento do Ensino Médio: documento orientador das ações de formação continuada de professores e coordenadores pedagógicos do Ensino Médio em 2014. [Brasília], 2014. Disponível em: <http://pactoensinomedio.mec.gov.br/images/pdf/documento_orientador.pdf>. Acesso em: 3 jan. 2017. BRASIL. Ministério da Educação (MEC). Secretaria de Educação Básica (SEB). Pacto Nacional pelo Ensino Médio: formação de professores do ensino médio: documento orientador preliminar. Brasília, 2013. Disponível em: <http:// www.observatoriodoensinomedio.ufpr.br/wp-content/uploads/2014/03/pnem- proposta-de-formac3a7c3a3o-continuada-de-professores.pdf>. Acesso em: 3 mar. 2017 BRASIL. Ministério da Educação (MEC). Secretaria de Educação Básica (SEB). Pacto Nacional pelo Ensino Médio: formação de professores do ensino médio: documento orientador preliminar. Brasília, 2013. Disponível em: <http:// www.observatoriodoensinomedio.ufpr.br/wp-content/uploads/2014/03/pnem- proposta-de-formac3a7c3a3o-continuada-de-professores.pdf>. Acesso em: 3 mar. 2017 38 BRZEZINSKI, I. (Org.). LDB dez anos depois: reinterpretação sob diversos olhares. 2. ed. São Paulo: Cortez, 2008. FREITAS, H. C. L. A (nova) política de formação de professores: a prioridade postergada. Educação & Sociedade, Campinas, v. 28, n. 100 – Especial, p. 1203-1230, out. 2007. Disponível em: <http://www.scielo.br/pdf/es/ v28n100/a2628100>. Acesso em: 31 dez. 2016. FREITAS, H. C. L. Neotecnicismo e formação do educador. In: ALVES, N. (Org.). Formação de professores: pensar e fazer. 11. ed. São Paulo: Cortez, 2011. GATTI, B. A. Análise das políticas públicas para formação continuada no Brasil, na última década. Revista Brasileira de Educação, Rio de Janeiro, v. 13, n. 37, p. 57-70, jan./abr. 2008. Disponível em: <http://www.scielo.br/ pdf/rbedu/v13n37/06.pdf>. Acesso em: 2 mar. 17. GATTI, B. A.; BARRETTO, E. S. S.; ANDRÉ, M. E. D. A. Políticas docentes no Brasil: um estado da arte. Brasília: Unesco, 2011. Disponível em: <http:// unesdoc.unesco.org/images/0021/002121/212183por.pdf>. Acesso em: 30 dez. 2016. IMBERNÓN, F. Formação docente e profissional. 3. ed. São Paulo: Cortez, 2000. IMBERNÓN, F. Formação docente e profissional. 3. ed. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 Referências bibliográficas São Paulo: Cortez, 2000. Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017 MAZZEU, L. T. A política de formação docente no Brasil: fundamentos teóricos e metodológicos. In: REUNIÃO ANUAL DA ANPED, 32., 2009, Caxambu. Anais... Caxambu: Anped, 2009. Disponível em: <http://32reuniao.anped.org.br/arquivos/ trabalhos/GT05-5789-Int.pdf>. Acesso em: 31 dez. 2016. MONTANDON, M. I. Políticas públicas para a formação de professores no Brasil: os programas Pibid e Prodocência. Revista da ABEM, Londrina, v. 20, n. 28, p. 47-60, 2012. Disponível em: <http://www.abemeducacaomusical.com.br/ revistas/revistaabem/index.php/revistaabem/article/viewFile/103/86>. Acesso em: 2 mar. 2017. NÓVOA, A. Formação de professores e profissão docente. In: NÓVOA, A. (org.). Os professores e sua formação. Lisboa: Publicações Dom Quixote, 1992. SAVIANI, D. A nova lei da educação: trajetória, limites e perspectivas. Campinas: Autores Associados, 1997. SAVIANI, D. Formação de professores no Brasil: dilemas e perspectivas. Poíesis Pedagógica, Catalão, v. 9, n. 1, p. 7-19, jan./jun. 2011. Disponível em: <https://www.revistas.ufg.br/poiesis/article/view/15667>. Acesso em: 5 jan. 2017. TOMMASI, L.; WARDE, M. J; HADDAD, S. O Banco Mundial e as políticas educacionais. São Paulo: Cortez, 2009. WEBER, S. O Plano Nacional de Educação Docente: confluências do debate nacional. Cadernos Cedes, Campinas, v. 35, n. 97, p. 495-515, set./dez. 2015. Disponível em: <http://www.scielo.br/pdf/ccedes/v35n97/1678-7110- ccedes-35-97-00495.pdf>. Acesso em 5 jan. 2017. 39 Marinalva Vieira Barbosa, doutora em Linguística pela Universidade Estadual de Campinas (Unicamp), com estágio no Laboratório Interactions, Corpus, Apprentissages, Représentations (ICAR), ligado à Université Lyon 2 e à Ecole Normale Superieure de Sciences Humaines et Lettres de Lyon, e pós-doutorado na Faculdade de Educação da Unicamp, é professora adjunta do Departamento de Linguística e Língua Portuguesa da Universidade Federal do Triângulo Mineiro (UFTM). Foi coordenadora institucional do Pibid no período de julho de 2011 a agosto de 2015, docente permanente do Programa de Mestrado em Educação e do Profletras da UFTM. marinalva.barbosa@uftm.edu.br Natália Aparecida Morato Fernandes, doutora em Sociologia, é professora do Departamento de Filosofia e Ciências Sociais da Universidade Federal do Triângulo Mineiro (UFTM). Atua nos cursos de Licenciatura e Serviço Social e realiza atividades de pesquisa e extensão sobre as temáticas: políticas educacionais; formação de professores; práticas de ensino interdisciplinares; mídia, educação e novas cidadanias. natmorato@gmail.com natmorato@gmail.com natmorato@gmail.com Em Aberto, Brasília, v. 30, n. 98, p. 23-39, jan./abr. 2017
https://openalex.org/W4280623332	https://ltu.diva-portal.org/smash/get/diva2:1657425/FULLTEXT01	English	null	Evaluation of Machine Learning versus Empirical Models for Monthly Reference Evapotranspiration Estimation in Uttar Pradesh and Uttarakhand States, India	Sustainability	2,022	cc-by	11,032	Priya Rai 1, Pravendra Kumar 1, Nadhir Al-Ansari 2 and Anurag Malik 3,* Evaluation of Machine Learning Versus Empirical Models for Monthly Reference Evapotranspiration Estimation in Uttar Pradesh and Uttarakhand States, India. Sustainability 2022, 14, 5771. https:// doi.org/10.3390/su14105771 Academic Editor: Venkat Sridhar Received: 4 April 2022 Accepted: 9 May 2022 Published: 10 May 2022 Published: 10 May 2022 Keywords: evapotranspiration; machine learning models; empirical models; statistical indicators Publisher’s Note: MDPI stays neu- tral with regard to jurisdictional claims in published maps and institu- tional affiliations. Priya Rai 1, Pravendra Kumar 1, Nadhir Al-Ansari 2 and Anurag Malik 3,* 1 Department of Soil and Water Conservation Engineering, College of Technology, G.B. Pant University of Agriculture and Technology, Pantnagar 263145, Uttarakhand, India; priya.rai8713@gmail.com (P.R.); pravendrak_05@yahoo.co.in (P.K.) 2 Civil, Environmental and Natural Resources Engineering, Lulea University of Technology, 97187 Lulea, Sweden; nadhir.alansari@ltu.se 3 Regional Research Station, Punjab Agricultural University, Bathinda 151001, Punjab, India * C d lik19@ d 3 Regional Research Station, Punjab Agricultural University, Bathinda 151001, Punjab, India * Correspondence: amalik19@pau.edu 3 Regional Research Station, Punjab Agricultural University, Bathinda 151001, Punjab, India * Correspondence: amalik19@pau.edu 3 Regional Research Station, Punjab Agricultural University, Bathinda 151001 Abstract: Reference evapotranspiration (ETo) plays an important role in agriculture applications such as irrigation scheduling, crop simulation, water budgeting, and reservoir operations. There- fore, the accurate estimation of ETo is essential for optimal utilization of available water resources on regional and global scales. The present study was conducted to estimate the monthly ETo at Nagina (Uttar Pradesh State) and Pantnagar (Uttarakhand State) stations by employing the three ML (machine learning) techniques including the SVM (support vector machine), M5P (M5P model tree), and RF (random forest) against the three empirical models (i.e., Valiantzas-1: V-1, Valiantzas- 2: V-2, Valiantzas-3: V-3). Three different input combinations (i.e., C-1, C-2, C-3) were formulated by using 8-year (2009–2016) climatic data of wind speed (u), solar radiation (Rs), relative humidity (RH), and mean air temperature (T) recorded at both stations. The predictive efficacy of ML and the empirical models was evaluated based on five statistical indicators i.e., CC (correlation coefficient), WI (Willmott index), EC (efficiency coefficient), RMSE (root mean square error), and MAE (mean absolute error) presented through a heatmap along with graphical interpretation (Taylor diagram, time-series, and scatter plots). The results showed that the SVM-1 model corresponding to the C-1 input combination outperformed the other ML and empirical models at both stations. Moreover, the SVM-1 model had the lowest MAE (0.076, 0.047 mm/month) and RMSE (0.110, 0.063 mm/month), and highest EC (0.995, 0.999), CC (0.998, 0.999), and WI (0.999, 1.000) values during validation period at Nagina and Pantnagar stations, respectively, and closely followed by the M5P model. Consequently, the ML model (i.e., SVM) was found to be more robust, and reliable in monthly ETo estimation and can be used as a promising alternative to empirical models at both study locations. Citation: Rai, P.; Kumar, P.; Al-Ansari, N.; Malik, A. [10] applied six ML models including DNN (deep neural network), TCN (temporal convolution neural network), LSTM (long short-term memory), SVM, and RF, and seven empirical models, i.e., Hargreaves, modified Hargreaves, Ritchie, Priestley- Taylor, Makkink, Romanenko, and Schendel, to estimate the daily ETo on the Northeast Plain of China. The results of the investigation demonstrate that the ML models per- formed superior to the empirical models. Mehdizadeh et al. [23] coupled ANFIS with SFLA (shuffled frog leaping algorithm) and IWO (invasive weed optimization) algorithms for estimation of daily ETo at the Tabriz and Shiraz stations of Iran. The performances of the ANFIS-SFLA and ANFIS-IWO models were compared with the Priestley–Taylor, Har- greaves–Samani, Romanenko, and Valiantzas models, and noted that the ANFIS-IWO model provides better estimates than the other models. Adnan et al. [24] estimated monthly ETo at the Dhaka and Mymensing stations of south-central Bangladesh using the ANFIS-MFO (moth flame optimization), ANFIS-WCA (water cycle algorithm), and AN- FIS-WCOMFO models. Results of the evaluations reveal that the hybrid ANFIS- WCOMFO model performed superior to the other models. In a related context recently, several nature-inspired algorithms have been embedded with ML models to optimize their performance in ETo estimation [25]. Alizamir et al. [1] estimated monthly ETo at two sites (Antalya and Isparta) placed in Turkey by employing the hybrid of the ANFIS-PSO (particle swarm optimization) and ANFIS-GA (genetic al- gorithm) against the classical CART (classification and regression tree), ANN, and ANFIS models. They reported that the hybrid ANFIS-PSO and ANFIS-GA models produce better estimates than other models at both stations. Maroufpoor et al. [26] applied the hybrid ANN-GWO (grey wolf optimizer) for estimating the monthly ETo in five different climates (i.e., arid, semi-arid, hyper-arid, humid, and sub-humid) of Iran. The efficacy of ANN- GWO was compared against the ANN and LSSVR (least square support vector regression) models, and found that the hybrid ANN-GWO model was more efficient than other mod- els in all climates. Rezaabad et al. [27] predicted the daily ETo in the Kerman province of Iran by coupling the ANFIS with the IWO (weed optimization algorithm), ICA (imperial- ist competitive algorithm), TLBO (teaching-learning-based optimization), and BBO (bio- geography-based optimization) algorithms. They found that the ANFIS-ICA model with EC = 0.98, RMSE = 0.50 mm/day and CC = 0.99 was superior to other models. Chia et al. For optimal utilization of scarce water resources, an accurate estimation of crop evap- otranspiration (ETc) is crucial for running large irrigation systems by enhancing the water application efficiency [1,2]. Moreover, the ETc plays an important role in acquiring knowledge about the appropriate management of water resources, irrigation scheduling, crop water use, crop production, and water conservation [2]. Usually, ETc is estimated by computing the reference evapotranspiration (ETo) and then multiplying ETo with Kc (crop coefficient) [3,4]. Therefore, the ETo is the key factor to improve irrigation and water use efficiencies [5]. Accordingly, the Penman–Monteith (PM) model was introduced by the Copyright: © 2022 by the authors. Li- censee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and con- ditions of the Creative Commons At- tribution (CC BY) license (https://cre- ativecommons.org/licenses/by/4.0/). www.mdpi.com/journal/sustainability Sustainability 2022, 14, 5771. https://doi.org/10.3390/su14105771 Sustainability 2022, 14, 5771 2 of 21 2 of 21 FAO (Food and Agriculture Organization) and is considered a benchmark model for ETo computation [6]. The FAO-56 PM model requires numerous climatic parameters to esti- mate ETo, which are often incomplete or unavailable, especially in developing countries [7,8]. Hence, the best possible alternatives are necessary to implement, requiring less cli- matic data for ETo estimation [9]. In the last decade, the aforesaid issues have been tackled by the ML models to esti- mate ETo with limited climatic variables on different time scales [10–13]. Some of the ML models such as SVM, RF, M5P, ELM (extreme learning machine), ANN (artificial neural network), ANFIS (adaptive neuro-fuzzy inference system), XGBoost (extreme gradient boosting), MARS (multivariate adaptive regression splines), and GEP (gene expression programming) received the massive application in ETo estimation [14–19]. The results of these studies report the better performance of ML models in comparison to empirical models. Apart from that, the ML models have become popular in modelling watershed hydrology [20,21]. Furthermore, Ashrafzadeh et al. [22] employed the SVM, GMDH (group method of data handling), and SARIMA (seasonal autoregressive integrated mov- ing average) techniques to estimate the monthly ETo in the Guilan Plain of Northern Iran. They noted the better feasibility of the SVM, GMDH, and SARIMA models in the study region. Chen et al. [28] optimized the ELM with three nature-inspired algorithms, namely PSO, MFO (moth– flame optimization), and WOA (whale optimization algorithm) for estimating daily ETo at the Sibu, Miri, and Sandakan sites (Malaysia). Results showed that the ELM-WOA mod- els outperformed the other models at all locations with RMSE of 0.0011 to 0.1927 mm/day, MAE of 0.0007 to 0.1443 mm/day, and R2 (determination coefficient) of 0.9486 to 1.0000. Sustainability 2022, 14, 5771 3 of 21 However, these studies also support the better viability of the ML models enhanced with numerous nature-inspired algorithms. From the above-mentioned literature, it was noted that several studies have been conducted on ETo estimation on different time scales in different climates. However, ac- cording to our knowledge, so far, the support vector machine (SVM), M5P model tree (M5P), random forest (RF), and empirical models (i.e., Valiantzas-1, Valiantzas-2, Valian- tzas-3) were not used for monthly ETo estimation at the Nagina and Pantnagar stations. Thus, this study was optimized with the specific objectives as (i) to formulate the three ML models, i.e., SVM, M5P, and RF, for monthly ETo estimation at both locations, and (ii) to compare the efficacy of three ML models against the empirical models based on statis- tical and graphical investigations. Moreover, the SVM model has better generalization ability than other ML models [29]. It is also highly robust to outliers [30]. Therefore, it is expected that the SVM can provide a better estimation of ETo, which is highly complex and contains a large number of outliers. ETo is one of the complex and vital hydrological variables, so this way of simulation will improve the estimation accuracy of ETo and will help in maintaining the agricultural water resources management operation for control- ling the increasing water stress in agriculture caused by global ecological fluctuations. 2.1. Study Site and Data Information 2.1. Study Site and Data Information Figure 1 demonstrates the location map of the Nagina and Pantnagar stations posi- tioned in Uttar Pradesh and Uttarakhand States of India. The monthly climatic data of mean air temperature (T, °C), relative humidity (RH, %), wind speed (u, m/s), and solar radiation (Rs, MJ/m2/month) of Nagina and Pantnagar from 2009 to 2016 (8-year) were collected from the Rice Research Station of Bijnor district in Uttar Pradesh State (India), and the CRC (Crop Research Centre, Pantnagar, India) of the G.B. Pant University of Ag- riculture and Technology, Uttarakhand State. The 8-year monthly climatic data of both sites was portioned into two phases: (i) a calibration phase that includes 60% data from 2009–2013, and (ii) a validation phase that contains 40% data from 2014–2016 for evalua- tion of machine learning against empirical models. Likewise, Table 1 summarizes the in- formation about the geographical coordinates and descriptive statistics, i.e., minimum, maximum, mean, standard deviation, skewness, and kurtosis, of both stations from 2009 to 2016. It was noted from Table 1 that the maximum ETo variation was of 6.76 mm/month at Nagina and 7.68 mm/month at Pantnagar. 4 of 21 Sustainability 2022, 14, 5771 Figure 1. Location map of study sites. Table 1. Statistical and geographical information of study sites. Station Statistical Properties Climatic Variables Geographical Properties T (oC) RH (%) u (m/s) Rs (MJ/m2/month) ETo (mm/month) Longitude (E) Latitude (N) Altitude (m) Climatic Data (Year) Nagina Minimum 10.900 24.600 0.278 8.300 1.140 78°25′59″ 29°26′35″ 282.0 2009–2016 Maximum 33.000 88.000 1.946 25.000 6.760 Mean 22.994 71.965 1.049 17.089 3.572 Standard deviation 6.412 11.891 0.422 4.611 1.573 Skewness −0.393 −1.235 0.210 −0.035 0.255 Kurtosis −1.274 1.667 −0.925 −0.989 −0.958 Pantnagar Minimum 11.450 41.500 0.584 8.200 1.240 79°38′00″ 29°00′00″ 243.8 2009–2016 Maximum 33.000 86.500 2.752 24.700 7.680 Mean 23.609 67.839 1.415 16.634 3.831 Standard deviation 6.204 11.303 0.527 4.413 1.692 Skewness −0.433 −0.721 0.375 −0.064 0.459 Kurtosis −1.255 −0.473 −0.435 −0.818 −0.629 Figure 1. Location map of study sites. Figure 1. Location map of study sites. Table 1. Statistical and geographical information of study sites. Table 1. Statistical and geographical information of study sites. 2.1. Study Site and Data Information Model Equation Reference V-1 𝐸𝑇𝑜= 0.0393𝑅𝑠√𝑇+ 9.5 −0.19𝑅𝑠 0.6𝜑0.15 + 0.048(𝑇+ 20) (1 −𝑅𝐻 100) 𝑢2 0.7 [31,32] V-2 𝐸𝑇𝑜= 0.0393𝑅𝑠√𝑇+ 9.5 −0.19𝑅𝑠 0.6𝜑0.15 + 0.078(𝑇+ 20) (1 −𝑅𝐻 100) [31,32] V-3 𝐸𝑇𝑜= 0.0393𝑅𝑠√𝑇+ 9.5 −0.19𝑅𝑠 0.6𝜑0.15 + 0.0061(𝑇+ 20)(1.12𝑇−𝑇𝑚𝑖𝑛−2)0.7 [31,32] Note: T = mean air temperature (°C), 𝑢2 = wind speed at 2 m height above ground (m/s), 𝜑 = lati- tude of site (rad), and 𝑇𝑚𝑖𝑛 = minimum temperature (°C). Table 2. Formulas of empirical models used at study sites. Model Equation Reference V-1 𝐸𝑇𝑜= 0.0393𝑅𝑠√𝑇+ 9.5 −0.19𝑅𝑠 0.6𝜑0.15 + 0.048(𝑇+ 20) (1 −𝑅𝐻 100) 𝑢2 0.7 [31,32] V-2 𝐸𝑇𝑜= 0.0393𝑅𝑠√𝑇+ 9.5 −0.19𝑅𝑠 0.6𝜑0.15 + 0.078(𝑇+ 20) (1 −𝑅𝐻 100) [31,32] V-3 𝐸𝑇𝑜= 0.0393𝑅𝑠√𝑇+ 9.5 −0.19𝑅𝑠 0.6𝜑0.15 + 0.0061(𝑇+ 20)(1.12𝑇−𝑇𝑚𝑖𝑛−2)0.7 [31,32] Note: T = mean air temperature (°C), 𝑢2 = wind speed at 2 m height above ground (m/s), 𝜑 = lati- tude of site (rad), and 𝑇𝑚𝑖𝑛 = minimum temperature (°C). Table 2. Formulas of empirical models used at study sites. Table 2. Formulas of empirical models used at study sites. 2.3. Penman-Monteith Model The present study utilized the Penman–Monteith (PM) model given by the Food and Agricultural Organization, with No. 56 designated as FAO-56 PM to compute the monthly ETo values at both study sites and written as [6]: 𝐸𝑇𝑜= 0.408∆(𝑅𝑛−𝐺) + 𝛾 900 𝑇+ 273 𝑢2(𝑒𝑠−𝑒𝑎) ∆+ 𝛾(1 + 0.34𝑢2) (1) (1) where 𝐸𝑇𝑜 = reference evapotranspiration in mm/month, ∆ = slope of saturation vapor pressure in kPa/°C, 𝑅𝑛 = net radiation in MJ/m2/month, 𝐺 = soil heat flux density in MJ/m2/month, 𝛾 = psychrometric constant in kPa/°C, and 𝑒𝑠 and 𝑒𝑎 = saturation and ac- tual vapor pressures in kPa. The computed time-series values of monthly ETo by the FAO- 56 PM model were considered as reference data to appraise the performance of the em- pirical models (i.e., V-1 to V-3) and ML models (i.e., SVM, M5P, and RF). 2.1. Study Site and Data Information The computed time-series values of monthly ETo by the FAO- 56 PM model were considered as reference data to appraise the performance of the em- pirical models (i.e., V-1 to V-3) and ML models (i.e., SVM, M5P, and RF). 2.4. Support Vector Machine Over time, for optimizing the nonlinear problems, the ML models, including the sup- port vector machine (SVM), have been utilized in numerous fields such as for predicting the penetration rate of tunnel-boring machines [33], solar radiation prediction [34], streamflow forecasting [35], landslide hazard modelling [36–38], seawater level simula- tion [39], forecasting electric load [40], and infiltration simulation [41,42]. The SVM ap- proach was recommended by Vapnik [43] and derived from statistical learning theory to solve classification and regression problems [44]. Figure 2 displays the typical assembly of the SVM model. The SVM technique applied the SRM (structural risk minimization) principle [45]. The SVM model utilized a nonlinear mapping function (𝜙(𝑥)) to project the calibration (or training) data points into a high-dimensional feature space, and the following linear regression function is obtained in the feature space [45]: 𝑧 = 𝑓(𝑥) = 𝑤∙𝜙(𝑥) + 𝑏 (2) where 𝑧 = output of SVM, 𝑥 = input of SVM (𝑥1, 𝑥2, … , 𝑥𝑙), 𝑓(𝑥) = loss function, 𝑤 = weight vector of high-dimensional feature space, and 𝑏 = constant. Following the princi- ple of SRM, accepting the ε-insensitive loss function, the minimal 𝑤 is updated for re- 2.2. Empirical Models Valiantzas [31,32] proposed three versions of empirical models, namely (i) Valian- tzas-1 (V-1) with a complete set of climatic data i.e., T, RH, u, and Rs, (ii) Valiantzas-2 (V- 2) without wind speed data i.e., T, RH, and Rs, and (iii) Valiantzas-3 (V-3) without relative humidity and wind speed data, i.e., T, and Rs for computation of reference evapotranspi- ration (ETo). The mathematical expression of V-1 to V-3 is given in Table 2. p Valiantzas [31,32] proposed three versions of empirical models, namely (i) Valian- tzas-1 (V-1) with a complete set of climatic data i.e., T, RH, u, and Rs, (ii) Valiantzas-2 (V- 2) without wind speed data i.e., T, RH, and Rs, and (iii) Valiantzas-3 (V-3) without relative humidity and wind speed data, i.e., T, and Rs for computation of reference evapotranspi- ration (ETo). The mathematical expression of V-1 to V-3 is given in Table 2. Table 2. Formulas of empirical models used at study sites. 2.1. Study Site and Data Information a e g g p y Station Statistical Properties Climatic Variables Geographical Properties T (oC) RH (%) u (m/s) Rs (MJ/m2/month) ETo (mm/month) Longitude (E) Latitude (N) Altitude (m) Climatic Data (Year) Nagina Minimum 10.900 24.600 0.278 8.300 1.140 78°25′59″ 29°26′35″ 282.0 2009–2016 Maximum 33.000 88.000 1.946 25.000 6.760 Mean 22.994 71.965 1.049 17.089 3.572 Standard deviation 6.412 11.891 0.422 4.611 1.573 Skewness −0.393 −1.235 0.210 −0.035 0.255 Kurtosis −1.274 1.667 −0.925 −0.989 −0.958 Pantnagar Minimum 11.450 41.500 0.584 8.200 1.240 79°38′00″ 29°00′00″ 243.8 2009–2016 Maximum 33.000 86.500 2.752 24.700 7.680 Mean 23.609 67.839 1.415 16.634 3.831 Standard deviation 6.204 11.303 0.527 4.413 1.692 Skewness −0.433 −0.721 0.375 −0.064 0.459 Kurtosis −1.255 −0.473 −0.435 −0.818 −0.629 Sustainability 2022, 14, 5771 5 of 21 2.2. Empirical Models Valiantzas [31,32] proposed three versions of empirical models, namely (i) Valian- tzas-1 (V-1) with a complete set of climatic data i.e., T, RH, u, and Rs, (ii) Valiantzas-2 (V- 2) without wind speed data i.e., T, RH, and Rs, and (iii) Valiantzas-3 (V-3) without relative humidity and wind speed data, i.e., T, and Rs for computation of reference evapotranspi- ration (ETo). The mathematical expression of V-1 to V-3 is given in Table 2. Table 2. Formulas of empirical models used at study sites. Model Equation Reference V-1 𝐸𝑇𝑜= 0.0393𝑅𝑠√𝑇+ 9.5 −0.19𝑅𝑠 0.6𝜑0.15 + 0.048(𝑇+ 20) (1 −𝑅𝐻 100) 𝑢2 0.7 [31,32] V-2 𝐸𝑇𝑜= 0.0393𝑅𝑠√𝑇+ 9.5 −0.19𝑅𝑠 0.6𝜑0.15 + 0.078(𝑇+ 20) (1 −𝑅𝐻 100) [31,32] V-3 𝐸𝑇𝑜= 0.0393𝑅𝑠√𝑇+ 9.5 −0.19𝑅𝑠 0.6𝜑0.15 + 0.0061(𝑇+ 20)(1.12𝑇−𝑇𝑚𝑖𝑛−2)0.7 [31,32] Note: T = mean air temperature (°C), 𝑢2 = wind speed at 2 m height above ground (m/s), 𝜑 = lati- tude of site (rad), and 𝑇𝑚𝑖𝑛 = minimum temperature (°C). 2.3. Penman-Monteith Model The present study utilized the Penman–Monteith (PM) model given by the Food and Agricultural Organization, with No. 56 designated as FAO-56 PM to compute the monthly ETo values at both study sites and written as [6]: 𝐸𝑇𝑜= 0.408∆(𝑅𝑛−𝐺) + 𝛾 900 𝑇+ 273 𝑢2(𝑒𝑠−𝑒𝑎) ∆+ 𝛾(1 + 0.34𝑢2) (1) where 𝐸𝑇𝑜 = reference evapotranspiration in mm/month, ∆ = slope of saturation vapor pressure in kPa/°C, 𝑅𝑛 = net radiation in MJ/m2/month, 𝐺 = soil heat flux density in MJ/m2/month, 𝛾 = psychrometric constant in kPa/°C, and 𝑒𝑠 and 𝑒𝑎 = saturation and ac- tual vapor pressures in kPa. 2.4. Support Vector Machine Over time, for optimizing the nonlinear problems, the ML models, including the sup- port vector machine (SVM), have been utilized in numerous fields such as for predicting the penetration rate of tunnel-boring machines [33], solar radiation prediction [34], streamflow forecasting [35], landslide hazard modelling [36–38], seawater level simula- tion [39], forecasting electric load [40], and infiltration simulation [41,42]. The SVM ap- proach was recommended by Vapnik [43] and derived from statistical learning theory to solve classification and regression problems [44]. Figure 2 displays the typical assembly of the SVM model. The SVM technique applied the SRM (structural risk minimization) principle [45]. The SVM model utilized a nonlinear mapping function (𝜙(𝑥)) to project the calibration (or training) data points into a high-dimensional feature space, and the following linear regression function is obtained in the feature space [45]: 𝑧 = 𝑓(𝑥) = 𝑤∙𝜙(𝑥) + 𝑏 (2) (2) where 𝑧 = output of SVM, 𝑥 = input of SVM (𝑥1, 𝑥2, … , 𝑥𝑙), 𝑓(𝑥) = loss function, 𝑤 = weight vector of high-dimensional feature space, and 𝑏 = constant. Following the princi- ple of SRM, accepting the ε-insensitive loss function, the minimal 𝑤 is updated for re- solving the convex optimization problem as follows [29,45]: Sustainability 2022, 14, 5771 6 of 21 { 𝑚𝑖𝑛𝑖𝑚𝑖𝑧𝑒 = 1 2 ‖𝑤‖2 + 𝐶∑(𝑖+ 𝑖 ∗) 𝑙 𝑖=1 𝑠𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜: { 𝑦𝑖−(𝑤∙𝜙(𝑥) + 𝑏) ≤𝜀+ 𝑖 (𝑤∙𝜙(𝑥) + 𝑏) −𝑦𝑖≤𝜀+ 𝑖 ∗ 𝑖, 𝑖 ∗≥0, 𝑖= 1, 2, … , 𝑙 (3) (3) where 𝐶 = penalty factor, i, and i ∗ = slack variables, and ε = tube size insensitive con- stant (Equation (3)). Next, the Lagrangian multiplier is used to resolve the dual convex optimization problem in Equation (3), and the following solution is obtained: 𝐿(𝑤, 𝑏, 𝑖, 𝑖 ∗, 𝑎𝑖, 𝑎𝑖 ∗, 𝜂𝑖, 𝜂𝑖 ∗) = 1 2 ‖𝑤‖2 + 𝐶∑(𝑖+ 𝑖 ∗) 𝑙 𝑖=1 −∑𝑎𝑖(𝑖+ 𝜀−𝑦𝑖+ 𝑤∙𝜙(𝑥𝑖) + 𝑏) 𝑙 𝑖=1 −∑𝑎𝑖 ∗(𝑖 ∗+ 𝜀+ 𝑦𝑖−𝑤∙𝜙(𝑥𝑖) −𝑏) 𝑙 𝑖=1 −∑(𝜂𝑖𝑖+ 𝜂𝑖 ∗𝑖 ∗) 𝑙 𝑖=1 (4) (4) where 𝑎𝑖, 𝑎𝑖 ∗, 𝜂𝑖, and 𝜂𝑖 ∗ = Lagrangian multipliers, which satisfy the non-negative con- straints. 2.4. Support Vector Machine The Lagrangian function (𝐿) minimizes 𝑤, 𝑏, 𝑖, 𝑖 ∗ and maximizes 𝑎𝑖, 𝑎𝑖 ∗, 𝜂𝑖, and 𝜂𝑖 ∗ according to the Karush–Kuhn–Tucker condition, and finally the regression func- tion of SVM can be obtained as: 𝑓(𝑥) = ∑(𝑎𝑖−𝑎𝑖 ∗ 𝑙 𝑖=1 ) 𝑘(𝑥𝑖, 𝑥𝑗) + 𝑏 (5) (5) where 𝑘(𝑥𝑖, 𝑥𝑗) = kernel function (KF), i.e., 𝑘(𝑥𝑖, 𝑥𝑗) = (𝜙(𝑥𝑖) ∙𝜙(𝑥𝑗). The choice of an ap- propriate kernel function improves the performance of the SVM model. A variety of KF are available but the most reliable and efficient is the RBF (radial basis function) [37,46]. The RBF is expressed as [40]: 𝐾(𝑥𝑖, 𝑥𝑗) = 𝑒𝑥𝑝(−𝛾‖𝑥𝑖−𝑥𝑗‖ 2) , 𝛾> 0 (6) (6) where 𝑥𝑖 and 𝑥𝑗 = input space vectors, and 𝛾 = kernel parameter. The 𝐶 and 𝛾 are the two most significant factors, which influence the accuracy of the SVM model. In the pre- sent study, both factors were optimized through the hit-and-trail procedure (C = 2, and  = 0.1) for predicting monthly ETo at two study sites. Further exhaustive background about the SVM can be gained from Vapnik [43], and Smola and Schölkopf [47]. 7 of 21 Sustainability 2022, 14, 5771 Figure 2. The architecture of the SVM model. Figure 2. The architecture of the SVM model. 2.5. M5P Tree The M5P tree is a data-mining technique, projected by Quinlan [48]. The association among output (dependent)-input (independent) variables is established based on a binary decision tree having a linear regression function at the leaf (terminal nodes). The divide- and-conquer approach is applied to produce the tree-based models [49]. Figure 3 displays the well-organized topology of the M5P tree model. The construction of a decision tree involves two stages: step-1: splitting the data into subgroups to create the decision tree by utilizing the principle of the standard deviation (std) and reducing the model training er- ror at node [50], and step-2: pruning of the overfitted tree (sample) and swapping the subtrees with linear regression functions [49]. Finally, the SDR (standard deviation reduc- tion) is computed as [49,50]: 𝑆𝐷𝑅= 𝑠𝑡𝑑(𝑀) −∑\|𝑀𝑗\| \|𝑀\| 𝑠𝑡𝑑(𝑀𝑗) (7) (7) where 𝑀 defines a group of samples that grasps the nodes and 𝑀𝑖 signifies the sub- group of samples that have the jth consequence of the latent set. In recent times, research- ers have explored the successful application of the M5P model in the simulation of several hydrological processes like drought forecasting [50], infiltration simulation [51], river dis- charge forecasting [52,53], reference evapotranspiration estimation [49,54], stage-dis- charge forecasting [55], and groundwater level prediction [56]. For comprehensive infor- mation about the M5P tree, readers refer to Quinlan [48]. 8 of 21 Sustainability 2022, 14, 5771 Figure 3. Structure of the M5P model. 2.6. Random Forest The random forest (RF) algorithm was designed by Breiman [57] for solving high- dimension classification and regression problems. Recently, the RF model received popu- larity in diverse fields of sciences such as, for instance, infiltration rate prediction [51], land use/land cover classification [58], and soil temperature estimation [59]. Figure 4 illus- trates the hierarchical network of the RF classifier. The construction of the RF model com- prises two steps: (i) an ensemble of decision trees (or classifiers) used to build the “RF” through supervised learning, and (ii) making predictions of each decision tree formed in the first step. The RF algorithm is comparatively insensitive to features of the training set and can achieve high prediction accuracy [57]. In the present study, the RF model was built by using a trial-and-error process in WEKA 3.9 software for the prediction of monthly ETo at both study locations. Figure 3. Structure of the M5P model. Figure 3. Structure of the M5P model. 2.6. Random Forest 2.6. Random Forest The random forest (RF) algorithm was designed by Breiman [57] for solving high- dimension classification and regression problems. Recently, the RF model received popu- larity in diverse fields of sciences such as, for instance, infiltration rate prediction [51], land use/land cover classification [58], and soil temperature estimation [59]. Figure 4 illus- trates the hierarchical network of the RF classifier. The construction of the RF model com- prises two steps: (i) an ensemble of decision trees (or classifiers) used to build the “RF” through supervised learning, and (ii) making predictions of each decision tree formed in the first step. The RF algorithm is comparatively insensitive to features of the training set and can achieve high prediction accuracy [57]. In the present study, the RF model was built by using a trial-and-error process in WEKA 3.9 software for the prediction of monthly ETo at both study locations. The random forest (RF) algorithm was designed by Breiman [57] for solving high- dimension classification and regression problems. Recently, the RF model received popu- larity in diverse fields of sciences such as, for instance, infiltration rate prediction [51], land use/land cover classification [58], and soil temperature estimation [59]. Figure 4 illus- trates the hierarchical network of the RF classifier. The construction of the RF model com- prises two steps: (i) an ensemble of decision trees (or classifiers) used to build the “RF” through supervised learning, and (ii) making predictions of each decision tree formed in the first step. The RF algorithm is comparatively insensitive to features of the training set and can achieve high prediction accuracy [57]. In the present study, the RF model was built by using a trial-and-error process in WEKA 3.9 software for the prediction of monthly ETo at both study locations. 9 of 21 9 of 21 Sustainability 2022, 14, 5771 Figure 4. Typical structure of the RF model. Figure 4. Typical structure of the RF model. Figure 4. Typical structure of the RF model. 2.7. Model Formulation and Statistical Indicators 2.7. Model Formulation and Statistical Indicators Different combinations of four climatic variables, namely T, RH, u, and Rs, were used for the estimation of monthly ETo at two locations in the present research. Three combina- tions of four inputs were formulated based on Valiantzas’ [31,32] concept and presented in Table 3. All inputs are used for C-1, three inputs for C-2, and two inputs for C-3. All three input combinations were used to train and test the ML models. Table 3. Different input combinations for the formulation of ML models at study sites. Table 3. Different input combinations for the formulation of ML models at study sites. Combination Inputs Output ML Models C-1 T, RH, u, Rs ETo SVM, M5P, RF C-2 T, RH, Rs ETo SVM, M5P, RF C-3 T, Rs ETo SVM, M5P, RF Afterward, five statistical indicators, i.e., MAE (mean absolute error), RMSE (root- mean-square error), EC (efficiency coefficient), CC (correlation coefficient), and WI (Will- mott index), were utilized to evaluate the predictive efficacy of the empirical (i.e., V-1 to V-3) and ML (i.e., SVM, M5P, RF) models used in the present study. Also, the graphical inspection includes temporal variation graphs, scatter plots, and Taylor diagrams that were used to make a clear interpretation of results yielded by the empirical and ML mod- els. Table 4 shows the formulas of MAE, RMSE, EC, CC, and WI along with their range. Table 3. Different input combinations for the formulation of ML models at study sites. Combination Inputs Output ML Models C-1 T, RH, u, Rs ETo SVM, M5P, RF C-2 T, RH, Rs ETo SVM, M5P, RF C-3 T, Rs ETo SVM, M5P, RF Afterward, five statistical indicators, i.e., MAE (mean absolute error), RMSE (root- mean-square error), EC (efficiency coefficient), CC (correlation coefficient), and WI (Will- mott index), were utilized to evaluate the predictive efficacy of the empirical (i.e., V-1 to V-3) and ML (i.e., SVM, M5P, RF) models used in the present study. Also, the graphical inspection includes temporal variation graphs, scatter plots, and Taylor diagrams that were used to make a clear interpretation of results yielded by the empirical and ML mod- els. Table 4 shows the formulas of MAE, RMSE, EC, CC, and WI along with their range. 10 of 21 Sustainability 2022, 14, 5771 Table 4. Formulas of different performance indicators. 2.7. Model Formulation and Statistical Indicators Equation Range Reference 𝑀𝐴𝐸= ∑ \|𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖−𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖\| 𝑁 𝑖=1 𝑁 (0 < MAE < ∞) [60,61] 𝑅𝑀𝑆𝐸= √∑ (𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖− 𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖)2 𝑁 𝑖=1 𝑁 (0 < RMSE < ∞) [62,63] 𝐸𝐶= 1 −[∑ (𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖− 𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖)2 𝑁 𝑖=1 ∑ (𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖− 𝐸𝑇𝑜𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅)2 𝑁 𝑖=1 ] (-∞ < EC < 1) [64,65] 𝐶𝐶 = ∑ (𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖− 𝐸𝑇𝑜𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅) (𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖 − 𝐸𝑇𝑜𝑒𝑠𝑡 ̅̅̅̅̅̅̅) 𝑁 𝑖=1 √∑ (𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖− 𝐸𝑇𝑜𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅)2 ∑ (𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖 − 𝐸𝑇𝑜𝑒𝑠𝑡 ̅̅̅̅̅̅̅)2 𝑁 𝑖=1 𝑁 𝑖=1 (−1 < CC < 1) [65,66] 𝑊𝐼 = 1 −[ ∑ (𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖−𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖)2 𝑁 𝑖=1 ∑ (\|𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖 − 𝐸𝑇𝑜𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅\| + \|𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖− 𝐸𝑇𝑜𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅\|)2 𝑁 𝑖=1 ] (0 < WI ≤ 1) [67,68] Note: 𝐸𝑇𝑜 𝑝𝑟𝑒,𝑖, and 𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖 = estimated and observed monthly reference evapotranspiration values at an ith time step, N = number of observations. 𝐸𝑇𝑜 𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅, and 𝐸𝑇𝑜 𝑝𝑟𝑒 ̅̅̅̅̅̅̅̅ = mean of observed and pre- dicted monthly reference evapotranspiration. Table 4. Formulas of different performance indicators. Equation Range Reference 𝑀𝐴𝐸= ∑ \|𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖−𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖\| 𝑁 𝑖=1 𝑁 (0 < MAE < ∞) [60,61] 𝑅𝑀𝑆𝐸= √∑ (𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖− 𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖)2 𝑁 𝑖=1 𝑁 (0 < RMSE < ∞) [62,63] 𝐸𝐶= 1 −[∑ (𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖− 𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖)2 𝑁 𝑖=1 ∑ (𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖− 𝐸𝑇𝑜𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅)2 𝑁 𝑖=1 ] (-∞ < EC < 1) [64,65] 𝐶𝐶 = ∑ (𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖− 𝐸𝑇𝑜𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅) (𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖 − 𝐸𝑇𝑜𝑒𝑠𝑡 ̅̅̅̅̅̅̅) 𝑁 𝑖=1 √∑ (𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖− 𝐸𝑇𝑜𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅)2 ∑ (𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖 − 𝐸𝑇𝑜𝑒𝑠𝑡 ̅̅̅̅̅̅̅)2 𝑁 𝑖=1 𝑁 𝑖=1 (−1 < CC < 1) [65,66] 𝑊𝐼 = 1 −[ ∑ (𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖−𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖)2 𝑁 𝑖=1 ∑ (\|𝐸𝑇𝑜 𝑒𝑠𝑡,𝑖 − 𝐸𝑇𝑜𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅\| + \|𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖− 𝐸𝑇𝑜𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅\|)2 𝑁 𝑖=1 ] (0 < WI ≤ 1) [67,68] Note: 𝐸𝑇𝑜 𝑝𝑟𝑒,𝑖, and 𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖 = estimated and observed monthly reference evapotranspiration values at an ith time step, N = number of observations. 𝐸𝑇𝑜 𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅, and 𝐸𝑇𝑜 𝑝𝑟𝑒 ̅̅̅̅̅̅̅̅ = mean of observed and pre- dicted monthly reference evapotranspiration. Table 4. Formulas of different performance indicators. Table 4. Formulas of different performance indicators. Note: 𝐸𝑇𝑜 𝑝𝑟𝑒,𝑖, and 𝐸𝑇𝑜 𝑜𝑏𝑠,𝑖 = estimated and observed monthly reference evapotranspiration values at an ith time step, N = number of observations. 𝐸𝑇𝑜 𝑜𝑏𝑠 ̅̅̅̅̅̅̅̅, and 𝐸𝑇𝑜 𝑝𝑟𝑒 ̅̅̅̅̅̅̅̅ = mean of observed and pre- dicted monthly reference evapotranspiration. 3.1. Model Evaluation Based on Statistical Indicators 3.1. Model Evaluation Based on Statistical Indicators The potential of the ML models, i.e., SVM, M5P, and RF, was investigated against the empirical models (i.e., V-1, V-2, and V-3) at the Nagina and Pantnagar stations based on statistical indicators. These models were trained with 60% data (2009–2013) and tested with 40% data (2014–2016) for both locations. The values of the statistical indicators (i.e., MAE, RMSE, EC, CC, and WI) of the empirical and ML models during the validation phase on the Nagina and Pantnagar stations are presented through a heatmap (see Figures 5 and 6), respectively. For Nagina, from Figure 5 for input combination C-1 corresponding to SVM-1, M5P-1, RF-1, and V-1 models, the values of MAE ranges from 0.076 to 0.210 mm/month, RMSE from 0.110 to 0.269 mm/month, EC from 0.995 to 0.970, CC from 0.998 to 0.986, and WI from 0.999 to 0.993 during the validation phase. For C-2 combination equivalent to SVM-2, M5P-2, RF-2, and V-2 models the MAE = 0.106 to 0.392 mm/month, RMSE = 0.201 to 0.504 mm/month, EC = 0.983 to 0.895, CC = 0.993 to 0.975, and WI = 0.996 to 0.975, and for C-3 combination matching to SVM-3, M5P-3, RF-3, and V-3 models the MAE = 0.111 to 0.434 mm/month, RMSE = 0.148 to 0.500 mm/month, EC = 0.991 to 0.897, CC = 0.996 to 0.988, and WI = 0.998 to 0.972 in the validation phase. It was noted from Figure 5 that the performance of kernel-based models, i.e., SVM-1, SVM-2, and SVM-3, was found to be more optimal than other models. Overall, on Nagina, the best estimation was achieved by the SVM-1 model followed by the M5P-1 model. In addition, the perfor- mance of the empirical models, i.e., V-1, V-2, and V-3, was found to be poor in comparison to the ML models at the Nagina site. The ranking of the models from best to worst is assigned as SVM-1, 2, 3 > M5P-1, 2, 3 > RF-1, 2, 3 > V-1, 2, 3, which are equivalent to C-1, 2, and 3 input combinations. 11 of 21 Sustainability 2022, 14, 5771 Figure 5. Heatmap of statistical indicators values produced by ML and empirical models corre- sponding to C-1 to C-3 input combinations in the validation phase at the Nagina station. Figure 5. 3.1. Model Evaluation Based on Statistical Indicators Heatmap of statistical indicators values produced by ML and empirical models cor sponding to C-1 to C-3 input combinations in the validation phase at the Nagina station. Figure 5. Heatmap of statistical indicators values produced by ML and empirical models corre- sponding to C-1 to C-3 input combinations in the validation phase at the Nagina station. 12 of 21 Sustainability 2022, 14, 5771 Figure 6. Heatmap of statistical indicators values produced by ML and empirical models corre- sponding to C-1 to C-3 input combinations in the validation phase at the Pantnagar station. Figure 6. Heatmap of statistical indicators values produced by ML and empirical models corre- sponding to C-1 to C-3 input combinations in the validation phase at the Pantnagar station. Figure 6. Heatmap of statistical indicators values produced by ML and empirical models corre- sponding to C-1 to C-3 input combinations in the validation phase at the Pantnagar station. Similarly, Figure 6 illustrates the statistical indicators of the SVM, M5P, RF, and Val- iantzas models in monthly ETo estimation on Pantnagar during the validation period. Here also the SVM-1, SVM-2, and SVM-3 had the lowest values of MAE (0.047, 0.141, 0.168 mm/month) and RMSE (0.063, 0.180, 0.226 mm/month), and the highest values of EC (0.999, 0.988, 0.981), CC (0.999, 0.995, 0.991), and WI (1.000, 0.997, 0.995), followed by the M5P, RF, and Valiantzas models corresponding to input combinations C-1, C-2, and C-3. The SVM model corresponding to C-1 produces better estimates than other models. In addition, the worst estimates were produced by the Valiantzas models. The models ranked from best to worst as SVM-1 > M5P-1 > V-1 > RF-1 for C-1 (i.e., T, RH, u, Rs); SVM- 2 > M5P-2 > RF-2 > V-2 for C-2 (i.e., T, RH, Rs), and SVM-3 > M5P-3 > RF-3 > V-3 for C-3 (i.e., T, Rs). From this analysis, it is renowned that the SVM model with C-1 input combi- nation including T, RH, u, and Rs climatic variables performed in a superior manner at both study stations. 3.2. Performance Evaluation using Graphical Inspection The graphical inspection was another goodness-of-fit criterion for evaluating the rel- ative performance of the ML and empirical models during the validation phases at both stations. Figures 7a–c and 8a–c illustrates the temporal variation and scatter plots of ob- served versus estimated monthly ETo values by the SVM, M5P, RF, and Valiantzas models equivalent to C-1, C-2, and C-3 input combinations at the Nagina and Pantnagar sites, respectively, during the validation phase. In these figures, the outputs of the four models were fitted with a 1:1 line (best-fit line) with relative error bands of ±10%, and the coeffi- cient of determination (R2) between the observed and model outputs was also presented Sustainability 2022, 14, 5771 13 of 21 on the plots. If the data are concentrated or close to the 1:1 line (black line) within ±10% relative error bands, this indicates better performance of a model. These figures clearly show the higher performance of the SVM model compared to other models during the validation phase on both stations. In addition, the R2 value was found highest in the SVM- 1 model (0.995) for Nagina (Figure 7a), and 0.999 for Pantnagar (Figure 8a) in the valida- tion stage, compared to other ML and empirical models. Overall, the SVM model can be considered optimal in estimating monthly ETo in terms of the results presented in Figures 7a–c and 8a–c. ( ) g ( g ) g ( g ) tion stage, compared to other ML and empirical models. Overall, the SVM model can be considered optimal in estimating monthly ETo in terms of the results presented in Figures 7a–c and 8a–c. Figure 7. Comparison of observed (FAO-56 PM) against estimated ETo values by the ML and Val- iantzas models corresponding to (a) C-1, (b) C-2, and (c) C-3 input combinations during the valida- tion stage at Nagina station. Figure 7. Comparison of observed (FAO-56 PM) against estimated ETo values by the ML and Val- iantzas models corresponding to (a) C-1, (b) C-2, and (c) C-3 input combinations during the valida- tion stage at Nagina station. 14 of 21 Sustainability 2022, 14, 5771 Figure 8. Comparison of observed (FAO-56 PM) against predicted ETo values by the ML and Val- iantzas models corresponding to (a) C-1, (b) C-2, and (c) C-3 input combinations during the valida- tion stage at Pantnagar station. 3.2. Performance Evaluation using Graphical Inspection The performance of the ML and empirical models was also evaluated using the Tay- lor diagram [66]. The obtained result during the validation period is presented in Figures 9a–c and 10a–c for Nagina and Pantnagar, respectively. The red circle on the x-axis of the Figure 8. Comparison of observed (FAO-56 PM) against predicted ETo values by the ML and Val- iantzas models corresponding to (a) C-1, (b) C-2, and (c) C-3 input combinations during the valida- tion stage at Pantnagar station. Figure 8. Comparison of observed (FAO-56 PM) against predicted ETo values by the ML and Val- iantzas models corresponding to (a) C-1, (b) C-2, and (c) C-3 input combinations during the valida- tion stage at Pantnagar station. The performance of the ML and empirical models was also evaluated using the Tay- lor diagram [66]. The obtained result during the validation period is presented in Figures 9a–c and 10a–c for Nagina and Pantnagar, respectively. The red circle on the x-axis of the Taylor diagram represents the observed monthly ETo. A model is considered better if it is near the observed point. Taylor’s diagram compares three statistics (i.e., RMSE, Std, and CC) together in a graphical way and, therefore, provides a reliable assessment of the Sustainability 2022, 14, 5771 15 of 21 15 of 21 relative performance of different models. The Taylor diagram of the models during vali- dation showed a much better performance of SVM compared to other models at the Nagina (Figure 9a–c) and Pantnagar (Figure 10a–c) stations. In addition, the SVM-pre- dicted monthly ETo was found better-correlated with the observed monthly ETo with less RMSE compared to other models on both stations during the validation phase. Likewise, the Std of SVM-predicted ETo was found much closer to observed ETo in comparison to other models on both stations during the validation. Therefore, SVM can be ranked as the best model in terms of the results presented in the Taylor diagram followed by the M5P, RF, and Valiantzas models at both study sites. Figure 9. Taylor’s diagram of ML and empirical models corresponding to (a) C-1, (b) C-2, and (c) C- 3 input combinations during the validation stage at Nagina station. Figure 9. Taylor’s diagram of ML and empirical models corresponding to (a) C-1, (b) C-2, and (c) C- 3 input combinations during the validation stage at Nagina station. 16 of 21 Sustainability 2022, 14, 5771 Figure 10. 3.2. Performance Evaluation using Graphical Inspection Taylor’s diagram of ML and empirical models corresponding to (a) C-1, (b) C-2, and (c) C-3 input combinations during the validation stage at Pantnagar station. Figure 10. Taylor’s diagram of ML and empirical models corresponding to (a) C-1, (b) C-2, and (c) C-3 input combinations during the validation stage at Pantnagar station. 3.3. Discussion Evapotranspiration is a complex hydrological process that depends on the integrated effect of several climatic variables [69]. It also governs the soil moisture, surface runoff, plant growth, and groundwater recharge for optimizing the available water resources [70]. Furthermore, it determines the processes responsible for land–atmosphere interac- tion or formation of the geographical environment, and weather and climate change through ground heat and moisture balance, and water balance and surface heat balance studies [70–72]. Similarly, Seong et al. [73] projected the implications of different potential evapotranspiration (PET) methods on streamflow under climate change in the Susque- hanna River basin of the northeastern United States. They found that the streamflow pro- jections are sensitive to the selection of the PET methods. So, the formulation of a reliable and robust model of ETo estimation is necessary for maintaining water resources and ag- ricultural operations on farmland under a changing climate. The ML models can handle this issue very well. In this study, three ML models such as SVM, M5P, and RF were de- veloped for monthly ETo estimation on two sites (Nagina and Pantnagar) and their out- comes were compared with empirical models. The appraisal of results shows the better Sustainability 2022, 14, 5771 17 of 21 17 of 21 feasibility of the SVM over other models at both sites. Similarly, Kaya et al. [74] estimated daily ETo in the Kosice City area of Slovakia by employing three ML models, i.e., MLP (multilayer perceptron), SVR (support vector regression), and MLR (multi-linear regres- sion). The daily data of wind speed, relative humidity, air temperature, and solar radia- tion were supplied as input to these models. The performance of the ML models was eval- uated against the empirical models (Hargreaves–Samani, Ritchie, & Turc), and it was found that the ML-based models provide better results than the empirical models. Kisi et al. [75] hybridized the M5 model tree with a radial basis function (RM5Tree) for estimating daily ETo at three stations (Antalya, Adana, and Isparta) in Turkey using the daily record of wind speed, relative humidity, air temperature, and solar radiation. The estimates of the RM5Tree model were compared with M5Tree, MLP, RSM (response surface method), and RBFNN (radial basis function neural network). Overall, they found the RM5Tree model provides more optimal results than the other models. 3.3. Discussion Furthermore, the findings of this research were equated to other studies conducted on ETo estimation by exploiting the ML techniques, for instance [1,9,22,76,77]. Tikhama- rine et al. [11] optimized the SVR model with the WOA, MVO (multi-verse optimizer), and ALO (ant-lion optimizer) algorithms to predict the monthly ETo at the Algiers and Tlemcen weather stations located in north Algeria. They found better performance of the SVR-WOA model with WI = 0.9987, 0.9997, CC = 0.9975, 0.9995, EC = 0.9949, 0.9989, RMSE = 0.0808, 0.0617 mm/month, and MAE = 0.0658, 0.0489 for the Algiers and Tlemcen sites, respectively. Gonzalez del Cerro et al. [78] compared the predictive performance of the ANFIS against the radiation and temperature-based empirical models for estimating the daily ETo in Tamil Nadu and the Coimbatore provinces of India. Results reveal that the ANFIS-based model (MAE = 0.0008 mm/day, WI = 0.9999, and CC = 0.9999) with all cli- matic data, i.e., mean air temperature, relative humidity, wind speed, and solar radiation produce better estimates than the empirical models. Ahmadi et al. [79] estimated monthly ETo on six stations located in Iran by exploiting three ML models, namely the SVR, GEP (gene expression programming), SVR-IWD (intelligent water drops) against the Priestley– Taylor, and H-S Hargreaves–Samani models. A comparison of results shows that the SVR- IWD model outperformed the other models at all stations. To this end, the aforementioned studies also recommend the effectiveness of machine learning models over the empirical models in predicting monthly ETo at both study loca- tions. 4. Conclusions and A.M.; visualization, P.R., P.K., N.A.-A. and A.M.; supervision, P.K., N.A.-A. and A.M.; project administration, N.A.-A.; funding acquisition, N.A.-A. All authors have read and agreed to the published version of the manuscript. Author Contributions: Conceptualization, P.R. and A.M.; methodology, P.R. and A.M.; software, P.R. and A.M.; validation, P.R., P.K., N.A.-A. and A.M.; formal analysis, P.R. and A.M.; investiga- tion, P.R., P.K., N.A.-A. and A.M.; writing—original draft preparation, P.R., P.K., N.A.-A. and A.M.; writing—review and editing, P.R., P.K., N.A.-A. and A.M.; visualization, P.R., P.K., N.A.-A. and A.M.; supervision, P.K., N.A.-A. and A.M.; project administration, N.A.-A.; funding acquisition, N.A.-A. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Funding: This research received no external funding. Conflicts of Interest: The authors declare no conflicts of interest. Conflicts of Interest: The authors declare no conflicts of interest. References Crop Evapotranspiration: Guidelines for Computing Crop Requirements. FAO Irrig. Drain. Pap. 56 1998, 300, D05109. 7. Abdullah, S.S.; Malek, M.A.; Abdullah, N.S.; Kisi, O.; Yap, K.S. Extreme Learning Machines: A New Approach for Prediction of Reference Evapotranspiration. J. Hydrol. 2015, 527, 184–195. https://doi.org/10.1016/j.jhydrol.2015.04.073. 8. Tabari, H.; Hosseinzadeh Talaee, P. Multilayer Perceptron for Reference Evapotranspiration Estimation in a Semiarid Region. Neural Comput. Appl. 2013, 23, 341–348. https://doi.org/10.1007/s00521-012-0904-7. 9. Tikhamarine, Y.; Malik, A.; Kumar, A.; Souag-Gamane, D.; Kisi, O. Estimation of Monthly Reference E Novel Hybrid Machine Learning Approaches. Hydrol. Sci. J. 2019, 64, 1824–1842. https://doi.org/10.108 9. Tikhamarine, Y.; Malik, A.; Kumar, A.; Souag-Gamane, D.; Kisi, O. Estimation of Monthly Reference Evapotranspiration Using Novel Hybrid Machine Learning Approaches. Hydrol. Sci. J. 2019, 64, 1824–1842. https://doi.org/10.1080/02626667.2019.1678750. 10. Chen, Z.; Zhu, Z.; Jiang, H.; Sun, S. Estimating Daily Reference Evapotranspiration Based on Limited Meteorological Data Using Deep Learning and Classical Machine Learning Methods. J. Hydrol. 2020, 591, 125286. https://doi.org/10.1016/j.jhydrol.2020.125286. y g pp y J , , p // g/ / 10. Chen, Z.; Zhu, Z.; Jiang, H.; Sun, S. Estimating Daily Reference Evapotranspiration Based on Limited Meteorological Data Using Deep Learning and Classical Machine Learning Methods. J. Hydrol. 2020, 591, 125286. https://doi.org/10.1016/j.jhydrol.2020.125286. 11. Tikhamarine, Y.; Malik, A.; Pandey, K.; Sammen, S.S.; Souag-Gamane, D.; Heddam, S.; Kisi, O. Monthly Evapotranspiration Estimation Using Optimal Climatic Parameters: Efficacy of Hybrid Support Vector Regression Integrated with Whale Optimization Algorithm. Environ. Monit. Assess. 2020, 192, 696. https://doi.org/10.1007/s10661-020-08659-7. p g , , p g 12. Ferreira, L.B.; da Cunha, F.F. New Approach to Estimate Daily Reference Evapotranspiration Based on Hourly Temperature and Relative Humidity Using Machine Learning and Deep Learning. Agric. Water Manag. 2020, 234, 106113. https://doi.org/10.1016/j.agwat.2020.106113. p g j g 13. Saggi, M.K.; Jain, S. Reference Evapotranspiration Estimation and Modeling of the Punjab Northern India Using Deep Learning. Comput. Electron. Agric. 2019, 156, 387–398. https://doi.org/10.1016/j.compag.2018.11.031. p g p g j p g 14. Khosravi, K.; Daggupati, P.; Alami, M.T.; Awadh, S.M.; Ghareb, M.I.; Panahi, M.; Pham, B.T.; Rezaie, F.; Qi, C.; Yaseen, Z.M. Meteorological Data Mining and Hybrid Data-Intelligence Models for Reference Evaporation Simulation: A Case Study in Iraq. Comput. Electron. Agric. 2019, 167, 105041. https://doi.org/10.1016/j.compag.2019.105041. p g p g j p g 15. Wu, L.; Fan, J. Comparison of Neuron-Based, Kernel-Based, Tree-Based and Curve-Based Machine Learning Models for Predicting Daily Reference Evapotranspiration. PLoS ONE 2019, 14, e0217520. https://doi.org/10.1371/journal.pone.0217520. 16. References 1. Alizamir, M.; Kisi, O.; Muhammad Adnan, R.; Kuriqi, A. Modelling Reference Evapotranspiration by Combining Neuro-Fuzzy and Evolutionary Strategies. Acta Geophys. 2020, 68, 1113–1126. https://doi.org/10.1007/s11600-020-00446-9. 1. Alizamir, M.; Kisi, O.; Muhammad Adnan, R.; Kuriqi, A. Modelling Reference Evapotranspiration by Combining Neuro-Fuzzy and Evolutionary Strategies. Acta Geophys. 2020, 68, 1113–1126. https://doi.org/10.1007/s11600-020-00446-9. 1. Alizamir, M.; Kisi, O.; Muhammad Adnan, R.; Kuriqi, A. Modelling Reference Evapotranspiration by Combining Neuro-Fuzzy and Evolutionary Strategies. Acta Geophys. 2020, 68, 1113–1126. https://doi.org/10.1007/s11600-020-00446-9. y g p y p g 2. Awal, R.; Habibi, H.; Fares, A.; Deb, S. Estimating Reference Crop Evapotranspiration under Limited Climate Data in West Texas. J. Hydrol. Reg. Stud. 2020, 28, 100677. https://doi.org/10.1016/j.ejrh.2020.100677. 2. Awal, R.; Habibi, H.; Fares, A.; Deb, S. Estimating Reference Crop Evapotranspiration under Limited Climate Data in West Texas. J. Hydrol. Reg. Stud. 2020, 28, 100677. https://doi.org/10.1016/j.ejrh.2020.100677. y g p g j j 3. Adamala, S.; Raghuwanshi, N.S.; Mishra, A.; Singh, R. Generalized Wavelet Neural Networks for Evap in India. ISH J. Hydraul. Eng. 2019, 25, 119–131. https://doi.org/10.1080/09715010.2017.1327825. y g p g j j 3. Adamala, S.; Raghuwanshi, N.S.; Mishra, A.; Singh, R. Generalized Wavelet Neural Networks for Evapotranspiration Modeling in India. ISH J. Hydraul. Eng. 2019, 25, 119–131. https://doi.org/10.1080/09715010.2017.1327825. 4 P L S All R G S h M R D C E E h FAO 6 P d F A i W 3. Adamala, S.; Raghuwanshi, N.S.; Mishra, A.; Singh, R. Generalized Wavelet Neural Networks for Evapotranspiration Modeling in India. ISH J. Hydraul. Eng. 2019, 25, 119–131. https://doi.org/10.1080/09715010.2017.1327825. y g p g 4. Pereira, L.S.; Allen, R.G.; Smith, M.; Raes, D. Crop Evapotranspiration Estimation with FAO56: P Manag. 2015, 147, 4–20. https://doi.org/10.1016/j.agwat.2014.07.031. 5. Feng, Y.; Cui, N.; Zhao, L.; Hu, X.; Gong, D. Comparison of ELM , GANN , WNN and Empirical Models for Estimating Reference Evapotranspiration in Humid Region of Southwest China Comparison of ELM , GANN , WNN and Empirical Models for Estimating Reference Evapotranspiration in Humid Region of Southwest. J. Hydrol. 2016, 536, 376–383. https://doi.org/10.1016/j.jhydrol.2016.02.053. 5. Feng, Y.; Cui, N.; Zhao, L.; Hu, X.; Gong, D. Comparison of ELM , GANN , WNN and Empirical Models for Estimating Reference Evapotranspiration in Humid Region of Southwest China Comparison of ELM , GANN , WNN and Empirical Models for Estimating Reference Evapotranspiration in Humid Region of Southwest. J. Hydrol. 2016, 536, 376–383. https://doi.org/10.1016/j.jhydrol.2016.02.053. p g j j y 6. Allen, R.G.; Pereira, L.S.; Raes, D.; Smith, M. 4. Conclusions The effectiveness of three machine learning models, such as the support vector ma- chine (SVM), M5P tree (M5P), and random forest (RF), was investigated in predicting monthly ETo on the Nagina and Pantnagar stations from 2009 to 2016 in the present study. From the available 8-year climatic data (2009–2016) at both stations, a total of three com- binations of different inputs were established to calibrate (train) and validate (test) the ML models over the empirical models (i.e., Valiantzas-1, Valiantzas-2, Valiantzas-3) based on statistical indicators and graphical inspection. The results of the evaluation demonstrate that the SVM models with the full set of climatic data, i.e., T, RH, u, and Rs, performed superior to the M5P, RF, and Valiantzas models during the validation period at both lo- cations under this study. In addition, the predictive accuracy of the SVM-1 to SVM-3 mod- els with respect to RMSE improved 32.9% to 59.1%, 4.3% to 60.1%, and 23.7% to 70.4%, for M5P-1 to M5P-3, RF-1 RF-3, and V-1 to V-3, respectively, on Nagina and 59.9% to 66.1%, 16.7% to 48.9%, and 19.6% to 47.6% for M5P-1 to M5P-3, RF-1 RF-3, and V-1 to V-3, respec- tively, on Pantnagar. This percentage analysis also reveals the supremacy of the SVM model in predicting monthly ETo at both sites under consideration. Furthermore, the per- formance of the empirical models was recorded as poor at both sites in comparison to the ML models. Overall, the findings of this research show that the ML models (i.e., SVM) had better efficacy and will support the irrigation engineers, agriculturists, and Sustainability 2022, 14, 5771 18 of 21 18 of 21 hydrologists to formulate smart intelligence systems for optimal planning and manage- ment of water resources at study sites. hydrologists to formulate smart intelligence systems for optimal planning and manage- ment of water resources at study sites. Future research will evaluate ensemble machine learning models with different ra- tios of training and testing datasets obtained from multi-locations of other climatic re- gions. In addition, the geospatial techniques will be considered for mapping the impact of reference evapotranspiration on a spatial scale. Author Contributions: Conceptualization, P.R. and A.M.; methodology, P.R. and A.M.; software, P.R. and A.M.; validation, P.R., P.K., N.A.-A. and A.M.; formal analysis, P.R. and A.M.; investiga- tion, P.R., P.K., N.A.-A. and A.M.; writing—original draft preparation, P.R., P.K., N.A.-A. and A.M.; writing—review and editing, P.R., P.K., N.A.-A. References Fan, J.; Yue, W.; Wu, L.; Zhang, F.; Cai, H.; Wang, X.; Lu, X.; Xiang, Y. Evaluation of SVM, ELM and Four Tree-Based Ensemble Models for Predicting Daily Reference Evapotranspiration Using Limited Meteorological Data in Different Climates of China. Agric. For. Meteorol. 2018, 263, 225–241. https://doi.org/10.1016/j.agrformet.2018.08.019. 19 of 21 19 of 21 Sustainability 2022, 14, 5771 17. Ferreira, L.B.; da Cunha, F.F.; de Oliveira, R.A.; Fernandes Filho, E.I. Estimation of Reference Evapotranspiration in Brazil with Limited Meteorological Data Using ANN and SVM—A New Approach. J. Hydrol. 2019, 572, 556–570. https://doi.org/10.1016/j.jhydrol.2019.03.028. p g j j y 18. Mehdizadeh, S.; Behmanesh, J.; Khalili, K. Using MARS, SVM, GEP and Empirical Equations for Estimation of Monthly Mean Reference Evapotranspiration. Comput. Electron. Agric. 2017, 139, 103–114. https://doi.org/10.1016/j.compag.2017.05.002. 19. Wang, S.; Lian, J.; Peng, Y.; Hu, B.; Chen, H. Generalized Reference Evapotranspiration Models with Limited Climatic Data Based on Random Forest and Gene Expression Programming in Guangxi , China. Agric. Water Manag. 2019, 221, 220–230. https://doi.org/10.1016/j.agwat.2019.03.027. p g j g 20. Sarker, S.; Veremyev, A.; Boginski, V.; Singh, A. Critical Nodes in River Networks. Sci https://doi.org/10.1038/s41598-019-47292-4. j g yev, A.; Boginski, V.; Singh, A. Critical Nodes in River Networks. Sci. Rep. 2019, 9, 11178. 38/s41598-019-47292-4. 21. Sarker, S. Investigating Topologic and Geometric Properties of Synthetic and Natural River Networks under Changing Climate. Dr. Diss. Univ. Cent. Fla. US 2021, 2020, 965. 22. Ashrafzadeh, A.; Kişi, O.; Aghelpour, P.; Biazar, S.M.; Masouleh, M.A. Comparative Study of Time Series Models, Support Vector Machines, and GMDH in Forecasting Long-Term Evapotranspiration Rates in Northern Iran. J. Irrig. Drain. Eng. 2020, 146, 04020010. https://doi.org/10.1061/(asce)ir.1943-4774.0001471. p g ( ) 23. Mehdizadeh, S.; Mohammadi, B.; Pham, Q.B.; Duan, Z. Development of Boosted Machine Learning Models for Estimating Daily Reference Evapotranspiration and Comparison with Empirical Approaches. Water 2021, 13, 3489. https://doi.org/10.3390/w13243489. p g 24. Adnan, R.M.; Mostafa, R.R.; Islam, A.R.M.T.; Kisi, O.; Kuriqi, A.; Heddam, S. Estimating Reference Evapotranspiration Using Hybrid Adaptive Fuzzy Inferencing Coupled with Heuristic Algorithms. Comput. Electron. Agric. 2021, 191, 106541. https://doi.org/10.1016/j.compag.2021.106541. p g j p g 25. Mohammadi, B.; Mehdizadeh, S. Modeling Daily Reference Evapotranspiration via a Novel Approach Based on Support Vector Regression Coupled with Whale Optimization Algorithm. Agric. Water Manag. 2020, 237, 106145. https://doi.org/10.1016/j.agwat.2020.106145. p g j g 26. Maroufpoor, S.; Bozorg-Haddad, O.; Maroufpoor, E. References Reference Evapotranspiration Estimating Based on Optimal Input Combination and Hybrid Artificial Intelligent Model: Hybridization of Artificial Neural Network with Grey Wolf Optimizer Algorithm. J. Hydrol. 2020, 588, 125060. https://doi.org/10.1016/j.jhydrol.2020.125060. g y p g j j y 27. Rezaabad, M.Z.; Ghazanfari, S.; Salajegheh, M. ANFIS Modeling with ICA, BBO, TLBO, and IWO Optimization Algorithms and Sensitivity Analysis for Predicting Daily Reference Evapotranspiration. J. Hydrol. Eng. 2020, 25, 04020038. https://doi.org/10.1061/(ASCE)HE.1943-5584.0001963. p g ( ) 28. Chia, M.Y.; Huang, Y.F.; Koo, C.H. Swarm-Based Optimization as Stochastic Training Strategy for Estimation of Reference Evapotranspiration Using Extreme Learning Machine. Agric. Water Manag. 2021, 243, 106447. https://doi.org/10.1016/j.agwat.2020.106447. p g j g 29. Panahi, M.; Sadhasivam, N.; Pourghasemi, H.R.; Rezaie, F.; Lee, S. Spatial Prediction of Groundwater Potential Mapping Based on Convolutional Neural Network (CNN) and Support Vector Regression (SVR). J. Hydrol. 2020, 588, 125033. https://doi.org/10.1016/j.jhydrol.2020.125033. 30. Borji, M.; Malekian, A.; Salajegheh, A.; Ghadimi, M. Multi-Time-Scale Analysis of Hydrological Drought Forecasting Using Support Vector Regression (SVR) and Artificial Neural Networks (ANN). Arab. J. Geosci. 2016, 9, 725. https://doi.org/10.1007/s12517-016-2750-x. p g 31. Valiantzas, J.D. Simple ET0 Forms of Penman’s Equation without Wind and/or Humidity Data. I: Theoretical Development. J. Irrig. Drain. Eng. 2013, 139, 1–8. https://doi.org/10.1061/(ASCE)IR.1943-4774.0000520. g g p g 32. Valiantzas, J.D. Simple ET0 Forms of Penman’s Equation without Wind and/or Humidity Data. II: Comparisons with Reduced Set-FAO and Other Methodologies. J. Irrig. Drain. Eng. 2013, 139, 9–19. https://doi.org/10.1061/(ASCE)IR.1943-4774.0000502. 33. Afradi, A.; Ebrahimabadi, A. Comparison of Artificial Neural Networks (ANN), Support Vector Machine (SVM) and Gene 32. Valiantzas, J.D. Simple ET0 Forms of Penman’s Equation without Wind and/or Humidity Data. II: Comparisons with Reduced Set-FAO and Other Methodologies. J. Irrig. Drain. Eng. 2013, 139, 9–19. https://doi.org/10.1061/(ASCE)IR.1943-4774.0000502. 33. Afradi, A.; Ebrahimabadi, A. Comparison of Artificial Neural Networks (ANN), Support Vector Machine (SVM) and Gene Expression Programming (GEP) Approaches for Predicting TBM Penetration Rate. SN Appl. Sci. 2020, 2, 2004. https://doi.org/10.1007/s42452-020-03767-y. p g y 34. Biazar, S.M.; Rahmani, V.; Isazadeh, M.; Kisi, O.; Dinpashoh, Y. New Input Selection Procedure for Machine Learning Methods in Estimating Daily Global Solar Radiation. Arab. J. Geosci. 2020, 13, 431. https://doi.org/10.1007/s12517-020-05437-0. 35. Hadi, S.J.; Tombul, M. Forecasting Daily Streamflow for Basins with Different Physical Characteristics through Data-Driven Methods. Water Resour. Manag. 2018, 32, 3405–3422. https://doi.org/10.1007/s11269-018-1998-1. 36. Hong, H.; Pradhan, B.; Jebur, M.N.; Bui, D.T.; Xu, C.; Akgun, A. Spatial Prediction of Landslide Hazard at the Luxi Area (China) Using Support Vector Machines. Environ. 39. Khaledian, M.R.; Isazadeh, M.; Biazar, S.M.; Pham, Q.B. Simulating Caspian Sea Surface Water Level by Artificial Neural Network and Support Vector Machine Models. Acta Geophys. 2020, 68, 553–563. https://doi.org/10.1007/s11600-020-00419-y. References Ali, M.; Deo, R.C.; Downs, N.J.; Maraseni, T. An Ensemble-ANFIS Based Uncertainty Assessment Model for Forecasting Multi- Scalar Standardized Precipitation Index. Atmos. Res. 2018, 207, 155–180. https://doi.org/10.1016/j.atmosres.2018.02.024. p p g j 51. Yaseen, Z.M.; Sihag, P.; Yusuf, B.; Al-Janabi, A.M.S. Modelling Infiltration Rates in Permeable Stormwater Channels Using Soft Computing Techniques. Irrig. Drain. 2020, 70, 117–130. https://doi.org/10.1002/ird.2530. 52. Kisi, O.; Khosravinia, P.; Nikpour, M.R.; Sanikhani, H. Hydrodynamics of River-Channel Confluence: Toward Modeling Separation Zone Using GEP, MARS, M5 Tree and DENFIS Techniques. Stoch. Environ. Res. Risk Assess. 2019, 33, 1089–1107. https://doi.org/10.1007/s00477-019-01684-0. p g 53. Taghi Sattari, M.; Pal, M.; Apaydin, H.; Ozturk, F. M5 Model Tree Application in Daily River Flow Forecasting in Sohu Stream, Turkey. Water Resour. 2013, 40, 233–242. https://doi.org/10.1134/S0097807813030123. y p g 54. Rahimikhoob, A. Comparison between M5 Model Tree and Neural Networks for Estimating Reference Evapotranspiration in an Arid Environment. Water Resour. Manag. 2014, 28, 657–669. 55. Bhattacharya, B.; Solomatine, D.P. Neural Networks and M5 Model Trees in Modelling Water Level–Discharge Relationship. Neurocomputing 2005, 63, 381–396. https://doi.org/10.1016/j.neucom.2004.04.016. p g p g j 56. Sattari, M.T.; Mirabbasi, R.; Sushab, R.S.; Abraham, J. Prediction of Groundwater Level in Ardebil Plain Using Support Vector Regression and M5 Tree Model. Groundwater 2018, 56, 636–646. https://doi.org/10.1111/gwat.12620. g p g g 57. Breiman, L. Random Forests. Mach. Learn. 2001, 45, 5–32. https://doi.org/10.1023/A:1010933404324. 58. Sarker, T. Role of Climatic and Non-Climatic Factors on Land Use and Land Cover Change in the Arctic: A Comparative Analysis of Vorkuta and Salekhard. Masters Diss. Georg. Washingt. Univ. US 2020. Available online: https://scholarspace.library.gwu.edu/etd/6969z1516 (accessed on 5 May 2022). p // p y g / / ( y ) 59. Sihag, P.; Esmaeilbeiki, F.; Singh, B.; Pandhiani, S.M. Model-Based Soil Temperature Estimation Using Climatic Parameters: The Case of Azerbaijan Province, Iran. Geol. Ecol. Landscapes 2020, 4, 203–215. https://doi.org/10.1080/24749508.2019.1610841. 60. Legates, D.R.; McCabe, G.J. Evaluating the Use of “Goodness-of-Fit” Measures in Hydrologic and Hydroclimatic Model Validation. Water Resour. Res. 1999, 35, 233–241. https://doi.org/10.1029/1998WR900018. 61. Malik, A.; Kumar, A.; Salih, S.Q.; Kim, S.; Kim, N.W.; Yaseen, Z.M.; Singh, V.P. Drought Index Prediction Using Advanced Fuzzy Logic Model: Regional Case Study over Kumaon in India. PLoS ONE 2020, 15, e0233280. https://doi.org/10.1371/journal.pone.0233280. p g j p 62. Malik, A.; Kumar, A. Meteorological Drought Prediction Using Heuristic Approaches Based on Effective Drought Index: A Case Study in Uttarakhand. Arab. J. Geosci. 2020, 13, 276. https://doi.org/10.1007/s12517-020-5239-6. 63. Willmott, C.; Matsuura, K. References Earth Sci. 2016, 75, 40. https://doi.org/10.1007/s12665-015-4866-9. 37. Hong, H.; Pradhan, B.; Bui, D.T.; Xu, C.; Youssef, A.M.; Chen, W. Comparison of Four Kernel Functions Used in Support Vector Machines for Landslide Susceptibility Mapping: A Case Study at Suichuan Area (China). Geomatics Nat. Hazards Risk 2017, 8, 544–569. https://doi.org/10.1080/19475705.2016.1250112. 38. Naghibi, S.A.; Pourghasemi, H.R.; Dixon, B. GIS-Based Groundwater Potential Mapping Using Boosted Regression Tree, Classification and Regression Tree, and Random Forest Machine Learning Models in Iran. Environ. Monit. Assess. 2016, 188, 44. g g 39. Khaledian, M.R.; Isazadeh, M.; Biazar, S.M.; Pham, Q.B. Simulating Caspian Sea Surface Water Level by Artificial Neural Network and Support Vector Machine Models. Acta Geophys. 2020, 68, 553–563. https://doi.org/10.1007/s11600-020-00419-y. 20 of 21 Sustainability 2022, 14, 5771 40. Zhang, X.; Wang, J.; Zhang, K. Short-Term Electric Load Forecasting Based on Singular Spectrum Analysis and Support Vector Machine Optimized by Cuckoo Search Algorithm. Electr. Power Syst. Res. 2017, 146, 270–285. https://doi.org/10.1016/j.epsr.2017.01.035. p g j p 41. Sihag, P.; Tiwari, N.K.; Ranjan, S. Support Vector Regression-Based Modeling of Cumulative Infiltration of Sandy Soil. ISH J. Hydraul. Eng. 2018, 26, 138–152. https://doi.org/10.1080/09715010.2018.1439776. y g p g 42. Sihag, P.; Singh, V.P.; Angelaki, A.; Kumar, V.; Sepahvand, A.; Golia, E. Modelling of Infiltration Using Artificial Intelligence Techniques in Semi-Arid Iran. Hydrol. Sci. J. 2019, 64, 1647–1658. https://doi.org/10.1080/02626667.2019.1659965. q y p g 43. Vapnik, V.N. The Nature of Statistical Learning Theory; Springer: New York, NY, USA, 1995; p. 314. 44. Pourghasemi, H.R.; Jirandeh, A.G.; Pradhan, B.; Xu, C.; Gokceoglu, C. Landslide Susceptibility Mappi Machine and GIS at the Golestan Province, Iran. J. Earth Syst. Sci. 2013, 122, 349–369. https://doi.org/10 45. Su, H.; Li, X.; Yang, B.; Wen, Z. Wavelet Support Vector Machine-Based Prediction Model of Dam Deformation. Mech. Syst. Signal Process. 2018, 110, 412–427. https://doi.org/10.1016/j.ymssp.2018.03.022. p g j y p 46. Granata, F. Evapotranspiration Evaluation Models Based on Machine Learning Algorithms—A Comparative Study. Agric. Water Manag. 2019, 217, 303–315. https://doi.org/10.1016/j.agwat.2019.03.015. 47. Smola, A.J.; Schölkopf, B. A Tutorial on Support Vector Regression. Stat. Comput. 2004, 14, 199–222. https://doi.org/10.1023/B:STCO.0000035301.49549.88. p g 48. Quinlan, J.R. Learning with Continuous Classes. In Proceedings of the 5th Australian joint conference on artificial intelligence, Hobart, Australia, 16–18 November 1992; Volume 92, pp. 343–348. pp 49. Pal, M.; Deswal, S. M5 Model Tree Based Modelling of Reference Evapotranspiration. Hydrol. Proc https://doi.org/10.1002/hyp.7266. pp M5 Model Tree Based Modelling of Reference Evapotranspiration. Hydrol. Process. 2009, 23, 1437–1443 002/hyp.7266. p g yp 50. , J y g , , p // g/ / 68. Malik, A.; Kumar, A.; Rai, P.; Kuriqi, A. Prediction of Multi-Scalar Standardized Precipitation Index by Using Artificial Intelligence and Regression Models. Climate 2021, 9, 28. https://doi.org/10.3390/cli9020028. g n the Validation of Models. Phys. Geogr. 1981, 2, 184–194. https://doi.org/10.1080/02723646.1981.10642213. A R i P K i i A P di i f M l i S l S d di d P i i i I d b U i A if References Advantages of the Mean Absolute Error (MAE) over the Root Mean Square Error (RMSE) in Assessing Average Model Performance. Clim. Res. 2005, 30, 79–82. https://doi.org/10.3354/cr030079. 64. Nash, J.E.; Sutcliffe, J.V. River Flow Forecasting through Conceptual Models Part I—A Discussion of Principles. J. Hydrol. 1970, 10, 282–290. https://doi.org/10.1016/0022-1694(70)90255-6. p g 65. Moriasi, D.N.; Gitau, M.W.; Pai, N.; Daggupati, P. Hydrologic and Water Quality Models: Perf Evaluation Criteria. Trans. ASABE 2015, 58, 1763–1785. https://doi.org/10.13031/trans.58.10715. 65. Moriasi, D.N.; Gitau, M.W.; Pai, N.; Daggupati, P. Hydrologic and Water Quality Models: Performance Measures and Evaluation Criteria. Trans. ASABE 2015, 58, 1763–1785. https://doi.org/10.13031/trans.58.10715. 66 Taylor K E Summarizing Multiple Aspects of Model Performance in a Single Diagram J Geophys Res Atmos 2001 106 7183 65. Moriasi, D.N.; Gitau, M.W.; Pai, N.; Daggupati, P. Hydrologic and Water Quality Models: Performance Measures and Evaluation Criteria. Trans. ASABE 2015, 58, 1763–1785. https://doi.org/10.13031/trans.58.10715. 66. Taylor, K.E. Summarizing Multiple Aspects of Model Performance in a Single Diagram. J. Geophys. Res. Atmos. 2001, 106, 7183– gg p y g y Evaluation Criteria. Trans. ASABE 2015, 58, 1763–1785. https://doi.org/10.13031/trans.58.10715. 66. Taylor, K.E. Summarizing Multiple Aspects of Model Performance in a Single Diagram. J. Geophys. Res. Atmos. 2001, 106, 7183– 7192. https://doi.org/10.1029/2000JD900719. 66. Taylor, K.E. Summarizing Multiple Aspects of Model Performance in a Single Diagram. J. Geophys. Re 7192. https://doi.org/10.1029/2000JD900719. rizing Multiple Aspects of Model Performance in a Single Diagram. J. Geophys. Res. Atmos. 2001, 106, 7183– /10.1029/2000JD900719. p g 67. Willmott, C.J. On the Validation of Models. Phys. Geogr. 1981, 2, 184–194. https://doi.org/10.1080/02723646.1981.10642213. 7. Willmott, C.J. On the Validation of Models. Phys. 68. Malik, A.; Kumar, A.; Rai, P.; Kuriqi, A. Prediction of Multi-Scalar Standardized Precipitation Index by Using Artificial Intelligence and Regression Models. Climate 2021, 9, 28. https://doi.org/10.3390/cli9020028. 21 of 21 Sustainability 2022, 14, 5771 69. Ye, L.; Zahra, M.M.A.; Al-Bedyry, N.K.; Yaseen, Z.M. Daily Scale Evapotranspiration Prediction over the Coastal Region of Southwest Bangladesh: New Development of Artificial Intelligence Model. Stoch. Environ. Res. Risk Assess. 2022, 36, 451–471. https://doi.org/10.1007/s00477-021-02055-4. p g 70. Gao, Z.; He, J.; Dong, K.; Li, X. Trends in Reference Evapotranspiration and Their Causative Factors in the West Liao River Basin, China. Agric. For. Meteorol. 2017, 232, 106–117. https://doi.org/10.1016/j.agrformet.2016.08.006. 71. Sridhar, V.; Hubbard, K.G.; Wedin, D.A. Assessment of Soil Moisture Dynamics of the Nebraska Sandhills Using Long-Term Measurements and a Hydrology Model. J. Irrig. Drain. Eng. 2006, 132, 463–473. https://doi.org/10.1061/(ASCE)0733- 9437(2006)132:5(463). References the Influence of Irrigation on Land Surface Fluxes and Boundary Layer Climatology. J. Contemp. Water Res. 3. https://doi.org/10.1111/j.1936-704X.2013.03170.x. 72. Sridhar, V. Tracking the Influence of Irrigation on Land Surface Fluxes and Boundary Layer Climatolo Educ. 2013, 152, 79–93. https://doi.org/10.1111/j.1936-704X.2013.03170.x. 73. Seong, C.; Sridhar, V.; Billah, M.M. Implications of Potential Evapotranspiration Methods for Streamflow Estimations under Changing Climatic Conditions. Int. J. Climatol. 2018, 38, 896–914. https://doi.org/10.1002/joc.5218. 74. Kaya, Y.Z.; Zelenakova, M.; Üneş, F.; Demirci, M.; Hlavata, H.; Mesaros, P. Estimation of Daily Evapotranspiration in Košice City (Slovakia) Using Several Soft Computing Techniques. Theor. Appl. Climatol. 2021, 144, 287–298. https://doi.org/10.1007/s00704-021-03525-z. p g 75. Kisi, O.; Keshtegar, B.; Zounemat-Kermani, M.; Heddam, S.; Trung, N.-T. Modeling Reference Evapotranspiration Using a Novel Regression-Based Method: Radial Basis M5 Model Tree. Theor. Appl. Climatol. 2021, 145, 639–659. https://doi.org/10.1007/s00704-021-03645-6. p g 76. Adnan, R.M.; Chen, Z.; Yuan, X.; Kisi, O.; El-Shafie, A.; Kuriqi, A.; Ikram, M. Reference Evapotransp New Heuristic Methods. Entropy 2020, 22, 547. https://doi.org/10.3390/E22050547. Z.; Yuan, X.; Kisi, O.; El-Shafie, A.; Kuriqi, A.; Ikram, M. Reference Evapotranspiration Modeling Using hods. Entropy 2020, 22, 547. https://doi.org/10.3390/E22050547. py p g 77. Malik, A.; Kumar, A.; Ghorbani, M.A.; Kashani, M.H.; Kisi, O.; Kim, S. The Viability of Co-Active Fuzzy Inference System Model for Monthly Reference Evapotranspiration Estimation: Case Study of Uttarakhand State. Hydrol. Res. 2019, 50, 1623–1644. https://doi.org/10.2166/nh.2019.059. p g 78. Gonzalez del Cerro, R.T.; Subathra, M.S..; Manoj Kumar, N.; Verrastro, S.; Thomas George, S. Modelling the Daily Reference Evapotranspiration in Semi-Arid Region of South India: A Case Study Comparing ANFIS and Empirical Models. Inf. Process. Agric. 2020, 8, 173–184. https://doi.org/10.1016/j.inpa.2020.02.003. 79. Ahmadi, F.; Mehdizadeh, S.; Mohammadi, B.; Pham, Q.B.; DOAN, T.N.C.; Vo, N.D. Application of an Artificial Intelligence Technique Enhanced with Intelligent Water Drops for Monthly Reference Evapotranspiration Estimation. Agric. Water Manag. 2021, 244, 106622. https://doi.org/10.1016/j.agwat.2020.106622.
https://openalex.org/W2953038115	https://escholarship.org/content/qt65t663ms/qt65t663ms.pdf?t=qaezau	English	null	Detecting the impact of temperature on transmission of Zika, dengue, and chikungunya using mechanistic models	PLoS neglected tropical diseases	2,017	cc-by	11,761	Permalink https://escholarship.org/uc/item/65t663ms UCLA UCLA Previously Published Works Title Detecting the impact of temperature on transmission of Zika, dengue, and chik using mechanistic models Permalink https://escholarship.org/uc/item/65t663ms Journal PLOS Neglected Tropical Diseases, 11(4) ISSN 1935-2727 Authors Mordecai, Erin A Cohen, Jeremy M Evans, Michelle V et al. Publication Date 2017 DOI 10.1371/journal.pntd.0005568 Peer reviewed UCLA UCLA Previously Published Works Title Detecting the impact of temperature on transmission of Zika, dengue, and chik using mechanistic models Permalink https://escholarship.org/uc/item/65t663ms Journal PLOS Neglected Tropical Diseases, 11(4) ISSN 1935-2727 Authors Mordecai, Erin A Cohen, Jeremy M Evans, Michelle V et al. Publication Date 2017 DOI 10.1371/journal.pntd.0005568 Peer reviewed UCLA UCLA Previously Published Works Title Detecting the impact of temperature on transmi using mechanistic models Permalink https://escholarship.org/uc/item/65t663ms Journal PLOS Neglected Tropical Diseases, 11(4) ISSN 1935-2727 Authors Mordecai, Erin A Cohen, Jeremy M Evans, Michelle V et al. Publication Date 2017 DOI 10.1371/journal.pntd.0005568 Peer reviewed UCLA UCLA Previously Published Works Title Title Detecting the impact of temperature on transmission of Zika, dengue, and chikungunya using mechanistic models RESEARCH ARTICLE Editor: Benjamin Althouse, Institute for Disease Modeling, UNITED STATES Editor: Benjamin Althouse, Institute for Disease Modeling, UNITED STATES Received: January 18, 2017 Accepted: April 12, 2017 Published: April 27, 2017 Copyright: © 2017 Mordecai et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Received: January 18, 2017 Accepted: April 12, 2017 Published: April 27, 2017 * emordeca@stanford.edu Detecting the impact of temperature on transmission of Zika, dengue, and chikungunya using mechanistic models Erin A. Mordecai1, Jeremy M. Cohen2, Michelle V. Evans3, Prithvi Gudapati1, Leah R. Johnson2,4, Catherine A. Lippi5, Kerri Miazgowicz6, Courtney C. Murdock3,6, Jason R. Rohr2, Sadie J. Ryan5,7,8,9, Van Savage10,11, Marta S. Shocket1,12, Anna Stewart Ibarra13, Matthew B. Thomas14, Daniel P. Weikel15 1 Biology Department, Stanford University, 371 Serra Mall, Stanford, CA, United States of America, 2 Department of Integrative Biology, University of South Florida, 4202 East Fowler Ave, SCA110 Tampa, FL, United States of America, 3 Odum School of Ecology, University of Georgia, Athens, GA, United States of America, 4 Department of Statistics, Virginia Polytechnic and State University, 250 Drillfield Drive Blacksburg, VA, United States of America, 5 Department of Geography, University of Florida, Turlington Hall, Gainesville, FL, United States of America, 6 Center for Tropical and Emerging Global Disease, Department of Infectious Diseases, University of Georgia College of Veterinary Medicine, 501 D.W. Brooks Drive, Athens, GA, United States of America, 7 Emerging Pathogens Institute, University of Florida, Gainesville, FL, United States of America, 8 Center for Global Health and Translational Science, Department of Microbiology and Immunology, Weiskotten Hall, SUNY Upstate Medical University, Syracuse, NY, United States of America, 9 School of Life Sciences, College of Agriculture, Engineering, and Science, University of KwaZulu Natal, Private Bag X01, Scottsville, KwaZulu Natal, South Africa, 10 Department of Ecology and Evolutionary Biology, University of California Los Angeles and Department of Biomathematics, University of California Los Angeles, Los Angeles, CA, United States of America, 11 Santa Fe Institute, Santa Fe, NM, United States of America, 12 Department of Biology, Indiana University, Jordan Hall 142, Bloomington, IN, United States of America, 13 Center for Global Health and Translational Sciences, SUNY Upstate Medical University, Syracuse, NY, United States of America, 14 Department of Entomology and Center for Infectious Disease Dynamics, Penn State University, 112 Merkle Lab, University Park, PA, United States of America, 15 Department of Biostatistics, University of Michigan, 1415 Washington Heights, Ann Arbor, MI, United States of America OPEN ACCESS Citation: Mordecai EA, Cohen JM, Evans MV, Gudapati P, Johnson LR, Lippi CA, et al. (2017) Detecting the impact of temperature on transmission of Zika, dengue, and chikungunya using mechanistic models. PLoS Negl Trop Dis 11 (4): e0005568. https://doi.org/10.1371/journal. pntd.0005568 Editor: Benjamin Althouse, Institute for Disease Modeling, UNITED STATES Powered by the California Digital Library University of California eScholarship.org Temperature predicts Zika, dengue, and chikungunya transmission https://www.nih.gov/), US Department of Agriculture (2009-35102-0543; https://www.usda. gov/wps/portal/usda/usdahome) and US Environmental Protection Agency (CAREER 83518801; https://www.epa.gov/). EAM and CCM were supported by the NSF (DEB-1640780; https:// nsf.gov/). EAM was supported by the Stanford Woods Institute for the Environment (https:// woods.stanford.edu/research/environmental- venture-projects) and the Stanford Center for Innovation in Global Health (http://globalhealth. stanford.edu/research/seed-grants.html). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Abstract Recent epidemics of Zika, dengue, and chikungunya have heightened the need to understand the seasonal and geographic range of transmission by Aedes aegypti and Ae. albopictus mos- quitoes. We use mechanistic transmission models to derive predictions for how the probability and magnitude of transmission for Zika, chikungunya, and dengue change with mean temper- ature, and we show that these predictions are well matched by human case data. Across all three viruses, models and human case data both show that transmission occurs between 18– 34˚C with maximal transmission occurring in a range from 26–29˚C. Controlling for population size and two socioeconomic factors, temperature-dependent transmission based on our mechanistic model is an important predictor of human transmission occurrence and inci- dence. Risk maps indicate that tropical and subtropical regions are suitable for extended sea- sonal or year-round transmission, but transmission in temperate areas is limited to at most three months per year even if vectors are present. Such brief transmission windows limit the likelihood of major epidemics following disease introduction in temperate zones. Copyright: © 2017 Mordecai et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All data, code, and outputs are available on Figshare: https://figshare. com/s/b79bc7537201e7b5603f, DOI: https://dx. doi.org/10.6084/m9.figshare.4563928 Funding: EAM, MBT, VS, SJR, LRJ, ASI, JRR, MS, JC, and DPW were supported by the National Science Foundation (DEB-1518681; https://nsf. gov/). JRR was supported by the NSF (EF- 1241889; https://nsf.gov/), National Institutes of Health (R01GM109499 and R01TW010286-01; 1 / 18 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 Author summary Understanding the drivers of recent Zika, dengue, and chikungunya epidemics is a major public health priority. Temperature may play an important role because it affects virus transmission by mosquitoes, through its effects on mosquito development, sur- vival, reproduction, and biting rates as well as the rate at which mosquitoes acquire and transmit viruses. Here, we measure the impact of temperature on transmission by two of the most common mosquito vector species for these viruses, Aedes aegypti and Ae. albopictus. We integrate data from several laboratory experiments into a mathematical model of temperature-dependent transmission, and find that transmission peaks at 26– 29˚C and can occur between 18–34˚C. Statistically comparing model predictions with recent observed human cases of dengue, chikungunya, and Zika across the Americas suggests an important role for temperature, and supports model predictions. Using the model, we predict that most of the tropics and subtropics are suitable for transmission in many or all months of the year, but that temperate areas like most of the United States are only suitable for transmission for a few months during the summer (even if the mosquito vector is present). Competing interests: The authors have declared that no competing interests exist. Temperature predicts Zika, dengue, and chikungunya transmission and Ae. albopictus that are (a) mechanistic, facilitating extrapolation beyond the current dis- ease distribution, (b) parameterized with biologically accurate unimodal thermal responses for all mosquito and virus traits that drive transmission, and (c) validated against human dengue, chikungunya, and Zika case data across the Americas. We synthesize available data to characterize the temperature-dependent traits of the mos- quitoes and viruses that determine transmission intensity. With these thermal responses, we develop mechanistic temperature-dependent virus transmission models for Ae. aegypti and Ae. albopictus. We then ask whether the predicted effect of temperature on transmission is consistent with patterns of actual human cases over space and time. To do this, we validate the models with DENV, CHIKV, and ZIKV human incidence data at the country scale in the Americas from 2014–2016. To isolate temperature dependence, we also statistically control for population size and two socioeconomic factors that may influence transmission. If tempera- ture fundamentally limits transmission potential, transmission should only occur at actual environmental temperatures that are predicted to be suitable, and conversely, areas with low predicted suitability should have low or zero transmission (i.e., false negative rates should be low). By contrast, low transmission may occur even when temperature suitability is high because other factors like vector control can limit transmission (i.e., the false positive rate should be higher than the false negative rate). Finally, if the simple mechanistic model accu- rately predicts climate suitability for transmission, then we can use it to map climate-based transmission risk of DENV, CHIKV, ZIKV, and other emerging pathogens transmitted by Ae. aegypti and Ae. albopictus seasonally and geographically. Introduction Epidemics of dengue, chikungunya, and Zika are sweeping through the Americas, and are part of a global public health crisis that places an estimated 3.9 billion people in 120 countries at risk [1]. Dengue virus (DENV) distribution and intensity in the Americas has increased over the last three decades, infecting an estimated 390 million people (96 million clinical) per year [2]. Chikungunya virus (CHIKV) emerged in the Americas in 2013, causing 1.8 million sus- pected cases from 44 countries and territories (www.paho.org). In the last two years, Zika virus (ZIKV) has spread throughout the Americas, causing 764,414 suspected and confirmed cases, with many more unreported (http://ais.paho.org/phip/viz/ed_zika_cases.asp, as of April 13, 2017). The growing burden of these diseases (including links between Zika infection and both microcephaly and Guillain-Barre´ syndrome [3]) and potential for spread into new areas cre- ates an urgent need for predictive models that can inform risk assessment and guide interven- tions such as mosquito control, community outreach, and education. Predicting transmission of DENV, CHIKV, and ZIKV requires understanding the ecology of the vector species. For these viruses the main vector is Aedes aegypti, a mosquito that prefers and is closely affiliated with humans, while Ae. albopictus, a peri-urban mosquito, is an impor- tant secondary vector [4,5]. We expect one of the main drivers of the vector ecology to be the climate, particularly temperature. For that reason, mathematical and geostatistical models that incorporate climate information have been valuable for predicting and responding to Aedes spp. spread and DENV, CHIKV, and ZIKV outbreaks [5–10]. The effects of temperature on ectotherms are largely predictable from fundamental meta- bolic and ecological processes. Survival, feeding, development, and reproductive rates predict- ably respond to temperature across a variety of ectotherms, including mosquitoes [11,12]. Because these traits help to determine transmission rates, the effects of temperature on trans- mission should also be broadly predictable from mechanistic models that incorporate temper- ature-dependent traits. Here, we introduce a model based on this framework that overcomes several major gaps that currently limit our understanding of climate suitability for transmis- sion. Specifically, we develop models of temperature-dependent transmission for Ae. aegypti 2 / 18 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 Temperature-dependent transmission Data gathered from the literature [9,13–30] revealed that all mosquito traits relevant to trans- mission—biting rate, egg-to-adult survival and development rate, adult lifespan, and fecundity —respond strongly to temperature and peak between 23˚C and 34˚C for the two mosquito species (Ae. aegypti in Fig 1 and Ae. albopictus in Fig A in S1 Text). DENV extrinsic incubation and vector competence peak at 35˚C [31–37] and 31–32˚C [31,32,34,38], respectively, in both mosquitoes—temperatures at which mosquito survival is low, limiting transmission potential (Fig 1, Fig A in S1 Text). Appropriate thermal response data were not available for CHIKV and ZIKV extrinsic incubation and vector competence. We estimated the posterior distribution of R0(T) and used it to calculate key temperature values that indicate suitability for transmission: the mean and 95% credible intervals (95% CI) on the critical thermal minimum, maximum, and optimum temperature for transmission by the two mosquito species. At constant temperatures, Ae. aegypti transmission peaked at 29.1˚C (95% CI: 28.4–29.8˚C), and declined to zero below 17.8˚C (95% CI: 14.6–21.2˚C) and above 34.6˚C (95% CI: 34.1–35.6˚C) (Fig 2). Ae. albopictus transmission peaked at 26.4˚C (95% CI: 25.2–27.4˚C) and declined to zero below 16.2˚C (95% CI: 13.2–19.9˚C) and above 31.6˚C (95% CI: 29.4–33.7˚C) (Fig 2). Overall, the thermal response curve for Ae. albopictus is shifted towards lower temperatures than Ae. aegypti, so Ae. albopictus transmission is better suited to cooler environments. For a more realistic scenario in which daily temperature ranged over 8˚C, the transmission peak, minimum, and maximum were slightly lower for both Ae. aegypti (28.5˚C, 13.5˚C, 34.2˚C, respectively) and Ae. albopictus (26.1˚C, 11.9˚C, and 28.3˚C, respec- tively). The lower thermal maximum under fluctuating temperatures occurs because we incor- porated empirically supported irreversible lethal effects of temperatures that exceed thermal maxima for survival (see Materials and Methods). PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 3 / 18 Temperature predicts Zika, dengue, and chikungunya transmission Fig 1. Thermal responses of Ae. aegypti and DENV traits that drive transmission (data sources listed in Table B in S2 Text). Informative priors based on data from additional Aedes spp. and flavivirus studies helped to constrain uncertainty in the model fits (see Materials and Methods; Table C in S2 Text). Points and error bars indicate the data means and standard errors (for display only; models were fit from the raw data). Temperature-dependent transmission Solid lines: mean posterior estimates; dashed lines: 95% credible intervals. R0 curves normalized to a 0–1 scale for ease of comparison and visualization Fig 2. Relative R0 across constant temperatures (˚C; top) for Ae. albopictus (light blue) and Ae. aegypti (dark blue), and histograms of the posterior distributions of the critical thermal minimum (bottom left), temperature at peak transmission (bottom middle), and critical thermal maximum (bottom right; all in ˚C). Solid lines: mean posterior estimates; dashed lines: 95% credible intervals. R0 curves normalized to a 0–1 scale for ease of comparison and visualization. https://doi.org/10.1371/journal.pntd.0005568.g002 https://doi.org/10.1371/journal.pntd.0005568.g002 Temperature-dependent transmission Black solid lines are the mean model fits; red dashed lines are the 95% credible intervals. Thermal responses for Ae. albopictus are shown in Fig A in S1 Text. https://doi.org/10.1371/journal.pntd.0005568.g001 Fig 1. Thermal responses of Ae. aegypti and DENV traits that drive transmission (data sources listed in Table B in S2 Text). Informative priors based on data from additional Aedes spp. and flavivirus studies helped to constrain uncertainty in the model fits (see Materials and Methods; Table C in S2 Text). Points and error bars indicate the data means and standard errors (for display only; models were fit from the raw data). Black solid lines are the mean model fits; red dashed lines are the 95% credible intervals. Thermal responses for Ae. albopictus are shown in Fig A in S1 Text. https://doi.org/10.1371/journal.pntd.0005568.g001 https://doi.org/10.1371/journal.pntd.0005568.g001 The posterior distribution of R0(T) allows us to evaluate uncertainty in key temperature val- ues that define the transmission range, including critical thermal minimum, maximum, and optimum. Uncertainty was higher for the critical thermal minimum for transmission than for the maximum or optimum, and the two mosquito species overlapped most for this outcome (Fig 2, bottom panels). This result occurred because several trait thermal responses increase gradually from low to mid temperatures but decline more steeply at high temperatures (Fig 1), so uncertainty is greatest at low temperatures. Ae. aegypti has a substantially higher optimum and maximum temperature than Ae. albopictus (Fig 2) due to its greater rates of adult survival at high temperatures (see Supplementary Materials for sensitivity analyses). PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 4 / 18 Temperature predicts Zika, dengue, and chikungunya transmission Fig 2. Relative R0 across constant temperatures (˚C; top) for Ae. albopictus (light blue) and Ae. aegypti (dark blue), and histograms of the posterior distributions of the critical thermal minimum (bottom left), temperature at peak transmission (bottom middle), and critical thermal maximum (bottom right; all in ˚C). Solid lines: mean posterior estimates; dashed lines: 95% credible intervals. R0 curves normalized to a 0–1 scale for ease of comparison and visualization. https://doi org/10 1371/journal pntd 0005568 g002 Fig 2. Relative R0 across constant temperatures (˚C; top) for Ae. albopictus (light blue) and Ae. aegypti (dark blue), and histograms of the posterior distributions of the critical thermal minimum (bottom left), temperature at peak transmission (bottom middle), and critical thermal maximum (bottom right; all in ˚C). Model validation Second we used a boot- strapping approach where models were fit on subsets of the case incidence data that were randomly sampled and then predictive accuracy of the competing models (Table D in S2 Text) was assessed on left-out data (“out of sample” analysis). and population size. In the models of the probability of autochthonous transmission occurring we used the product of the posterior probability that R0(T) > 0 (which we notate as GR0) and the log of population size (p) to give log(p)GR0. (Here, and throughout, log denotes the natural logarithm.) In the models of incidence, given that transmission does occur, we used the log of the product of the posterior mean of R0(T) and population size, log(pR0(T)). To control for several socioeconomic factors that might obscure the impact of temperature, we also included log of gross domestic product (GDP) and log of percent of GDP in tourism (using logs because the predictors were highly skewed, to stabilize variance). These are potential indicators of investment in and/or success of vector control and infrastructure improvements that prevent transmission. By comparing models that included the R0(T) metric alone, socioeconomic fac- tors alone, or both, we tested whether R0(T) was an important predictor of observed transmis- sion occurrence and incidence (see Table D in S2 Text). Note that R0(T) is out of sample for all validation analyses because it is derived and calculated strictly from laboratory data on mos- quitoes, and we perform a validation analyses for R0(T) using independent case incidence reports. For this validation step we assessed model adequacy for the transmission data in two ways. First we used the full dataset for case incidence reports to select the best model (Table D in S2 Text) and to determine whether or not our predicted value of relative R0(T) based on lab- oratory data was included in the model (“within sample” analysis). Second we used a boot- strapping approach where models were fit on subsets of the case incidence data that were randomly sampled and then predictive accuracy of the competing models (Table D in S2 Text) was assessed on left-out data (“out of sample” analysis). For the probability of autochthonous transmission occurring, the model that included both the R0(T) predictor and socioeconomic predictors had overwhelming support based on Bayes- ian Information Criterion (BIC; model PA5 relative probability = 1, Table D in S2 Text). Model validation B, log(pR0(T)) (log of R0(T) times the population size) versus the log of incidence, given that it exceeds the threshold for local transmission. Tick-marks and points: human transmission occurrence and incidence data, respectively, by country-week in the Americas and Caribbean. Lines and shaded areas: mean and 95% CI from GLM fits for DENV (blue) and CHIKV and ZIKV (red). For simplicity, we show the models that only include the covariates log(p)GR0 or log(pR0(T)), respectively, and do not include the socioeconomic covariates (models PA6 and IM4 in Table D in S2 Text). For each case report data point, log(p)GR0 and log(pR0(T)) were calculated at the mean temperature 10 weeks prior to the reporting week [40]. https://doi.org/10.1371/journal.pntd.0005568.g003 ig 3. Ae. aegypti R0(T) and population size predict the probability and magnitude of transmission of DENV, and population size. In the models of the probability of autochthonous transmission occurring we used the product of the posterior probability that R0(T) > 0 (which we notate as GR0) and the log of population size (p) to give log(p)GR0. (Here, and throughout, log denotes the natural logarithm.) In the models of incidence, given that transmission does occur, we used the log of the product of the posterior mean of R0(T) and population size, log(pR0(T)). To control for several socioeconomic factors that might obscure the impact of temperature, we also included log of gross domestic product (GDP) and log of percent of GDP in tourism (using logs because the predictors were highly skewed, to stabilize variance). These are potential indicators of investment in and/or success of vector control and infrastructure improvements that prevent transmission. By comparing models that included the R0(T) metric alone, socioeconomic fac- tors alone, or both, we tested whether R0(T) was an important predictor of observed transmis- sion occurrence and incidence (see Table D in S2 Text). Note that R0(T) is out of sample for all validation analyses because it is derived and calculated strictly from laboratory data on mos- quitoes, and we perform a validation analyses for R0(T) using independent case incidence reports. For this validation step we assessed model adequacy for the transmission data in two ways. First we used the full dataset for case incidence reports to select the best model (Table D in S2 Text) and to determine whether or not our predicted value of relative R0(T) based on lab- oratory data was included in the model (“within sample” analysis). Model validation We used generalized linear models (GLM) to ask whether the predicted relationship between temperature and transmission, R0(T), was consistent with observed human cases of DENV, CHIKV, and ZIKV. Specifically, we assessed whether R0(T) was an important predictor of the probability of autochthonous transmission occurring and of the incidence given that transmis- sion occurred. We also controlled for human population size, virus species, and two socioeco- nomic factors. (Note that we focused on testing the R0(T) model, rather than on constructing the best possible statistical model of human case data.) To do this, we used the version of the Ae. aegypti R0(T) model that includes 8˚C daily temperature range, along with country-scale weekly case reports of DENV, CHIKV, and ZIKV in the Americas and the Caribbean between 2014–2016. We first addressed the fact that countries with larger populations have greater opportunities for (large) epidemics by creating two predictors that incorporate scaled R0(T) PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 5 / 18 Temperature predicts Zika, dengue, and chikungunya transmission Fig 3. Ae. aegypti R0(T) and population size predict the probability and magnitude of transmission of DENV, CHIKV, and ZIKV across the Americas. A, log(p)GR0 (the posterior probability that R0(T) > 0 times the log of population size) versus the probability of local transmission in the data. B, log(pR0(T)) (log of R0(T) times the population size) versus the log of incidence, given that it exceeds the threshold for local transmission. Tick-marks and points: human transmission occurrence and incidence data, respectively, by country-week in the Americas and Caribbean. Lines and shaded areas: mean and 95% CI from GLM fits for DENV (blue) and CHIKV and ZIKV (red). For simplicity, we show the models that only include the covariates log(p)GR0 or log(pR0(T)), respectively, and do not include the socioeconomic covariates (models PA6 and IM4 in Table D in S2 Text). For each case report data point, log(p)GR0 and log(pR0(T)) were calculated at the mean temperature 10 weeks prior to the reporting week [40]. https://doi.org/10.1371/journal.pntd.0005568.g003 Fig 3. Ae. aegypti R0(T) and population size predict the probability and magnitude of transmission of DENV, CHIKV, and ZIKV across the Americas. A, log(p)*GR0 (the posterior probability that R0(T) > 0 times the log of population size) versus the probability of local transmission in the data. Temperature predicts Zika, dengue, and chikungunya transmission Fig B in S1 Text). In analyses of out-of-sample accuracy, models that included the R0(T) metric (with or without the socioeconomic factors) were surprisingly accurate. They predicted the probability of transmission with 86–91% out-of-sample accuracy for DENV (Table D in S2 Text). For CHIKV and ZIKV, models that included the R0(T) metric or population alone had 66–69% out-of-sample accuracy (Table D in S2 Text). There were no significant differences in out-of-sample accuracy between the top four models, but for both DENV and CHIKV/ZIKV the best model was significantly better than the worst model [see supplementary code in 39 for full results]. The lower out-of-sample accuracy for CHIKV and ZIKV likely reflects the much lower frequency of positive values and the lower total sample size of this dataset. All results were similar for a set of models that separated GR0 from population size, so for simplicity we show the model predictors that combines GR0 and population size here (see Table D in S2 Text and [39] for results of other models). Further, from a biological perspective, the combined model better describes what we know about disease systems: if either the probability of R0(T) being greater than zero is small or population size is very small, transmission is unlikely to occur. Together, these analyses suggest that R0(T) is an important predictor of transmission occurrence, but that CHIKV and ZIKV need further data to better explain the probability of transmission occurrence (Fig 3A, Fig B in S1 Text). R0(T) was also an important predictor of incidence, given that autochthonous transmission did occur. Within-sample, incidence was best predicted by the model that included both R0(T) and the socioeconomic predictors (model IM5 in Table D in S2 Text) based on BIC (relative probability = 1). The models that included R0(T) out-performed those that did not based on deviance explained (Table D in S2 Text). In out-of-sample validation, the models that included R0(T) explained the magnitude of incidence based on mean absolute percentage error (85– 86% accuracy versus 83% accuracy for models that did not include R0(T); Table D in S2 Text), but this difference was not statistically significant. For illustration, we show the simpler model that only contains the R0(T) predictor in the main text (Fig 3B; model IM1 in Table D in S2 Text). Notably, the models that contained R0(T) predicted incidence well for all three viruses, despite the lower incidence of CHIKV and ZIKV. Although predicted R0(T) correlated with the observed occurrence and magnitude of human incidence for all three viruses, these observed incidence metrics were higher for DENV than for CHIKV and ZIKV. While the reason for this difference is unclear, the most likely explanation is that DENV is much more established in the Americas, so it is more likely to be detected, diagnosed, and reported. Because ZIKV and CHIKV are newly emerging, they may not have fully saturated the region at this early stage. The ability of the model to explain the probability and magnitude of transmission is notable given the coarse scale of the human incidence versus mean temperature data (i.e., country- scale means), the lack of CHIKV- and ZIKV-specific trait thermal response data to inform the model, the nonlinear relationship between transmission and incidence, and all the well-docu- mented factors other than temperature that influence transmission. Together, these analyses show simple mechanistic models parameterized with laboratory data on mosquitoes and den- gue virus are consistent with observed temperature suitability for transmission. Moreover, the similar responses of human incidence of ZIKV, CHIKV, and DENV to temperature suggest that the thermal ecology of their shared mosquito vectors is a key determinant of outbreak location, timing, and intensity. Model validation Based on deviance explained, the models that included R0(T), with or without the socioeconomic predictors out-performed the model that did not include R0(T) (Table D in S2 Text; Fig 3A, 6 / 18 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 Mapping climate suitability for transmission Indeed, DENV, CHIKV, and/or ZIKV local transmission has occurred in Texas, Florida, Hawaii, and Puerto Rico (www.cdc.gov). On the other hand, many temper- ate regions experience temperatures suitable for transmission three months or less per year (Fig 4). Temperature thus limits the potential for the viruses to generate extensive epidemics in temperate areas even where the vectors are present. Moreover, many temperate regions with seasonally suitable temperatures currently lack Ae. aegypti and Ae. albopictus mosquitoes, mak- ing vector transmission impossible (Fig 4, black line). The posterior distribution of R0(T) also allows us to map months of risk with different degrees of uncertainty (e.g., 97.5%, 50%, and 2.5% posterior probability that that R0 > 0), ranging from the most to least conservative (Fig D in S1 Text). number of months per year at which mean temperatures do not prevent transmission, large areas of tropical and subtropical regions, including Puerto Rico and parts of Florida and Texas, are currently suitable year-round or seasonally (Fig 4). These regions are fundamentally at risk for DENV, CHIKV, ZIKV, and other Aedes arbovirus transmission during a substantial part of the year (Fig 4). Indeed, DENV, CHIKV, and/or ZIKV local transmission has occurred in Texas, Florida, Hawaii, and Puerto Rico (www.cdc.gov). On the other hand, many temper- ate regions experience temperatures suitable for transmission three months or less per year (Fig 4). Temperature thus limits the potential for the viruses to generate extensive epidemics in temperate areas even where the vectors are present. Moreover, many temperate regions with seasonally suitable temperatures currently lack Ae. aegypti and Ae. albopictus mosquitoes, mak- ing vector transmission impossible (Fig 4, black line). The posterior distribution of R0(T) also allows us to map months of risk with different degrees of uncertainty (e.g., 97.5%, 50%, and 2.5% posterior probability that that R0 > 0), ranging from the most to least conservative (Fig D in S1 Text). Mapping climate suitability for transmission The validated model can be used to predict where transmission is not excluded (posterior probability that R0(T) > 0, a conservative estimate of transmission risk). Considering the 7 / 18 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 Temperature predicts Zika, dengue, and chikungunya transmission Fig 4. Map of predicted temperature suitability for virus transmission by Ae. albopictus and Ae. aegypti. Color indicates the consecutive months in which temperature is permissive for transmission (predicted R0 > 0) for Aedes spp. transmission based on the minimum likely range (> 97.5% posterior probability that R0 > 0). Black lines indicate the CDC estimated range for the two Aedes spp. in the United States. Model suitability predictions combine temperature mean and 8˚C daily variation and are informed by laboratory data (Fig 1, Fig A in S1 Text) and validated against field data (Fig 3). https://doi.org/10.1371/journal.pntd.0005568.g004 Fig 4. Map of predicted temperature suitability for virus transmission by Ae. albopictus and Ae. aegypti. Color indicates the consecutive months in which temperature is permissive for transmission (predicted R0 > 0) for Aedes spp. transmission based on the minimum likely range (> 97.5% posterior probability that R0 > 0). Black lines indicate the CDC estimated range for the two Aedes spp. in the United States. Model suitability predictions combine temperature mean and 8˚C daily variation and are informed by laboratory data (Fig 1, Fig A in S1 Text) and validated against field data (Fig 3). https://doi.org/10.1371/journal.pntd.0005568.g004 Fig 4. Map of predicted temperature suitability for virus transmission by Ae. albopictus and Ae. aegypti. Color indicates the consecutive months in which temperature is permissive for transmission (predicted R0 > 0) for Aedes spp. transmission based on the minimum likely range (> 97.5% posterior probability that R0 > 0). Black lines indicate the CDC estimated range for the two Aedes spp. in the United States. Model suitability predictions combine temperature mean and 8˚C daily variation and are informed by laboratory data (Fig 1, Fig A in S1 Text) and validated against field data (Fig 3). https://doi.org/10.1371/journal.pntd.0005568.g004 https://doi.org/10.1371/journal.pntd.0005568.g004 number of months per year at which mean temperatures do not prevent transmission, large areas of tropical and subtropical regions, including Puerto Rico and parts of Florida and Texas, are currently suitable year-round or seasonally (Fig 4). These regions are fundamentally at risk for DENV, CHIKV, ZIKV, and other Aedes arbovirus transmission during a substantial part of the year (Fig 4). Temperature predicts Zika, dengue, and chikungunya transmission traits, and because the traits combine nonlinearly to drive transmission, the emergent relation- ship between temperature and transmission is difficult to infer directly from field data or from individual trait responses. Here, we present a model of temperature-dependent DENV, CHIKV, and ZIKV transmission that advances on previous models because it is mechanistic, fitted from experimental trait data (Fig 1, Fig A in S1 Text), and validated against independent human case data at a broad geographic scale (Fig 3). CHIKV, and ZIKV transmission that advances on previous models because it is mechanistic, fitted from experimental trait data (Fig 1, Fig A in S1 Text), and validated against independent human case data at a broad geographic scale (Fig 3). Mechanistic understanding is valuable for extrapolating beyond the current spatial and temporal range of transmission (Fig 4), as compared to environmental niche models, for example [5,42,43]. Of the six previous mechanistic temperature-dependent models of DENV, CHIKV, or ZIKV transmission by Ae. aegypti and Ae. albopictus that we were able to repro- duce, three had similar thermal optima [7,44,45] while the other three had dramatically higher optima (3–6˚C) [9,46] (Fig E in S1 Text). Two of the models were very similar to ours [44,45]; of the remaining four models, two predicted much greater suitability for transmission at low temperatures [46] and all four predicted greater suitability at high temperatures [7,9,46] (Fig E in S1 Text). Only one of these previous models was (like ours) statistically validated against independent data not used to estimate model parameters, and its predictions were very similar to those of our model [44]. Other mechanistic and environmental niche models could not be directly compared with ours [5,10,41–43], either because fully reproducible equations, param- eters, and/or code were not provided or because their predicted marginal effects of tempera- ture were not displayed. Visually, our maps are similar to maps based on a previous model of Ae. aegypti and Ae. albopictus persistence suitability indices [41]. Recent environmental niche models of Zika distribution have shown similar but more constrained predicted distributions of environmental suitability, in part because these models include not just temperature suit- ability but also further environmental, socioeconomic, and demographic constraints [5,42,43,47]. Discussion Temperature is an important driver of—and limitation on—vector transmission, so accurately describing the temperature range and optimum for transmission of DENV, CHIKV, and ZIKV is critical for predicting their geographic and seasonal patterns of spread [12,41]. We directly estimated the temperature–transmission relationship using mechanistic transmission models for each mosquito species (Fig 2). These models are built using empirical estimates of the (unimodal) effects of temperature on mosquito and pathogen traits that drive transmis- sion, including survival, development, reproduction, and biting rates (Fig 1, Fig A in S1 Text). Because these trait thermal responses are unimodal across the majority of ectotherm taxa and 8 / 18 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 Temperature-sensitive R0 models We constructed temperature-dependent models of transmission using a previously developed R0 framework. We modeled transmission rate as the basic reproduction rate, R0—the number of secondary infections that would originate from a single infected individual introduced to a fully susceptible population. In previous work on malaria, we adapted a commonly used expression for R0 for vector transmission to include the temperature-sensitive traits that drive mosquito population density [12]: R0 T ð Þ ¼ aðTÞ 2 bðTÞ cðTÞ emðTÞ=PDRðTÞ EFDðTÞ pEAðTÞ MDRðTÞ N r mðTÞ 3 !1=2 ð1Þ ð1Þ Here, (T) indicates that the trait is a function of temperature, T; a is the per-mosquito biting rate, b is the proportion of infectious bites that infect susceptible humans, c is the proportion of bites on infected humans that infect previously uninfected mosquitoes (i.e., bc = vector competence), μ is the adult mosquito mortality rate (lifespan, lf = 1/μ), PDR is the parasite de- velopment rate (i.e., the inverse of the extrinsic incubation period, the time required between a mosquito biting an infected host and becoming infectious), EFD is the number of eggs pro- duced per female mosquito per day, pEA is the mosquito egg-to-adult survival probability, MDR is the mosquito immature development rate (i.e., the inverse of the egg-to-adult development time), N is the density of humans, and r is the human recovery rate. For each temperature-sensi- tive trait in each mosquito species, we fit either symmetric (Quadratic, -c(T–T0)(T–Tm)) or asymmetric (Brière, cT(T–T0)(Tm−T)1/2) unimodal thermal response models to the available empirical data [48]. In both functions, T0 and Tm are respectively the minimum and maximum temperature for transmission, and c is a positive rate constant. We consider a normalized version of the R0 equation such that it is rescaled to range from zero to one with the value of one occurring at the unimodal peak. Although absolute values of R0 that are used to determine when transmission is stable depend on additional factors not captured in our model, the minimum and maximum temperatures for which R0 > 0 map exactly onto our normalized equations, allowing us to accurately calculate whether or not transmission should be possible at all. Empirical estimates of absolute values of R0 are difficult to obtain in any case, but it is much easier to determine whether transmission is occurring and for how long. Temperature predicts Zika, dengue, and chikungunya transmission yellow fever, Uganda S, or Ross River viruses). Efforts to extrapolate and to map temperature suitability (Fig 4) will be critical for improving management of both ongoing and future emerging epidemics. Mechanistic models like the one presented here are useful for extrapolat- ing the potential geographic range of transmission beyond the current envelope of environ- mental conditions in which transmission occurs (e.g., under climate change and for newly invading pathogens). Accurately estimating temperature-driven transmission risk in both highly suitable and marginal regions is critical for predicting and responding to future out- breaks of these and other Aedes-transmitted viruses. Even though the thermal response data are imperfect—for example, CHIKV and ZIKV thermal response data are missing—and the human case data are reported at a coarse spatial scale, the validation analyses suggest that R0(T) is an important predictor of both the probabil- ity of transmission occurring and the magnitude of incidence for DENV, CHIKV, and ZIKV. This has several key implications. First, temperature-dependent transmission is pervasive enough to be detected at a coarse spatial scale. Second, dynamics of the mosquito predict trans- mission for a suite of Ae. aegypti-transmitted viruses, without additional virus-specific infor- mation. Third, climate and socio-economic factors combine to shape variation in incidence across countries. Finally, these simple predictors explain a substantial proportion of the vari- ance in both the probability and intensity of transmission. Predicting arbovirus transmission at a higher spatial resolution and precision will require more detailed information on factors like the exposure and susceptibility of human popula- tions, environmental variation (e.g., oviposition habitat availability, seasonal and daily temper- ature variation), and socioeconomic factors. However, as a first step our mechanistic model provides valuable insight because it makes broad predictions about suitable environmental conditions for transmission, it is mechanistic and grounded in experimental trait data, it is val- idated against independent human case data, and its predictions are applicable across three different viruses. Using these thermal response models as a scaffold, additional drivers could be incorporated to obtain more precise and specific predictions about transmission dynamics, which could in turn be used for public health and vector control applications. For this purpose, all code and data used in the models are available on Figshare [39]. The socio-ecological conditions that enabled CHIKV, ZIKV, and DENV to become the three most important emerging vector-borne diseases in the Americas make the emergence of additional Aedes-transmitted viruses likely (potentially including Mayaro, Rift Valley fever, PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 9 / 18 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 Temperature predicts Zika, dengue, and chikungunya transmission choice of formula (Fig F in S1 Text). Similarly, because different expressions for R0, including the square of Eq (1), map monotonically onto our function, they will produce identical esti- mates for the temperatures at which transmission declines to zero and peaks (Fig F in S1 Text). Consequently, our use of relative R0 adequately describes the nonlinear relationship between mosquito and virus traits and transmission. We fit the trait thermal responses in Eq (1) based on an exhaustive search of published labo- ratory studies that fulfilled the criterion of measuring a trait at three or more constant temper- atures, ideally capturing both the rise and the fall of each unimodal curve (Tables S1-S2). Constant-temperature laboratory conditions are required to isolate the direct effect of temper- ature from confounding factors in the field and to provide a baseline for estimating the effects of temperature variation through rate summation [50]. We attempted to obtain raw data from each study, but if they were not available we collected data by hand from tables or digitized data from figures using WebPlotDigitizer [51]. We obtained raw data from Delatte [19] and Alto [21] for the Ae. albopictus egg-to-adult survival probability (pEA), mosquito development rate (MDR), gonotrophic cycle duration (GCD, which we assumed was equal to the inverse of the biting rate) and total fecundity (TFD) (Table D in S2 Text). Data did not meet the inclusion criterion for CHIKV or ZIKV vector competence (b, c) or extrinsic incubation period (EIP) in either Ae. albopictus or Ae. aegypti. Instead, we used DENV EIP and vector competence data, combined with sensitivity analyses. Following Johnson et al. [52], we fit a thermal response for each trait using Bayesian mod- els. We first fit Bayesian models for each trait thermal response using uninformative priors (T0 ~ Uniform (0, 24), Tm ~ Uniform (25, 45), c ~ Gamma (1, 10) for Brière and c ~ Gamma (1, 1) for Quadratic fits) chosen to restrict each parameter to its biologically realistic range (i.e., T0 < Tm and we assumed that temperatures below 0˚C and above 45˚C were lethal). Any negative values for all thermal response functions were truncated at zero, and thermal responses for probabilities (pEA, b, and c) were also truncated at one. Temperature-sensitive R0 models While different model formulations for predicting R0 versus temperature can produce results with different magnitudes and potentially different overall shapes [49], the temperatures for which R0 is above or below zero (or one) are mostly model independent. For instance, two competing models differ only by whether or not the formula in Eq (1) is squared, but the square of a number (e.g., an absolute R0 value) greater than one is always greater than one, and the square of a number less than one is always less than one. Therefore, the threshold temperatures at which absolute R0 > 0 or absolute R0 > 1 will be exactly the same for either PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 10 / 18 Model validation with DENV, CHIKV, and ZIKV incidence data To validate the model, we used data on human cases of DENV, CHIKV, and ZIKV at the country scale and mean temperature during the transmission window. Using statistical models (as described below), we estimated the effects of predicted R0(T) on the probability of local transmission and the magnitude of incidence, controlling for population size and several socioeconomic factors. We downloaded and manually entered Pan American Health Organi- zation (PAHO) weekly case reports for DENV and CHIKV for all countries in the Americas (North, Central, and South America and the Caribbean Islands) from week 1 of 2014 to week 8 of 2015 for CHIKV and from week 52 of 2013 to week 47 of 2015 for DENV (www.paho.org). ZIKV weekly case reports for reporting districts (e.g., provinces) within Colombia, Mexico, El Salvador, and the US Virgin Islands were available from the CDC Epidemic Prediction Initia- tive (https://github.com/cdcepi/) from November 28, 2015 to April 2, 2016. We aggregated the ZIKV data into country-level weekly case reports to match the spatial resolution of the DENV, CHIKV, and covariate data. Incorporating daily temperature variation in transmission models Because organisms do not typically experience constant temperature environments in nature, we incorporated the effects of temperature variation on transmission by calculating a daily average R0 assuming a daily temperature range of 8˚C, across the range of mean temperatures. This range is consistent with daily temperature variation in tropical and subtropical environ- ments but lower than in most temperate environments. At each mean temperature, we used a Parton-Logan model to generate hourly temperatures and calculate each temperature-sensitive trait on an hourly basis [56]. We assumed an irreversible high-temperature threshold above which mosquitoes die and transmission is impossible [57,58]. We set this threshold based on hourly temperatures exceeding the critical thermal maximum (Tm in Tables A-B in S1 Text) for egg-to-adult survival or adult longevity by any amount for five hours or by 3˚C for one hour. We averaged each trait over 24 hours to obtain a daily average trait value, which we used to calculate relative R0 across a range of mean temperatures. We used this model in the valida- tion against human cases (Fig 3) and the risk map (Fig 4). Temperature predicts Zika, dengue, and chikungunya transmission main model presented in the text (Fig 2). It is comparable to some but not all previous mecha- nistic models for Ae. aegypti and Ae. albopictus transmission (Fig E in S1 Text). Results of our main model, fit with informative priors, did not vary substantially from the model fit with uninformative priors (Figs G-H in S1 Text). We modeled the observed data as aris- ing from a normal distribution with the mean predicted by the thermal response function cal- culated at the observed temperature, and the precision τ, (τ = 1/σ), distributed as τ ~ Gamma (0.0001, 00001). We fit the models using Markov Chain Monte Carlo (MCMC) sampling in JAGS, using the R [53] package rjags [54]. For each thermal response, we ran five MCMC chains with a 5000-step burn-in and saved the subsequent 5000 steps. We thinned the poste- rior samples by saving every fifth sample and used the samples to calculate R0 from 15–40˚C, producing a posterior distribution of R0 versus temperature. We summarized the relationship between temperature and each trait or overall R0 by calculating the mean and 95% highest pos- terior density interval (HPD interval; a type of credible interval that includes the smallest con- tinuous range containing 95% of the probability, as implemented in the coda package [55]) for each curve across temperatures. We fit a second set of models for each mosquito species that used informative priors to reduce uncertainty in R0 versus temperature and in the trait thermal responses. In these mod- els, we used Gamma-distributed priors for each parameter T0, Tm, c, and τ fit from an addi- tional ‘prior’ dataset of Aedes spp. trait data that did not meet the inclusion criteria for the primary dataset (Table C in S2 Text). We found that these initial informative priors could have an overly strong influence on the posteriors, in some cases drawing the posterior distributions well away from the primary dataset, which was better controlled and met the inclusion criteria. We accounted for our lower confidence in this data set by increasing the variance in the infor- mative priors, by multiplying all hyperparameters (i.e., the parameters of the Gamma distribu- tions of priors for T0, Tm, and c) by a constant k to produce a distribution with the same mean but 1/k times larger variance. We chose the value of k based on our relative confidence in the prior versus main data. Thus we chose k = 0.5 for b, c, and PDR and k = 0.01 for lf. This is the PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 11 / 18 Temperature predicts Zika, dengue, and chikungunya transmission in each country that accurately represent weather conditions in the areas where transmission occurs, excluding extreme areas where transmission is unlikely. For the study period of Octo- ber 1, 2013 through April 30, 2016, we downloaded daily temperature data for each station from Weather Underground using the weatherData package in R [59]. We removed all data from Chile because it spans so much latitude and the terrain is so diverse that its country-level mean is unlikely to be very representative of the temperature where an outbreak occurred. Socioeconomic covariate data We accessed available data on projected 2016 gross domestic product (GDP) for countries of interest via the International Monetary Fund’s World Economic Outlook Database (http:// www.imf.org/external/ns/cs.aspx?id=28). The direct and total contributions of tourism to GDP in 2016 were compiled from World Travel and Tourism Council economic impact reports (http://www.wttc.org/research/economic-research/economic-impact-analysis/ country-reports/#undefined). We retrieved population size data for 2013–2015 from the United Nations Population Division (https://esa.un.org/unpd/wpp/Download/Standard/ Population/) and averaged them across the three years for each country. Throughout the anal- yses below, unless otherwise specified, we used the natural log of the population size and of GDP as our predictors. We have two reasons for this choice. The first is that, intuitively, the relative order of magnitude of the population/GDP is more important in determining observed outbreak sizes or probabilities than their absolute sizes. Second, population sizes and GDPs across countries tend to exhibit clumped patterns with a few outliers that are much larger than the others. From a statistical perspective, using the un-transformed populations (or GDPs) results in those few large/rich countries having very high leverage in the analysis, and thus potentially skewing the results. Taking a log of the population better balances these pre- dictors and is the standard accepted approach when using these kinds of predictors in regres- sion models. Temperature data collection We matched the DENV, CHIKV, and ZIKV incidence data with temperature using daily temperature data from METAR stations in each country, averaged at the country level by epi- demic week. A previous study found a six-week lagged relationship between temperature and oviposition for Aedes aegypti in Ecuador [40]. Assuming that the subsequent transmission, dis- ease development, medical care-seeking, and case reporting in humans takes an additional four weeks, we assumed a priori a ten-week lag between temperature and incidence (i.e., mean temperature for the week that is ten weeks prior to each case report). METAR stations are internationally standardized weather reporting stations that report hourly temperature and precipitation measures. Outlier weather stations were excluded if they reported a daily maxi- mum temperature below 5˚C or a daily minimum temperature above 40˚C during the study period, extremes that would certainly eliminate the potential for transmission in a local area. Because case data are reported at the country level, we needed a collection of weather stations PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 12 / 18 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 Temperature predicts Zika, dengue, and chikungunya transmission that we tested. As ZIKV is not as well established as either CHIKV or DENV at this time, smaller numbers of cases may indicate autochthonous transmission. Consequently, we chose a threshold of three cases for ZIKV (approximately half the CHIKV threshold). Further, the results were fairly sensitive to the ZIKV threshold as many locations have small numbers of cases. Since higher thresholds exclude a very large proportion of available case data making analysis impossible, we used the slightly less conservative threshold of three cases for autoch- thonous transmission of ZIKV. The resulting data consisted of zeros for no transmission and positive case counts when transmission is presumed to be occurring. To model these data, we used a hurdle model that first uses logistic regression on the presence/absence of local trans- mission data to understand the factors correlated with local transmission occurring or not (PA analysis). Then we modeled the log of incidence (number of new cases per reporting week) for positive values with a gamma generalized linear models (incidence analysis). We were interested in understanding whether R0(T) was an important predictor of human transmission occurrence and incidence, after controlling for potentially confounding factors like population size and socioeconomic conditions. To do this, we fit a series of models with different subsets of predictors that included R0(T) and population size, the socioeconomic var- iables, or both (see Table D in S2 Text for full models). To control for human population size, we created new metrics based on R0(T) and population size to use for validation against the PAHO incidence data. We define GR0, which is the posterior probability that R0(T) > 0. We use log(p)GR0, where p is the population size, as the relevant R0-based predictor for the PA analysis. For the incidence analysis, we instead use log(pR0(T)) as the predictor. In all cases log refers to the natural logarithm. For simplicity, we refer to these as the R0(T) metrics hereaf- ter and in the Results. In both the PA and incidence analyses, we first used the full data sets to examine which of the candidate models best described the data. Randomized quantile residuals indicated that the logistic and gamma GLM models were performing adequately. We compared the approxi- mate model probabilities, calculated from the BIC scores, as well as the proportion of deviance explained (D2) from each model. Validation analyses with human incidence versus temperature datasets To validate the R0(T) model while controlling for population and socio-economic factors, we used generalized linear models (GLMs) on the weekly case count data. Importantly, we focused on testing whether the case counts were consistent with the transmission–temperature relationship predicted from our model, rather than on maximizing the variation explained in the statistical model. We are more specifically interested in understanding autochthonous transmission (i.e., locally acquired, not just imported cases). We set country-level thresholds for the number of cases defining autochthonous transmission for our three diseases separately, based on current transmission understanding: seven cases of CHIKV, 70 cases of DENV, and three cases of ZIKV. We derived these thresholds in the following way. First, we looked for data on outbreaks of travel related cases in countries that are not expected to experience any local transmission. For instance, in 2014 Canada experienced 320 confirmed, travel-related cases of chikungunya (http://www.phac-aspc.gc.ca/publicat/ccdr-rmtc/15vol41/dr-rm41-01/ rapid-eng.php), equivalent to an average of more than six cases per week. Thus, to be conserva- tive in our estimates, we set the threshold of transmission as seven cases/week for CHIKV. The reported weekly cases of DENV transmission in our study sample are considerably higher than for CHIKV (mean DENV incidence was nearly 100 times higher mean CHIKV incidence). We chose a moderately high threshold of 70 cases in a week (i.e., 10 times higher than the CHIKV threshold based on Canadian cases) to reflect higher overall incidence and increased potential for travel related cases. We examined the sensitivity of the results to choice of thresh- old by varying it from 25 to 100, and we found qualitatively similar results for all thresholds PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 13 / 18 Temperature predicts Zika, dengue, and chikungunya transmission Supporting information S1 Text. Supplementary Results, References, and Figures A-O. (PDF) S2 Text. Supplementary Tables A-D. (PDF) S2 Text. Supplementary Tables A-D. (PDF) Mapping temperature suitability for transmission Using the validated model, we were interested in where the temperature was suitable for Ae. aegypti and/or Ae. albopictus transmission for some or all of the year to predict the potential geographic range of outbreaks in the Americas. We visualized the minimum, median, and maximum extent of transmission based on probability of occurrence thresholds from the R0 models for both mosquitoes. We calculated the number of consecutive months in which the posterior probability of R0 > 0 exceeds a threshold of 0.025, 0.5, or 0.975 for both mosquito species, representing the maximum, median, and minimum likely ranges, respectively. The minimum range is shown in Fig 4 and all three ranges are overlaid in Fig D in S1 Text. This analysis indicates the predicted seasonality of temperature suitability for transmission geo- graphically, but does not indicate its magnitude. To generate the maps, we cropped monthly mean temperature rasters from 1950–2000 for all twelve months (Worldclim; www.worldclim. org/) to the Americas (R, raster package, crop function) and assigned cells values of one or zero depending on whether the probability that R0 > 0 exceeded the threshold at the temperatures in those cells. We then synthesized the monthly grids into a single raster that reflected the max- imum number of consecutive months where cell values equaled one. The resulting rasters were plotted in ArcGIS 10.3, overlaying the three cutoffs (Fig D in S1 Text). We employed this process for both mosquito species. Acknowledgments Barry Alto, Krijn Paaijmans, Francis Ezeakacha, and Helene Delatte kindly provided raw data used in the analyses. We gratefully acknowledge the Centers for Disease Control and Preven- tion Epidemic Predictions Initiative (CDC EPI) for collating and sharing the Zika incidence data on GitHub (https://zenodo.org/record/48946#.Vz-EM2bb8ys). Next we examined the performance of the models in predict- ing out of sample, for both PA and incidence analyses. To do this we created 1000 random par- titions, where 90% of the data were used to train the model and 10% were used for testing. In the PA analyses we classified each partition based on presence/absence, with separate classifi- cation thresholds for DENV versus CHIKV/ZIKV as these grouping had much different prob- abilities of occurrence. We assessed the performance of the model for the PA analysis based on the mean misclassification rate. In the incidence analyses we assessed the model performance based on the predictive mean absolute percentage error (MAPE). Since differences in predic- tion success between the models in both the PA and incidence analyses were not statistically significant, we present the simpler models that only include the R0(T) metrics in the main text (Fig 3) and the models that additionally include socioeconomic covariates in the Supplemen- tary Information (Figs B-C in S1 Text). We plotted the model predictions as a function of the R0(T) metrics together with the observed data for the PA and incidence analyses using the R package visreg [60]. The residuals of the incidence model exhibit “inverse trumpeting,” in which residual varia- tion is larger at low than high predicted incidence (Fig I in S1 Text). This occurs in part because we forced the model to go through the origin, i.e., no transmission when R0(T) or the population size is equal to zero. However, the data did sometimes show transmission where we did not expect it, potentially because of imported cases, errors in reporting, or small pockets of transmission suitability in countries or times that are otherwise unsuitable on average. More local-scale case reporting that separates autochthonous from travel-associated cases would be needed to tease apart the source of this error. 14 / 18 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 References 1. Brady OJ, Gething PW, Bhatt S, Messina JP, Brownstein JS, Hoen AG, et al. Refining the global spatial limits of dengue virus transmission by evidence-based consensus. PLOS Negl Trop Dis. 2012; 6: e1760. https://doi.org/10.1371/journal.pntd.0001760 PMID: 22880140 2. Bhatt S, Gething PW, Brady OJ, Messina JP, Farlow AW, Moyes CL, et al. The global distribution and burden of dengue. Nature. 2013; 496: 504–507. https://doi.org/10.1038/nature12060 PMID: 23563266 3. Rasmussen SA, Jamieson DJ, Honein MA, Petersen LR. Zika virus and birth defects—reviewing the evidence for causality. N Engl J Med. 2016; 374: 1981–1987. https://doi.org/10.1056/NEJMsr1604338 PMID: 27074377 4. Scott TW, Takken W. Feeding strategies of anthropophilic mosquitoes result in increased risk of patho- gen transmission. Trends Parasitol. 2012; 28: 114–121. https://doi.org/10.1016/j.pt.2012.01.001 PMID: 22300806 5. Messina JP, Kraemer MU, Brady OJ, Pigott DM, Shearer FM, Weiss DJ, et al. Mapping global environ- mental suitability for Zika virus. eLife. 2016; 5: e15272. https://doi.org/10.7554/eLife.15272 PMID: 27090089 6. Magori K, Legros M, Puente ME, Focks DA, Scott TW, Lloyd AL, et al. Skeeter Buster: A stochastic, spatially explicit modeling tool for studying Aedes aegypti population replacement and population sup- pression strategies. PLOS Negl Trop Dis. 2009; 3: e508. https://doi.org/10.1371/journal.pntd.0000508 PMID: 19721700 7. Johansson MA, Powers AM, Pesik N, Cohen NJ, Staples JE. Nowcasting the spread of chikungunya virus in the Americas. PLoS ONE. 2014; 9: e104915. https://doi.org/10.1371/journal.pone.0104915 PMID: 25111394 8. Perkins TA, Metcalf CJE, Grenfell BT, Tatem AJ. Estimating drivers of autochthonous transmission of chikungunya virus in its invasion of the Americas. PLoS Curr. 2015; 7. 9. Morin CW, Monaghan AJ, Hayden MH, Barrera R, Ernst K. Meteorologically driven simulations of den- gue epidemics in San Juan, PR. PLoS Negl Trop Dis. 2015; 9: e0004002. https://doi.org/10.1371/ journal.pntd.0004002 PMID: 26275146 10. Zhang Q, Sun K, Chinazzi M, Pastore-Piontti A, Dean NE, Rojas DP, et al. Projected spread of Zika virus in the Americas. bioRxiv. 2016; 066456. 11. Dell AI, Pawar S, Savage VM. Systematic variation in the temperature dependence of physiological and ecological traits. Proc Natl Acad Sci. 2011; 108: 10591–10596. https://doi.org/10.1073/pnas. 1015178108 PMID: 21606358 12. Mordecai EA, Paaijmans KP, Johnson LR, Balzer CH, Ben-Horin T, de Moor E, et al. Optimal tempera- ture for malaria transmission is dramatically lower than previously predicted. Ecol Lett. 2013; 16: 22–30. https://doi.org/10.1111/ele.12015 PMID: 23050931 13. Focks DA, Haile DG, Daniels E, Mount GA. Dynamic life table model for Aedes aegypti (Diptera: Culici- dae): analysis of the literature and model development. Author Contributions Conceptualization: EAM MBT VS SJR LRJ ASI JRR. Data curation: JMC MVE PG KM CCM MS DPW EAM CAL. Formal analysis: EAM MVE PG KM MSS DPW LRJ CAL. Funding acquisition: EAM JRR SJR MBT ASI LRJ VS CCM. Investigation: JMC MVE PG KM CCM MS DPW CAL EAM. Methodology: EAM LRJ SJR MSS. Project administration: EAM. Software: LRJ EAM MSS MVE DPW. Supervision: EAM MBT JRR VS LRJ SJR ASI CCM. Validation: LRJ EAM. Project administration: EAM. 15 / 18 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 Temperature predicts Zika, dengue, and chikungunya transmission Visualization: JMC SJR LRJ EAM. Writing – original draft: EAM. Writing – review & editing: EAM JMC MVE PG LRJ CAL KM CCM JRR SJR VS MSS ASI MBT DPW. Temperature predicts Zika, dengue, and chikungunya transmission 18. Eisen L, Monaghan AJ, Lozano-Fuentes S, Steinhoff DF, Hayden MH, Bieringer PE. The impact of tem- perature on the bionomics of Aedes (Stegomyia) aegypti, with special reference to the cool geographic range margins. J Med Entomol. 2014; 51: 496–516. PMID: 24897844 19. Delatte H, Gimonneau G, Triboire A, Fontenille D. Influence of temperature on immature development, survival, longevity, fecundity, and gonotrophic cycles of Aedes albopictus, vector of chikungunya and dengue in the Indian Ocean. J Med Entomol. 2009; 46: 33–41. PMID: 19198515 20. Muturi EJ, Lampman R, Costanzo K, Alto BW. Effect of temperature and insecticide stress on life-his- tory traits of Culex restuans and Aedes albopictus (Diptera: Culicidae). J Med Entomol. 2011; 48: 243– 250. PMID: 21485359 21. Alto BW, Juliano SA. Temperature effects on the dynamics of Aedes albopictus (Diptera: Culicidae) populations in the laboratory. J Med Entomol. 2001; 38: 548–556. PMID: 11476335 22. Westbrook CJ, Reiskind MH, Pesko KN, Greene KE, Lounibos LP. Larval environmental temperature and the susceptibility of Aedes albopictus Skuse (Diptera: Culicidae) to chikungunya virus. Vector- Borne Zoonotic Dis. 2010; 10: 241–247. https://doi.org/10.1089/vbz.2009.0035 PMID: 19725768 23. Briegel H, Timmermann SE. Aedes albopictus (Diptera: Culicidae): Physiological aspects of develop- ment and reproduction. J Med Entomol. 2001; 38: 566–571. PMID: 11476337 24. Calado DC, Navarro-Silva MA. Influência da temperatura sobre a longevidade, fecundidade e atividade hematofa´gica de Aedes (Stegomyia) albopictus Skuse, 1894 (Diptera, Culicidae) sob condic¸ões de laborato´rio. Rev Bras Entomol. 2002; 46: 93–98. 25. Beserra EB, Fernandes CRM, Silva SA de O, Silva LA da, Santos JW dos. Efeitos da temperatura no ciclo de vida, exigências te´rmicas e estimativas do nu´mero de gerac¸ões anuais de Aedes aegypti (Dip- tera, Culicidae). Iheringia Se´r Zool. 2009; Available: http://agris.fao.org/agris-search/search.do? recordID=XS2010500501 26. Westbrook CJ. Larval ecology and adult vector competence of invasive mosquitoes Aedes albopictus and Aedes aegypti for Chikungunya virus [Internet]. University of Florida. 2010. Available: http://etd. fcla.edu/UF/UFE0041830/westbrook_c.pdf 27. Couret J, Dotson E, Benedict MQ. Temperature, Larval Diet, and Density Effects on Development Rate and Survival of Aedes aegypti (Diptera: Culicidae). PLoS ONE. 2014; 9: e87468. https://doi.org/10. 1371/journal.pone.0087468 PMID: 24498328 28. Ezeakacha N. Environmental impacts and carry-over effects in complex life cycles: the role of different life history stages. Dissertation. 2015; Available: http://aquila.usm.edu/dissertations/190 29. Teng H-J, Apperson CS. References J Med Entomol. 1993; 30: 1003–1017. PMID: 8271242 14. Yang HM, Macoris MLG, Galvani KC, Andrighetti MTM, Wanderley DMV. Assessing the effects of tem- perature on the population of Aedes aegypti, the vector of dengue. Epidemiol Infect. 2009; 137: 1188– 1202. https://doi.org/10.1017/S0950268809002040 PMID: 19192322 15. Rueda L, Patel K, Axtell R, Stinner R. Temperature-dependent development and survival rates of Culex quinquefasciatus and Aedes aegypti (Diptera: Culicidae). J Med Entomol. 1990; 27: 892–898. PMID: 2231624 16. Tun-Lin W, Burkot TR, Kay BH. Effects of temperature and larval diet on development rates and survival of the dengue vector Aedes aegypti in north Queensland, Australia. Med Vet Entomol. 2000; 14: 31–37. PMID: 10759309 17. Kamimura K, Matsuse IT, Takahashi H, Komukai J, Fukuda T, Suzuki K, et al. Effect of temperature on the development of Aedes aegypti and Aedes albopictus. Med Entomol Zool. 2002; 53: 53–58. 16 / 18 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 Temperature predicts Zika, dengue, and chikungunya transmission 39. Mordecai EA. Model code, data, and output for “Detecting the impact of temperature on transmission of Zika, dengue and chikungunya using mechanistic models.” [Internet]. 2017. 40. Stewart Ibarra AM, Ryan SJ, Beltra´n E, Mejı´a R, Silva M, Muñoz A´ . Dengue vector dynamics (Aedes aegypti) influenced by climate and social factors in Ecuador: implications for targeted control. PLoS ONE. 2013; 8: e78263. https://doi.org/10.1371/journal.pone.0078263 PMID: 24324542 41. Brady OJ, Golding N, Pigott DM, Kraemer MUG, Messina JP, Reiner RC Jr, et al. Global temperature constraints on Aedes aegypti and Ae. albopictus persistence and competence for dengue virus trans- mission. Parasit Vectors. 2014; 7: 338. https://doi.org/10.1186/1756-3305-7-338 PMID: 25052008 42. Carlson CJ, Dougherty ER, Getz W. An ecological assessment of the pandemic threat of Zika virus. PLoS Negl Trop Dis. 2016; 10: e0004968. https://doi.org/10.1371/journal.pntd.0004968 PMID: 27564232 43. Samy AM, Thomas SM, Wahed AAE, Cohoon KP, Peterson AT, Samy AM, et al. Mapping the global geographic potential of Zika virus spread. Mem Inst Oswaldo Cruz. 2016; 111: 559–560. https://doi.org/ 10.1590/0074-02760160149 PMID: 27653360 44. Wesolowski A, Qureshi T, Boni MF, Sundsøy PR, Johansson MA, Rasheed SB, et al. Impact of human mobility on the emergence of dengue epidemics in Pakistan. Proc Natl Acad Sci. 2015; 201504964. 45. Liu-Helmersson J, Stenlund H, Wilder-Smith A, Rocklo¨v J. Vectorial Capacity of Aedes aegypti: Effects of Temperature and Implications for Global Dengue Epidemic Potential. PLoS ONE. 2014; 9: e89783. https://doi.org/10.1371/journal.pone.0089783 PMID: 24603439 46. Caminade C, Turner J, Metelmann S, Hesson JC, Blagrove MSC, Solomon T, et al. Global risk model for vector-borne transmission of Zika virus reveals the role of El Niño 2015. Proc Natl Acad Sci. 2017; 114: 119–124. https://doi.org/10.1073/pnas.1614303114 PMID: 27994145 47. Bogoch II, Brady OJ, Kraemer MUG, German M, Creatore MI, Kulkarni MA, et al. Anticipating the inter- national spread of Zika virus from Brazil. The Lancet. 2016; 387: 335–336. 48. Briere J-F, Pracros P, Le Roux A-Y, Pierre J-S. A novel rate model of temperature-dependent develop- ment for arthropods. Environ Entomol. 1999; 28: 22–29. 49. Li J, Blakeley D, Smith? RJ. The Failure of R0. Comput Math Methods Med. 2011; 2011: e527610. 50. Lambrechts L, Paaijmans KP, Fansiri T, Carrington LB, Kramer LD, Thomas MB, et al. Impact of daily temperature fluctuations on dengue virus transmission by Aedes aegypti. Proc Natl Acad Sci. 2011; 108: 7460–7465. https://doi.org/10.1073/pnas.1101377108 PMID: 21502510 51. Rohatgi A. WebPlotDigitizer [Internet]. 2015. PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 Development and Survival of Immature Aedes albopictus and Aedes triseria- tus (Diptera: Culicidae) in the Laboratory: Effects of Density, Food, and Competition on Response to Temperature. J Med Entomol. 2000; 37: 40–52. PMID: 15218906 30. Wiwatanaratanabutr S, Kittayapong P. Effects of temephos and temperature on Wolbachia load and life history traits of Aedes albopictus. Med Vet Entomol. 2006; 20: 300–307. https://doi.org/10.1111/j.1365- 2915.2006.00640.x PMID: 17044881 31. Xiao F-Z, Zhang Y, Deng Y-Q, He S, Xie H-G, Zhou X-N, et al. The effect of temperature on the extrinsic incubation period and infection rate of dengue virus serotype 2 infection in Aedes albopictus. Arch Virol. 2014; 159: 3053–3057. https://doi.org/10.1007/s00705-014-2051-1 PMID: 24990415 32. Watts DM, Burke DS, Harrison BA, Whitmire RE, Nisalak A. Effect of temperature on the vector effi- ciency of Aedes aegypti for dengue 2 virus. Am J Trop Med Hyg. 1987; 36: 143–152. PMID: 3812879 33. McLean DM, Clarke AM, Coleman JC, Montalbetti CA, Skidmore AG, Walters TE, et al. Vector capabil- ity of Aedes aegypti mosquitoes for California encephalitis and dengue viruses at various temperatures. Can J Microbiol. 1974; 20: 255–262. PMID: 4132612 34. Carrington LB, Armijos MV, Lambrechts L, Scott TW. Fluctuations at a low mean temperature acceler- ate dengue virus transmission by Aedes aegypti. PLoS Negl Trop Dis. 2013; 7: e2190. https://doi.org/ 10.1371/journal.pntd.0002190 PMID: 23638208 35. Davis NC. The effect of various temperatures in modifying the extrinsic incubation period of the yellow fever virus in Aedes aegypti. Am J Epidemiol. 1932; 16: 163–176. 36. McLean DM, Miller MA, Grass PN. Dengue virus transmission by mosquitoes incubated at low tempera- tures. Mosq News. 1975; Available: http://agris.fao.org/agris-search/search.do?recordID= US19760088008 37. Focks DA, Daniels E, Haile DG, Keesling JE. A simulation model of the epidemiology of urban dengue fever: literature analysis, model development, preliminary validation, and samples of simulation results. Am J Trop Med Hyg. 1995; 53: 489–506. PMID: 7485707 38. Alto BW, Bettinardi D. Temperature and dengue virus infection in mosquitoes: independent effects on the immature and adult stages. Am J Trop Med Hyg. 2013; 88: 497–505. https://doi.org/10.4269/ajtmh. 12-0421 PMID: 23382163 17 / 18 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0005568 April 27, 2017 Available: http://arohatgi.info/WebPlotDigitizer 52. Johnson LR, Ben-Horin T, Lafferty KD, McNally A, Mordecai E, Paaijmans KP, et al. Understanding uncertainty in temperature effects on vector-borne disease: a Bayesian approach. Ecology. 2015; 96: 203–213. PMID: 26236905 53. R Development Core Team. R: A Language and Environment for Statistical Computing [Internet]. Vienna, Austria: R Foundation for Statistical Computing; 2014. Available: http://www.R-project.org 54. Plummer M. rjags: Bayesian Graphical Models using MCMC [Internet]. 2016. Available: http://CRAN.R- project.org/package=rjags 55. Plummer M, Best N, Cowles K, Vines K. CODA: Convergence Diagnosis and Output Analysis for MCMC. 2006. 56. Parton WJ, Logan JA. A model for diurnal variation in soil and air temperature. Agric Meteorol. 1981; 23: 205–216. 57. Paaijmans KP, Heinig RL, Seliga RA, Blanford JI, Blanford S, Murdock CC, et al. Temperature variation makes ectotherms more sensitive to climate change. Glob Change Biol. 2013; 19: 2373–2380. 58. Vasseur DA, DeLong JP, Gilbert B, Greig HS, Harley CDG, McCann KS, et al. Increased temperature variation poses a greater risk to species than climate warming. Proc R Soc Lond B Biol Sci. 2014; 281: 20132612. 59. Narasimhan R. weatherData: Get Weather Data from the Web [Internet]. 2014. Available: https://cran.r- project.org/web/packages/weatherData/index.html 60. Breheny P, Burchett W. visreg: Visualization of Regression Models [Internet]. 2016. Available: https:// cran.r-project.org/web/packages/visreg/index.html 18 / 18
https://openalex.org/W2912074262	https://europepmc.org/articles/pmc6379858?pdf=render	English	null	Gitelman Syndrome: A Rare Cause of Seizure Disorder and a Systematic Review	Case reports in medicine	2,019	cc-by	6,954	Muhammad Asim Shahzad ,1 Maryam Mukhtar,2 Asrar Ahmed,3 Waqas Ullah,3 Rehan Saeed,3 and Mohsin Hamid3 1Resident Physician, Louis Weiss Memorial Hospital, Chicago, IL, USA 2Independent Research Scholar, Fauji Foundation Hospital, Rawalpindi, Pakistan 3Resident Physician, Abington Hospital-Jeﬀerson Health, Abington, PA, USA 1Resident Physician, Louis Weiss Memorial Hospital, Chicago, IL, USA 2Independent Research Scholar, Fauji Foundation Hospital, Rawalpindi, Pakistan 3Resident Physician, Abington Hospital-Jeﬀerson Health, Abington, PA, USA Correspondence should be addressed to Muhammad Asim Shahzad; dr.asimshahzad786@gmail.com Received 26 August 2018; Revised 20 November 2018; Accepted 17 January 2019; Published 5 February 2019 Academic Editor: Masahiro Kohzuki Copyright © 2019 Muhammad Asim Shahzad et al. Tis isanopenaccessarticle distributed undertheCreative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Gitelman syndrome is one of the few inherited causes of metabolic alkalosis due to salt losing tubulopathy. It is caused by tubular defects at the level of distal convoluted tubules, mimicking a thiazide-like tumor. It usually presents in late childhood or in teenage as nonspeciﬁc weakness, fatigability, polyuria, and polydipsia but very rarely with seizures. It is classically associated with hypokalemia, hypomagnesemia, hypocalciuria, hyperreninemia, and hyperaldosteronism. However, less frequently, it can present with normal magnesium levels. It is even rarer to ﬁnd normomagnesemic patients of GS who develop seizures as the main complication since hypomagnesemia is considered the principal etiology of abnormal foci of seizure-related brain activity in GS cases. Interestingly, patients with GS are oftentimes diagnosed during pregnancy when the classic electrolyte pattern consistent with GS is noticed. Our case presents GS with normal serum magnesium in a patient, with seizures being the main clinical presentation. We also did a comprehensive literature review of 122 reported cases to show the prevalence of normal magnesium in GS cases and an overview of clinical and biochemical variability in GS. We suggest that further studies and in-depth analysis are required to understand the pathophysiology of seizures in GS patients with both normal and low magnesium levels. 1. Materials and Methods appeared to be in pain. Review of the system was negative for any previous episodes of seizures in the past, fever, diarrhea, abdominal pain, history of diuretic or laxative abuse, any periorbital puﬃness, and extremities swelling. She was given lorazepam followed by successful resolution of seizures. Two diﬀerent databases (PubMed and Scopus) were searched for all case reports and review articles previously published on GS syndrome. Moreover, after taking informed consent from our patient, we included data from the elec- tronic medical record system to use this information for publication purposes. On physical examination, she was having borderline low blood pressure close to her baseline (105/56) with HR of 80, RR 18, O2 sat. 100% on room air. Systemic examination was otherwise unremarkable without any overt signs of dehydration. Hindawi Case Reports in Medicine Volume 2019, Article ID 4204907, 8 pages https://doi.org/10.1155/2019/4204907 Hindawi Case Reports in Medicine Volume 2019, Article ID 4204907, 8 pages https://doi.org/10.1155/2019/4204907 2. Case Presentation EKG showed U waves and nonspeciﬁc T wave changes. Pertinent labs showed serum blood urea nitrogen (BUN) and creatinine (Cr) of 16 and 0.77, respectively. Serum electrolytes showed serum sodium (Na) of 150 mEq/L, se- rum potassium (K) of 1.4 mEq/L, serum magnesium (Mg) of 2.8 mg/dL, and serum bicarbonate (HCO3) of 35 mEq/L. Urine electrolytes included urine K 22 mEq/L, urine Na 121 mEq/L, and urine Cl 146 mEq/L. Her transtubular A 22-year-old female was brought to the hospital with the complaint of vomiting, generalized weakness, and two ep- isodes of witnessed generalized tonic-clonic seizures 24 hours prior to the time of admission. She had about 5 episodes of nonbloody nonbilious vomiting. She was non- verbal at baseline but was reported to be more lethargic than usual and had a poor oral intake for the last 2 days and Case Reports in Medicine 2 Table 1: Summary of the literature review. Table 1: Summary of the literature review. 2. Case Presentation y Demographics Total (%) Associations Total (%) Age, mean 31 Pregnancy 17 (14) Range 0.3–80 years Calcium pyrophosphate deposition disease (CPPD) 7 (5.7) Males 45 (36) Sjogren syndrome 5 (4) Females 77 (63) Chondrocalcinosis 4 (3.3) Presentation Tyrotoxicosis hypokalemic periodic paralysis (THPP) 2 (1.6) Weakness 52 (43) Empty sella syndrome 2 (1.6) Cramps 23 (19) Type 2 diabetes 2 (1.6) Carpopedal spasms 11 (9) Primary aldosteronism 2 (1.6) Nausea, vomiting 7 (6) Type 1 diabetes 1 (0.8) Nocturia 6 (5) Pemphigus vegetans 1 (0.8) Paralysis 5 (4) Mitochondrial encephalopathy 1 (0.8) Numbness 5 (4) Varicose veins 1 (0.8) Joint pain, arthritis 4 (3) each Fanconi syndrome 1 (0.8) Muscle pain 3 (2) Autosomal dominant familial neurohypophyseal diabetes insipidus 1 (0.8) Polydipsia 3 (2) Syndrome of inappropriate ADH secretion 1 (0.8) Sicca symptoms 3 (2) Familial Mediterranean fever 1 (0.8) Hypokalemic paralysis 2 (1.6) Parathyroid adenoma 1 (0.8) Syncope 2 (1.6) Pancreatic cancer 1 (0.8) Salt craving 2 (1.6) Hashimoto thyroiditis 1 (0.8) Tirst, palpitations, frequent micturition, somnolence 1 (0.8) each Scleroderma 1 (0.8) Nausea, vomiting 7 (6) Crowded lens syndrome 1 (0.8) Paralysis 5 (4) Transient hypophosphatemia of infancy 1 (0.8) Visual abnormalities 4 (3) Pseudotumor cerebri 1 (0.8) Failure to thrive 4 (3) Gout 1 (0.8) Loss of appetite 4 (3) Graves disease 1 (0.8) Respiratory distress 3 (2) Serum Electrolytes Total (%) Arthralgias 2 (1.6) Sodium Headache 2 (1.6) Normal range (135–145 mEq/L) 46 (38) Diarrhea 2 (1.6) Low 15 (12) Raynaud’s phenomenon 2 (1.6) Potassium Incontinence 1 (0.8) Normal range (3.5–5 mEq/L) 14 (11) Insomnia 1 (0.8) Lower limit (2.5–3 mEq/L) 57 (47) Tinnitus 1 (0.8) Low (<2.5 mEq/L) 41 (34%) Perspiration 1 (0.8) Calcium Constipation 1 (0.8) Normal range (2.2–2.7 mmol/L) 42 (34) Complications Low 13 (11) Metabolic alkalosis 7 (5.7) High 2 (1.6) Hypokalemic paralysis 6 (4.9) Magnesium Hypokalemia 5 (4) Normal range (0.70–1.0 mmol/L) 19 (16) Prolonged QT intervals 5 (4) Low 69 (57) Pseudogout 4 (3.3) High 6 (5) Rhabdomyolysis 4 (3.3) Urine analysis ST depression 2 (1.6) Sodium T wave changes on EKG 2 (1.6) mmol/24h Gestational diabetes mellitus 2 (1.6) Normal (40–220 mmol/24 h) 17 (14) Focal segmental glomerulosclerosis 2 (1.6) High 14 (11) Prominent U waves 2 (1.6) Spot (mmol/L) Tubulointerstitial nephritis 2 (1.6) Normal (<20 mmol/L) 1 (0.8) Brain calciﬁcation 1 (0.8) High 3 (2.5) Diabetic ketoacidosis 1 (0.8) Potassium Left ventricular dysfunction 1 (0.8) mmol/24h Prolonged PR interval 1 (0.8) Normal (25–125 mmol/24 h) 30 (25) Ventricular ﬁbrillation 1 (0.8) High 8 (6.5) MPGN 1 (0.8) Low 2 (1.6) Focal seizures 1 (0.8) Spot (mmol/L) Iron deﬁciency anemia 1 (0.8) Normal (20–40 mmol/L) 3 (2.5) Pericardial eﬀusion 1 (0 8) High 3 (2 5) Case Reports in Medicine 3 Table 1: Continued. 2. Case Presentation Table 1: Continued. Demographics Total (%) Associations Total (%) Neuropsychological symptoms 1 (0.8) Calcium Sclerochoroidal calciﬁcations 1 (0.8) mmol/24h Renal tubular acidosis 1 (0.8) Normal (15–20 mmol/24 h) 2 (1.6) High 3 (2.5) Diagnosis Low 58 (48) Based on electrolyte abnormality 68 (56) Spot (mmol/L) Genetic mutations Normal (20–40 mmol/L) SLC12A3 gene mutations 46 (38) High NCCT gene 3 (2.5) Magnesium TSC gene 3 (2.5) mmol/24h Screening 2 (1.6) Normal (3–5 mmol/24 h) 5 (4) CLCNKB gene 1 (0.8) High 13 (11) Low 7 (6) Management Spot (mmol/L) Electrolyte replacement (Mg, K supplements) 92 (75) Normal (8–152 mmol/L) 1 (0.8) Spironolactone 32 (26) High N/A Pain killers 13 (11) Low 1 (0.8) Angiotensin receptor blocker 7 (5.7) Chloride (140–250 mmol/24 h) Amiloride 7 (5.7) mmol/24 h Steroids 5 (4) Normal(140–250 mmol/24 h) 2 (1.6) Eplerenone 3 (2.5) High 3 (2.5) Colchicine (for gout) 2 (1.6) Low 3 (2.5) Desmopressin 2 (1.6) mmol/L Growth hormone (for empty sella syndrome) 2 (1.6) Normal (98–107 mmol/L) Febuxostat (for gout) 1 (0.8) Low 6 (4.9) Cyclophosphamide 1 (0.8) High 4 (3.3) Triamterene 1 (0.8) 24 hr urinary protein Phenytoin 1 (0.8) Normal (<80 mg/24 h) 1 (0.8) Amiodarone (for ventricular ﬁbrillation) 1 (0.8) Low Metoclopramide 1 (0.8) High 7 (6) Antithyroid drugs 1 (0.8) Calcium creatinine ratio Outcome Normal (<0.14) Recovery 86 (70) High Low Total (%) Gitelman syndrome. Tere was no language ﬁlter placed, and articles were collected from their inception till May 2018, using the MEDLINE, Cochrane, Embase, and Scopus databases. Diﬀerent MeSH terminologies such as “Gitel- man,” “Gitelman syndrome,” “Gitelman disease,” and “GS” were combined using the Boolean operators “AND” and “OR” with the terms “hypomagnesemia,” “low mag- nesium,” “serum magnesium,” “plasma magnesium,” and “magnesium levels.” Another author collected few articles through manual search using the reference list of all re- trieved publications through the aforementioned search strategy. potassium gradient (TTKG) was 6.82. Complete blood count and liver function panel were within normal limits. Plasma renin activity (PRA) was 0.33 ng/ml/hr, serum aldosterone/ K ratio of 1/1.4, and aldosterone/plasma renin ratio of 3. Diﬀerential included primary hyperaldosteronism, vomit- ing, and Bartter/Gitelman syndrome. EEG showed abnormal epileptiform activity in the brain consistent with seizure. Low normal BP, high urine Cl with urine Ca, and history negative for laxative/diuretic intake made GS the more likely diﬀerential. Later on, biallelic identiﬁcation of inactivating SLC12A3 mutation conﬁrmed the diagnosis of GS. 2. Case Presentation Patient’s condition improved with aggressive K re- plenishment and antiepileptics in the medical ICU. She was later discharged in a medically stable condition and advised to follow-up with nephrologist and neurologist as an outpatient. 4. Results and Statistical Analysis 4.1. Literature Retrieval and the Results. After a thorough computer literature search, careful veriﬁcation of references, and screening based upon the titles and abstracts, 122 cases of GS patients from 100 articles were identiﬁed for selection [1–100]. It was ensured that repetitive cases in these articles were excluded. Out of these 100 articles, data were also extracted from articles published in languages other than English. 3. Literature Search About 7 cases had no mention of the presenting complaints. Rest of the patients had their own unique features as seen in Table 1. Our patient had a unique presentation of generalized tonic-clonic seizure despite normal serum Mg levels, which has not been pre- viously reported in the literature. GS was found to be most commonly associated with pseudogout and CPPD crystal deposition in about 10% of patients. Other associations included but not limited to Sjogren’s syndrome in 4%, chondrocalcinosis in 3%, and diabetes mellitus (both type 1 and type 2) and primary hyperaldosteronism in about 2% each. A less common association is seen with empty sella syndrome in 2 patients. Seizure disorder as a possible as- sociation with GS was previously reported in only one case by Beltagi et al., most likely due to hypomagnesemia [15]. Our patient, however, was unique with no prior history of epilepsy and had a seizure as the very ﬁrst presentation with normal magnesium levels. 4.5. Diagnosis and Management with Outcomes. Except for one case (n 1/122), where there is no mention of the diagnostic method, genetic testing was utilized in 42% (n 52/122) cases, to deﬁnitively diagnose GS. Te speciﬁc mutations to help make the diagnosis can be seen in Table 1. Almost 56% of patients (n 68/122) were diagnosed based on the presenting electrolytes abnormalities including serum and urine Na, K, Mg, and Ca used adjunctively with PAR concentration. Although the supportive testing with elec- trolytes and supplementary tests were highly suggestive of GS in these 68 cases, genetic tests were not done for various reasons. Tese included lack of resources, nonavailability of genetic test, and loss of follow-up by the patients to be the major ones. Of note is the serum Mg level in the reported cases. Considering the normal range to be between 0.7 and 1 mmol/L (1.5–2 mEq/L; 1.7–2.4 mg/dL), 55% (n 66/122) patients had hypomagnesemia, i.e., <0.7 mmol/L, whereas 20% (n 25/122) had levels 0.7 mmol/L and above. In 31 cases, serum magnesium levels were not reported. Tese levels were important as the clinical severity of presentation was reﬂected by the degree of hypomagnesemia. y g yp g Electrolytes replacement, NSAIDs, and potassium- sparing diuretics with and without ACE In/ARB’s were the mainstay of treatment in almost all of the cases. 3. Literature Search Te available literature was systematically searched by three authors independently to retrieve all available ma- terial on variable clinical and metabolic presentations in 4 Case Reports in Medicine 4.2. Patients Description. Tere were a total of 122 patients including 45% (n 55) males and 65% (n 77) females. Te age of female patients ranged from 4.8 months to 79 years (mean age 28.5 years), whereas for males, it ranged from 7 months to 80 years (mean age of 27.8 years). Te de- scription of patients included in this study is listed in Table 1. and renal tubular acidosis (RTA). Tyrotoxic periodic pa- ralysis and hypokalemic periodic paralysis were also seen in a few cases. However, it must be noted that it is rare for two diﬀerent renal entities to occur at the same time, and several of the studies did not conﬁrm the diagnosis of GS by identifying the inactivation gene mutation leaving open the possibility that underlying pathology may not have been actually Gitelman’s. Long-term follow-up is usually required to observe for these complications; our patient, however, had no further follow-up in our hospital and was referred to the neurologist care. 4.3. Spectrum of Clinical Presentation and Associations. Clinical presentation of Gitelman syndrome was found to be highly variable in the reported patient population. About 30% (n 36/122) of the patients, including 14% (n 17/122) pregnant patients, were having nonspeciﬁc muscle cramps, weakness, fatigability, and anorexia, as the main presentation. Tese were likely due in part to hypo- kalemia and hypomagnesemia. About 12% (n 15/122) of the patients had extremities weakness out of which 7% (n 9/122) presented with bilateral lower limb weakness/ paralysis and the rest of them had quadriplegia as initial presentation. Interestingly, 10.6% (n 12/122) of patients had perioral numbness and symptoms related to tetany/ carpopedal spasm as ﬁrst signs of Gitelman. About 6% of patients had polydipsia, polyuria/enuresis, and salt craving as presenting complaint; however, almost half of the total reported patients had some degree of polydipsia and polyuria in addition to main presenting clinical symptoms. Seven percent (n 9/122) of patients were completely asymptomatic and were diagnosed with routine lab work, either during routine clinical visits or perioperatively. Only 5.7% of patients (n 7/122) had GI-related issues such as anorexia, vomiting, constipation, abdominal pain, and weight loss as the main complaint. 3. Literature Search Out- comes and prognosis were remarkable, and patients fully recovered from their acute presenting symptoms with ex- ception of a few cases. Tese few cases reported persistent electrolytes abnormalities such as hypokalemia, metabolic alkalosis, hypomagnesemia, occasional paralysis and neu- rological symptoms, and treatment-related complications (indomethacin-related GI upset and bleeding). Recovery in the other cases is being deﬁned as a sustained increase in electrolytes with magnesium >2, potassium >4, and sig- niﬁcant improvement in the symptoms. Around 22% (n 28/122) cases did not comment on the outcomes. 4.4. Complications Related to Gitelman Syndrome. Complications related to renal, cardiac, and endocrine systems have frequently been reported in the previous cases. Cardiac manifestations ranged from electrolytes related, asymptomatic ECG changes including prolonged Qtc, nonspeciﬁc T and U waves to pericardial eﬀusion, and ventricular ﬁbrillation. Reported renal pathologies included glomerulonephritides such as MPGN, FSGS, membranous nephropathy, and also cases of tubulointerstitial nephritis 4.4. Complications Related to Gitelman Syndrome. Complications related to renal, cardiac, and endocrine systems have frequently been reported in the previous cases. Cardiac manifestations ranged from electrolytes related, asymptomatic ECG changes including prolonged Qtc, nonspeciﬁc T and U waves to pericardial eﬀusion, and ventricular ﬁbrillation. Reported renal pathologies included glomerulonephritides such as MPGN, FSGS, membranous nephropathy, and also cases of tubulointerstitial nephritis References [1] C. Z. Molin and D. J. Trevisol, “Persistent severe hypoka- lemia: gitelman syndrome and diﬀerential diagnosis,” Jornal Brasileiro de Nefrologia, vol. 39, no. 3, pp. 337–340, 2017. [2] X. Gu, Z. Su, M. Chen, Y. Xu, and Y. Wang, “Acquired Gitelman syndrome in a primary Sj¨ogren syndrome patient with a SLC12A3 heterozygous mutation: a case report and literature review,” Nephrology, vol. 22, no. 8, pp. 652–655, 2017. [3] C. J. Subasinghe, N. D. Sirisena, C Herath et al., “Novel mutation in the SLC12A3 gene in a Sri Lankan family with Gitelman syndrome & coexistent diabetes: a case report,” BMC Nephrology, vol. 18, no. 1, p. 140, 2017. [4] Q. U. Mustafa, Z. H. Haroon, A. Ijaz, M. T. Sajid, and M. Ayyub, “Gitelman syndrome,” Journal of College of Physicians and Surgeons–Pakistan: JCPSP, vol. 27, no. 3, pp. 30–32, 2017. 5.1. Treatment. Most patients with GS remain untreated. Te observation that chondrocalcinosis is due to magnesium deﬁciency argues clearly in favor of magnesium supple- mentation [15]. Most asymptomatic patients with GS remain untreated and undergo ambulatory monitoring, once a year, generally by nephrologists. Lifelong supplementation of magnesium and potassium is mandatory [105]. Cardiac workup should be performed to screen for risk factors of cardiac arrhythmias. All GS patients are encouraged to maintain a high-sodium diet. In general, the long-term prognosis of GS is excellent. Health education with annual regular nephrologist follow-up to evaluate for any developing complications seems to be a reasonable approach. As men- tioned in the abstract, GS can be ﬁrst identiﬁed during pregnancy when classic electrolyte abnormalities are noticed on the lab work [106]. Successful pregnancy is possible in majority of the patients; however, miscarriages have also been reported in the literature, which alludes to regular nephrol- ogist follow-up during pregnancy. [5] T. Kusuda, T. Hosoya, T. Mori et al., “Acquired gitelman syndrome in an anti-SSA antibody-positive patient with a,” Internal Medicine, vol. 55, no. 21, pp. 3201–3204, 2016. [6] S. M. Troster, J. E. Raizman, and L. Rubin, “An unusual case of gout in a young woman with gitelman syndrome,” Journal of Rheumatology, vol. 43, no. 11, pp. 2085–2087, 2016. [7] K. Nozu, Y. Nozu, K Nakanishi et al., “Cryptic exon acti- vation in SLC12A3 in Gitelman syndrome,” Journal of Human Genetics, vol. 62, no. 2, p. 335, 2017. [8] Y. Zhang, F. 5. Discussion In our literature review, only one patient who was reported by Beltagi et al. [15] presented with somnolence and altered mental status and had a focal seizure as a complication. Even in that case, hypomagnesemia can be considered as the cause of epileptiform activity on EEG. Tis observation prompts us to consider causes other than hy- pomagnesemia as a culprit of seizure disorder, whenever evaluating the patient with GS. Te ﬁnal diagnosis of GS is based on the triad of clinical symptoms, biochemical ab- normalities, and genetic testing [103]. Genetic testing is recommended for all patients, and the diagnosis is con- ﬁrmed with the biallelic identiﬁcation of inactivating SLC12A3 mutations [104]. We emphasize after this litera- ture review that contrary to common clinical practice, overall clinical picture with more emphasis on genetic testing is a better strategy to clinch the diagnosis, and the diagnosis of GS can still be made even with normal serum magnesium levels. (v) Successful pregnancy is possible in majority of the patients; however, miscarriages have also been re- ported in the literature, which alludes to regular nephrologist follow-up in the pregnant GS patient References Zhang, D Chen et al., “A novel homozygous mutation in the solute carrier family 12 member 3 gene in a Chinese family with Gitelman syndrome,” Brazilian Journal of Medical and Biological Research, vol. 49, no. 11, 2016. [9] K. Gandhi, D. Prasad, V. Malhotra, and D. Agrawal, “Gitelman’s syndrome presenting with hypocalcemic tetany and hypokalemic periodic paralysis,” Saudi Journal of Kidney Diseases and Transplantation, vol. 27, no. 5, p. 1026, 2016. [10] S. Skalova and S. Kutilek, “Transient hyperphosphatemia: a benign laboratory disorder in a boy with Gitelman syn- drome,” Jornal Brasileiro de Nefrologia, vol. 38, no. 3, pp. 363–365, 2016. Conflicts of Interest Te authors declare that they have no conﬂicts of interest. 5. Discussion However, GS can also present with normal serum magne- sium levels, and in one case, it has been reported to be in around 20–40% of GS cases [101]. From our review of around 122 cases, 20% (n 25/122) patients had serum magnesium levels >0.7 mmol/L. Both the groups of GS patients with normal and low magnesium levels largely stay asymptomatic and present later in life. Most present in Case Reports in Medicine 5 teenage or adulthood with nonspeciﬁc generalized weakness or muscle cramps/fatigability, polyuria, and polydipsia [103]. However, seizure disorder has very rarely been re- ported as one of the main presenting complaints. Hypo- magnesemia and metabolic alkalosis have been proposed as the pathophysiological basis of these rarely reported seizure disorders. Our case reports are unique in this sense that the patient of GS presented with seizure despite having normal serum magnesium levels. In our literature review, only one patient who was reported by Beltagi et al. [15] presented with somnolence and altered mental status and had a focal seizure as a complication. Even in that case, hypomagnesemia can be considered as the cause of epileptiform activity on EEG. Tis observation prompts us to consider causes other than hy- pomagnesemia as a culprit of seizure disorder, whenever evaluating the patient with GS. Te ﬁnal diagnosis of GS is based on the triad of clinical symptoms, biochemical ab- normalities, and genetic testing [103]. Genetic testing is recommended for all patients, and the diagnosis is con- ﬁrmed with the biallelic identiﬁcation of inactivating SLC12A3 mutations [104]. We emphasize after this litera- ture review that contrary to common clinical practice, overall clinical picture with more emphasis on genetic testing is a better strategy to clinch the diagnosis, and the diagnosis of GS can still be made even with normal serum magnesium levels. (iv) Further studies are recommended to better un- derstand the pathophysiology of abnormal epilep- tiform activity in GS teenage or adulthood with nonspeciﬁc generalized weakness or muscle cramps/fatigability, polyuria, and polydipsia [103]. However, seizure disorder has very rarely been re- ported as one of the main presenting complaints. Hypo- magnesemia and metabolic alkalosis have been proposed as the pathophysiological basis of these rarely reported seizure disorders. Our case reports are unique in this sense that the patient of GS presented with seizure despite having normal serum magnesium levels. 6. Conclusion Kaniuka-Jakubowska, M Nagel et al., “A case report of Gitelman syndrome resulting from two novel mutations in SLC12A3 gene,” Nefrolog´ıa (English Edition), vol. 36, no. 3, pp. 304–309, 2016. [33] A. Ali, Q. Masood, S. Yaqub, and W. Kashif, “A case of Gitelman syndrome with severe hyponatraemia and hypo- phosphataemia,” Singapore Medical Journal, vol. 54, no. 1, pp. 18–20, 2013. [17] Q. Lü, Y. Zhang, C Song et al., “A novel SLC12A3 gene homozygous mutation of Gitelman syndrome in an Asian pedigree and literature review,” Journal of Endocrinological Investigation, vol. 39, no. 3, pp. 333–340, 2016. [34] S. Skalova, D. Neuman, P. Lnenicka, and J. Stekrova, “Case report: gitelman syndrome as a cause of psychomotor re- tardation in a toddler,” Arab Journal of Nephrology and Transplantation, vol. 6, no. 1, pp. 37–39, 2013. [18] D. R. Waguespack, R. Kasekar, K. Abdel-Kader, and R. B. Fissell, “Two cases of successful pregnancy in patients with Gitelman’s syndrome,” Clinical nephrology, vol. 84, no. 11, pp. 301–306, 2015. [35] M. Yildiz, B. S. Yildiz, S. Karakoyun, S. Cakal, A. Sahin, and N. B. Aladag, “Te eﬀects of serum potassium and mag- nesium levels in a patient with Gitelman’s syndrome on the timing of ventricular wall motion and the pattern of ven- tricular strain and torsion,” Echocardiography, vol. 30, no. 2, pp. 47–50, 2013. [19] V. Martin-Miguel, M. A. Lafarga-Giribets, L. Garcia-Esteve, and M. D. Rodrigo-Claverol, “Casual diagnosis of Gitelman’s syndrome,” Semergen, vol. 40, no. 7, pp. e95–e98, 2014. [20] M. Guedes-Marques, C. Silva, E. Ferreira, P. Maia, A. Carreira, and M. Campos, “Gitelman syndrome with hiponatremia, a rare presentation,” Nefrolog´ıa (Madrid), vol. 34, no. 2, pp. 266–268, 2014. [36] S. K. Das, A. Ghosh, N. Banerjee, and S. Khaskil, “Gitelman’s syndrome presenting with hypocalcaemia, basal ganglia calciﬁcation and periodic paralysis,” Singapore Medical Journal, vol. 53, no. 10, pp. 222–224, 2012. [21] N. Demoulin, S. Aydin, J. P Cosyns et al., “Gitelman syn- drome and glomerular proteinuria: a link between loss of sodium-chloride cotransporter and podocyte dysfunction?,” Nephrology Dialysis Transplantation, vol. 29, no. 4, pp. 117–120, 2014. [37] S¸ Erten, G. G. Ceylan, and A. Altuno˘glu, “Concomitance of Gitelman syndrome and familial mediterranean fever: a rare case presentation,” Renal Failure, vol. 34, no. 10, pp. 1333-1334, 2012. [22] M. Yoshihara, A. Sayo, M. Mayama, and H. Oguchi, “Pseudo gitelman syndrome associated with pregnancy,” Obstetrics & Gynecology, vol. 126, no. 4, pp. 877–880, 2015. 6. Conclusion (i) GS with variable biochemical presentation, i.e., normal serum magnesium level is a rare but potentially possible ﬁnding seen in various clinical settings (i) GS with variable biochemical presentation, i.e., normal serum magnesium level is a rare but potentially possible ﬁnding seen in various clinical settings pp [11] A. Zabotti, P. Della Siega, L. Picco, L. Quartuccio, M. Bassetti, and S. De Vita, “Gitelman syndrome disclosed by calcium pyrophosphate deposition disease: early diagnosis by ultrasonographic study,” Reumatismo, vol. 68, no. 1, pp. 53–55, 2016. (ii) Although exceedingly rare, seizure disorder can be the main clinical presentation of GS (ii) Although exceedingly rare, seizure disorder can be the main clinical presentation of GS pp [12] S. Mukhopadhyay, S. Jana, A. Chatterjee, M. Roy, A. Sarkar, and J. Mukhopadhyay, “Quadriparesis due to Gitelman’s syndrome diagnosed with thiazide diuretic test response,” Saudi Journal of Kidney Diseases and Transplantation, vol. 27, no. 2, p. 407, 2016. (iii) Causes other than low magnesium levels should be sought for the explanation of seizure disorder in GS 6 Case Reports in Medicine [29] P. Cotovio, C. Silva, N. Oliveira, and F. Costa, “Gitelman syndrome,” BMJ Case Reports, vol. 2013, article bcr2013009095, 2013. [13] L. Koudsi, S. Nikolova, and V. Mishra, “Management of a severe case of Gitelman syndrome with poor response to standard treatment,” BMJ Case Reports, vol. 2016, article bcr2015212375, 2016. [30] G. Brambilla, M. Perotti, S. Perra, R. Dell’Oro, G. Grassi, and A. I. Pincelli, “It is never too late for a genetic disease: a case of a 79-year-old man with persistent hypokalemia,” Journal of Nephrology, vol. 26, no. 3, pp. 594–598, 2013. [14] H. Akkari, M. Belkahla, M Youssef et al., “Pemphigus vegetans associated with Gitelman syndrome,” Indian Journal of Dermatology, Venereology, and Leprology, vol. 81, no. 6, p. 655, 2015. f p gy pp [31] S. Mathen, M. Venning, and J. Gillham, “Outpatient man- agement of Gitelman’s syndrome in pregnancy,” BMJ Case Reports, vol. 2013, article bcr2012007927, 2013. [15] A. E. Beltagi, A. Norbash, and S. Vattoth, “Novel brain MRI abnormalities in Gitelman syndrome,” Neuroradiology Journal, vol. 28, no. 5, pp. 523–528, 2015. [32] A. J. Cruz and A. Castro, “Gitelman or bartter type 3 syn- drome? A case of distal convoluted tubulopathy caused by CLCNKB gene mutation,” BMJ Case Reports, vol. 2013, article bcr2012007929, 2013. [16] W. Wolyniec, S. 6. Conclusion pp [38] N. Takahashi, H. Kimura, S Mizuno et al., “Severe intra- glomerular detachment of podocytes in a Gitelman syn- drome patient,” Clinical and Experimental Nephrology, vol. 16, no. 3, pp. 495–500, 2012. [23] K. Bouchireb, O. Boyer, L Mansour-Hendili et al., “Fanconi syndrome and severe polyuria: an uncommon clinicobio- logical presentation of a Gitelman syndrome,” BMC pedi- atrics, vol. 14, no. 1, p. 201, 2014. [39] A. M. Tom`as, A. R. Cid, C. C. Fern´andez, and J. C. Catal´a, “Carpal arthritis as the initial manifestation of gitelman’s syndrome,” Reumatolog´ıa Cl´ınica (English Edition), vol. 8, no. 3, p. 159, 2012. [24] F. J. de Vargas Garcip´erez, J. Mendoza, C. Ortiz, and P. S´anchez-Calder´on, “Gitelman’s syndrome: a wolf in sheep’s clothing,” Revista Clinica Espanola, vol. 214, no. 4, pp. 229-230, 2014. [40] Y. K. Wen, “An unusual case of Gitelman’s syndrome with hypercalcemia,” Renal Failure, vol. 34, no. 2, pp. 241–243, 2012. [25] M. Brugnara, R. Gaudino, S Tedeschi et al., “Type III Bartter- like syndrome in an infant boy with Gitelman syndrome and autosomal dominant familial neurohypophyseal diabetes insipidus,” Journal of Pediatric Endocrinology and Meta- bolism, vol. 27, no. 9-10, pp. 971–975, 2014. [41] J. D. Farmer, G. M. Vasdev, and D. P. Martin, “Perioperative considerations in patients with Gitelman syndrome: a case series,” Journal of Clinical Anesthesia, vol. 24, no. 1, pp. 14–18, 2012. [26] A. S¸ Koçkara, F. Candan, C. Hüzmeli, M. Kayatas¸, and D. Alaygut, “Gitelman’s syndrome associated with chon- drocalcinosis: a case report,” Renal Failure, vol. 35, no. 9, pp. 1285–1288, 2013. [42] F. Nakhoul, N. Nakhoul, E. Dorman, L. Berger, K. Skorecki, and D. Magen, “Gitelman’s syndrome: a pathophysiological and clinical update,” Endocrine, vol. 41, no. 1, pp. 53–57, 2012. [43] L. A. Cal`o and P. Caielli, “Gitelman’s syndrome and preg- nancy: new potential pathophysiological inﬂuencing factors, therapeutic approach and materno-fetal outcome,” Journal of Maternal-Fetal & Neonatal Medicine, vol. 25, no. 8, pp. 1511–1513, 2012. [27] E. Schneck, S. Schaumberg, C. Koch, M. Rickert, and C. Lichtenstern, “An¨asthesiologisches management des Gitelman-syndroms anesthesiological management of Gitelman syndrome,” Der Anaesthesist, vol. 62, no. 9, pp. 728–733, 2013. pp [28] B. Murienne, P. Pointet, and G. Beaune, “Gitelman syn- drome: a crucial role of laboratory medicine for the di- agnosis,” InAnnales de Biologie Clinique, vol. 71, no. 2, pp. 235–239, 2013. [44] A. Sinha, P. Lnˇeniˇcka, B. Basu, A. Gulati, P. Hari, and A. 6. Conclusion 893–897, 2009. [51] S. K. Bandyopadhyay, S. Datt, S. K. Pal, and A. K. Saha, “Gitelman’s syndrome: a diﬀerential diagnosis of normo- calcemic tetany,” Journal of Association of Physicians of India, vol. 58, p. 395, 2010. [68] G. de Arriba, M. S´anchez-Heras, and M. A. Basterrechea, “Gitelman syndrome during pregnancy: a therapeutic challenge,” Archives of Gynecology and Obstetrics, vol. 280, no. 5, pp. 807–809, 2009. [52] R. Enriquez, V. Adam, A. E. Sirvent, A. B. Garc´ıa-Garc´ıa, I. Mill´an, and F. Amor´os, “Gitelman syndrome due to p. A204T mutation in CLCNKB gene,” International Urology and Nephrology, vol. 42, no. 4, pp. 1099–1102, 2010. [69] T. Kasifoglu, A. Akalin, D. U. Cansu, and C. Korkmaz, “Hypokalemic paralysis due to primary hyperaldosteronism simulating Gitelman’s syndrome,” Saudi Journal of Kidney Diseases and Transplantation, vol. 20, no. 2, p. 285, 2009. [53] O. Hinschberger, L. Martzolﬀ, G. Ioannou, D. Baumann, F. Jaeger, and P. Kieﬀer, “Acquired Gitelman syndrome associated with Sj¨ogren’s syndrome and scleroderma,” La Revue de Medecine Interne, vol. 32, no. 8, pp. 96–98, 2011. [70] A. G. P´erez, T. Olea, C. Caramelo, E. Coto, and F. Santos, “Compound heterozygosis for intr´on 9 + 1 G > T and Leu 850pro mutations in the SLC12A3 gene in Gitelman’s syndrome,” Nefrolog´ıa, vol. 28, no. 6, 2008. [54] S. Shanbhag, J. Neil, and C. Howell, “Anaesthesia for cae- sarean section in a patient with Gitelman’s syndrome,” In- ternational Journal of Obstetric Anesthesia, vol. 19, no. 4, pp. 451–453, 2010. [71] Y. K. Kim, H. C. Song, W. Y Kim et al., “Acquired Gitelman syndrome in a patient with primary Sj¨ogren syndrome,” American Journal of Kidney Diseases, vol. 52, no. 6, pp. 1163–1167, 2008. [55] M. Ceri, S. Unverdi, M Altay et al., “Focal segmental glo- merulosclerosis in association with Gitelman syndrome,” International Urology and Nephrology, vol. 43, no. 3, pp. 905–907, 2011. [72] J. de Haan, T. Geers, and A. Berghout, “Gitelman syndrome in pregnancy,” International Journal of Gynecology & Ob- stetrics, vol. 103, pp. 69–71, 2008. pp [56] N. Lakhi, J. Jones, and A. Govind, “Fetal demise despite normalisation of serum potassium in Gitelman syndrome case report and literature review,” Australian and New Zealand Journal of Obstetrics and Gynaecology, vol. 50, no. 3, pp. 301-302, 2010. [73] B. Akinci, A. Celik, F. Saygili, and S. Yesil, “A case of Gitelman’s syndrome presenting with extreme hypokalaemia and paralysis,” Experimental and Clinical Endocrinology & Diabetes, vol. 6. Conclusion Bagga, “Gitelman syndrome: novel mutation and long- term follow-up,” Clinical and Experimental Nephrology, vol. 16, no. 2, pp. 306–309, 2012. 7 Case Reports in Medicine [45] J. G. Dis, “Pancreatic tumor and Gitelman syndrome,” Journal of Gastrointestinal and Liver Diseases, vol. 20, no. 3, pp. 329–332, 2011. of the literature,” Nephrology Dialysis Transplantation, vol. 25, no. 4, pp. 1338–1340, 2010. [62] H. Kumagai, S. Matsumoto, and K. Nozu, “Hypokalemic rhabdomyolysis in a child with Gitelman’s syndrome,” Pe- diatric Nephrology, vol. 25, no. 5, pp. 953–955, 2010. pp [46] M. Biagioni, M. Marigliano, A Iannilli et al., “Diabetic ketoacidosis complicated with previously unknown Gitel- man syndrome in a Tunisian child,” Diabetes Care, vol. 34, no. 6, p. 107, 2011. [63] M. Gutierrez, F. Silveri, C Bertolazzi et al., “Gitelman syn- drome, calcium pyrophosphate dihydrate deposition disease and crowned dens syndrome. A new association?,” Rheu- matology, vol. 49, no. 3, pp. 610–613, 2009. [47] A. Azak, B. Huddam, G. Koçak, L. Ortabozkoyun, M. Uzel, and M. Duranay, “Gitelman syndrome complicated with dysglycemia,” Acta Diabetologica, vol. 48, no. 3, pp. 249-250, 2011. [64] K. Nozu, K. Iijima, Y Nozu et al., “A deep intronic mutation in the SLC12A3 gene leads to Gitelman syndrome,” Pediatric Research, vol. 66, no. 5, pp. 590–593, 2009. [48] S. Ahmed, M. Qayyum, and F. Farooq, “Quadriparesis in an adult--Gitelman syndrome,” JPMA. Journal of Pakistan Medical Association, vol. 61, no. 2, pp. 182–184, 2011. [65] Z. Miao, Y. Gao, R. J Bindels et al., “Coexistence of nor- motensive primary aldosteronism in two patients with Gitelman’s syndrome and novel thiazide-sensitive Na–Cl cotransporter mutations,” European Journal of Endocrinol- ogy, vol. 161, no. 2, pp. 275–283, 2009. [49] T. K. Kwan and M. C. Falk, “Second pregnancy outcome in a patient with Gitelman syndrome without the use of par- enteral electrolyte supplementation,” Australian and New Zealand Journal of Obstetrics and Gynaecology, vol. 51, no. 1, pp. 94-95, 2011 1. gy pp [66] N. Akhtar and F. Hafeez, “A rare case of Gitelman’s syn- drome with hypophosphatemia,” Journal of the College of Physicians and Surgeons–Pakistan : JCPSP, vol. 19, pp. 257– 259, 2009. [50] C. S. Quinlan, J. C. Walsh, A. M. Moran, C. Moran, and S. K. O’Rourke, “Gitelman’s syndrome,” Journal of Bone and Joint Surgery. British, vol. 93, no. 2, pp. 266–268, 2011. [67] M. Roser, N. Eibl, B Eisenhaber et al., “Gitelman syndrome,” Hypertension, vol. 53, no. 6, pp. 6. Conclusion [95] Z. Dimitrijevi´c, B. Miti´c, and V. ÐorCevi´c, “Gitelman syn- drome as a rare cause of hypokalemia: case report,” Acta Medica Medianae, vol. 53, no. 3, pp. 54–57, 2014. [80] N. Godefroid, E. Riveira-Munoz, C. Saint-Martin, M. C. Nassogne, K. Dahan, and O. Devuyst, “A novel splicing mutation in SLC12A3 associated with Gitelman syndrome and idiopathic intracranial hypertension,” American Journal of Kidney Diseases, vol. 48, no. 5, pp. 73–79, 2006. [96] Y. Takeuchi, E. Mishima, H Shima et al., “Exonic mutations in the SLC12A3 gene cause exon skipping and premature termination in Gitelman syndrome,” Journal of American Society of Nephrology, vol. 26, no. 2, pp. 271–279, 2015. [81] H. Y. Ng, S. H. Lin, C. Y. Hsu, Y. Z. Tsai, H. C. Chen, and C. T. Lee, “Hypokalemic paralysis due to Gitelman syn- drome: a family study,” Neurology, vol. 67, no. 6, pp. 1080–1082, 2006. [97] B. Zha, P. Zheng, J. Liu, and X. Huang, “Coexistence of graves’ disease in a 14-year-old young girl with gitelman syndrome,” Clinical Endocrinology, vol. 83, no. 6, pp. 995– 997, 2015. [82] T. Hashida, M. Yamada, K Hashimoto et al., “Loss of consciousness and hypokalemia in an elderly man with a mutation of the thiazide-sensitive Na-Cl cotransporter gene,” Endocrine Journal, vol. 53, no. 6, pp. 859–863, 2006. [98] J. Bolton and J. F. Mayhew, “Anesthesia in a patient with Gitelman syndrome,” Anesthesiology, vol. 105, no. 5, pp. 1064-1065, 2006. [99] C. Li, X. Zhou, W Han et al., “Identiﬁcation of two novel mutations in SLC12A3 gene in two Chinese pedigrees with Gitelman syndrome and review of literature,” Clinical En- docrinology, vol. 83, no. 6, pp. 985–993, 2015. [83] C. Hanevold, A. Mian, and R. Dalton, “C1q nephropathy in association with Gitelman syndrome: a case report,” Pedi- atric Nephrology, vol. 21, no. 12, pp. 1904–1908, 2006. [84] Y. T. Lee, I. F. Wang, T. H. Lin, and C. T. Huang, “Gitelman syndrome: report of three cases and literature review,” Kaohsiung Journal of Medical Sciences, vol. 22, no. 7, pp. 357–362, 2006. [100] F. Tosi, N. D. Bianda, A. C Truttmann et al., “Normal plasma total magnesium in Gitelman syndrome,” American Journal of Medicine, vol. 116, no. 8, pp. 573-574, 2004. f [101] M. Saeed, J. S. Bhandohal, P. Nepal, and S. Chaudhari, “Gitelman syndrome with a normal magnesium level,” Consultant, vol. 58, no. 6, p. 188, 2018. [85] C. Schwarz, T. Barisani, E. 6. Conclusion Bauer, and W. Druml, “A woman with red eyes and hypokalemia: a case of acquired Gitelman syndrome “Rote Augen” and Hypokali¨amie,” Wiener klinische Wochenschrift, vol. 118, no. 7-8, pp. 239–242, 2006. [102] I. Kurtz, J. J. Cohen, J. T. Harrington, N. E. Madias, and C. J. Zusman, “Molecular pathogenesis of Bartter’s and Gitelman’s syndromes,” Kidney International, vol. 54, no. 4, pp. 1396–1410, 1998. [86] K. Panichpisal, F. Angulo-Pernett, S. Selhi, and K. M. Nugent, “Gitelman-like syndrome after cisplatin therapy: a case report and literature review,” BMC Ne- phrology, vol. 7, no. 1, p. 10, 2006. [103] D. N. Cruz, A. J. Shaer, M. J. Bia, R. P. Lifton, and D. B. Simon, “Gitelman’s syndrome revisited: an evaluation of symptoms and health-related quality of life,” Kidney In- ternational, vol. 59, no. 2, pp. 710–717, 2001. [87] R. Gupta, V. Hu, T. Reynolds, and R. Harrison, “Scle- rochoroidal calciﬁcation associated with Gitelman syndrome and calcium pyrophosphate dihydrate deposition,” Journal of Clinical Pathology, vol. 58, no. 12, pp. 1334-1335, 2005. [104] W. Ji, J. N. Foo, B. J. O’Roak et al., “Rare independent mutations in renal salt handling genes contribute to blood pressure variation,” Nature Genetics, vol. 40, no. 5, pp. 592–599, 2008. [88] J. A. Riancho, G. Saro, C. Sanudo, M. J. Izquierdo, and M. T. Zarrabeitia, “Gitelman syndrome: genetic and ex- pression analysis of the thiazide-sensitive sodium-chloride transporter in blood cells,” Nephrology Dialysis Trans- plantation, vol. 21, no. 1, pp. 217–220, 2005. pp [105] L. Jiang, C. Chen, T Yuan et al., “Clinical severity of Gitelman syndrome determined by serum magnesium,” American Journal of Nephrology, vol. 39, no. 4, pp. 357–366, 2014. [106] J. De Haan, T. Geers, and A. Berghout, “Gitelman syndrome in pregnancy,” International Journal of Gynecology & Ob- stetrics, vol. 103, no. 1, pp. 69–71, 2008. [89] G. S. Talaulikar and M. C. Falk, “Outcome of pregnancy in a patient with Gitelman syndrome: a case report,” Nephron Physiology, vol. 101, no. 2, pp. 35–38, 2005. [90] T. Ring, N. Knoers, M. S. Oh, and M. L. Halperin, “Reevaluation of the criteria for the clinical diagnosis of Gitelman syndrome,” Pediatric Nephrology, vol. 17, no. 8, pp. 612–616, 2002. [91] T. Bourcier, P. Blain, P. Massin, J. P. Grünfeld, and A. Gaudric, “Sclerochoroidal calciﬁcation associated with Gitelman syndrome,” American Journal of Ophthalmology, vol. 128, no. 6, pp. 767-768, 1999. [92] L. C. Liaw, K. Banerjee, and M. G. 6. Conclusion 117, no. 02, pp. 69–71, 2009. [74] K. Aoki, T. Tajima, Y Yabushita et al., “A novel initial codon mutation of the thiazide-sensitive Na-Cl cotransporter gene in a Japanese patient with Gitelman’s syndrome,” Endocrine Journal, vol. 55, no. 3, pp. 557–560, 2008. [57] F. Tammaro, A. Bettinelli, D Cattarelli et al., “Early ap- pearance of hypokalemia in Gitelman syndrome,” Pediatric Nephrology, vol. 25, no. 10, pp. 2179–2182, 2010. pp [75] A. Morton, B. Panitz, and A. Bush, “Eplerenone for gitelman syndrome in pregnancy,” Nephrology, vol. 16, no. 3, p. 349, 2011. [58] T. Bansal, S. Abeygunasekara, and V. Ezzat, “An unusual presentation of primary renal hypokalemia– hypomagnesemia (Gitelman’s syndrome),” Renal Failure, vol. 32, no. 3, pp. 407–410, 2010. [76] A. Volpe, P. Caramaschi, U Talheimer et al., “Familiar association of Gitelman’s syndrome and calcium pyro- phosphate dihydrate crystal deposition disease—a case re- port,” Rheumatology, vol. 46, no. 9, pp. 1506–1508, 2007. [59] G. Daskalakis, S. Marinopoulos, A. Mousiolis, S. Mesogitis, N. Papantoniou, and A. Antsaklis, “Gitelman syndrome- associated severe hypokalemia and hypomagnesemia: case report and review of the literature,” Journal of Maternal-Fetal & Neonatal Medicine, vol. 23, no. 11, pp. 1301–1304, 2010. [77] G. Ducarme, C. Davitian, M. Uzan, X. Belenfant, and C. Poncelet, “Pregnancy in a patient with Gitelman syn- drome: a case report and review of literature,” Journal de Gynecologie, Obstetrique et Biologie de la Reproduction, vol. 36, no. 3, pp. 310–313, 2007. [60] R. Bansal and V. K. Ranga, “Acquired Gitelman’s syndrome: an oxymoron?,” International Urology and Nephrology, vol. 43, no. 1, pp. 233–236, 2011. [78] H. Kaito, K. Nozu, X. J. Fu et al., “Detection of a transcript abnormality in mRNA of the SLC12A3 gene extracted from [61] F. P. McCarthy, C. N. Magee, W. D. Plant, and L. C. Kenny, “Gitelman’s syndrome in pregnancy: case report and review 8 Case Reports in Medicine urinary sediment cells of a patient with Gitelman’s syn- drome,” Pediatric Research, vol. 61, no. 4, pp. 502–505, 2007. [94] F. Ferraro, P. Debruxelles, A. Massard, B. Dor´emus, and A. Blondiaux, “Gitelman syndrome: a rare cause of hypokalemia-hypomagnesemia in children,” Archives de Pediatrie: Organe Oﬃciel de la Societe Francaise de Pediatrie, vol. 3, no. 3, pp. 293-294, 1996. [79] M. Sartori, E. Parotto, E Bonso et al., “Autonomic nervous system function in chronic hypotension associated with Bartter and Gitelman syndromes,” American Journal of Kidney Diseases, vol. 49, no. 2, pp. 330–335, 2007. 6. Conclusion Coulthard, “Dose related growth response to indometacin in Gitelman syndrome,” Archives of Disease in Childhood, vol. 81, no. 6, pp. 508–510, 1999. [93] A. Bettinelli, R. Rusconi, S Ciarmatori et al., “Gitelman disease associated with growth hormone deﬁciency, dis- turbances in vasopressin secretion and empty sella: a new hereditary renal tubular-pituitary syndrome?,” Pediatric Research, vol. 46, no. 2, pp. 232–238, 1999.
https://openalex.org/W3207071915	https://www.researchsquare.com/article/rs-966059/v1.pdf?c=1634097198000	English	null	Computer-assisted individualized hemodynamic management reduces intraoperative hypotension	null	2,021	cc-by	551	Computer-assisted individualized hemodynamic management reduces intraoperative hypotension Alexandre Joosten Joseph Rinehart Philippe Van der Linden Brenton Alexander Christophe Penna Jacques De Montblanc Maxime Cannesson Jean-Louis Vincent Eric Vicaut Jacques Duranteau Abstract A new study published in the journal Anesthesiology suggests that computer-assisted individualized hemodynamic management is a promising strategy to minimize hypotension during certain surgeries. Intraoperative hypotension is common but can increase the risk of postoperative complications. During surgery, individualized hemodynamic management can help mitigate hypotension and decrease such complications. In this strategy, clinicians use both intravenous fluids and vasopressors to maintain mean arterial pressure and blood volume near a patient’s personalized baseline. Despite its reported success, this approach requires constant monitoring and adjustment, which can be particularly challenging during complex and prolonged surgeries. Automated systems that can deliver either vasopressors or fluids have recently been developed, but a single closed-loop system that can deliver both with minimal clinician intervention isn’t yet available. To pave the way for such technology, researchers at a hospital in Bicêtre, France, designed a computer-assisted system with two main parts: one for closed-loop vasopressor administration and another for decision-supported fluid administration. The researchers then tested their system against traditional fluid and vasopressor management with respect to intraoperative hemodynamics and postoperative complications. The researchers randomized thirty-eight patients scheduled for elective intermediate- to high-risk abdominal or orthopedic surgery to either a computer- assisted goal-directed therapy group or a manually adjusted goal-directed therapy group and compared the outcomes. The same hemodynamic management protocol was used in both groups; only the presence of automation differed. Compared with the manual group, the computer-assisted group had significantly less hypotension, indicating better blood pressure management. This was likely due to the much higher number of vasopressor infusion rate modifications in this group. However, there was no difference in the postoperative complication rate between groups, likely because of the small sample size. Additionally, all flow-based hemodynamic variables were kept closer to the target value in the computer-assisted group than in the manual group. Finally, the mean stroke volume index and cardiac index were significantly greater in the computer-assisted group, further supporting the superior hemodynamic control of the computerized system. Notably, the computer-assisted surgeries were supervised by the principal investigator, and provider experience and type weren’t controlled. Blood loss and aspirin use were greater in the computer-assisted group than in the manual group. Additionally, hypotension during anesthesia induction couldn’t be assessed. Finally, the findings may not apply to types of surgeries beyond those examined. Despite these limitations, this study demonstrates that computer-assisted individualized hemodynamic management is superior to goal-directed manual adjustments in minimizing intraoperative hypotension. Video Abstract Keywords: hemodynamic management, individualized hemodynamic management, vasopressor, fluid titration, intraoperative hypotension, randomized controlled trial, mini-fluid challenge, stroke volume index, cardiac output, decision support system, closed-loop system, norepinephrine infusion, mean arterial pressure, bolus, assisted fluid management, perioperative medicine, Edwards Lifesciences, intermediate- and high-risk surgery, anesthesia, surgery, anesthesiology Posted Date: October 12th, 2021 DOI: https://doi.org/10.21203/rs.3.rs-966059/v1 DOI: https://doi.org/10.21203/rs.3.rs-966059/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Page 1/2 Page 1/2 Abstract This work is yet another steppingstone in the development of automated systems to optimize patient management during surgery. Page 2/2 Page 2/2
https://openalex.org/W4294243331	https://zaguan.unizar.es/record/124017/files/texto_completo.pdf	English	null	Predictive study of pharmacological reversal for residual neuromuscular blockade and postoperative pulmonary complications: a prospective, observational, cohort study	Scientific reports	2,022	cc-by	7,201	www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Scientific Reports \| (2022) 12:14955 Methods Study design and setting. A single-center, prospective, observational cohort study was designed that included patients undergoing elective or emergency surgery at Miguel Servet University Hospital in Zaragoza from January 2016 to December 2019. The reporting of this study conforms to the STROBE statement. Ethics. The study was first approved by the Ethical and Research Committee of Miguel Servet University Hospital, Zaragoza, Spain, with registration code 06/2014 (Chairperson J.M. Larrosa Poves). Subsequently, it was reauthorized by the Regional Ethics Committee of Aragón (CEICA), with the number CAB-SUG-2019-01 (Chairperson M. Gonzalez Hinjos) as requested by regional guidelines. This study was performed in line with the principles of the Declaration of Helsinki and written informed consent was obtained from all subjects. Inclusion/exclusion criteria. The inclusion criteria were: patients with ASA physical status I to III, aged over 18 years, who were to undergo general anaesthesia with neuromuscular blocking agents and signed informed consent. Exclusion criteria included patients with ASA physical status IV to V, known neuromuscular disease, diabetes mellitus with diagnosed neuropathy, pregnancy or lactation, known allergy to neuromuscular blocking agents, cardiac surgery, or planned admission to surgical ICU with mechanical ventilation. The patients were selected before the surgery, having signed the consent form for inclusion in the study. Primary and secondary outcomes. The primary outcome was the presence of postoperative RNMB, defined as a TOF ratio < 0.9 at admission to the PACU.hi i The secondary outcomes were the POPC, as defined in other studies like ARISCAT15 or PERISCOPE16. Early POPC were considered as at least one of the following respiratory events in the PACU: upper airway obstruction, desaturation below 92%, bronchoaspiration, or need for reintubation for the severe respiratory failure of the patient. Late POPC were defined as at least one event of pneumonia or atelectasis in the 30 days following surgery. Patient population and anaesthesia. The recruited patients were those who were to receive neuromuscular blocking agents under balanced general anaesthesia. Neuromuscular blockade was performed according to routine clinical practice and usual department protocol with cisatracurium (0.1–0.2 mg/kg) or rocuronium (0.6–1.2 mg/kg) for anaesthetic induction at the choice of the anaesthesiologist in charge of the patient who was blinded to the patient’s inclusion in the study. Similarly, anaesthetic maintenance, intraoperative quantitative NMM, repeated doses of the neuromuscular blocking agent, or pharmacological reversal at the end of surgery depended on the clinical criteria of the same anaesthesiologist. Predictive study of pharmacological reversal for residual neuromuscular blockade and postoperative pulmonary complications: a prospective, observational, cohort study OPEN Cristian Aragón‑Benedí1,3, Ana Pascual‑Bellosta1,3, Sonia Ortega‑Lucea1,3, Sara Visiedo‑Sánchez2,3, Javier Martínez‑Ubieto1,3 & Research Group in Anaesthesia, Resuscitation, and Perioperative Medicine of Institute for Health Research Aragón (ISS Aragón)3 In recent years, some studies have generated controversy since they conclude that intraoperatively pharmacological reversal of neuromuscular blockade does not contribute to the reduction of postoperative residual neuromuscular blockade or pulmonary complications. Therefore, the main objective of this study was to assess the incidence of residual neuromuscular blockade and postoperative pulmonary complications according to spontaneous or pharmacological neuromuscular reversal. The secondary aim was to present a prognostic model to predict the probability of having postoperative residual neuromuscular blockade depending on a patient’s comorbidities and intraoperative neuromuscular blocking agents management. A single-center, prospective, observational cohort study including patients undergoing surgical procedures with general anesthesia was designed. A total of 714 patients were analyzed. Patients were divided into four groups: cisatracurium with spontaneous reversal, cisatracurium with neostigmine antagonism, rocuronium with spontaneous reversal, and rocuronium with sugammadex antagonism. According to our binomial generalized linear model, none of the studied comorbidities was a predisposing factor for an increase in the residual neuromuscular blockade. However, in our study, pharmacological reversal of rocuronium with sugammadex and, particularly, neuromuscular monitoring during surgery were the factors that most effectively reduced the risk of residual neuromuscular blockade as well as early and late postoperative pulmonary complications. Abbreviations RNMB Residual neuromuscular blockade POPC Postoperative pulmonary complications NMM Neuromuscular monitoring NMB Neuromuscular blockade PACU Post-anaesthesia care unit TOF Train of four 1Department of Anesthesia, Resuscitation and Pain Therapy, Miguel Servet University Hospital, 50009 Zaragoza, Spain. 2Department of Anaesthesia, Resuscitation and Pain Therapy, University Clinical Hospital Lozano Blesa, Zaragoza, Spain. 3Institute for Health Research Aragón (IIS Aragón), Zaragoza, Spain. A list of authors and their affiliations appears at the end of the paper. email: cristianaragon@outlook.com \| https://doi.org/10.1038/s41598-022-18917-y Scientific Reports \| (2022) 12:14955 www.nature.com/scientificreports/ www.nature.com/scientificreports/ More than 400 million people receive neuromuscular blocking agents annually to paralyze skeletal muscle groups, facilitate tracheal intubation, allow for controlled mechanical ventilation and achieve optimum relaxation conditions for surgery1,2.ht g y The possibility of residual neuromuscular blockade (RNMB) after using neuromuscular blocking agents has been known for some time. However, in recent years, there has been an increase in the number of publications showing its high incidence, its relationship to postoperative pulmonary complications (POPC), and increased potential healthcare costs3–6. POPC include upper airway obstruction, oxygen desaturation, bronchoaspiration, pneumonia, atelectasis, and reintubation for severe respiratory failure requiring an unplanned admission to an intensive care unit (ICU)2,3,5–7. Numerous studies and multiple international organizations have suggested that every patient receiving non-depolarising neuromuscular blocking drugs should have at least qualitative, and preferably quantitative intraoperative monitoring of the neuromuscular blockade (NMB) and assessment of the pharmacologic antagonism of NMB8–11. Nevertheless, anesthesia professionals have not widely utilized quantitative measurements of drug-induced NMB and the adequacy of pharmacologic reversal3–5. Furthermore, intraoperative neuromuscular monitoring (NMM) rates vary according to each center and do not cover the entire surgical patients with general anaesthesia and neuromuscular blocking agents11–13. p g g g In addition, in recent years, some studies have generated controversy since they contradict most previous studies, concluding that intraoperative monitoring of the NMB and pharmacological reversal do not contribute in any way to the reduction of postoperative RNMB or pulmonary complications2,14. In this regard, the primary objective of this study was to assess the incidence of RNMB and POPC according to spontaneous or pharmacological neuromuscular reversal, based on current clinical practice. The secondary objective was to present a prognostic model to predict the probability of having RNMB depending on the patient’s comorbidities and the intraoperative management of the neuromuscular blocking agents. www.nature.com/scientificreports/ restrictive lung disease, asthma, acute myocardial infarction, heart failure, high blood pressure, anaemia, chronic renal failure, diabetes mellitus, dyslipidemia, hyperthyroidism, hypothyroidism, chronic liver disease, dementia, and fragility). restrictive lung disease, asthma, acute myocardial infarction, heart failure, high blood pressure, anaemia, chronic renal failure, diabetes mellitus, dyslipidemia, hyperthyroidism, hypothyroidism, chronic liver disease, dementia, and fragility).h g y The type of surgery (general surgery, maxillofacial surgery, otolaryngology, urology, vascular surgery, and others), emergency or elective procedure, intraoperative quantitative NMM during surgery, and repeated doses of the neuromuscular blocking agent were recorded as intraoperative data. To measure the postoperative RNMB, we used, in 100% of the patients at admission to the PACU, a single TOF measurement (four stimuli of 0.2 ms in duration at a frequency of 2 Hz) with an intensity of 40 mA using a TOF-Watch-SX® acceleromyography device [Organon, Oss, The Netherlands] calibrated in the operating room before the first dose of NMB. It was performed by the research staff, who was blinded and was not involved in the intraoperative care of the patient. p p To assess the POPC, we consulted the patient’s electronic clinical history, recording any clinical event in PACU or on the hospital ward, laboratory test, radiological study, and primary care or emergency room consultation reports during hospital admission or 30 days after surgery confirming the type of POPC. Sample size. Assuming an incidence of residual neuromuscular blockade with rocuronium and cisatracurium of 13%17 and 34.1%7, respectively, with a significant level of 5% and 95% of power, a sample size of 103 patients was calculated using the EPIDAT v. 4.1. software. To account for dropouts, we included at least 110 patients per group. Patient recruitment was performed through a sequential review of cases in a recruitment period from January 2016 to December 2019. Statistical analysis. A descriptive analysis was completed to perform data analysis using the mean, standard deviation, and quartiles to summarize quantitative data according to normal distribution. For qualitative data, frequency and percentages were used. A χ2 test and a Fisher’s test were used for qualitative variables, and when proportions were compared for different groups, a difference in proportions test was used. A Kruskal–Wallis test and an analysis of variance (ANOVA) test were used to study the relationship of a qualitative variable with a quantitative variable. www.nature.com/scientificreports/ q A Binomial Generalized Linear Model was performed to predict RNMB using the demographic data (age, weight, gender, ASA), the comorbidities (chronic obstructive pulmonary disease, obstructive sleep apnea syn- drome, restrictive lung disease, asthma, acute myocardial infarction, heart failure, high blood pressure, anaemia, chronic renal failure, diabetes mellitus, dyslipidemia, hyperthyroidism, hypothyroidism, chronic liver disease, dementia, and fragility) and the variables of the neuromuscular blockade management (neuromuscular blocking agent, intraoperative NMM, pharmacological reversal) as previously detailed. g g y For this, the Likelihood-Ratio test was used to select the variables of the Binomial Generalized Linear Model with Logistic Regression (logit link) that were part of the final model. The modeling process was carried out in stages, eliminating the variables with a lower significance or equivalently with a higher p-value for the Likelihood- Ratio Test in each stage.fih Differences for which the p-value was < 0.05 were considered significant. The analysis has been developed with R version 3.4.4 (R Foundation for Statistical Computing, Vienna, Austria). The statistical analysis and the data review were developed by Jorge Luis Ojeda Cabrera Ph.D. (Dept. of Statistical Methods of the University of Zaragoza). Results During the study period, 735 patients were included, 21 of whom were excluded, as detailed in Fig. 1. STROBE patient flow diagram. Patients were divided into the four groups detailed in the methodology by type of neuromuscular blocking agent and spontaneous or pharmacological reversal (Fig. 1). The groups were homogeneous, and there were no differences between the groups in patient demographic data or comorbidities (Table 1). Residual neuromuscular blockade between groups. We found that 28.3% (n = 202) of all patients had RNMB. According to the four groups, the incidence of RNMB was: group 1 cisatracurium without pharmacological reversal 33.52% (n = 59), group 2 cisatracurium with neostigmine antagonism 30.35% (n = 34), group 3 rocuronium without pharmacological reversal 35.87% (n = 99), and group 4 rocuronium with sugammadex antagonism 5.33% (n = 8), with p < 0.001, χ2 test (Table 2). Methods If pharmacological antagonism was performed, patients with rocuronium received sugammadex (2–4 mg/kg), and those with cisatracurium were administered neostigmine (0.03–0.05 mg/kg) and atropine (0.02 mg/kg) according to routine clinical practice. In our department protocol, we did not assess the reversal of rocuronium with neostigmine, given the current evidence of increased postoperative complications with this combination6,7. According to the neuromuscular blocking agents and spontaneous or pharmacological reversal, the patients were then categorized into four groups: group 1 cisatracurium without pharmacological reversal, group 2 cisatracurium with neostigmine antagonism, group 3 rocuronium without pharmacological reversal, and group 4 rocuronium with sugammadex antagonism. Measurements and data handling. Patient demographic data included age, weight, gender, ASA physical status, and comorbidities (chronic obstructive pulmonary disease, obstructive sleep apnea syndrome, https://doi.org/10.1038/s41598-022-18917-y Scientific Reports \| (2022) 12:14955 \| www.nature.com/scientificreports/ Intraoperative neuromuscular monitoring and residual neuromuscular blockade. Intraoperative NMM was used in 30.3% (n = 216) of patients, with no statistically significant differences in the four groups (p=0.98, χ2 test). Intraoperative neuromuscular monitoring and residual neuromuscular blockade. Intraoperative NMM was used in 30.3% (n = 216) of patients, with no statistically significant differences in the four groups (p = 0.98, χ2 test). If we analyze the influence of intraoperative NMM and RNMB, patients not monitored intraoperatively had an incidence of RNMB of 35.7% (n = 178). However, when monitored, the incidence decreased to 10.2% (n = 22) with p < 0.001, χ2 test (Table 3). (p χ ) If we analyze the influence of intraoperative NMM and RNMB, patients not monitored intraoperatively had an incidence of RNMB of 35.7% (n = 178). However, when monitored, the incidence decreased to 10.2% (n = 22) with p < 0.001, χ2 test (Table 3). Postoperative pulmonary complications between groups. Concerning the respiratory events, a total of 15.27% (n = 109) of all patients had some early POPC in the PACU. Of the total patients, 10.92% (n = 78) presented oxygen desaturation and 4.34% (n = 31) presented upper airway obstruction. There were no cases of bronchoaspiration or reintubation for severe respiratory. https://doi.org/10.1038/s41598-022-18917-y Scientific Reports \| (2022) 12:14955 \| www.nature.com/scientificreports/ www.nature.com/scientificreports/ Excluded n = 8 Unexpected critical care n = 3 Missing information n = 5 Missing information n = 13 Patients quantitative TOF PACU n = 727 Cisatracurium - No Pharmacological Reversal n = 176 Cisatracurium - Neostigmine n = 112 Rocuronium - No Pharmacological Reversal n = 276 Rocuronium - Sugammadex n = 150 Patients recruited n = 735 Patients included in the study n = 714 Figure 1. STROBE patient flow diagram. TOF, train of four; PACU, post-anaesthesia care unit. Patients recruited n = 735 Patients included in the study n = 714 Figure 1. STROBE patient flow diagram. TOF, train of four; PACU, post-anaesthesia care unit. On the other hand, the incidence of late POPC at 30 days after surgery was 8.12% (n = 58): 6.44% (n = 46) had atelectasis and 1.68% (n = 12) had pneumonia (Tables 2 and 3). On the other hand, the incidence of late POPC at 30 days after surgery was 8.12% (n = 58): 6.44% (n = 46) had atelectasis and 1.68% (n = 12) had pneumonia (Tables 2 and 3). Predictive model for residual neuromuscular blockade. All those variables with a positive coefficient estimate contributed to increasing the incidence, while those with a negative coefficient decreased it (Table 4). Intraoperative neuromuscular monitoring and residual neuromuscular blockade. Intraoperative NMM was used in 30.3% (n = 216) of patients, with no statistically significant differences in the four groups (p=0.98, χ2 test). None of the demographic data and comorbidities added to the model predisposed to having more RNMB. However, the type of neuromuscular blocking agent used, TOF monitoring during surgery, and pharmacological reversal did have a significant effect.i gif Specifically, as can be seen from the following data (Table 4), the pharmacological combination of rocuronium (− 0.44, coefficient estimate) with sugammadex (− 0.88, coefficient estimate) and, particularly, intraoperative NMM (− 1.46, coefficient estimate) significantly reduced the incidence of RNMB. fi gi y When using rocuronium, avoidance of intraoperative NMM and neuromuscular blockers antagonism led to an incidence of RNMB of 41.27%; Conversely, the use of monitoring and pharmacological reversal decreased the probability to 2.17% (Table 5). Discussionh This prospective, observational cohort study was intended to clarify certain questions arising in recent years from several international studies on RNMB. According to the available literature, this article is one of the few analyzing the patient’s demographic data, comorbidities, and the current clinical practice of intraoperative man- agement of the neuromuscular blocking agents in a single predictive model for RNMB. Postoperative residual neuromuscular blockade. We showed that the incidence of RNMB in our study was 28.3%, i.e. approximately 1 out of every 3 patients under balanced general anaesthesia presented this complication. This figure still appears to be high today; however, it did not differ from the data from the most recent studies, where the incidence ranges from 14 to 32%18–22.i g According to our results, RNMB was significantly decreased when intraoperative monitoring was performed nd when rocuronium was reverted with sugammadex. g In cases where intraoperative NMM was used, the expected probability of RNMB decreased by a little over 25%. On the other hand, this probability decreased by 17% when neuromuscular management was done with rocuronium and sugammadex. However, more importantly, when we performed both techniques in the same intervention, the probability decreased by more than 30%. p y y Nevertheless, the incidence of RNMB in the group with sugammadex was 5.3% since it could probably be explained by clinical error. The dosing of sugammadex should be based on actual body weight23–25. For moderate NMB, defined as 1 to 2 twitches, the dose is 2 mg/kg, but for deep NMB, defined as a post-tetanic count of 1 to 2, it is up to 4 mg/kg, and if no monitoring is performed, the degree of NMB cannot be known. Many specialists routinely use 200 mcg of sugammadex12, which probably often leads to overdosing, but also to underdosing, particularly in patients weighing more than 100 kg25. The same occurs during emergency surgery. The dose of rocuronium is usually doubled, i.e. 1.2 mg/kg, when a rapid sequence intubation is used. Discussionh Especially in these cases, https://doi.org/10.1038/s41598-022-18917-y Scientific Reports \| (2022) 12:14955 \| www.nature.com/scientificreports/ Quantitative variables (n) Cisatracurium—no reversal group n = 176 Cisatracurium + neostigmine group n = 112 Rocuronium— no reversal group n = 276 Rocuronium + sugammadex group n = 150 P-value Mean SD Mean SD Mean SD Mean SD ANOVA Age; years (714) 59.8 17.3 60.8 16.7 60.7 16.3 59.4 15.5 0.81 Weight; kg (714) 70.2 11.9 72.6 12.5 72.1 13.2 75.4 16.8 0.11 ASA Score (714) 2.10 0.69 2.16 0.71 2.11 0.67 2.14 0.66 0.89 Qualitative variables (n) Percent % (n) Percent % (n) Percent % (n) Percent % (n) χ2 test Male (474) 23.2% (110) 16.2% (77) 40.1% (190) 20.4% (97) 0.49 Female (240) 27.5% (66) 14.5% (35) 35.8% (86) 22.1% (53) COPD (85) 21.1% (18) 16.4% (14) 49.4% (42) 12.9% (11) 0.09 OSAS (32) 40.6% (13) 3.10% (1) 34.3% (11) 21.8% (7) 0.07 Restrictive lung disease (8) 12.5% (1) 25.0% (2) 25.0% (2) 37.5% (3) 0.49 Asthma (14) 35.7% (5) 0.00% (0) 42.8% (6) 21.4% (3) 0.38 AMI (67) 22.3% (15) 23.8% (16) 35.8% (24) 17.9% (12) 0.28 Heart failure (20) 20.0% (4) 15.0% (3) 40.0% (8) 25.0% (5) 0.94 High blood pressure (329) 26.4% (87) 17.6% (58) 37.9% (125) 17.9% (59) 0.16 Anaemia (62) 30.6% (19) 17.7% (11) 35.4% (22) 16.1% (10) 0.54 Chronic renal failure (40) 37.5% (15) 17.5% (7) 27.5% (11) 17.5% (7) 0.20 DM (131) 23.6% (31) 21.3% (28) 36.6% (48) 18.3% (24) 0.25 Dyslipidemia (165) 24.2% (40) 18.7% (31) 37.5% (62) 19.3% (62) 0.64 Hyperthyroid- ism (4) 0.00% (0) 25.0% (1) 50.0% (2) 25.0% (1) 0.71 Hypothyroidism (2) 27.2% (6) 13.6% (3) 45.4% (10) 13.6% (3) 0.80 Chronic liver disease (20) 35.0% (7) 20.0% (4) 20.0% (4) 25.0% (5) 0.36 Dementia (2) 0.00% (0) 0.00% (0) 0.00% (0) 100% (2) 0.05 Fragility (114) 20.1% (23) 17.5% (20) 43.8% (50) 18.4% (21) 0.42 Table 1. Homogeneity and comparison of demographic data and comorbidities between groups. Basic descriptives and tests for the demographic and comorbidity variables for each group. As can be seen, there was no significant relationship between the demographic and comorbidity variables and each group. Absolute (N) and relative (%) frequencies along with independence tests (χ2) for the qualitative variables, and mean and standard deviation (SD) along with comparing means tests (analysis of variance [ANOVA]) for the quantitative variables. Significance defined as p-value < 0.05. Table 1. Homogeneity and comparison of demographic data and comorbidities between groups. Basic descriptives and tests for the demographic and comorbidity variables for each group. As can be seen, there was no significant relationship between the demographic and comorbidity variables and each group. Absolute (N) and relative (%) frequencies along with independence tests (χ2) for the qualitative variables, and mean and standard deviation (SD) along with comparing means tests (analysis of variance [ANOVA]) for the quantitative variables. Significance defined as p-value < 0.05. ASA, American Society of Anesthesiologists score; COPD, chronic obstructive pulmonary disease; OSAS, obstructive sleep apnea syndrome; AMI, acute myocardial infarction; DM, diabetes mellitus. Discussionh ASA, American Society of Anesthesiologists score; COPD, chronic obstructive pulmonary disease; OSAS, obstructive sleep apnea syndrome; AMI, acute myocardial infarction; DM, diabetes mellitus. Table 2. Incidence of residual neuromuscular blockade, early and late postoperative pulmonary complications between groups. Absolute (N) and relative (%) frequencies for each group along with independence tests (χ2) Significance defined as p-value < 0.05. RNMB, residual neuromuscular blockade; POPC, postoperative pulmonary complications. Qualitative variables Cisatracurium—no reversal group n = 176 Cisatracurium + neostigmine group n = 112 Rocuronium—no reversal group n = 276 Rocuronium + sugammadex group n = 150 P-value Percent % (n) Percent % (n) Percent % (n) Percent % (n) χ2 test RNMB 33.5% (59) 30.3% (34) 35.8% (99) 5.33% (8) < 0.001 Early POPC 26.7% (47) 18.7% (21) 12.3% (34) 4.67% (7) < 0.001* Late POPC 7.39% (13) 8.93% (10) 9.78% (27) 2.67% (4) 0.038* Qualitative variables Cisatracurium—no reversal group n = 176 Cisatracurium + neostigmine group n = 112 Rocuronium—no reversal group n = 276 Rocuronium + sugammadex group n = 150 P-value Percent % (n) Percent % (n) Percent % (n) Percent % (n) χ2 test RNMB 33.5% (59) 30.3% (34) 35.8% (99) 5.33% (8) < 0.001* Early POPC 26.7% (47) 18.7% (21) 12.3% (34) 4.67% (7) < 0.001* Late POPC 7.39% (13) 8.93% (10) 9.78% (27) 2.67% (4) 0.038* Table 2. Incidence of residual neuromuscular blockade, early and late postoperative pulmonary complications between groups. Absolute (N) and relative (%) frequencies for each group along with independence tests (χ2) Significance defined as p-value < 0.05. RNMB, residual neuromuscular blockade; POPC, postoperative pulmonary complications. Table 2. Incidence of residual neuromuscular blockade, early and late postoperative pulmonary complications between groups. Absolute (N) and relative (%) frequencies for each group along with independence tests (χ2) Significance defined as p-value < 0.05. RNMB, residual neuromuscular blockade; POPC, postoperative pulmonary complications. as always, it is critical to use NMM and, if necessary, to use the correct dose of sugammadex per the patient’s real weight and degree of the blockade24–27. s always, it is critical to use NMM and, if necessary, to use the correct dose of sugammadex per the patient’s real weight and degree of the blockade24–27. https://doi.org/10.1038/s41598-022-18917-y Scientific Reports \| (2022) 12:14955 \| www.nature.com/scientificreports/ Table 3. Incidence of residual neuromuscular blockade and postoperative pulmonary complications if exists both intraoperative neuromuscular monitoring and pharmacological reversal. Discussionh Absolute (N) and relative (%) frequencies for each group along with independence tests (χ2) Significance defined as p-value < 0.05. RNMB, residual neuromuscular blockade; POPC, postoperative pulmonary complications; NMM, neuromuscular monitoring. Qualitative variables No intraoperative NMM Intraoperative NMM P-value No pharmacological reversal Pharmacological reversal P-value Percent % (n) Percent % (n) χ2 test Percent % (n) Percent % (n) χ2 test RNMB 35.7% (178) 10.2% (22) < 0.001 35.0% (158) 16.0% (42) < 0.001* Early POPC 17.2% (86) 10.6% (23) 0.023* 17.9% (81) 10.7% (28) 0.009* Hypoxaemia 12.2% (61) 7.87% (17) 0.084 12.8% (58) 7.63% (20) 0.031* Airway obstruction 5.02% (25) 2.78% (6) 0.176 5.09% (23) 3.03% (8) 0.198 Late POPC 9.24% (46) 3.70% (8) 0.011* 8.85% (40) 5.34% (14) 0.087 Pneumonia 2.01% (10) 0.92% (2) 0.301 1.99% (9) 1.15% (3) 0.396 Atelectasis 7.63% (38) 3.70% (8) 0.049* 7.30% (33) 4.96% (13) 0.219 Table 3. Incidence of residual neuromuscular blockade and postoperative pulmonary complications if exists both intraoperative neuromuscular monitoring and pharmacological reversal. Absolute (N) and relative (%) frequencies for each group along with independence tests (χ2) Significance defined as p-value < 0.05. RNMB, residual neuromuscular blockade; POPC, postoperative pulmonary complications; NMM, neuromuscular monitoring. Table 4. Variables and coefficients of the generalized linear model with likelihood-ratio test to predict residual neuromuscular blockade. Significant coefficients of the Generalized Linear Model along with the standard error (Std. Error), the corresponding Z value and p-values [Pr ( >\|z\|)]. The sign of the coefficients of each variables indicates the direction of the influence in the residual neuromuscular blockade. In our case, all the coefficients were negative, so they were factors that reduced the probability of the residual neuromuscular blockade; Significance defined as p-value < 0.05. NMM, neuromuscular monitoring. Estimate Std. error Z value Pr ( >\|z\|) (Intercept) − 0.08 0.15 − 0.56 0.57 Rocuronium − 0.44 0.17 − 2.50 0.010* Intraoperative NMM − 1.46 0.24 − 5.91 < 0.001* Sugammadex − 0.88 0.20 − 4.33 < 0.001* Estimate Std. error Z value Pr ( >\|z\|) (Intercept) − 0.08 0.15 − 0.56 0.57 Rocuronium − 0.44 0.17 − 2.50 0.010* Intraoperative NMM − 1.46 0.24 − 5.91 < 0.001* Sugammadex − 0.88 0.20 − 4.33 < 0.001* Table 4. Variables and coefficients of the generalized linear model with likelihood-ratio test to predict residual neuromuscular blockade. Significant coefficients of the Generalized Linear Model along with the standard error (Std. Discussionh Error), the corresponding Z value and p-values [Pr ( >\|z\|)]. The sign of the coefficients of each variables indicates the direction of the influence in the residual neuromuscular blockade. In our case, all the coefficients were negative, so they were factors that reduced the probability of the residual neuromuscular blockade; *Significance defined as p-value < 0.05. NMM, neuromuscular monitoring. Table 5. Probability of residual neuromuscular blockade according to the neuromuscular blocking agent, neuromuscular monitoring and pharmacological reversal according to the generalized linear model. NMB, neuromuscular blockade; NMM, neuromuscular monitoring; RNMB, residual neuromuscular blockade. NMB agent Intraoperative NMM Pharmacological reversal Probability RNMB (%) Cisatracurium No No 40.5 Rocuronium No No 41.2 Cisatracurium Yes No 14.2 Rocuronium Yes No 14.6 Cisatracurium No Neostigmine 39.5 Rocuronium No Sugammadex 8.33 Cisatracurium Yes Neostigmine 13.7 Rocuronium Yes Sugammadex 2.17 Table 5. Probability of residual neuromuscular blockade according to the neuromuscular blocking agent, neuromuscular monitoring and pharmacological reversal according to the generalized linear model. NMB, neuromuscular blockade; NMM, neuromuscular monitoring; RNMB, residual neuromuscular blockade. NMB agent Intraoperative NMM Pharmacological reversal Probability RNMB (%) Cisatracurium No No 40.5 Rocuronium No No 41.2 Cisatracurium Yes No 14.2 Rocuronium Yes No 14.6 Cisatracurium No Neostigmine 39.5 Rocuronium No Sugammadex 8.33 Cisatracurium Yes Neostigmine 13.7 Rocuronium Yes Sugammadex 2.17 NMB agent Intraoperative NMM Pharmacological reversal Probability RNMB (%) Cisatracurium No No 40.5 Rocuronium No No 41.2 Cisatracurium Yes No 14.2 Rocuronium Yes No 14.6 Cisatracurium No Neostigmine 39.5 Rocuronium No Sugammadex 8.33 Cisatracurium Yes Neostigmine 13.7 Rocuronium Yes Sugammadex 2.17 Table 5. Probability of residual neuromuscular blockade according to the neuromuscular blocking agent, neuromuscular monitoring and pharmacological reversal according to the generalized linear model. NMB, neuromuscular blockade; NMM, neuromuscular monitoring; RNMB, residual neuromuscular blockade. Table 5. Probability of residual neuromuscular blockade according to the neuromuscular blocking agent, neuromuscular monitoring and pharmacological reversal according to the generalized linear model. NMB, neuromuscular blockade; NMM, neuromuscular monitoring; RNMB, residual neuromuscular blockade. and Postoperative pulmonary complication and intraoperative neuromuscular monitoring. Concerning the percentage of intraoperative NMM at our center, we found that in 30.3% of the surgical procedures, quantitatively monitoring of the NMB was used as routine clinical practice; this figure is similar to those of other studies and centers7,12,13. As stated by Naguib et al., the percentage of anaesthesiologists who rely solely on clinical signs for extubation remains very high28. According to our results, this lack of intraoperative NMM increased in both early and late POPC. However, this statement is only valid for desaturation and atelectasis since, in our sample, we have not been able to demonstrate that intraoperative NMM decreased the incidence of postoperative pneumonia and obstruction. The incidence for early POPC was 15.27%, and for late POPC was 8.12%, similar to those reported by Kheterpal et al. or Ledowsky et al.3,29,30. https://doi.org/10.1038/s41598-022-18917-y Scientific Reports \| (2022) 12:14955 \| www.nature.com/scientificreports/ Postoperative pulmonary complication and pharmacological reversal. After analyzing the NMM, we should see what happened with the neuromuscular reversal and its influence on POPC. The controversy31,32 lies with some studies, such as Grosse-Sundrup et al.2, POPULAR14, or Li et al.33, which have reported that reversal, with one reversal agent and another, was not able to decrease these postoperative complications. Moreover, they questioned the utility of quantitative monitoring “the use of reversal agents or neuromuscular monitoring could not decrease this risk.”14.f g We found that the use of rocuronium of sugammadex was associated with a lower risk of suffering early POPC in the PACU, which can be seen in Tables 2 and 3. In addition, when rocuronium and sugammadex were used instead of cisatracurium and neostigmine, the incidence of desaturation decreased by approximately 12% and, in the case of upper airway obstruction, by up to 2%. y y Regarding the late POPC and pharmacological reversal, we also showed that the combination of rocuronium with sugammadex reduced them by up to 7%. In fact, the use of cisatracurium with neostigmine did not appear to decrease but subtly increase the incidence of these complications, which paradoxically aligned with the con- clusions of recent studies2,22,34–36. Limitations. One of our limitations was that detection of late respiratory complications, both pneumonia and atelectasis, was based on clinical and laboratory criteria, and it may underestimate the complication rate. Conclusionh Thus, based on the results of our study, it may be concluded that intraoperative NMM was one of the factors that most effectively reduced the risk of all these postoperative complications. Furthermore, the use of rocuronium with a pharmacological reversal with sugammadex was associated with a lower risk of RNMB and postoperative desaturation in the PACU and atelectasis during hospitalization. and In addition, we reviewed the patient’s clinical history without performing systematic X-ray in all cases since the patients can develop well-tolerated clinical postoperative atelectasis, implying unnecessary radiological exposure to all study patients. As described by Chen et al.37, a more effective and improved method for future research would be systematic examination with pulmonary ultrasound. It currently provides similar results to chest CT and chest X-ray for evaluating pneumonia and atelectasis37,38. y g p Moreover, our results were based on clinical management under real-life conditions. We have not analyzed other factors, such as mechanical ventilation parameters, recruitment maneuvers, opioid doses, fluid therapy, and others that are known to increase these complications and probably need to be assessed in subsequent studies39,40. Data availabilityh y The datasets used and analyzed during the current study are available from the corresponding author on reasonable request. Received: 14 April 2022; Accepted: 22 August 2022 References 2. Grosse-Sundrup, M. et al. Intermediate acting non-depolarizing neuromuscular blocking agents and risk of postoperati respiratory complications: Prospective propensity score matched cohort study. BMJ 345, e6329 (2012). 3. Kheterpal, S. et al. Sugammadex versus neostigmine for reversal of neuromuscular blockade and postoperative pulmonary complications (STRONGER): A multi-center matched cohort analysis. Anesthesiology 132, 1371–1381 (2020).l p ( ) y gy ( ) 4. Kim, N. Y. et al. Influence of reversal of neuromuscular blockade with sugammadex or neostigmine on postoperative quality of recovery following a single bolus dose of rocuronium: A prospective, randomized, double-blinded, controlled study. J. Clin. Anesth. 57, 97–102 (2019). p y gy 4. Kim, N. Y. et al. Influence of reversal of neuromuscular blockade with sugammadex or neostigmine on postoperative quality of recovery following a single bolus dose of rocuronium: A prospective, randomized, double-blinded, controlled study. J. Clin. Anesth. 57, 97–102 (2019). 5. Togioka, B. M. et al. Randomised controlled trial of sugammadex or neostigmine for reversal of neuromuscular block on the incidence of pulmonary complications in older adults undergoing prolonged surgery. Br. J. Anaesth. 124, 553–561 (2020). lk l d l l bl k l d k f h l 5. Togioka, B. M. et al. Randomised controlled trial of sugammadex or neostigmine for reversal of neuromuscular block on the incidence of pulmonary complications in older adults undergoing prolonged surgery. Br. J. Anaesth. 124, 553–561 (2020). 6 B lk C M t l N d l i i l bl ki t l d i k f t ti i A th i l 6. Bulka, C. M. et al. Nondepolarizing neuromuscular blocking agents, reversal, and risk of postoperative pneumonia. Anesthesiology 125, 647–655 (2016). 7. Martinez-Ubieto, J. et al. Prospective study of residual neuromuscular block and postoperative respiratory complications in patients reversed with neostigmine versus sugammadex. Minerva Anestesiol. 82, 735–742 (2016). 7. Martinez-Ubieto, J. et al. Prospective study of residual neuromuscular block and postoperative respiratory complications in patients reversed with neostigmine versus sugammadex. Minerva Anestesiol. 82, 735–742 (2016). b l f l h l ( ) g g 8. Naguib, M. et al. Consensus statement on perioperative use of neuromuscular monitoring. Anesth. Analg. 127, 71–80 (2018). 9 Pl d B l G id li l l d l i h i A h C i C P i M d 39(1) 125 142 (2020 g g 8. Naguib, M. et al. www.nature.com/scientificreports/ et al. Effects of neostigmine reversal of nondepolarizing neuromuscular blocking agents on postoperative respiratory outcomes: A prospective study. Anesthesiology 121, 959–968 (2014). p p y gy 36. Abad-Gurumeta, A. et al. A systematic review of sugammadex vs neostigmine for reversal of neuromuscular blockade. Anaest 70, 1441–1452 (2015).t 36. Abad-Gurumeta, A. et al. A systematic review of sugammadex vs neostigmine for reversal of neuromuscular blockade. Anaesthesia 70, 1441–1452 (2015). 37. Chen, Y., Zhang, Y. G. & Yi, J. Evaluation of postoperative residual curarisation after administration of neostigmine or sugammadex by diaphragmatic ultrasonography: A randomized double-blind controlled trial. Zhongguo Yi Xue Ke Xue Yuan Xue Bao 44(3), 415–421 (2022). 37. Chen, Y., Zhang, Y. G. & Yi, J. Evaluation of postoperative residual curarisation after administration of neostigmine or sugammadex by diaphragmatic ultrasonography: A randomized double-blind controlled trial. Zhongguo Yi Xue Ke Xue Yuan Xue Bao 44(3), 415–421 (2022). ( ) 8. De la Quintana-Gordon, F. B., Nacarino-Alcorta, B. & Fajardo-Pérez, M. Basic lung ultrasound. Part 2. Parenchymal diseases. Rev Esp. Anestesiol. Reanim. 62, 337–349 (2015).l ( ) 38. De la Quintana-Gordon, F. B., Nacarino-Alcorta, B. & Fajardo-Pérez, M. Basic lung ultrasound. Part 2. Parenchymal diseases. Rev. Esp. Anestesiol. Reanim. 62, 337–349 (2015).l p 9. Colomina, M. J. et al. Observational study on fluid therapy management in surgical adult patients. BMC Anesthesiol. 21(1), 316 (2021). ( ) 40. Ferrando, C. et al. Individualised perioperative open-lung approach versus standard protective ventilation in abdominal surgery (iPROVE): A randomised controlled trial. Lancet Respir. Med. 6, 193–203 (2018). 0. Ferrando, C. et al. Individualised perioperative open-lung approach versus standard protective ventilation in abdominal surgery (iPROVE): A randomised controlled trial. Lancet Respir. Med. 6, 193–203 (2018). 40. Ferrando, C. et al. Individualised perioperative open-lung approach versus standard protective ventilation in abdominal surgery (iPROVE): A randomised controlled trial. Lancet Respir. Med. 6, 193–203 (2018). Acknowledgements g Contributing authors from the Research Group in Anaesthesia, Resuscitation, and Perioperative Medicine of Institute for Health Research Aragón (IIS Aragón) for this study are cited above. The authors would also like to thank Dr. Pablo Oliver-Forniés for his invaluable comments and suggestions. Statistical analysis was developed by Jorge Luis Ojeda Cabrera (Dept. of Statistical Methods of University of Zaragoza). References Consensus statement on perioperative use of neuromuscular monitoring. Anesth. Analg. g , p p g g , ( ) 9. Plaud, B. et al. Guidelines on muscle relaxants and reversal in anaesthesia. Anaesth. Crit. Care Pain Med. 39(1), 125–142 (2020). 9. Plaud, B. et al. Guidelines on muscle relaxants and reversal in anaesthesia. Anaesth. Crit. Care Pain Med. 39(1), 125–142 (2020). 9. Plaud, B. et al. Guidelines on muscle relaxants and reversal in anaesthesia. Anaesth. Crit. Care Pain Med. 39(1), 125–142 (2020). ( ) ( ) 0. Checketts, M. R. et al. Recommendations for standards of monitoring during anaesthesia and recovery 2015: Association o Anaesthetists of Great Britain and Ireland. Anaesthesia 71, 85–93 (2016). 11. Stoelting, R. K. Monitoring of neuromuscular blockade: What would you expect if you were the patient?. APSF Newslett. 30, 45–47 (2016).i 11. Stoelting, R. K. Monitoring of neuromuscular blockade: What would you expect if you were the patient?. APSF Newslett. 30, 45–47 (2016).i 2. Hyman, E. C. & Brull, S. J. Clarification: Current status of neuromuscular reversal and monitoring, challenges and opportunities Anesthesiology 127, 730 (2017).h gy 3. Todd, M. M., Hindman, B. J. & King, B. J. The implementation of quantitative electromyographic neuromuscular monitoring in an academic anesthesia department. Anesth. Analg. 119, 323–331 (2014).t 4. Kirmeier, E. et al. Post-anaesthesia pulmonary complications after use of muscle relaxants (POPULAR): A multicentre, prospective observational study. Lancet Respir. Med. 7, 129–140 (2019). 15. Mazo, V. et al. Prospective external validation of a predictive score for postoperative pulmonary complications. Anesthesiology 121, 219–231 (2014). 6. Canet, J. et al. PERISCOPE study: Predicting post-operative pulmonary complications in Europe. Eur. J. Anaesthesiol. 28(6) 459–461 (2011). 17. Kotake, Y. et al. Reversal with sugammadex in the absence of monitoring did not preclude residual neuromuscular block. Anesthesia Analg. 117(2), 345–351 (2013). https://doi.org/10.1038/s41598-022-18917-y Scientific Reports \| (2022) 12:14955 \| www.nature.com/scientificreports/ Keating, G. M. Sugammadex: A review of neuromuscular blo g, g g ( ), ( ) 27. Alday, E. et al. Effects of neuromuscular block reversal with sugammadex versus neostigmine on postoperative respiratory ou after major abdominal surgery: A randomized-controlled trial. Canad. J. Anesth. 66, 1328–1337 (2019). t 28. Naguib, M., Brull, S. J. & Johnson, K. B. Conceptual and technical insights into the basis of neuromuscular monitoring. Anaest 72, 16–37 (2017). 29. Ledowsky, T. et al. Introduction of sugammadex as reversal agent: Impact on the incidence of residual neuromuscular blockade and postoperative patient outcome. Indian J. Anaesth. 57, 46–51 (2013). p p p 0. Ledowski, T. et al. Reversal of residual neuromuscular block with neostigmine or sugammadex and postoperative pulmonary li i A i d i d d bl bli d i l i hi h i k ld i B J A h 127(2) 316 323 (2021) 0. Ledowski, T. et al. Reversal of residual neuromuscular block with neostigmine or sugammadex and postoperative pulmonary complications: A prospective, randomised, double-blind trial in high-risk older patients. Br. J. Anaesth. 127(2), 316–323 (2021). g g p p p complications: A prospective, randomised, double-blind trial in high-risk older patients. Br. J. Anaesth. 127(2), 316–32 p p p g p 31. Unterbuchner, C. Neuromuscular block and blocking agents in 2018. Turk. J. Anaesthesiol. Reanim. 46, 75–80 (2018).t Unterbuchner, C. Neuromuscular block and blocking agents in 20 32. Nemes, R. et al. Impact of reversal strategies on the incidence of postoperative residual paralysis after rocuronium relaxation without neuromuscular monitoring: A partially randomised placebo controlled trial. Eur. J. Anaesthesiol. 34, 609–616 (2017). 33. Li, G. et al. Postoperative pulmonary complications’ association with sugammadex versus neostigmine: A retrospective registry analysis. Anesthesiology 134(6), 862–873 (2021). g p y p ( ) 33. Li, G. et al. Postoperative pulmonary complications’ association with sugammadex versus neostigmine: A retrospective registry analysis. Anesthesiology 134(6), 862–873 (2021). y gy 34. McLean, D. J. et al. Dose-dependent association between intermediate-acting neuromuscular-blocking agents and postoperative respiratory complications. Anesthesiology 122, 1201–1213 (2015).f y gy 34. McLean, D. J. et al. Dose-dependent association between intermediate-acting neuromuscular-blocking agents and postoperative respiratory complications. Anesthesiology 122, 1201–1213 (2015).f p y p gy ( ) 35. Sasaki, N. et al. Effects of neostigmine reversal of nondepolarizing neuromuscular blocking agents on postoperative respira outcomes: A prospective study. Anesthesiology 121, 959–968 (2014). p y p gy ( ) 5. Sasaki, N. Author contributions C.A.: methodology, formal analysis, investigation, writing-original draft preparation; A.P.: conceptualization, methodology, investigation, writing—reviewing and editing, supervision; S.O.: conceptualization, methodology, investigation, writing—reviewing and editing, project administration; S.V.: investigation, visualization, validation; J.M.: conceptualization, methodology, investigation, writing-reviewing and editing, supervision. All authors read and approved the final version of the manuscript. Funding Supported by ERDF (FEDER) Operational Programme of Aragon Health Research Institute (IIS Aragón) No. B26_17D. www.nature.com/scientificreports/ www.nature.com/scientificreports/ 18. Cammu, G. Residual neuromuscular blockade and postoperative pulmonary complications: What does the recent evidence demonstrate?. Curr. Anesthesiol. Rep. 27, 1–6 (2020).hfi p 19. Hristovska, A. M., Duch, P., Allingstrup, M. & Afshari, A. The comparative efficacy and safety of sugammadex and neostigm in reversing neuromuscular blockade in adults. A cochrane systematic review with meta-analysis and trial sequential ana Anaesthesia 73, 631–641 (2018). 19. Hristovska, A. M., Duch, P., Allingstrup, M. & Afshari, A. The comparative efficacy and safety of sugammadex and neostigmine in reversing neuromuscular blockade in adults. A cochrane systematic review with meta-analysis and trial sequential analysis. Anaesthesia 73, 631–641 (2018). 20 Errando C L Garutti I Mazzinari G Díaz Cambronero Ó & Bebawy J F Residual neuromuscular blockade in the 20. Errando, C. L., Garutti, I., Mazzinari, G., Díaz-Cambronero, Ó. & Bebawy, J. F. Residual neuromuscular blockade in postanesthesia care unit: Observational cross-sectional study of a multicenter cohort. Minerva Anestesiol. 82, 1267–1277 (2f 20. Errando, C. L., Garutti, I., Mazzinari, G., Díaz-Cambronero, Ó. & Bebawy, J. F. Residual neuromuscular blockade in the postanesthesia care unit: Observational cross-sectional study of a multicenter cohort. Minerva Anestesiol. 82, 1267–1277 (2016). 21. Brueckmann, B. et al. Effects of sugammadex on incidence of postoperative residual neuromuscular blockade: A randomized, controlled study. Br. J. Anaesth. 115, 743–751 (2015). p y ( ) 21. Brueckmann, B. et al. Effects of sugammadex on incidence of postoperative residual neuromuscular blockade: A randomized, controlled study. Br. J. Anaesth. 115, 743–751 (2015). y 2. Fuchs-Buder, T., Nemes, R. & Schmartz, D. Residual neuromuscular blockade: Management and impact on postoperative pulmonary outcome. Curr. Opin. Anaesthesiol. 29, 662–667 (2016). p y p ( ) 3. Merck & Co. Bridion® (sugammadex) injection, for intravenous use: US prescribing information (2015) https://www.accessdata fda.gov/drugsatfda_docs/label/2015/022225lbl.pdf (Accessed 14 Apr 2022). g j fda.gov/drugsatfda_docs/label/2015/022225lbl.pdf (Accessed 14 g g p ( p ) 4. European Medicines Agency. Bridion® (sugammadex) 100 mg/ml solution for injection: EU summary of product characteristics (2015) https://www.ema.europa.eu/en/medicines/human/EPAR/bridion (Accessed 14 Apr 2022). 24. European Medicines Agency. Bridion® (sugammadex) 100 mg/ml solution for injection: EU summ (2015) https://www.ema.europa.eu/en/medicines/human/EPAR/bridion (Accessed 14 Apr 2022). p p p 25. De Boer, H. D., Carlos, R. V. & Brull, S. J. Is lower-dose sugammadex a cost-saving strategy for reversal of deep neurom block? Facts and fiction. BMC Anesthesiol. 18, 159 (2018). i 26. Keating, G. M. Sugammadex: A review of neuromuscular blockade reversal. Drugs 76(10), 1041–1052 (2016).f i 26. Funding unding upported by ERDF (FEDER) Operational Programme of Aragon Health Research Institute (IIS Aragón) No 26_17D. Competing interests h dd t o a o at o Correspondence and requests for materials should be addressed to C.A.-B. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. https://doi.org/10.1038/s41598-022-18917-y Scientific Reports \| (2022) 12:14955 \| www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2022 © The Author(s) 2022 Research Group in Anaesthesia, Resuscitation, and Perioperative Medicine of Institute for Health Research Aragón (ISS Aragón) Research Group in Anaesthesia, Resuscitation, and Perioperative Medicine of Institute for Health Research Aragón (ISS Aragón) Cristian Aragón‑Benedí1, Ana Pascual‑Bellosta1, Sonia Ortega‑Lucea1, Javier Martínez‑Ubieto1, Luis Alfonso Muñoz‑Rodríguez1, Guillermo Pérez‑Navarro1, Natividad Quesada‑Gimeno1, Lucía Tardós‑Ascaso1, Sara Visiedo‑Sánchez2, Teresa Jiménez‑Bernadó3, Berta Pérez‑Otal3 & Francisco Romero‑Caro3 Cristian Aragón‑Benedí1, Ana Pascual‑Bellosta1, Sonia Ortega‑Lucea1, Javier Martínez‑Ubieto1, Luis Alfonso Muñoz‑Rodríguez1, Guillermo Pérez‑Navarro1, Natividad Quesada‑Gimeno1, Lucía Tardós‑Ascaso1, Sara Visiedo‑Sánchez2, Teresa Jiménez‑Bernadó3, Berta Pérez‑Otal3 & Francisco Romero‑Caro3 https://doi.org/10.1038/s41598-022-18917-y Scientific Reports \| (2022) 12:14955 \|
https://openalex.org/W3013752240	https://europepmc.org/articles/pmc7180861?pdf=render	English	null	Bacillus sp. FPF-1 Produced Keratinase with High Potential for Chicken Feather Degradation	Molecules/Molecules online/Molecules annual	2,020	cc-by	11,794	Received: 3 February 2020; Accepted: 27 February 2020; Published: 26 March 2020 Abstract: Chicken feathers are predominantly composed of keratin; hence, valorizing the wastes becomes an imperative. In view of this, we isolated keratinase-producing bacteria and identiﬁed them through the 16S rDNA sequence. The process condition for keratinase activity was optimized, and electron micrography of the degradation timelines was determined. Keratinolytic bacteria were isolated and identiﬁed as Bacillus sp. FPF-1, Chryseobacterium sp. FPF-8, Brevibacillus sp. Nnolim-K2, Brevibacillus sp. FPF-12 and Brevibacillus sp. FSS-1; and their respective nucleotide sequences were deposited in GenBank, with the accession numbers MG214993, MG214994, MG214995, MG214996 and MG214999. The degree of feather degradation and keratinase concentration among the isolates ranged from 62.5 ± 2.12 to 86.0 ± 1.41(%) and 214.55 ± 5.14 to 440.01 ± 20.57 (U/mL), respectively. In the same vein, 0.1% (w/v) xylose, 0.5% (w/v) chicken feather, an initial fermentation pH of 5.0, fermentation temperature of 25 ◦C and an agitation speed of 150 rpm, respectively, served as the optimal physicochemical conditions for keratinase activity by Bacillus sp. FPF-1. The time course showed that Bacillus sp. FPF-1 yielded a keratinase concentration of 1698.18 ± 53.99(U/mL) at 120 h. The electron microscopic imaging showed completely structural dismemberment of intact chicken feather. Bacillus sp. FPF-1 holds great potential in the valorization of recalcitrant keratinous biomass from the agro sector into useful products. Keywords: biodegradation; chicken feathers; keratinase; valorization Keywords: biodegradation; chicken feathers; keratinase; valorization Molecules 2020, 25, 1505; doi:10.3390/molecules25071505 molecules molecules molecules molecules molecules Bacillus sp. FPF-1 Produced Keratinase with High Potential for Chicken Feather Degradation Nonso E. Nnolim 1,2,* , Anthony I. Okoh 1,2 and Uchechukwu U. Nwodo 1,2 1 SA-MRC Microbial Water Quality Monitoring Centre, University of Fort Hare, Alice, 5700 Eastern Cape, South Africa; AOkoh@ufh.ac.za (A.I.O.); UNwodo@ufh.ac.za (U.U.N.) 2 Applied and Environmental Microbiology Research Group (AEMREG), Department of Biochemistry and Microbiology, University of Fort Hare, Private Bag X1314, Alice, 5700 Eastern Cape, South Africa * Correspondence: nnnolim@ufh.ac.za; Tel.: +27-406-022-693 1 SA-MRC Microbial Water Quality Monitoring Centre, University of Fort Hare, Alice, 5700 Eastern Cape, South Africa; AOkoh@ufh.ac.za (A.I.O.); UNwodo@ufh.ac.za (U.U.N.) 2 Applied and Environmental Microbiology Research Group (AEMREG), Department of Biochemistry and Microbiology, University of Fort Hare, Private Bag X1314, Alice, 5700 Eastern Cape, South Africa * Correspondence: nnnolim@ufh.ac.za; Tel.: +27-406-022-693 1. Introduction Chicken feathers are a signiﬁcant waste product from poultry processing farms, and the disposal process has remained incineration, commuting to landﬁlls and limited use as non-nutritive additives of feed. The composition of chicken feathers includes keratin, which bestows the mechanical stability associated with chicken feathers, leading to the recalcitrance to degradation [1]. The estimated abundance of crude protein in the avian feather has been approximated to about 82.36% [2], and this signiﬁcant protein content portends an excellent opportunity for the valorization of the agro-waste biomass to high-value, cost-eﬀective and renewable-source proteins for various applications, including livestock-feed supplementation. Notwithstanding the potentials presented by avian feathers as a high-protein source, the ability to dismember the feathers into functional components becomes the key for the valorization of the agro-waste. The traditional approach for the valorization of feathers involves endergonic reactions and chemical treatments [3]; these methods are capital intensive, and they produce a poorly digestible protein which is unsatisfactory as a feed supplement. Besides, the application of the methods also incorporates other compounds with no known biological importance [4]. Molecules 2020, 25, 1505; doi:10.3390/molecules25071505 www.mdpi.com/journal/molecules Molecules 2020, 25, 1505 2 of 16 Conversely, the use of the microbial-based technique for the dismemberment of the complex structures constituting the avian feather for high-value products has gained considerable traction. The bio-based approach has been adjudged as eco-friendly, energy-conserving and the protein yields are of high quality [5]. Consequently, the biodegradation of feather, therefore, represents a cost-eﬀective and renewable approach for the valorization of agro-waste to high-value products. An account of bacterial and fungal species showing the potentials for the degradation of avian feathers abound. However, most of the microbes with excellent keratinolytic properties have been reported to exist among the dermatophytes, and this creates limitations for useful application [5,6]. Keratinolytic bacteria, on the other hand, present a litany of advantage over the fungal counterpart, including a high rate of extracellular proteolytic enzyme production [7]. The bio-catalytic eﬃciency of the keratinases produced by bacteria species tends toward a broad spectrum of proteinaceous substrates and robust at extreme conditions [8]. Considering the promising applicability of keratinase in green technology, it became imperative to explore some environment for bacterial diversity with the potentials to produce novel keratinases [6]. Nonetheless, the footprint of keratinases in the market is relatively low, and this further indicates that an opportunity abounds in the endeavor of obtaining keratinases with novel function. 1. Introduction Therefore, this study was undertaken to assess the keratinolytic potentials of some bacteria isolated from poultry-composting sites and to optimize the process conditions for keratinolytic activity. 2.1. Identiﬁcation of the Keratinase-Producing Bacteria The 16S rDNA nucleotide sequences BLAST showed the isolate coded as FPF-1 to have a high sequence homology (100%) with Bacillus cereus AB1 (MF800922) and Bacillus thuringiensis WG20 (KY085971); hence, the isolate was identiﬁed as Bacillus sp. FPF-1, with an accession number MG214993. The other isolates coded as FPF-10, FPF-12 and FSS-1 showed 99% sequence homology with Brevibacillus parabrevis C20 (KX832699), Brevibacillus parabrevis NAP3 (KJ872854) and Brevibacillus sp. BAB-6437 (KY672924), respectively. Consequently, they were identiﬁed as Brevibacillus sp. Nnolim-K2, Brevibacillus sp. FPF-12 and Brevibacillus sp. FSS-1. The nucleotide sequences of the isolates were deposited in GenBank, with the respective accession numbers MG214995, MG214996 and MG214999, as shown in Table 1. Lastly, the isolate coded as FPF-8 showed 98% sequence homology with Chryseobacterium culicis R4-1A (NR_117008); therefore, it was identiﬁed as Chryseobacterium sp. FPF-8, and the nucleotide sequence was deposited in GenBank, with accession number MG214994. Table 1. Identiﬁcation of keratinolytic bacterial isolates, using 16S rDNA sequence. S/N Isolate Code Reference Sequence Sequence Similarity (%) Sequence Identity NCBI Accession Number 1. FPF-1 Bacillus cereus AB1 (MF800922) Bacillus thuringiensis WG20 (KY085971) 100 Bacillus sp. FPF-1 MG214993 2. FPF-8 Chryseobacterium culicis R4-1A (NR_117008) 98 Chryseobacterium sp. FPF-8 MG214994 3. FPF-10 Brevibacillus parabrevis C20 (KX832699) 99 Brevibacillus sp. Nnolim-K2 MG214995 4. FPF-12 Brevibacillus parabrevis NAP3 (KJ872854) 99 Brevibacillus sp. FPF-12 MG214996 5. FSS-1 Brevibacillus sp. BAB-6437 (KY672924) 99 Brevibacillus sp. FSS-1 MG214999 able 1. Identiﬁcation of keratinolytic bacterial isolates, using 16S rDNA sequence. 2.2. Evaluation for Keratinolytic Activity The hydrolysis of casein in skimmed milk (halo zones formation on skimmed milk agar) was an indication that the isolates produced extracellular proteases. Protease activity is relevant and 3 of 16 was an Molecules 2020, 25, 1505 f The hydrolysis a signiﬁcant indication of the presences of keratinases. The screening for protease activity as a potential indicator for keratinase activity is a standard approach for screening and selection of potential keratinolytic bacteria [9]. The isolates coded as FPF-1, FPF-8, FPF-10, FPF-12 and FSS-1 showed signiﬁcant feather degradation in basal medium with chicken feathers as a sole source of carbon and nitrogen (Figure 1). Consequently, these isolates were used in subsequent studies. The ﬂask containing isolate coded FPF-8 showed orange-colored broth, which is characteristic of ﬂexirubin-producing Chryseobacterium species [10]. significant indication of the presences of keratinases. The screening for protease activity as a potential indicator for keratinase activity is a standard approach for screening and selection of potential keratinolytic bacteria [9]. The isolates coded as FPF-1, FPF-8, FPF-10, FPF-12 and FSS-1 showed significant feather degradation in basal medium with chicken feathers as a sole source of carbon and nitrogen (Figure 1). Consequently, these isolates were used in subsequent studies. The flask containing isolate coded FPF-8 showed orange-colored broth, which is characteristic of flexirubin- producing Chryseobacterium species [10]. Figure 1. Biodegradation of intact chicken feathers by the keratinolytic bacterial isolates: (a) before incubation; and (b) after 96 h of incubation Figure 1. Biodegradation of intact chicken feathers by the keratinolytic bacterial isolates: (A) before incubation; and (B) after 96 h of incubation. Figure 1. Biodegradation of intact chicken feathers by the keratinolytic bacterial isolates: (a) before incubation; and (b) after 96 h of incubation Figure 1. Biodegradation of intact chicken feathers by the keratinolytic bacterial isolates: (A) before incubation; and (B) after 96 h of incubation. The residual feathers were harvested and quantified by using the formula in Equation (1), at the end of fermentation duration, and the degree of degradation was 86.0 ± 1.41 (%) against Bacillus sp. FPF-1, 82.0 ± 1.41 (%) against Chryseobacterium sp. FPF-8, 66.0 ± 2.83 (%) against Brevibacillus sp. Nnolim-K2, 62.5 ± 2.12 (%) against Brevibacillus sp. FPF-12 and 69.5 ± 2.12 (%) against Brevibacillus sp. FSS-1 (Table 2). The pH of the cell-free broth, the concentrations of keratinase, total protein and thiol of the same cell-free broth were analyzed. 2.2. Evaluation for Keratinolytic Activity The extracellular keratinase activity by the isolates ranged from 214.55 ± 5.14 (U/mL) against Brevibacillus sp. Nnolim-K2 to 440.01 ± 20.57 (U/mL) against Bacillus sp. FPF-1. Perhaps it would be prudent to indicate, at this point, that keratinases mediated the degradation of chicken feathers, as ample quantities of the enzymes were present in the cell-free broth. The bacterial isolates thrived on the keratinous biomass as the sole carbon and nitrogen source. In no small degree of certainty, it would be safe to indicate that the isolates produced keratinases [11]. Keratin-degrading efficiency of the test bacteria may be a reflection of the effectiveness of the t ll l k ti d d d th b t i l i l t d i ti ti h d i t The residual feathers were harvested and quantiﬁed by using the formula in Equation (1), at the end of fermentation duration, and the degree of degradation was 86.0 ± 1.41 (%) against Bacillus sp. FPF-1, 82.0 ± 1.41 (%) against Chryseobacterium sp. FPF-8, 66.0 ± 2.83 (%) against Brevibacillus sp. Nnolim-K2, 62.5 ± 2.12 (%) against Brevibacillus sp. FPF-12 and 69.5 ± 2.12 (%) against Brevibacillus sp. FSS-1 (Table 2). The pH of the cell-free broth, the concentrations of keratinase, total protein and thiol of the same cell-free broth were analyzed. The extracellular keratinase activity by the isolates ranged from 214.55 ± 5.14 (U/mL) against Brevibacillus sp. Nnolim-K2 to 440.01 ± 20.57 (U/mL) against Bacillus sp. FPF-1. Perhaps it would be prudent to indicate, at this point, that keratinases mediated the degradation of chicken feathers, as ample quantities of the enzymes were present in the cell-free broth. The bacterial isolates thrived on the keratinous biomass as the sole carbon and nitrogen source. In no small degree of certainty, it would be safe to indicate that the isolates produced keratinases [11]. Keratin-degrading eﬃciency of the test bacteria may be a reﬂection of the eﬀectiveness of the extracellular keratinase produced; and the bacterial isolates under investigation showed consistency in that regard [12]. extracellular keratinase produced; and the bacterial isolates under investigation showed consistency in that regard [12]. The concentration of thiols in the fermentation broth ranged from 461.69 ± 5.53 (μM) against Brevibacillus sp. FPF-12 to 2206.59 ± 49.79 (μM) against Bacillus sp. FPF-1. 2.2. Evaluation for Keratinolytic Activity Disulfide bond reduction was identified to promote complete decomposition of keratin through structural modification significantly; and such structural alteration increases the vulnerability of keratin to proteolytic p g y g The concentration of thiols in the fermentation broth ranged from 461.69 ± 5.53 (µM) against Brevibacillus sp. FPF-12 to 2206.59 ± 49.79 (µM) against Bacillus sp. FPF-1. Disulﬁde bond reduction was identiﬁed to promote complete decomposition of keratin through structural modiﬁcation signiﬁcantly; and such structural alteration increases the vulnerability of keratin to proteolytic hydrolysis [13]. The pH of the fermented broth drifted from slightly acidic (initial fermentation pH) to alkaline; 7.35 ± 0.00 against Brevibacillus sp. FPF-12 to 8.53 ± 0.02 against Chryseobacterium sp. FPF-8 (Table 2). Biodegradation of keratinous biomass has been reported to instigate ammoniﬁcation of the fermentation medium [14], which causes a drift in the medium pH toward the alkaline spectrum. extracellular keratinase produced; and the bacterial isolates under investigation showed consistency in that regard [12]. The concentration of thiols in the fermentation broth ranged from 461.69 ± 5.53 (μM) against Brevibacillus sp. FPF-12 to 2206.59 ± 49.79 (μM) against Bacillus sp. FPF-1. Disulfide bond reduction was identified to promote complete decomposition of keratin through structural modification significantly; and such structural alteration increases the vulnerability of keratin to proteolytic The concentration of thiols in the fermentation broth ranged from 461.69 ± 5.53 (µM) against Brevibacillus sp. FPF-12 to 2206.59 ± 49.79 (µM) against Bacillus sp. FPF-1. Disulﬁde bond reduction was identiﬁed to promote complete decomposition of keratin through structural modiﬁcation signiﬁcantly; and such structural alteration increases the vulnerability of keratin to proteolytic hydrolysis [13]. The pH of the fermented broth drifted from slightly acidic (initial fermentation pH) to alkaline; 7.35 ± 0.00 against Brevibacillus sp. FPF-12 to 8.53 ± 0.02 against Chryseobacterium sp. FPF-8 (Table 2). Biodegradation of keratinous biomass has been reported to instigate ammoniﬁcation of the fermentation medium [14], which causes a drift in the medium pH toward the alkaline spectrum. 4 of 16 Molecules 2020, 25, 1505 Table 2. Keratinolytic activities of the bacterial fermentation from chicken feathers. S/N Isolate Keratinase Activity (U/mL) Protein Concentration (µg/mL) Thiol Concentration (µM) Final pH Feather Weight Loss (%) 1. Bacillus sp. FPF-1 440.01 ± 20.57 c 759.97 ± 22.29 d 2206.59 ± 49.79 d 8.09 ± 0.01 d 86.0 ± 1.41 c 2. Chryseobacterium sp. 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FP 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FPF-1 Upon evaluating the inﬂuence of initial fermentation pH on keratinase activity, respectively, fermentation occurred at pH 4.0 to 10.0, and the observation was that Bacillus sp. FPF-1 produced keratinase at all the pH for which fermentation was initiated. However, Bacillus sp. FPF-1 optimally produced keratinase at weak acidic pH 5.0. Conversely, the total protein concentration in the broth increased as the pH tended toward alkaline and was markedly high with the initial formation pH 10.0 (Figure 2). The reason behind the observed phenomenon is not apparent. However, a logical interpretation may be that the test bacteria produced other enzymes, which cleaved the chicken feather into protein subunits, diﬀerent from keratinases. A few studies have reported an initial ambient pH from weak acidic to neutral [16,17], and this observation is in accord with our ﬁndings. Molecules 2020, 25, x FOR PEER REVIEW 5 of 17 Figure 2. Effect of initial cultivation pH on keratinase activity by Bacillus sp. FPF-1. 0 100 200 300 400 500 600 700 800 900 0 200 400 600 800 1000 1200 1400 1600 4 5 6 7 8 9 10 Protein concentration (µg/mL) Keratinase activity (U/mL) Initial medium pH Keratinase activity Protein concentration Figure 2. Eﬀect of initial cultivation pH on keratinase activity by Bacillus sp. FPF-1. Figure 2. Effect of initial cultivation pH on keratinase activity by Bacillus sp. FPF-1. Figure 2. Eﬀect of initial cultivation pH on keratinase activity by Bacillus sp. FPF-1. In consideration of the effect of carbon sources, other than the chicken feather, on the production of keratinases by Bacillus sp. FPF-1, xylose showed a considerable influence with keratinase yield at 1155.45 ± 52.71 (U/mL) (Figure 3). Subsequently, xylose concentration was varied from 0.08% to 1.0% ( / ) d 0 1% h i i i i ld d k i i f 1096 36 With Bacillus species in perspective, keratinase activity by some has been reported to occur between neutral and alkaline pH condition. A good example would be B. pumilus GRK, which actively produced keratinase between pH 8.0 and 12.0, achieving the maximum keratinase activity at optimum In consideration of the effect of carbon sources, other than the chicken feather, on the production of keratinases by Bacillus sp. FPF-1, xylose showed a considerable influence with keratinase yield at 1155.45 ± 52.71 (U/mL) (Figure 3). 2.2. Evaluation for Keratinolytic Activity FPF-8 260.0 ± 25.71 b 279.35 ± 24.51 b 1028.98 ± 88.53 c 8.53 ± 0.02 e 82.0 ± 1.41 c 3. Brevibacillus sp. Nnolim-K2 214.55 ± 5.14 a 160.38 ± 38.99 a 465.59 ± 11.07 a 7.43 ± 0.02 b 66.0 ± 2.83 ab 4. Brevibacillus sp. FPF-12 233.18 ± 12.21ab 287.23 ± 20.06 b 461.69 ± 5.53 a 7.35 ± 0.00 a 62.5 ± 2.12 a 5. Brevibacillus sp. FSS-1 252.73 ± 5.14 ab 639.42 ± 34.54 c 755.11 ± 0.00 b 7.51 ± 0.03 c 69.5 ± 2.12 b The values are presented as the mean and standard deviation of triplicate experiments; the values without the same superscript letters down the column are signiﬁcantly diﬀerent (P < 0.05). Table 2. Keratinolytic activities of the bacterial fermentation from chicken feathers. Consequently, the process of alkalization of the medium causes substrate swelling, which ultimately weakens the mechanical stability of keratin, hence, yielding to more eﬃcient keratinolytic attack [15]—considering the keratinolytic activities of the bacterial isolates, Bacillus sp. FPF-1 displayed unusual keratinolytic activity; therefore, it was selected for the rest of the studies reported afterward. 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FP Subsequently, xylose concentration was varied from 0.08% to 1.0% ( / ) d 0 1% th ti t ti it i ld d k ti t ti f 1096 36 ± With Bacillus species in perspective, keratinase activity by some has been reported to occur between neutral and alkaline pH condition. A good example would be B. pumilus GRK, which actively produced keratinase between pH 8.0 and 12.0, achieving the maximum keratinase activity at optimum 5 of 16 Molecules 2020, 25, 1505 initial pH 10.0 [18]. Bacillus licheniformis ER-15 similarly produced keratinase within the pH range of 7.0 to 10.0, and the optimum initial fermentation pH was 7.0 [14]. Reports abound on the alkalophilic nature of keratinase-producing Bacillus species. However, the acidophilic nature of Bacillus sp. FPF-1, which is the focus of the study, might be an embodiment of unique biotechnological potentials this bacterial strain might hold in the valorization value chain. In consideration of the eﬀect of carbon sources, other than the chicken feather, on the production of keratinases by Bacillus sp. FPF-1, xylose showed a considerable inﬂuence with keratinase yield at 1155.45 ± 52.71 (U/mL) (Figure 3). Subsequently, xylose concentration was varied from 0.08% to 1.0% (w/v), and 0.1% was the optimum concentration, as it yielded a keratinase concentration of 1096.36 ± 120.85 (U/mL) (Figure 4). Beyond 0.1%, the concentration of xylose in the medium the eﬀect was not positive, as the keratinase concentration in the broth did not increase. The mechanism for which xylose in the fermentation medium inﬂuenced/enhanced keratinase activity is unclear; however, it may have upregulated the expression of keratinase gene in Bacillus sp. FPF-1. In a similar study, 0.5% xylose was used as the fermentation medium, and Bacillus amyloliquefaciens B6 achieved a higher concentration of keratinase [19]. Molecules 2020, 25, x FOR PEER REVIEW 6 of 17 Molecules 2020 25 x FOR PEER REVIEW 6 of 17 Figure 3. The influence of saccharide supplementation of medium for enhanced keratinase activity by Bacillus sp. FPF-1. Figure 4. The effect of xylose concentrations on keratinase activity by Bacillus sp. FPF-1. 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FP The supplementation of the fermentation medium with organic and inorganic nitrogen sources 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 Protein concentration (μg/mL) Keratinase activity (U/mL) Supplemented carbon source (0.1%, w/v) Keratinase activity Protein concentration 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 0.08 0.09 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Protein concentration (μg/mL) Keratinase activity (U/mL) Xylose concentration (%, w/v) Keratinase activity Protein concentration Figure 3. The inﬂuence of saccharide supplementation of medium for enhanced keratinase activity by Bacillus sp. FPF-1. , , Figure 3. The influence of saccharide supplementation of medium for enhanced keratinase activity by Bacillus sp. FPF-1. Figure 4. The effect of xylose concentrations on keratinase activity by Bacillus sp. FPF-1. The supple e tatio of the fe e tatio ediu ith o ga ic a d i o ga ic it oge sou ces 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 Protein concentration (μg/mL) Keratinase activity (U/mL) Supplemented carbon source (0.1%, w/v) Keratinase activity Protein concentration 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 0.08 0.09 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Protein concentration (μg/mL) Keratinase activity (U/mL) Xylose concentration (%, w/v) Keratinase activity Protein concentration Figure 4. The eﬀect of xylose concentrations on keratinase activity by Bacillus sp. FPF-1. Figure 3. The influence of saccharide supplementation of medium for enhanced keratinase activity by Bacillus sp. FPF-1. 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 Protein concentration (μg/mL) Keratinase activity (U/mL) Supplemented carbon source (0.1%, w/v) Keratinase activity Protein concentration Figure 3. The inﬂuence of saccharide supplementation of medium for enhanced keratinase activity by Bacillus sp. FPF-1. Figure 3. The influence of saccharide supplementation of medium for enhanced keratinase activity by Bacillus sp. FPF-1. 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 Protein concentration (μg/mL) Keratinase activity (U/mL) Supplemented carbon source (0.1%, w/v) Keratinase activity Protein concentration Figure 3. The influence of saccharide supplementation of medium for enhanced keratinase activity by Bacillus sp. FPF-1. 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FP 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 Protein concentration (μg/mL) Keratinase activity (U/mL) Supplemented carbon source (0.1%, w/v) Keratinase activity Protein concentration Figure 3. The inﬂuence of saccharide supplementation of medium for enhanced keratinase activity by Bacillus sp. FPF-1. Figure 3. The influence of saccharide supplementation of medium for enhanced keratinase activity by Bacillus sp. FPF-1. 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 Protein concentration (μg/mL) Keratinase activity (U/mL) Supplemented carbon source (0.1%, w/v) Keratinase activity Protein concentration Figure 3. The influence of saccharide supplementation of medium for enhanced keratinase activity by Bacillus sp. FPF-1. Figure 3. The inﬂuence of saccharide supplementation of medium for enhanced keratinase activity by Bacillus sp. FPF-1. Figure 3. The influence of saccharide supplementation of medium for enhanced keratinase activity by Bacillus sp. FPF-1. Figure 4. The effect of xylose concentrations on keratinase activity by Bacillus sp. FPF-1. Th l i f h f i di i h i d i i i 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 0.08 0.09 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Protein concentration (μg/mL) Keratinase activity (U/mL) Xylose concentration (%, w/v) Keratinase activity Protein concentration Figure 4. The effect of xylose concentrations on keratinase activity by Bacillus sp. FPF-1. 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 0.08 0.09 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Protein concentration (μg/mL) Keratinase activity (U/mL) Xylose concentration (%, w/v) Keratinase activity Protein concentration Figure 4. The eﬀect of xylose concentrations on keratinase activity by Bacillus sp. FPF-1. Figure 4. The effect of xylose concentrations on keratinase activity by Bacillus sp. FPF-1. Figure 4. The effect of xylose concentrations on keratinase activity by Bacillus sp. FPF-1. Figure 4. The eﬀect of xylose concentrations on keratinase activity by Bacillus sp. FPF-1. Molecules 2020, 25, 1505 6 of 16 Other carbon sources showed a diﬀerent pattern of inﬂuence on keratinase activity by Bacillus sp. FPF-1. Glucose decreased keratinase activity, as the yield was at 367.27 ± 53.99 (U/mL) as compared to the control with yield of 793.64 ± 42.43 (U/mL). 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FP Perhaps it would be prudent to indicate that catabolite repression may have been responsible for the phenomenon observed, as it is a known factor with the utilization of simple sugars as against a complex substrate [20,21]. Keratinase activity with the addition of the fructose, sucrose, and galactose, respectively were 889.99 ± 26.99, 935.45 ± 42.43 and 923.63 ± 10.28 (U/mL). Keratinase yield with these carbon sources were slightly higher than the control (P > 0.05); hence, they may serve as good inducers for keratinase activity (Figure 3). Mannitol, maltose and lactose neither signiﬁcantly inﬂuence keratinase activity nor repressed the production. The supplementation of the fermentation medium with organic and inorganic nitrogen sources showed that Bacillus sp. FPF-1 produced lower quantities of keratinase across all the spectrum of supplemented nitrogen sources (Figure 5). Bacillus sp. FPF-1 would have preferentially utilized these nitrogen sources against the degradation of the keratinous biomass, and this concept may have accounted for the lower quantity of keratinases measured in the fermentation broth. Molecules 2020, 25, x FOR PEER REVIEW 7 of 17 B. licheniformis MZK-3 were influenced to produce keratinases in the presence of additional nitrogenous source [22,23]. However, Bacillus sp. FPF-1 produced keratinases at a higher concentration in the absence of a nitrogen source other than chicken feathers, and an industrial prospect abound for the bacterial strain. Figure 5. The effect of nitrogen supplementation on keratinase activity by Bacillus sp. FPF-1. 0 200 400 600 800 1000 1200 1400 1600 0 200 400 600 800 1000 1200 1400 1600 Protein concentration (μg/mL) Keratinase activity (U/mL) Supplemented nitrogen source (0.02%, w/v) Keratinase activity Protein concentration Figure 5. The eﬀect of nitrogen supplementation on keratinase activity by Bacillus sp. FPF-1. Figure 5. The effect of nitrogen supplementation on keratinase activity by Bacillus sp. FPF-1. Figure 5. The eﬀect of nitrogen supplementation on keratinase activity by Bacillus sp. FPF-1. The chicken feather concentration that was optimum for keratinase activity was at 0.5% (w/v) in the fermentation medium containing a 2% (v/v) starter culture with cell concentration of about 1 × 108 CFU/mL (Figure 6). A further increase in the concentration of the chicken feather beyond 0.5% resulted in a constant decline in keratinase activity by Bacillus sp. FPF-1. 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FP However, the converse was the case with the total protein content measured in the fermentation broth; as the concentration of keratinase declined from 790.91 ± 10.29 to 714.55 ± 77.14 to 582.73 ± 26.99 (U/mL) with the respective chicken feather concentrations 0.75%, 1.0% and 1.25% (w/v), the total protein concentration in the fermentation medium increased as follows; 277.77 ± 2.23, 385.72 ± 10.03 and 408.57 ± 2.23 (µg/mL), respectively. The presence of a high concentration of total protein in the fermentation broth may be understood, as high potent keratinases breaking down the keratinous biomass faster than the producing bacteria is able to utilize, hence, the high concentration. Consequently, if the needed it il bl f i il ti th t ll th i ld d l t Again, the total protein concentration in the fermentation broth was considerably higher with the ammonium salts; (NH4)2HPO4, (NH4)H2PO4, NH4Cl, NH4NO3 and (NH4)2SO4 yielded 717.42 ± 6.69, 746.57±1.11, 779.66 ± 25.63, 596.08 ± 191.6 and 736.33 ± 6.69 (µg/mL), respectively (Figure 5). It may be inferred that these ammonium salts inﬂuenced the production of some sorts of proteases, which are not keratinases. Consequently, the presence of more proteins in the fermentation supplemented with the salts compared to the control. These proteases may have acted upon cleaved units from the complex chicken feather structure. Complex nitrogen sources, including yeast extract, showed negligible eﬄuence, which was statistically insigniﬁcant (P > 0.05). Xanthomonas sp. P5 and B. licheniformis MZK-3 were inﬂuenced to produce keratinases in the presence of additional nitrogenous source [22,23]. However, Bacillus sp. FPF-1 produced keratinases at a higher concentration in the absence of a nitrogen source other than chicken feathers, and an industrial prospect abound for the bacterial strain. nitrogen source were available for assimilation, then naturally the organisms would downregulate the keratinase-expressing gene. Perhaps other enzymes that are different from keratinases are responsible for the cleavage of some peptides or proteins to forms assimilable by Bacillus sp. FPF-1. A similar phenomenon as described above has been reported previously. However, the reason advanced for the process was a reduction in dissolved oxygen concentration as a result of high viscosity of basal medium, hence, limiting the optimal activity of the microbial species [19,24]. 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FP A similar pattern was observed with the total protein concentration of the fermentation broth; at 25 °C (1189.37 ± 23.39 µg/mL), 30 °C (617.36 ± 5.57 µg/mL), 35 °C (242.32 ± 16.71 µg/mL) and 40 °C (124.92 ± 2.29 µg/mL). The findings only buttress the supposition that keratinases have a high substrate-to-product conversion rate and that other enzymes, which may be heat labile, are also in play in the conversion of polymers to proteins. The presence of a high concentration of total protein in the fermentation broth may be understood, as high potent keratinases breaking down the keratinous biomass faster than the producing bacteria is able to utilize, hence, the high concentration. Consequently, if the needed nitrogen source were available for assimilation, then naturally the organisms would downregulate the keratinase-expressing gene. Perhaps other enzymes that are diﬀerent from keratinases are responsible for the cleavage of some peptides or proteins to forms assimilable by Bacillus sp. FPF-1. y y p y p y p However, organisms are unique and so are their characteristics; some keratinase-producing bacteria have functioned well at temperatures ranging from 25 to 40 °C [17,25,26]. Thermophilic keratinase-producing bacteria with optimum temperatures ranging from 50 to 70 °C have been reported [27–29]. However, Bacillus sp. FPF-1 may be graded as mesophilic, and the biotechnological significance would include a low energy input in a large-scale fermentation. A similar phenomenon as described above has been reported previously. However, the reason advanced for the process was a reduction in dissolved oxygen concentration as a result of high viscosity of basal medium, hence, limiting the optimal activity of the microbial species [19,24]. Although the pattern of results was similar, we submit that a more logical explanation would lie on the simple concept of enzyme–substrate–gene regulation. 1000 1200 1400 1600 1800 2000 1000 1200 1400 1600 1800 2000 2200 entration (μg/mL) activity (U/mL) The variation of the fermentation temperature showed that Bacillus sp. FPF-1 optimally produced keratinase at 25 ◦C (1918.18 ± 64.28 U/mL). As the fermentation temperature increased, keratinase activity decreased consistently (Figure 7). A similar pattern was observed with the total protein concentration of the fermentation broth; at 25 ◦C (1189.37 ± 23.39 µg/mL), 30 ◦C (617.36 ± 5.57 µg/mL), 35 ◦C (242.32 ± 16.71 µg/mL) and 40 ◦C (124.92 ± 2.29 µg/mL). 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FP Although the pattern of results was similar, we submit that a more logical explanation would lie on p p The chicken feather concentration that was optimum for keratinase activity was at 0.5% (w/v) in the fermentation medium containing a 2% (v/v) starter culture with cell concentration of about 1 × 108 CFU/mL (Figure 6). A further increase in the concentration of the chicken feather beyond 0.5% resulted in a constant decline in keratinase activity by Bacillus sp. FPF-1. However, the converse was the case with the total protein content measured in the fermentation broth; as the concentration of keratinase declined from 790.91 ± 10.29 to 714.55 ± 77.14 to 582.73 ± 26.99 (U/mL) with the respective chicken 7 of 16 Molecules 2020, 25, 1505 feather concentrations 0.75%, 1.0% and 1.25% (w/v), the total protein concentration in the fermentation medium increased as follows; 277.77 ± 2.23, 385.72 ± 10.03 and 408.57 ± 2.23 (µg/mL), respectively. Molecules 2020, 25, x FOR PEER REVIEW 8 of 17 feather concentrations 0.75%, 1.0% and 1.25% (w/v), the total protein concentration in the fermentation medium increased as follows; 277.77 ± 2.23, 385.72 ± 10.03 and 408.57 ± 2.23 (µg/mL), respectively. Molecules 2020, 25, x FOR PEER REVIEW 8 of 17 Figure 6. The effect of chicken feather concentration on keratinase activity by Bacillus sp. FPF-1. 0 200 400 600 800 1000 0 100 200 300 400 500 600 700 800 900 1000 0.1 0.25 0.5 0.75 1 1.25 1.5 1.75 Protein concentration (μg/mL) Keratinase activity (U/mL) Chicken feather (%; w/v) Keratinase activity Protein concentration Figure 6. The eﬀect of chicken feather concentration on keratinase activity by Bacillus sp. FPF-1. 0.75 Chicken feather (%; w/v) Figure 6. The effect of chicken feather concentration on keratinase activity by Bacillus sp. FPF-1. Figure 6. The eﬀect of chicken feather concentration on keratinase activity by Bacillus sp. FPF-1. Figure 6. The effect of chicken feather concentration on keratinase activity by Bacillus sp. FPF-1. Figure 6. The eﬀect of chicken feather concentration on keratinase activity by Bacillus sp. FPF-1. The variation of the fermentation temperature showed that Bacillus sp. FPF-1 optimally produced keratinase at 25 °C (1918.18 ± 64.28 U/mL). As the fermentation temperature increased, keratinase activity decreased consistently (Figure 7). 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FP The ﬁndings only buttress the supposition that keratinases have a high substrate-to-product conversion rate and that other enzymes, which may be heat labile, are also in play in the conversion of polymers to proteins. 0 200 400 600 800 0 200 400 600 800 1000 Protein conce Keratinase However, organisms are unique and so are their characteristics; some keratinase-producing bacteria have functioned well at temperatures ranging from 25 to 40 ◦C [17,25,26]. Thermophilic keratinase-producing bacteria with optimum temperatures ranging from 50 to 70 ◦C have been reported [27–29]. However, Bacillus sp. FPF-1 may be graded as mesophilic, and the biotechnological signiﬁcance would include a low energy input in a large-scale fermentation. Molecules 2020, 25, 1505 reported [27–29]. H 8 of 16 ological Figure 7. The effect of incubation temperature on keratinase activity by Bacillus sp. FPF-1. 0 200 400 600 800 1000 1200 1400 1600 1800 2000 0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 25 30 35 40 Protein concentration (μg/mL) Keratinase activity (U/mL) Incubation temperature (°Ϲ) Keratinase activity Protein concentration Figure 7. The eﬀect of incubation temperature on keratinase activity by Bacillus sp. FPF-1. Molecules 2020, 25, x FOR PEER REVIEW 9 of 17 The impact of aeration on keratinase activity by Bacillus sp. FPF-1 showed that an agitation speed of 150 rpm was optimal (1914.54 ± 205.70 U/mL), and at higher agitation speed (200 rpm), keratinase activity drastically reduced (Figure 8). The total protein concentration followed a similar pattern as keratinase activity. Although agitation of a fermentation process suffices as a measure for aeration, it also serves other purposes, including homogeneity in nutrient access for cells in the fermentation process. h d l k b b d h h d b h h h Incubation temperature (°Ϲ) y y p 05.70 U/mL), and at higher ag Figure 7. The effect of incubation temperature on keratinase activity by Bacillus sp. FPF-1. Figure 7. The eﬀect of incubation temperature on keratinase activity by Bacillus sp. FPF-1. p p , g g y The impact of aeration on keratinase activity by Bacillus sp. FPF-1 showed that an agitation speed of 150 rpm was optimal (1914.54 ± 205.70 U/mL), and at higher agitation speed (200 rpm), keratinase activity drastically reduced (Figure 8). The total protein concentration followed a similar pattern as keratinase activity. 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FP The pH of the fermentation broth decreased from the initial pH 5.0 to 4.5 ± 0.00 after 24 h of cultivation, and then it increased until 192 h of i ubatio ith a a i u alue of 8 5 ± 0 01 (Fi u e 9) The lau ible e la atio fo the ob e ed 0 500 1000 1500 2000 0 500 1000 1500 2000 2500 0 50 100 150 200 Protein concentration (μg/mL) Keratinase activity (U/mL) Agitation speed (rpm) Keratinase activity Protein concentration Figure 8. The eﬀect of agitation speed on keratinase activity by Bacillus sp. FPF-1. The decline in keratinase activity may be attributed to the shear stress occasioned by the high agitation speed, which may have disrupted the integrity of the bacterial cells and, ultimately, aﬀected the productivity [30]. However, if the integrity of the bacterial cells was not compromised, the conditions may not have been suitable for the desired metabolic processes. The peculiarity of organisms also inﬂuences their needs, and as such, Pseudomona stutzeri K4 [31], Arthrobacter sp. NFH5 [32] and Bacillus Figure 8. The effect of agitation speed on keratinase activity by Bacillus sp. FPF-1. 0 500 1000 1500 2000 0 500 1000 1500 2000 2500 0 50 100 150 200 Protein concentration (μg/mL) Keratinase activity (U/mL) Agitation speed (rpm) Keratinase activity Protein concentration Figure 8. The eﬀect of agitation speed on keratinase activity by Bacillus sp. FPF-1. Figure 8. The effect of agitation speed on keratinase activity by Bacillus sp. FPF-1. Figure 8. The eﬀect of agitation speed on keratinase activity by Bacillus sp. FPF-1. 2.4. Time Course Profile of the Keratinolytic Activity by Bacillus sp. FPF-1 Keratinase activity time course by Bacillus sp. FPF-1 studied over 192 h showed that the active growth stage of the organism is very crucial, as keratinase concentration in the fermentation medium remained high; 1698.18 ± 53.99 (U/mL) was achieved in 120 h, and that was the optimum. Beyond 120 h, the keratinase concentration significantly declined. The pH of the fermentation broth decreased The decline in keratinase activity may be attributed to the shear stress occasioned by the high agitation speed, which may have disrupted the integrity of the bacterial cells and, ultimately, aﬀected the productivity [30]. However, if the integrity of the bacterial cells was not compromised, the conditions may not have been suitable for the desired metabolic processes. 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FP Although agitation of a fermentation process suﬃces as a measure for aeration, it also serves other purposes, including homogeneity in nutrient access for cells in the fermentation process. The decline in keratinase activity may be attributed to the shear stress occasioned by the high agitation speed, which may have disrupted the integrity of the bacterial cells and, ultimately, affected the productivity [30]. However, if the integrity of the bacterial cells was not compromised, the conditions may not have been suitable for the desired metabolic processes. The peculiarity of organisms also influences their needs, and as such, Pseudomona stutzeri K4 [31], Arthrobacter sp. NFH5 [32] and Bacillus altitudinis GVC11 [33] were reported to optimally produce keratinases at an agitation speed peculiar to the respective organism. The impact of aeration on keratinase activity by Bacillus sp. FPF-1 showed that an agitation speed of 150 rpm was optimal (1914.54 ± 205.70 U/mL), and at higher agitation speed (200 rpm), keratinase activity drastically reduced (Figure 8). The total protein concentration followed a similar pattern as keratinase activity. Although agitation of a fermentation process suﬃces as a measure for aeration, it also serves other purposes, including homogeneity in nutrient access for cells in the fermentation process. The decline in keratinase activity may be attributed to the shear stress occasioned by the high agitation speed, which may have disrupted the integrity of the bacterial cells and, ultimately, affected the productivity [30]. However, if the integrity of the bacterial cells was not compromised, the conditions may not have been suitable for the desired metabolic processes. The peculiarity of organisms also influences their needs, and as such, Pseudomona stutzeri K4 [31], Arthrobacter sp. NFH5 [32] and Bacillus altitudinis GVC11 [33] were reported to optimally produce keratinases at an agitation speed peculiar to the respective organism. Figure 8. The effect of agitation speed on keratinase activity by Bacillus sp. FPF-1. 2.4. Time Course Profile of the Keratinolytic Activity by Bacillus sp. FPF-1 Keratinase activity time course by Bacillus sp. FPF-1 studied over 192 h showed that the active growth stage of the organism is very crucial, as keratinase concentration in the fermentation medium remained high; 1698.18 ± 53.99 (U/mL) was achieved in 120 h, and that was the optimum. Beyond 120 h, the keratinase concentration significantly declined. 2.4. Time Course Proﬁle of the Keratinolytic Activity by Bacillus sp. FPF-1 Keratinase activity time course by Bacillus sp. FPF-1 studied over 192 h showed that the active growth stage of the organism is very crucial, as keratinase concentration in the fermentation medium remained high; 1698.18 ± 53.99 (U/mL) was achieved in 120 h, and that was the optimum. Beyond 120 h, the keratinase concentration signiﬁcantly declined. The pH of the fermentation broth decreased from the initial pH 5.0 to 4.5 ± 0.00 after 24 h of cultivation, and then it increased until 192 h of incubation, with a maximum value of 8.5 ± 0.01 (Figure 9). The plausible explanation for the observed trend would be the metabolites secreted by the microbe, as well the products of enzyme activities. The metabolites secreted at the early stages that made the medium more acidic may include the organic acid. At subsequent stages, alkaline-based metabolites would have dominated, and they would include nitrogen-related metabolites. Alkalization is brought about, predominantly, by the deamination of amino acids that emanate from keratin fragmentation [24]. Molecules 2020, 25, x FOR PEER REVIEW 10 of 17 period may be due to the utilization of the thiol groups’ containing products faster than the enzymes converted the products, or the thiol groups were not liberated. The concentration of thiol groups liberated by Bacillus sp. FPF-1 is significantly higher than has been reported for Bacillus sp. MBRL 575 [34] and Stenotrophomonas maltophilia R13 [25]. Consequently, Bacillus sp. FPF-1 presents a potential for exploitation as an industrially relevant microbe. Figure 9. Time courses profile of keratinolytic activity by Bacillus sp. FPF-1 in an optimized fermentation medium at 25 °C and 150 rpm. 0 200 400 600 800 1000 1200 1400 1600 1800 2000 0 1 2 3 4 5 6 7 8 9 0 24 48 72 96 120 144 168 192 Keratinase activity (U/mL); Thiol concentration (µM) & Protein concentration (µg/mL) pH & OD600nm Incubation time (h) pH Cell growth Keratinase activity Protein concentration Thiol concentration Figure 9. Time courses proﬁle of keratinolytic activity by Bacillus sp. FPF-1 in an optimized fermentation medium at 25 ◦C and 150 rpm. Figure 9. Time courses profile of keratinolytic activity by Bacillus sp. FPF-1 in an optimized fermentation medium at 25 °C and 150 rpm. Figure 9. Time courses proﬁle of keratinolytic activity by Bacillus sp. FPF-1 in an optimized fermentation medium at 25 ◦C and 150 rpm. 2.5. 2.3. Optimization of Extracellular Keratinase Activity and Total Protein Production by Bacillus sp. FP The peculiarity of organisms also inﬂuences their needs, and as such, Pseudomona stutzeri K4 [31], Arthrobacter sp. NFH5 [32] and Bacillus 2.4. Time Course Profile of the Keratinolytic Activity by Bacillus sp. FPF-1 Keratinase activity time course by Bacillus sp. FPF-1 studied over 192 h showed that the active growth stage of the organism is very crucial, as keratinase concentration in the fermentation medium remained high; 1698.18 ± 53.99 (U/mL) was achieved in 120 h, and that was the optimum. Beyond 120 h, the keratinase concentration significantly declined. The pH of the fermentation broth decreased The decline in keratinase activity may be attributed to the shear stress occasioned by the high agitation speed, which may have disrupted the integrity of the bacterial cells and, ultimately, aﬀected the productivity [30]. However, if the integrity of the bacterial cells was not compromised, the conditions may not have been suitable for the desired metabolic processes. The peculiarity of organisms also inﬂuences their needs, and as such, Pseudomona stutzeri K4 [31], Arthrobacter sp. NFH5 [32] and Bacillus Molecules 2020, 25, 1505 9 of 16 altitudinis GVC11 [33] were reported to optimally produce keratinases at an agitation speed peculiar to the respective organism. 2.5. Biodegradation of Intact Chicken Feather—Structural Evaluation 2.5. Biodegradation of Intact Chicken Feather—Structural Evaluation The electron microscopic imaging of the chicken feather degradation showed the dismemberment of the structurally intact feather (Figure 10). As would be seen with Figure 10a, the feather is intact with all shafts and vane in place, and the integrity of the barbules were intact on the vane. However, as fermentation progressed, the barbules were completed degraded from the vane (Figure 10b), and the vane likewise was completely degraded in Figure 10c. All of these events occurred within 72 h of fermentation with Bacillus sp. FPF-1. In addition, beyond 72 h, it was clear that the organism had the potential to bio-convert the shaft, as a complete degradation was also achieved (Figure 10d). The imaging of chicken feather degradation by for B. licheniformis RG1 showed a similar pattern [35]. However, considerable chicken feather degradation was implemented by extensively studied Bacillus licheniformis PWD-1 after 10 days of incubation [36]. Again, the signiﬁcance of the imaging process includes the understanding of the degradation pattern. Therefore, at a scale-up process, the nature of feedstock is essential for optimal product yields. Molecules 2020, 25, x FOR PEER REVIEW 11 of 17 Figure 10. Electron micrography of feather degradation by Bacillus sp. FPF-1: (a) control (un- inoculated feather); (b) chicken feather at 24 h of fermentation; (c) feathers after 48 h fermentation; and (d) feathers after 72 h of cultivation. Figure 10. Electron micrography of feather degradation by Bacillus sp. FPF-1: (a) control (un-inoculated feather); (b) chicken feather at 24 h of fermentation; (c) feathers after 48 h fermentation; and (d) feathers after 72 h of cultivation. Figure 10. Electron micrography of feather degradation by Bacillus sp. FPF-1: (a) control (un- inoculated feather); (b) chicken feather at 24 h of fermentation; (c) feathers after 48 h fermentation; and (d) feathers after 72 h of cultivation. Figure 10. Electron micrography of feather degradation by Bacillus sp. FPF-1: (a) control (un-inoculated feather); (b) chicken feather at 24 h of fermentation; (c) feathers after 48 h fermentation; and (d) feathers after 72 h of cultivation. 2.4. Time Course Proﬁle of the Keratinolytic Activity by Bacillus sp. FPF-1 Biodegradation of Intact Chicken Feather—Structural Evaluation The electron microscopic imaging of the chicken feather degradation showed the dismemberment of the structurally intact feather (Figure 10). As would be seen with Figure 10a, the feather is intact with all shafts and vane in place, and the integrity of the barbules were intact on the vane. However, as fermentation progressed, the barbules were completed degraded from the vane (Figure 10b), and the vane likewise was completely degraded in Figure 10c. All of these events occurred within 72 h of fermentation with Bacillus sp. FPF-1. In addition, beyond 72 h, it was clear that the organism had the potential to bio-convert the shaft, as a complete degradation was also achieved (Figure 10d). The imaging of chicken feather degradation by for B. licheniformis RG1 showed a similar pattern [35]. However, considerable chicken feather degradation was implemented by The thiol concentration was negligible at 48 h of fermentation (Figure 9). Nevertheless, as fermentation progressed, thiol production gradually increased, and a maximum concentration of 379.53 ± 22.13 (µM) was observed after 96 h of fermentation. The thiol concentration declined after 168 h of fermentation, and it may be understood to reﬂect the reduction in keratinase activity occasioned by reduced disulﬁde bond breakage in the polymer. Moreover, the presence of thiol groups suggests eﬀective disulﬁde bonds reduction and the liberation of free-sulfhydryl-group-containing products. The negligible presence of thiol groups during the early stages of fermentation period may be due to the utilization of the thiol groups’ containing products faster than the enzymes converted the products, or the thiol groups were not liberated. The concentration of thiol groups liberated by Bacillus sp. FPF-1 is signiﬁcantly higher than has been reported for Bacillus sp. MBRL 575 [34] and Stenotrophomonas Molecules 2020, 25, 1505 10 of 16 maltophilia R13 [25]. Consequently, Bacillus sp. FPF-1 presents a potential for exploitation as an industrially relevant microbe. 3.4. Feather Hydrolysis Assay The degree of feather degradation by the test bacteria was determined by the weight-loss approach, as previously described by Reddy et al. [18]. The fermentation broth was ﬁltered (Whatman®no. 1, Johannesburg, Gauteng, South Africa) to recover undegraded feathers, and subsequently washed with distilled water to remove cell biomass. The residual feathers were oven-dried at 60 ◦C for 24 h, and the constant weight was achieved. The degree of feather degradation was calculated and expressed as a percentage, as shown in Equation (1) below: Percentage feather hydrolysis (%) = (1 −RF/WF) × 100 (1) (1) where RF is the weight of residual feathers after fermentation; and WF is the weight of intact feathers before fermentation. where RF is the weight of residual feathers after fermentation; and WF is the weight of intact feathers before fermentation. 3.3. Preliminary Assessment of Isolates for Proteolytic and Keratinolytic Activities The proteolytic activity of the isolates was evaluated by inoculating skimmed milk agar (SMA) plates containing (g/L); K2HPO4, 0.3; KH2PO4, 0.4; MgCl2, 0.2; CaCl2, 0.22; NH4Cl, 0.5; skimmed milk, 10 and bacteriological agar, 15 [9], with 10 µL of the standardized bacterial suspension (equivalent to 1 × 108 CFU/mL). The inoculum was prepared and adjusted to 0.1 at 600 nm. The standardization approach was used for all inoculum preparation, unless stated otherwise. The plates were incubated at 30 ◦C for 24 h, and halo zone formation was an indication of casein hydrolysis. Isolates positive for proteolytic activity were then evaluated for keratinolytic activity. A 2% (v/v) standardized culture was inoculated into 98 mL of sterile basal media containing a 1% (w/v) intact chicken feather as sole carbon and nitrogen source [38]. The culture was incubated at 30 ◦C for 96 h in a rotary shaker (130 rpm). The ﬂasks were monitored for observable feather degradation, and fermentation broth was analyzed accordingly. Stock culture of potent chicken feather degrading bacteria was maintained on PCF agar slants at 4 ◦C, for a fresh inoculum preparation, and another culture maintained in 20% glycerol at −86 ◦C for long-term storage. 3 1 Keratin—Substrate Preparation 3.1. Keratin—Substrate Preparation p The keratin substrate was prepared from chicken feathers obtained from a local poultry processing farm. The feathers were thoroughly rinsed with distilled water and dried at 60 °C for 48 h. Dried feathers were milled with a pulverizer fitted with a 2 mm mesh and stored at room temperature in an airtight container The keratin substrate was prepared from chicken feathers obtained from a local poultry processing farm. The feathers were thoroughly rinsed with distilled water and dried at 60 ◦C for 48 h. Dried feathers were milled with a pulverizer ﬁtted with a 2 mm mesh and stored at room temperature, in an airtight container. 11 of 16 Molecules 2020, 25, 1505 11 of 16 3.2. Sample Collection and Bacteria Isolation Soil samples were collected from composting sites (32◦43’48’S 27◦1’32’E) at the Fort Cox College of Agriculture and Forestry, in the Raymond Mhlaba Local Municipality, Eastern Cape Province of South Africa. The samples were aseptically transported to the laboratory and processed within 6 h of collection. Basal salt media (BSM) was prepared as follows (g/L): K2HPO4, 0.3; KH2PO4, 0.4; MgCl2, 0.2; CaCl2, 0.22; NH4Cl, 0.5; powdered chicken feather (PCF), 10 [37]. About 1 g of the soil samples was inoculated in 99 mL of BSM and incubated at 30 ◦C, over a 5-day timeline, in an orbital shaker incubator (Labotec, South Africa), at 130 rpm. About 100 µL of the BSM-culture was spread on a PCF agar plate supplemented with 50 mg/L of nystatin. The plates were incubated at 30 ◦C for 48 h. 3.3. Preliminary Assessment of Isolates for Proteolytic and Keratinolytic Activities 3.6. Extracellular Keratinase and Total Protein Production 3.6. Extracellular Keratinase and Total Protein Production The extracellular keratinase and total protein was produced in a submerged fermentation with media constituents as (g/L): K2HPO4, 0.3; KH2PO4, 0.4; MgCl2, 0.2; CaCl2, 0.22; and CFP, 10. The fermentation ﬂasks were inoculated with a 2% (v/v) starter culture (comparable to 1 × 108 CFU/mL) and incubated at 30 ◦C for 120 h, in an orbital shaker (Labotec IncoShake (Pty) Ltd, Midrand, Gauteng, South Africa). After the triplicate submerged state fermentation, the culture broth was centrifuged at 15,000× g for 10 min, using micro-centrifuge (HERMLE Labortechnik GmbH, Wehingen, Germany) and the cell-free ﬁltrate served as crude enzyme for analytical assays. 3.5. Identiﬁcation of Keratinolytic Bacteria The genomic DNA of the keratinolytic bacterial isolates was extracted by using the ZR Fungal/Bacterial DNA KitTM (Zymo Research, Irvine, CA, USA). The 16S target region was ampliﬁed, using the polymerase chain reaction (PCR), under standard conditions. The set of universal primers used for the 16S rRNA gene sequence ampliﬁcation were 27f (5′-AGAGTTTGATCMTGGCTCAG-3′) and 1492r (5′-CGGTTACCTTGTTACGACTT-3′) as forward and reverse primers, respectively [39]. The amplicons were gel extracted (Zymo Research, ZymocleanTM Gel DNA Recovery Kit, Irvine, CA, USA) and sequenced in forward and reverse directions, on the ABI PRISMTM 3500xl Genetic Analyzer (Foster city, CA, USA). The sequence comparison with the reference sequences in the database was conducted by using the basic local alignment search tool (BLAST). The nucleotide sequences were 12 of 16 Molecules 2020, 25, 1505 submitted to the GenBank, with the following accession numbers: MG214993, MG214994, MG214995, MG214996 and MG214999. submitted to the GenBank, with the following accession numbers: MG214993, MG214994, MG214995, MG214996 and MG214999. 3.7. Assay for Keratinase Activity and Total Protein Concentration Keratinase activity assay was based on the method of Jaouadi et al. [40], with slight modiﬁcation. The reaction mixture contained 0.5 mL of 10 g/L of keratin azure (Sigma-Aldrich, St. Louis, MO, USA) in 0.1 M Tris-HCl buﬀer, pH 7.5, and 0.5 mL of suitably diluted crude enzyme solution. The mixture was incubated at 37 ◦C for 1 h, with shaking at 220 rpm; after that, the reaction was stopped by placing the assay mixture in ice-cooled water for 10 min. The unutilized substrates were removed by centrifugation at 15,000× g for 10 min, and subsequently ﬁltered (Millipore cellulose ﬁlters; 0.45 µm). The azo dye released in the ﬁltrate was determined at 595 nm, using a SYNERGYMx 96 well microplate reader (BioTek Instrument Inc., Winooski, VT, USA). The control was treated at the same condition which contained the enzyme solution and buﬀer without the substrate. One keratinase unit was deﬁned as the amount of enzyme causing an increase in absorbance of 0.01 per hour under the standard assay condition. y The total protein concentration was estimated by using the Bradford method [41], with bovine serine albumin as a standard protein. The respective assays were done in triplicate, and the results presented were mean plus standard deviation. 3.8. Determination of Medium Thiol Concentration The thiol concentration was determined in the ﬁltrate, using the methods described by Ellman [42]. Brieﬂy, 50 µL of 4 mg/mL of 5,5′-dithiobis (2-nitrobenzoic acid) (DTNB) (Sigma-Aldrich, St. Louis, MO, USA) in 100 mM phosphate buﬀer, pH 8.0, was mixed with 500 µL of distilled water. Then, the cell-free ﬁltrate (250 µL) was added in the mixture and allowed to stand at room temperature for 5 min, for stable color development. The absorbance of yellow-colored 2-nitro-5-thiobenzoic acid (TNB) that formed upon reduction of DTNB was measured at 412 nm. Un-inoculated broth treated in the same manner as above served as the control. Triplicate assays were carried out, and the results presented were mean plus standard deviation. 3.9. Eﬀect of Physicochemical Conditions on Keratinase Activity and Total Protein Production 3.10. Structural Studies: Scanning Electron Microscopy A scan electron microscopy of the biodegraded feather was conducted in line with the described procedures of Gupta and Singh [43]. Brieﬂy, degraded feather at diﬀerent interval of fermentation was recovered, washed with distilled water and oven-dried at 60 ◦C, for 24 h. The residual feather was ﬁxed on carbon-sided-tape mounted on the stub and coated with gold palladium for 5 min, using an ion coater (Eiko Engineering Co. Ltd, Hitachinaka, Ibaraki, Japan), and observed under scanning electron microscope (JEOL Ltd, Tokyo Japan). 3.11. Statistical Analysis Experiments were performed in triplicates. The data were subjected to analysis of variance, and the degree of freedom was set at P < 0.05 signiﬁcance level. The analysis was conducted with the IBM Statistical Package for Social Science version 23 (Armonk, NY, USA). 3.9. Eﬀect of Physicochemical Conditions on Keratinase Activity and Total Protein Production The eﬀect of physicochemical conditions on keratinase activity and total protein production was assessed by using the one-variable-at-a-time (OVAT) approach. The eﬀect of pH was assessed through varying the initial fermentation pH from 4.0 to 10, at intervals of one unit. The incubation temperature was also varied from 25 to 40 ◦C, at intervals of 5 ◦C. The inﬂuence of the carbon sources, xylose, mannitol, glucose, fructose, sucrose, maltose, soluble starch, galactose, lactose and sorbitol, (Merck chemicals (Pty) Ltd, Modderfontein, Gauteng, South Africa) was likewise evaluated. Moreover, the concentrations of the carbon sources were initially at 0.1% (w/v) of the fermentation medium. The carbon sources with the optimum activity were further evaluated with concentration variation from 0.08% to 1% (w/v). The eﬀects of the nitrogen sources, gelatin, yeast extract, malt extract, casein, peptone, tryptone, urea, beef extract, NH4SO4, NH4NO3, KNO3, NaNO3, NH4Cl, (NH4)H2PO4 and (NH4)2HPO4 (Merck chemicals (Pty) Ltd., Modderfontein, Gauteng, South Africa), were also 13 of 16 Molecules 2020, 25, 1505 13 of 16 investigated at a concentration of 0.02% (w/v). The eﬀect of chicken feather concentration (0.1–1.75%, w/v) on keratinase activity and total protein production was also investigated. Lastly, the eﬀect of agitation speed on keratinase activity and total protein production was determined at speeds ranging from 0 to 200 rpm, at intervals of 50 units. The optimum conditions were used to conduct the time course assay, and the formulated basal media contained the following (w/v): 0.03% K2HPO4, 0.04% KH2PO4, 0.02% MgCl2, 0.022% CaCl2, 0.1% xylose and 0.5% PCF. The ﬂasks were inoculated with 2% (v/v) starter culture and incubated at 25 ◦C, for 192 h and 150 rpm. An aliquot of the fermentation broth was aseptically and periodically withdrawn to determine the keratinase activity, protein concentration, cell growth, thiol concentration and pH change. In each case, triplicate experiments were carried out, and the results presented were mean plus standard deviation. Funding: This research was funded by The Department of Science and Innovation (DSI) and the Technology Innovation Agency, grant number DST/CON/0177/2018. Author Contributions: Conceptualization, N.E.N. and U.U.N.; methodology, N.E.N.; formal analysis, N.E.N.; investigation, N.E.N.; resources, A.I.O. and U.U.N.; data curation, N.E.N.; writing—original draft preparation, N.E.N.; writing—review and editing, A.I.O. and U.U.N.; supervision, A.I.O. and U.U.N.; project administration, U.U.N.; funding acquisition, U.U.N. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by The Department of Science and Innovation (DSI) and the Technology Innovation Agency, grant number DST/CON/0177/2018. 4. Conclusions Valorization of waste to valuable products is an integral part of the exploitation of renewable resources for sustainable development. The approach of value addition to wastes has fed extensively into the bio-economy of many nations. Thus, the potentials presented by Bacillus sp. FPF-1 as indicative of the excellent production of keratinases which bio-converts chicken feathers into proteins and related products. The process conditions showed an optimal keratinase yield at a weak acidic initial fermentation medium, a mesophilic temperature condition and moderate agitation speed. Spiking the fermentation medium with xylose enhanced the production of the enzyme of interest, and the opposite eﬀect was achieved with the inclusion of additional nitrogen sources. The capacity to eﬀectively utilize chicken feathers as a sole source of carbon and nitrogen by Bacillus sp. FPF-1 is indicative of the fact that a battery of enzymes is produced to eﬀectively degrade the complex pertinacious polymer. The high concentration of total proteins in the fermentation broth is a strong indicator to the assertion. In addition, the potency of the keratinase was laid bare in the electron micrograph indicating the degradation pattern over time. Bacillus sp. FPF-1 presents a very good prospects vis-à-vis the unique enzymes, and enzyme products and the organism as a whole serving as an industrially relevant strain. Conclusively, the complete and eﬀective degradation of chicken feathers by Bacillus sp. FPF-1 underpins its potential in the valorization of recalcitrant keratinous waste biomass from the agro sector into useful products. Author Contributions: Conceptualization, N.E.N. and U.U.N.; methodology, N.E.N.; formal analysis, N.E.N.; investigation, N.E.N.; resources, A.I.O. and U.U.N.; data curation, N.E.N.; writing—original draft preparation, N.E.N.; writing—review and editing, A.I.O. and U.U.N.; supervision, A.I.O. and U.U.N.; project administration, U.U.N.; funding acquisition, U.U.N. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by The Department of Science and Innovation (DSI) and the Technology Innovation Agency, grant number DST/CON/0177/2018. 14 of 16 14 of 16 Molecules 2020, 25, 1505 Acknowledgments: The Department of Science and Innovation (DSI) and the Technology Innovation Agency (TIA) supported this work under SIIP enzyme and microbial technologies (grant number: DST/CON/0177/2018). We also acknowledge the support of the South African Medical Research Council (SAMRC). Conﬂicts of Interest: The authors declare no conﬂict of interest. References 1. Wang, B.; Yang, W.; McKittrick, J.; Meyers, M.A. Keratin: Structure, mechanical properties, occurrence in biological organisms, and eﬀorts at bioinspiration. Prog. Mater. Sci. 2016, 76, 229–318. [CrossRef] 1. Wang, B.; Yang, W.; McKittrick, J.; Meyers, M.A. Keratin: Structure, mechanical properties, occurrence in biological organisms, and eﬀorts at bioinspiration. Prog. Mater. Sci. 2016, 76, 229–318. [CrossRef] 2. Tesfaye, T.; Sithole, B.; Ramjugernath, D.; Chunilall, V. Valorisation of chicken feathers: Characterisation of chemical properties. Waste Manag. 2017, 68, 626–635. [CrossRef] 2. Tesfaye, T.; Sithole, B.; Ramjugernath, D.; Chunilall, V. Valorisation of chicken feathers: Characterisation of chemical properties. Waste Manag. 2017, 68, 626–635. [CrossRef] 3. Onifade, A.A.; Al-Sane, N.A.; Al-Musallam, A.A.; Al-Zarban, S. A review: Potentials for biotechnological applications of keratin-degrading microorganisms and their enzymes for nutritional improvement of feathers and other keratins as livestock feed resources. Bioresour. Technol. 1998, 66, 1–11. [CrossRef] 3. Onifade, A.A.; Al-Sane, N.A.; Al-Musallam, A.A.; Al-Zarban, S. A review: Potentials for biotechnological applications of keratin-degrading microorganisms and their enzymes for nutritional improvement of feathers and other keratins as livestock feed resources. Bioresour. Technol. 1998, 66, 1–11. [CrossRef] 4. Kang, D.; Herschend, J.; Al-Soud, W.A.; Mortensen, M.S.; Gonzalo, M.; Jacquiod, S.; Sørensen, S.J. Enrichment and characterization of an environmental microbial consortium displaying eﬃcient keratinolytic activity. Bioresour. Technol. 2018, 270, 303–310. [CrossRef] [PubMed] 5. Brouta, F.; Descamps, F.; Fett, T.; Losson, B.; Gerday, C.; Mignon, B. Puriﬁcation and characterization of a 43·5 kDa keratinolytic metalloprotease from Microsporum canis. Med. Mycol. 2001, 39, 269–275. [CrossRef] [PubMed] 6. Singh, V.; Haque, S.; Niwas, R.; Srivastava, A.; Pasupuleti, M.; Tripathi, C.K.M. Strategies for fermentation medium optimization: An in-depth review. Front. Microbiol. 2017, 7, 2087. [CrossRef] 7. Brandelli, A.; Daroit, D.J.; Riﬀel, A. Biochemical features of microbial keratinases and their production and applications. Appl. Microbiol. Biotechnol. 2010, 85, 1735–1750. [CrossRef] 8. Verma, A.; Singh, H.; Anwar, S.; Chattopadhyay, A.; Tiwari, K.K.; Kaur, S.; Dhilon, G.S. Microbial keratinases: Industrial enzymes with waste management potential. Crit. Rev. Biotechnol. 2017, 37, 476–491. [CrossRef] 8. Verma, A.; Singh, H.; Anwar, S.; Chattopadhyay, A.; Tiwari, K.K.; Kaur, S.; Dhilon, G.S. Microbial keratinases: Industrial enzymes with waste management potential. Crit. Rev. Biotechnol. 2017, 37, 476–491. [CrossRef] 9. Riﬀel, A.; Brandelli, A. Keratinolytic bacteria isolated from feather waste. Braz. J. Microbiol. 2006, 37, 395–399. Industrial enzymes with waste management potential. Crit. Rev. Biotechnol. 2017, 37, 476–491. [CrossRef] 9. Riﬀel, A.; Brandelli, A. Keratinolytic bacteria isolated from feather waste. Braz. J. Acknowledgments: The Department of Science and Innovation (DSI) and the Technology Innovation Agency (TIA) supported this work under SIIP enzyme and microbial technologies (grant number: DST/CON/0177/2018). We also acknowledge the support of the South African Medical Research Council (SAMRC). References Microbiol. 2006, 37, 395–399. [CrossRef] 9. Riﬀel, A.; Brandelli, A. Keratinolytic bacteria isolated from feather waste. Braz. J. Microbiol. 2006, 37, 395–399. [CrossRef] 10. Venil, C.K.; Zakaria, Z.A.; Usha, R.; Ahmad, W.A. Isolation and characterization of ﬂexirubin type pigment from Chryseobacterium sp. UTM-3T. Biocatal. Agric. Biotechnol. 2014, 3, 103–107. [CrossRef] 11. Daroit, D.J.; Brandelli, A. A current assessment on the production of bacterial keratinases. Crit. Rev. Biotechnol. 2014, 34, 372–384. [CrossRef] [PubMed] 12. Rieger, T.J.; De Oliveira, C.T.; Pereira, J.Q.; Brandelli, A.; Daroit, D.J. Proteolytic system of Bacillus sp. CL18 is capable of extensive feather degradation and hydrolysis of diverse protein substrates. Br. Poult. Sci. 2017, 58, 329–335. [CrossRef] [PubMed] 13. He, Z.; Sun, R.; Tang, Z.; Bu, T.; Wu, Q.; Li, C.; Chen, H. Biodegradation of Feather Waste Keratin by the Keratin-Degrading Strain Bacillus subtilis 8. J. Microbiol. Biotechnol. 2018, 28, 314–322. [CrossRef] [PubMed] 14. Tiwary, E.; Gupta, R. Medium optimization for a novel 58 kDa dimeric keratinase from Bacillus licheniformis ER-15: Biochemical characterization and application in feather degradation and dehairing of hides. Bioresour. Technol. 2010, 101, 6103–6110. [CrossRef] [PubMed] 15. Daroit, D.J.; Corrêa, A.P.F.; Brandelli, A. Keratinolytic potential of a novel Bacillus sp. P45 isolated from the Amazon basin ﬁsh Piaractus mesopotamicus. Int. Biodeterior. Biodegrad. 2009, 63, 358–363. [CrossRef] 16. Kim, J.M.; Lim, W.J.; Suh, H.J. Feather-degrading Bacillus species from poultry waste. Process Biochem. 2001, 37, 287–291. [CrossRef] 17. Sangali, S.; Brandelli, A. Feather keratin hydrolysis by a Vibrio sp. strain kr2. J. Appl. Microbiol. 2000, 89, 735–743. [CrossRef] 18. Reddy, M.R.; Reddy, K.S.; Chouhan, Y.R.; Bee, H.; Reddy, G. Eﬀective feather degradation and keratinase production by Bacillus pumilus GRK for its application as bio-detergent additive. Bioresour. Technol. 2017, 243, 254–263. [CrossRef] 19. Bose, A.; Pathan, S.; Pathak, K.; Keharia, H. Keratinolytic protease production by Bacillus amyloliquefaciens 6B using feather meal as substrate and application of feather hydrolysate as organic nitrogen input for agricultural soil. Waste Biomass Valoriz. 2014, 5, 595–605. [CrossRef] 15 of 16 Molecules 2020, 25, 1505 15 of 16 20. Lin, X.; Lee, S.W.; Bae, H.D.; Shelford, J.A.; Cheng, K.J. Comparison of two feather-degrading Bacillus licheniformis strains. Asian-Australas. J. Anim. Sci. 2001, 14, 1769–1774. [CrossRef] 21. Gupta, R.; Ramnani, P. Microbial keratinases and their prospective applications: An overview. Appl. Microbiol. Biotechnol. 2006, 70, 21. [CrossRef] [PubMed] 22. Hossain, M.S.; Azad, A.K.; Sayem, S.A.; Mostafa, G.; Hoq, M.M. References Production and partial characterization of feather-degrading keratinolytic serine protease from Bacillus licheniformis MZK-3. J. Biol. Sci. 2007, 7, 599–606. [CrossRef] 23. Jeong, J.H.; Park, K.H.; Oh, D.J.; Hwang, D.Y.; Kim, H.S.; Lee, C.Y.; Son, H.J. Keratinolytic enzyme-mediated biodegradation of recalcitrant feather by a newly isolated Xanthomonas sp. P5. Polym. Degrad. Stab. 2010, 95, 1969–1977. [CrossRef] 24. Park, G.T.; Son, H.J. Keratinolytic activity of Bacillus megaterium F7-1, a feather-degrading mesophilic bacterium. Microbiol. Res. 2009, 164, 478–485. [CrossRef] [PubMed] 25. Jeong, J.H.; Lee, O.M.; Jeon, Y.D.; Kim, J.D.; Lee, N.R.; Lee, C.Y.; Son, H.J. Production of keratinolytic enzyme by a newly isolated feather-degrading Stenotrophomonas maltophilia that produces plant growth-promoting activity. Process Biochem. 2010, 45, 1738–1745. [CrossRef] 26. Pereira, J.Q.; Lopes, F.C.; Petry, M.V.; da Costa Medina, L.F.; Brandelli, A. Isolation of three novel Antarctic psychrotolerant feather-degrading bacteria and partial puriﬁcation of keratinolytic enzyme from Lysobacter sp. A03. Int. Biodeterior. Biodegrad. 2014, 88, 1–7. [CrossRef] 27. Chitte, R.R.; Nalawade, V.K.; Dey, S. Keratinolytic activity from the broth of a feather-degrading thermophilic Streptomyces thermoviolaceus strain SD8. Lett. Appl. Microbiol. 1999, 28, 131–136. [CrossRef] 28. Sivakumar, T.; Shankar, T.; Vijayabaskar, P.; Ramasubramanian, V. Optimization for keratinase enzyme production using Bacillus thuringiensis TS2. Acad J. Plant Sci. 2012, 5, 102–109. [CrossRef] 29. Gurav, R.G.; Jadhav, J.P. Biodegradation of keratinous waste by Chryseobacterium sp. RBT isolated from soil contaminated with poultry waste. J. Basic Microbiol. 2013, 53, 128–135. [CrossRef] 30. Kim, S.W.; Hwang, H.J.; Xu, C.P.; Choi, J.W.; Yun, J.W. Eﬀect of aeration and agitation on the production of mycelial biomass and exopolysaccharides in an enthomopathogenic fungus Paecilomyces sinclairii. Lett. Appl. Microbiol. 2003, 36, 321–326. [CrossRef] 31. Chaturvedi, V.; Bhange, K.; Bhatt, R.; Verma, P. Production of kertinases using chicken feathers as substrate by a novel multifunctional strain of Pseudomonas stutzeri and its dehairing application. Biocatal. Agric. Biotechnol. 2014, 3, 167–174. [CrossRef] 32. Barman, N.C.; Zohora, F.T.; Das, K.C.; Mowla, M.G.; Banu, N.A.; Salimullah, M.; Hashem, A. Production, partial optimization and characterization of keratinase enzyme by Arthrobacter sp. NFH5 isolated from soil samples. AMB Express 2017, 7, 181. [CrossRef] [PubMed] 33. Kumar, E.V.; Srijana, M.; Chaitanya, K.; Reddy, Y.; Reddy, G. Biodegradation of poultry feathers by a novel bacterial isolate Bacillus altitudinis GVC11. Indian, J. Biotechnol. 2011, 10, 502–507. 34. Kshetri, P.; Ningthoujam, D.S. Keratinolytic activities of alkaliphilic Bacillus sp. MBRL 575 from a novel habitat, limestone deposit site in Manipur, India. SpringerPlus 2016, 5, 595. p p p y g 2. Ellman, G.L. Tissue sulfhydryl groups. Arch. Biochem. Biophys. 1959, 82, 70–77. [CrossRef] 41. Bradford, M.M. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 1976, 72, 248–254. [CrossRef] 1. Bradford, M.M. A rapid and sensitive method for the quantitation of microgram quantities of protein the principle of protein-dye binding. Anal. Biochem. 1976, 72, 248–254. [CrossRef] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). 42. Ellman, G.L. Tissue sulfhydryl groups. Arch. Biochem. Biophys. 1959, 82, 70–77. [CrossRef] 41. Bradford, M.M. A rapid and sensitive method for the quantitation of microgram quantities of pr the principle of protein-dye binding. Anal. Biochem. 1976, 72, 248–254. [CrossRef] 43. Gupta, S.; Singh, R. Hydrolyzing proﬁciency of keratinases in feather degradation. Indian J. Microbiol. 2014, 54, 466–470. [CrossRef] [PubMed] Bradford, M.M. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing References [CrossRef] 35. Ramnani, P.; Singh, R.; Gupta, R. Keratinolytic potential of Bacillus licheniformis RG1: Structural and biochemical mechanism of feather degradation. Can. J. Microbiol. 2005, 51, 191–196. [CrossRef] 36. Williams, C.M.; Richter, C.S.; Mackenzie, J.M.; Shih, J.C. Isolation, identiﬁcation, and characterization of a feather-degrading bacterium. Appl. Environ. Microbiol. 1990, 56, 1509–1515. [CrossRef] 37. Letourneau, F.; Soussotte, V.; Bressollier, P.; Branland, P.; Verneuil, B. Keratinolytic activity of Streptomyces sp. S. K1-02: A new isolated strain. Lett. Appl. Microbiol. 1998, 26, 77–80. [CrossRef] 38. Tatineni, R.; Doddapaneni, K.K.; Potumarthi, R.C.; Vellanki, R.N.; Kandathil, M.T.; Kolli, N.; Mangamoori, L.N. Puriﬁcation and characterization of an alkaline keratinase from Streptomyces sp. Bioresour. Technol. 2008, 99, 1596–1602. [CrossRef] 39. Turner, S.; Pryer, K.M.; Miao, V.P.; Palmer, J.D. Investigating deep phylogenetic relationships among cyanobacteria and plastids by small subunit rRNA sequence analysis. J. Eukaryot. Microbiol. 1999, 46, 327–338. [CrossRef] 40. Jaouadi, B.; Abdelmalek, B.; Fodil, D.; Ferradji, F.Z.; Rekik, H.; Zaraî, N.; Bejar, S. Puriﬁcation and characterization of a thermostable keratinolytic serine alkaline proteinase from Streptomyces sp. strain AB1 with high stability in organic solvents. Bioresour. Technol. 2010, 101, 8361–8369. [CrossRef] Molecules 2020, 25, 1505 16 of 16 16 of 16
https://openalex.org/W4317988067	https://zenodo.org/records/7567943/files/3.%20Malarvizhi.pdf	Tamil	null	Su. Tamilselvi's Female narrators in the novel of Manickam	Zenodo (CERN European Organization for Nuclear Research)	2,023	cc-by	3,472	Su. Tamilselvi’s Female narrators in the novel of Manickam முதைவர் சவ.மலர்விழி, இதணப் பெராெிாியர், தமிழாய்வுத்துதை, ஸ்ரீ வித்யா மந்திர் கதல மற்றும் அைிவியல் கல்லூாி(தன்ைாட்ெி), காட்போி, ஊத்தங்கதர. லர்விழி, இதணப் பெராெிாியர், தமிழாய்வுத்துதை, ஸ்ரீ வித்யா மந்திர் கதல மற்றும் அைிவியல் கல்லூாி(தன்ைாட்ெி), காட்போி, ஊத்தங்கதர. Dr V. Malarvizhi, Associate Professor, Department of Tamil, Sri Vidya Mandir Arts and Science College, Katteri, Uthangarai. DOI : 10.5281/zenodo.7567943 Abstract The main characters are the lifeblood of every innovation. The main characters are those whom the story features in the novel. They will be placed towards the end of the story. They will be displayed as the hero of the story. Both Thalaivan and Thalavi featured in the novel are Chief Mandars. They are the lifeblood of the story. As the mother and father of a family, the leader of the story excels. The main character of the story is seen continuously from the beginning to the end of the story. Karu is the leader of the story. Without them, the story would not be a story. All the events that are essential for the story to take place are often centered around the main characters. The study is carried out with the main female characters appearing in the seven novels of S.Tamilchelvi, which have been studied because the novel author expresses the ideas he has to say to the society through the main characters.    ISSN:2583-0139 Vol.3: Issue 1, January 2022    ISSN:2583-0139 Vol.3: Issue 1, January 2022    Vol.3: Issue 1, January 2022 ISSN:2583-0139 Vol.3: Issue 1, January 2022 ஆய்வுச்சுருக்கம் ்ுு ஒவ்சவாரு புதிைத்திற்கும் உயிர்நாடியாக விளங்குெவர்கள் முதன்தமப் ொத்திரங்கபள ஆவர். புதிைத்தில் இேம் செறும் கதத மாந்தர்களில் யாதரப் ெற்ைி ெிைப்ொக இேம்செைச் செய்கிைபதா, அவர்கதளபய முதன்தமப் ொத்திரங்கள். இவர்கள் கததயின் இறுதிவதர இேம் செறுவர். கததயின் கதா நாயகர் பொல காட்ெிெடுத்தப்ெடுவர். புதிைத்தில் இேம்செறும் ததலவன், ததலவி இருவரும் ததலதம மாந்தர்கள் ஆவர். கததயின் உயிபராட்ேபம இவர்கள்தாம். ஒரு குடும்ெத்திற்குத் தாய், தந்ததயர் பொன்று கததக்குத் ததலதம மாந்தர்கள் ெிைந்து விளங்குவார்கள். கததயின் ததலதம ொத்திரம் கததயின் சதாேக்கம் முதல் இறுதிவதர சதாேர்ச்ெியாகக் காணப்ெடுவார்கள். கததக்குக் கருபவ ததலதம மாந்தர்கள்தாம். இவர்கள் இல்தல என்ைால் கதத கததயாக அதமயாது. கதத முழுவதும் நதேசெறுவதற்கு இன்ைியதமயாத நிகழ்ச்ெிகள் யாவும் செரும்ொலும் முதன்தமப் ொத்திரங்கதள தமயமாகக் சகாண்பே அதமயும். புதிை ஆெிாியர் முதன்தம மாந்தர்களின் மூலம் ெமூகத்திற்குத் தாம் கூைவிதளயும் கருத்துகதள சவளிப்ெடுத்துகிைார் என்ெதைால் ஆய்வுக்குட்ெடுத்தப்ெட்ே சு.தமிழ்ச்செல்வியின் ஏழு புதிைங்களில் இேம்செறும் முதன்தம செண்ொத்திரங்கதளக் சகாண்டு ஆய்வு பமற்சகாள்ளப்ெடுகிைது. ு்்்்்்்்ு்்்்் Keywords: சு.தமிழ்ச்செல்வி, கததமாந்தர்கள், செண், நாவல், முதன்தமக் கதத மாந்தர்கள் 18 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES  Vol.3: Issue 1, January 2022 To Discern the Truth is Wisdom Quarterly Journal ISSN:2583-0139 18    Vol.3: Issue 1, January 2022 ISSN:2583-0139 ISSN:2583-0139 Vol.3: Issue 1, January 2022 முன்னுதர சுயநலம், அன்பு, ொகெ உணர்வு, மூர்க்கம், இவற்ைின் கலதவயாகத் திகழும் மாணிக்கத்தத தன் வாழ்வில் எதிர்சகாள்கிைாள் செல்லாயி என்னும் எளிய கிராமத்துப் செண் இந்நாவலில் முதன்தம செண் கதத மாந்தராகத் திகழ்கிைாள். தாயின்ைி அப்ொவின் வளர்ப்ெில் வளர்ந்துக் சகாண்டிருக்கும் செல்லாயிதய தங்கத்தாச்ெி தன் மகன் மாணிக்கத்திற்கு செண் பகட்டு திருமணம் முடிக்கிைாள். செல்லாயி நல்ல உேல் வாகுேன் உதழக்கும் திைன் சகாண்ேவள். ஆைால் தங்கத்தாச்ெியின் மகன் மாணிக்கபமா விருப்ெமின்ைி செல்லாயிதய திருமணம் செய்து சகாள்கிைான். மாமியார், நாத்தைார் சகாடுதமயுேன் கணவன் சகாடுதமயும் சொிதாகிைது. செல்லாயிக்கு ெமூகம், குழந்ததகள் எைப் ெல்பவறு ெிக்கல்கதள எதிர் சகாள்கிைாள் செல்லாயி. இரவு சவகு பநரம் ஆைாலும் வீட்டிற்கு வரமாட்ோன் மாணிக்கம். ஓமிபயாெதி ெடித்துவிட்டு சுற்ைியுள்ள கிராம மக்களுக்கு மருத்துவச் பெதவ செய்ய தைியாக ஒரு மருத்துவமதை தவத்திருந்தான். செண் உதழப்பு, ெிக்கைம், பெமிப்பு இளம் வயது முதபல தாயின்ைி வளர்ந்தாலும் சொறுப்ொைவளாகவும் உதழப்ொளியாகவும் இருக்கும் செல்லாயி தத மாெி மாதங்களில் கதிரறுக்கும் பவதலக்குப் பொவாள், கிதேக்கும் கூலி சநல்தல சொிய ொதைகளில் அவெரம் ஆத்திரத்துக்கு உதவுசமன்று சகாட்டி தவப்ொள். பமலும் கிதேக்கும் சநல்தல அவித்துப் பொட்டு அதரத்துச் ொப்ெிடுவது உண்டு. ஒரு நாதளக்கு மூன்று அல்லது நான்கு மரக்கால் கூலியாகக் கிதேக்கும். பவதல நாளில் ஒரு நாள் கூே வீட்டில் தங்க மாட்ோள். கதிரறுக்கும்பொது ெில ெமயங்களில் அாிவாள் தகதயயும் ெதம் ொர்த்து விடும். இரத்தம் வந்து தகதயபய ெிவப்ொக்கிவிடும். பெற்தையும் சுண்ணாம்தெயும் காயத்தில் அப்புவாள் ெீதலத்துணிதய கிழித்துப்பொட்டு பெர்த்துக் கட்டிக்சகாள்வாள். அந்த காயத்துேபை மறுநாள் பவதலக்குப் பொய்விடுவாள். இப்ெடி ஓடி ஓடி ெம்ொதித்து வரும் சநல்தலசயல்லாம் குைவைிேம் வாங்கிய சதாம்தெக் கூட்டில் சகாட்டிதவத்துக்சகாண்டு நான்கு மாதங்களுக்காவது நல்லபொறு ொப்ெிே பவண்டும் என்று ஆதெப்ெடுவாள். இவ்வாைாக உதழத்து பெர்த்து தவத்த செல்லாயி சதாேர்ந்து அப்ெடி செய்யவில்தல காரணம் மாணிக்கம் தெக்கிளுக்கு ேயர் மாத்தனும், செயின் வாங்கி மாட்ேனும் என்று சொல்லிவிட்டு தெக்கிள் கதேயில் ொதியும் ொராயக் கதேயில் ொதியும் விற்றுவிடுவான். சதாேர்ந்து இப்ெடி உதழப்செல்லாம் வீணாகப் பொகிைபத என்று நிதைத்த செல்லாயி, சநல்லு வியாொாியிேம் விற்றுவிட்டு ெணமாகப் செற்று ெத்திரப்ெடுத்திக் சகாள்கிைாள். அப்ெணத்திற்கு, வீட்டிற்குத் பததவயாைப் சொருட்கதள வாங்கியதாக கணக்கு காட்டிவிடுவாள். இதவ மட்டுமன்ைி ெயத்தம், சநற்று, உளுந்து சொறுக்கவும் சென்று கூலியாக குழம்புக்குத் பததவயாை ெருப்பு வதககதளயும் பெமித்து தவத்துக் சகாள்ளும் செல்லாயி திைமும் 19 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES  Vol.3: Issue 1, January 2022 To Discern the Truth is Wisdom Quarterly Journal ISSN:2583-0139 19 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES 2022 To Discern the Truth is Wisdom Quarterly Journal Vol.3: Issue 1, January 2022 ISSN:2583-0139    ISSN:2583-0139 Vol.3: Issue 1, January 2022      ISSN:2583-0139 Vol.3: Issue 1, January 2022 பொைாக்க வாங்கிவரும் அாிெியில் ஒவ்சவாரு ெிடி அள்ளி ொதையில் பொட்டு விடுவாள். ெிடியாிெியால் நிதைந்திருந்தது ஒரு ொதை ெல நாள் ெிடியாிெி பெர்த்து ஒரு நாதளக்கு ெடியாிெியாகக் கூே ஆகிவிடும். அன்தைக்கு அாிெி வாங்க சகாடுக்கும் காசு செல்லாயியின் ெிறுவாட்டில் பெர்ந்துவிடும் இவ்வளவு சொறுப்புணர்பவாடும், ெிக்கைமாகவும் குடும்ெம் நேத்தும் செல்லாயி கணவன், மாமியார், நாத்தைார், குழந்ததகள், ெமூகம் எை ெல்பவறு துன்ெங்கதளச் ெந்திக்கிைாள்.    ISSN:2583-0139 Vol.3: Issue 1, January 2022    ISSN:2583-0139 பொைாக்க வாங்கிவரும் அாிெியில் ஒவ்சவாரு ெிடி அள்ளி ொதையில் பொட்டு விடுவாள். ெிடியாிெியால் நிதைந்திருந்தது ஒரு ொதை ெல நாள் ெிடியாிெி பெர்த்து ஒரு நாதளக்கு ெடியாிெியாகக் கூே ஆகிவிடும். அன்தைக்கு அாிெி வாங்க சகாடுக்கும் காசு செல்லாயியின் ெிறுவாட்டில் பெர்ந்துவிடும் இவ்வளவு சொறுப்புணர்பவாடும், ெிக்கைமாகவும் குடும்ெம் நேத்தும் செல்லாயி கணவன், மாமியார், நாத்தைார், குழந்ததகள், ெமூகம் எை ெல்பவறு துன்ெங்கதளச் ெந்திக்கிைாள். பொைாக்க வாங்கிவரும் அாிெியில் ஒவ்சவாரு ெிடி அள்ளி ொதையில் பொட்டு விடுவாள். ெிடியாிெியால் நிதைந்திருந்தது ஒரு ொதை ெல நாள் ெிடியாிெி பெர்த்து ஒரு நாதளக்கு ெடியாிெியாகக் கூே ஆகிவிடும். அன்தைக்கு அாிெி வாங்க சகாடுக்கும் காசு செல்லாயியின் ெிறுவாட்டில் பெர்ந்துவிடும் இவ்வளவு சொறுப்புணர்பவாடும், ெிக்கைமாகவும் குடும்ெம் நேத்தும் செல்லாயி கணவன், மாமியார், நாத்தைார், குழந்ததகள், ெமூகம் எை ெல்பவறு துன்ெங்கதளச் ெந்திக்கிைாள். கணவைால் செண்ெடும் துன்ெம் தன் கணவன் திருமணத்திற்கு முன்பு ெக்குொதய காதலித்ததாகவும் அவள் தன்தை நிதைத்துக் சகாண்டு இன்னும் திருமணம் செய்து சகாள்ளாமல் இருப்ெதாகவும் என்ைால் ெக்குொதய மைக்க முடியவில்தல அதைால், ெக்குொதய திருமணம் செய்துசகாள்கிபைன் என்றுக் கூைியததக் பகட்ேவுேன் செல்லாயியின் மைது கலங்கியது. மரவள்ளிக் சகால்தலக்குள்பளபய உட்கார்ந்து ஒப்ொாி தவக்கிைாள். “கருவ மரத்தடிபய நான் காயாடும் பொதல்லபவா காயாட்ேத்தத்தாம் மைந்து - இந்த கருணருக்கு மாலயிட்பேன் பவப்ெ மரத்தடிபய நான் சவளயாடும் பொதல்லபவா சவளயாட்ேத்தாம் மைந்து - இந்த வீமருக்கு மாலயிட்பேன் ஆை வுழந்த ெள்ளம் ஆவாதுன்னு சொன்பைபை ஆவுமுன்னு சொல்லியல்பலா – என்ை அழச்ெி வச்ெி மாலயிட்பே குதுர வுழந்த ெள்ளம் கூோதுன்னு சொன்பைபை கூடுமுன்னு சொல்லியல்பலா – என்ை கூட்டிவச்ெி மாலயிட்பே ஏலக்காப் சொடியின்னு ஏந்திபைன் கய்சநையா – நீ இருமைசுக் காரருன்னு – எைக்கு இருந்த ெைம் சொல்லலபய வால்மிளகு சொட்டியின்னு வாங்கிபைங் தகசநையா – நீ 20 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES To Discern the Truth is Wisdom Quarterly Journal    ISSN:2583-0139 Vol.3: Issue 1, January 2022    Vol.3: Issue 1, January 2022    Vol.3: Issue 1, January 2022 ISSN:2583-0139 Vol.3: Issue 1, January 2022 மாணிக்கம், ெக். 79-86.) மாணிக்கம், ெக். 79-86.) மாணிக்கம், ெக். 79-86.) என்று தன்மைக்குதைதய ஒப்ொாியாகப் ொடிைாள். கணவன் கண்டு சகாள்ளாமல் சென்றுவிடுகிைான்; செல்லாயி நான்கு மாத கர்ப்ெமாக இருக்கிைாள். இரவு சவகு பநரம் ஆைாலும் மாணிக்கம் வீட்டிற்கு வரமாட்ோன் செல்லாயி தன்தை தன் அப்ொவீட்டில் சகாண்டுபொய் விேச்சொல்லி அழுகிைாள். அப்ொவீட்டில் விடுவதாகக் கூைிவிட்டு கர்ப்ெிணி என்றும் ொராமல் நேந்பத நீண்ேதூரம் அதழத்துச்சென்ைான்; நேக்கமுடியாமல் செல்லாயி உட்கார்ந்பதவிட்ோள். மணலி சென்று அங்கிருந்து மதுக்கூருக்குத் தைிக்குடித்தைம் அதழத்து வந்துவிட்ோன் மாணிக்கம். நிதைமாத கர்ப்ெிணியாை செல்லாயிதய அவள் அக்கா ெரசு அதழத்துக்சகாண்டு பொைாள். ெரசு செல்லாயிதய நன்ைாக கவைித்துக் சகாண்ோள். அவள் கணவனும் அப்ெடிபய; குழந்தத ெிைக்கும் அைிகுைிகள் சதாிந்தவுேன் ெரசுவின் கணவன் சுக்குரு ஓடிப்பொய் குழந்ததப் பெறுொர்க்க அரும்ொயிதயயும், சரத்தைாதவயும் வண்ணாத்தி எை அதைவதரயும் அதழத்து வந்தான். செல்லாயிக்கு செண்குழந்தத ெிைந்தது குழந்ததப்பெற்ைின்சொழுது செண் உேல் ெடும் துன்ெங்கதள, என்று தன்மைக்குதைதய ஒப்ொாியாகப் ொடிைாள். கணவன் கண்டு சகாள்ளாமல் சென்றுவிடுகிைான்; செல்லாயி நான்கு மாத கர்ப்ெமாக இருக்கிைாள். இரவு சவகு பநரம் ஆைாலும் மாணிக்கம் வீட்டிற்கு வரமாட்ோன் செல்லாயி தன்தை தன் அப்ொவீட்டில் சகாண்டுபொய் விேச்சொல்லி அழுகிைாள். அப்ொவீட்டில் விடுவதாகக் கூைிவிட்டு கர்ப்ெிணி என்றும் ொராமல் நேந்பத நீண்ேதூரம் அதழத்துச்சென்ைான்; நேக்கமுடியாமல் செல்லாயி உட்கார்ந்பதவிட்ோள். மணலி சென்று அங்கிருந்து மதுக்கூருக்குத் தைிக்குடித்தைம் அதழத்து வந்துவிட்ோன் மாணிக்கம். ்்்்்ு்ு்்ு நிதைமாத கர்ப்ெிணியாை செல்லாயிதய அவள் அக்கா ெரசு அதழத்துக்சகாண்டு பொைாள். ெரசு செல்லாயிதய நன்ைாக கவைித்துக் சகாண்ோள். அவள் கணவனும் அப்ெடிபய; குழந்தத ெிைக்கும் அைிகுைிகள் சதாிந்தவுேன் ெரசுவின் கணவன் சுக்குரு ஓடிப்பொய் குழந்ததப் பெறுொர்க்க அரும்ொயிதயயும், சரத்தைாதவயும் வண்ணாத்தி எை அதைவதரயும் அதழத்து வந்தான். செல்லாயிக்கு செண்குழந்தத ெிைந்தது குழந்ததப்பெற்ைின்சொழுது செண் உேல் ெடும் துன்ெங்கதள, “என்ை இப்ெடி, நாலாப்ெக்கமும் கிழிஞ்ெி பொய்க் சகேக்கு “புள்ள சொிொருக்குல்ல, கிழியாம என்ை செய்யும்” “தகய வச்ெித் சதாேக்கபவ ெயமாருக்கு ோரு……. ோரா கிழிஞ்ெி பொயிக் சகேக்கு. இந்த புண்ண சயல்லாம் எப்ெடி ஆத்துை? “புண்ண ஆத்துைதாப் சொிெி. இந்தப் சொிய புள்ள முழுொ உசுபராே சொைந்துச்பெ… அதச் சொல்லு” (பமலது, ெக். 108, 109) எனும் உதரயாேல்கள் ெதிவு செய்கின்ைை. செண் ெிரெவத்தின்சொழுது எவ்வாசைல்லாம் துன்ெப்ெடுகிைாள் அவள் உேல் உறுப்புக்கள் எப்ெடிசயல்லாம் பெதமாகின்ைை என்ெததக் கூைி மைதத ெலைப் ெேதவக்கும் சு. தமிழ்ச்செல்வி அக்குழந்தததயப் ொராட்டி தாலாட்டி தூங்கதவப்ெதத, யாங்கண்பணராராபரா… ராாிரபரா பமக…த்து மின்சைாளிபயா….ஓ 21 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES To Discern the Truth is Wisdom Quarterly Journal    ISSN:2583-0139 Vol.3: Issue 1, January 2022    Vol.3: Issue 1, January 2022    ISSN:2583-0139 Vol.3: Issue 1, January 2022 வாதாடி… வாதாடி… நீ வாதாடி… வாதாடி… நீ வந்த சதரவியபமா…ஓ” (பமலது, ெக். 142, 143) வந்த சதரவியபமா…ஓ (பமலது, ெக். 142, 143) என்று தாலாட்டிப் ெிள்தளதயத் தூங்கதவக்கும் செல்லாயி தன் ெிள்தளகதள வளர்க்க ெடும்ொடு சொல்லி மாளாதது. தைிக் குடித்தைம் சென்ை மாணிக்கம் மருத்துவத் சதாழிதல விட்டுவிட்டு ெர்க்கஸ்பொல தெக்கிள் ஓட்டி வித்தத காட்டுகிைான். நிதைய ெணம் பெர்ந்த உேன் அவ்வளதவயும் குடி ெீட்டு எை அதில் விட்டு விடுகிைான். அம்தம பநாயால் தாக்கப்ெட்டு நேக்கவும் முடியாமல் இருக்கும் கணவனுேன் தகக்குழந்தததயயும் அதழத்துக் சகாண்டு தன் சொந்த ஊராை கற்ெகநாதர் குளத்திற்குப் பொய்விே பவண்டும் என்று நிதைத்த செல்லாயி, “நம்மளுக்கு ஊருல சநலம் ெலம் இல்லயா சொத்து சொவம் இல்லயா? ொப்ொரும் பகப்ொரும் இல்லாம இஞ்ெ யாங்சகேத்து தவிக்கணும்”( பமலது, ெ. 137) என்று எண்ணி எப்ெடியாவது கணவதை ெதழயெடி சதம்புதேயவைாக மாற்ைி விே பவண்டும் எை நிதைத்து மாமியார் வீோை கற்ெகநாதர் குளம் வந்தாள். மாமியார் தங்கத்தாச்ெி தைக்கு மகனுமில்தல மருமகளுமில்தல என் உயிர் இருக்கும் வதர வீட்டிற்குள் யாதரயும் விேமாட்பேன் எைக் கூைி மாட்டுக் சகாட்ேதகயில் இருக்கும்ெடி கூைிவிட்ோள். பமற்கூதர மட்டும் பொேப்ெட்டிருந்தது. ெக்கத்தில உள்ள செண்ணிேம் வயிற்றுப்ெெிக்காக யாெகம் பகட்கிைாள். பொளத்ததயும் ஒரு ஓட்தேச்ெட்டிதயயும் செற்று வந்து மாவதரத்து உப்ெில்லா உணவு ெதமத்து ெடுக்தகயில் இருக்கும் கணவனுக்கு பொடுகிைாள். என்று தாலாட்டிப் ெிள்தளதயத் தூங்கதவக்கும் செல்லாயி தன் ெிள்தளகதள வளர்க்க ெடும்ொடு சொல்லி மாளாதது. தைிக் குடித்தைம் சென்ை மாணிக்கம் மருத்துவத் சதாழிதல விட்டுவிட்டு ெர்க்கஸ்பொல தெக்கிள் ஓட்டி வித்தத காட்டுகிைான். நிதைய ெணம் பெர்ந்த உேன் அவ்வளதவயும் குடி ெீட்டு எை அதில் விட்டு விடுகிைான். அம்தம பநாயால் தாக்கப்ெட்டு நேக்கவும் முடியாமல் இருக்கும் கணவனுேன் தகக்குழந்தததயயும் அதழத்துக் சகாண்டு தன் சொந்த ஊராை கற்ெகநாதர் குளத்திற்குப் பொய்விே பவண்டும் என்று நிதைத்த செல்லாயி, “நம்மளுக்கு ஊருல சநலம் ெலம் இல்லயா சொத்து சொவம் இல்லயா? ொப்ொரும் “நம்மளுக்கு ஊருல சநலம் ெலம் இல்லயா சொத்து சொவம் இல்லயா? ொப்ொரும் பகப்ொரும் இல்லாம இஞ்ெ யாங்சகேத்து தவிக்கணும்”( பமலது, ெ. 137) என்று எண்ணி எப்ெடியாவது கணவதை ெதழயெடி சதம்புதேயவைாக மாற்ைி விே பவண்டும் எை நிதைத்து மாமியார் வீோை கற்ெகநாதர் குளம் வந்தாள். மாமியார் தங்கத்தாச்ெி தைக்கு மகனுமில்தல மருமகளுமில்தல என் உயிர் இருக்கும் வதர வீட்டிற்குள் யாதரயும் விேமாட்பேன் எைக் கூைி மாட்டுக் சகாட்ேதகயில் இருக்கும்ெடி கூைிவிட்ோள். பமற்கூதர மட்டும் பொேப்ெட்டிருந்தது. ெக்கத்தில உள்ள செண்ணிேம் வயிற்றுப்ெெிக்காக யாெகம் பகட்கிைாள். பொளத்ததயும் ஒரு ஓட்தேச்ெட்டிதயயும் செற்று வந்து மாவதரத்து உப்ெில்லா உணவு ெதமத்து ெடுக்தகயில் இருக்கும் கணவனுக்கு பொடுகிைாள். இததக் களியாய் கிண்டிைால் அப்ெடிபய பொட்டு கிண்ே பவண்டியது தான் கூழாய்க் காய்ச்ெிைால் ொதி மாவு பொட்ோல் பொதும். இததபய இரண்டு பவதளக்கு தவத்துக் சகாள்ளலாம் எை நிதைத்தாள். கூழாய்க் காய்ச்ெிைால் ஊற்ைி குடிக்க ேம்ளர் பவண்டுபம. ேம்ளருக்கு எங்பக பொவது? களியாகபவ கிண்டிவிேலாம் எை பயாெித்தவளாய் எல்லாமாதவயும் சகாட்டி கிண்டிைாள். வாய்க்கால் ஓரத்தில் பெத்தத எருக்தகபொல குத்து குத்தாய் வளர்ந்திருந்தது. பெத்ததயின் அடித்தண்டில் ெிமிழ் பொன்ை ெச்தெக்காய்கள் காய்த்திருந்தை. அகலமாை இதலயாய் ொர்த்து ஏசழட்டு பெத்ததயிதலகதள இணிக்கி வந்தாள். இதலகதள கழுவி தவத்தாள். மாணிக்கத்தத எழுப்ெிைாள். மாணிக்கத்தின் முன்ைால் இன்சைாரு கீற்தை எடுத்து வந்து சொட்டு அதில் ஈர இதலகதள இதேசவளியில்லாமல் வட்ேமாய் அடுக்கிைாள். அதில் பொளக்களிதயப் பொட்டு ஆப்தெயால் இழுத்துவிட்டு ஆைச் செய்தாள். PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES மாணிக்கம் ஒருவாய் எடுத்துப்பொட்ோன் களி உப்ெில்லாமல் ெப்சென்ைிருந்தது. உப்ெில்லாமல் எப்ெடி ொப்ெிடுவது? என்று பகட்ோன். இவற்தைசயல்லாம் நான் யாெகம் செற்று செய்பதன் ஒவ்சவான்ைாக பகட்டுப் செற்ைதவ இரவாைதால் உப்தெ பகட்கவில்தல என்று செல்லாயி கூை, உப்ெில்லா பொளக்களிதய ெெியின் சகாடுதமயால் சதன்ைங்குரும்ெதயக் கடித்துக்சகாண்டு ொப்ெிட்ோர்கள் குழந்ததக்கும் அததபய 22 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES  Vol.3: Issue 1, January 2022 To Discern the Truth is Wisdom Quarterly Journal ISSN:2583-0139 22 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES ISSN:2583-0139      ISSN:2583-0139 Vol.3: Issue 1, January 2022 ஊட்டிைாள். வாைம் இருட்டியது எங்கும் ஒபர அதமதி. ஊர் முழுவதும் தூங்கிக் சகாண்டிருக்க மதழ பொசவைப் செய்ததைால் மாட்டுக் சகாட்ேதக முழுவதும் தண்ணீர் செருக்சகடுத்து ஓடியது. ெிள்தள கிேந்த சதாட்டிலுக்கு மதலச்ொரலடித்தது சதாட்டில் கட்டியிருந்த ெீதல ஈரமாகியது. இருந்தாலும் ெிள்தள அழாமல் தூங்கிக் சகாண்டிருந்தது. நாலாப்ெக்கமும் அடித்த ொரலும் பமற்கூதையிலிருந்து ஒழுகிய தண்ணீரும் செல்லாயி மாணிக்கம் இரண்டு பெதரயும் முழுதாய் நதைத்தது. மாணிக்கத்திற்கு எடுத்த நடுக்கம் விடியும் வதர நிற்கவில்தல. குளிருக்கு ஈரத்துணிதயபய பொர்த்திக் சகாண்ோன். தன் கணவனுக்கு எதுவும் ஆகக் கூோது என்று மாவிளக்குப் பொடுவதாக மாாியம்மதை பவண்டுகிைாள். ்்ு்்ுு்      ISSN:2583-0139 Vol.3: Issue 1, January 2022    ISSN:2583-013 Vol.3: Issue 1, January 2022 ISSN:2583-0139 ஊட்டிைாள். வாைம் இருட்டியது எங்கும் ஒபர அதமதி. ஊர் முழுவதும் தூங்கிக் சகாண்டிருக்க மதழ பொசவைப் செய்ததைால் மாட்டுக் சகாட்ேதக முழுவதும் தண்ணீர் செருக்சகடுத்து ஓடியது. ெிள்தள கிேந்த சதாட்டிலுக்கு மதலச்ொரலடித்தது சதாட்டில் கட்டியிருந்த ெீதல ஈரமாகியது. இருந்தாலும் ெிள்தள அழாமல் தூங்கிக் சகாண்டிருந்தது. நாலாப்ெக்கமும் அடித்த ொரலும் பமற்கூதையிலிருந்து ஒழுகிய தண்ணீரும் செல்லாயி மாணிக்கம் இரண்டு பெதரயும் முழுதாய் நதைத்தது. மாணிக்கத்திற்கு எடுத்த நடுக்கம் விடியும் வதர நிற்கவில்தல. குளிருக்கு ஈரத்துணிதயபய பொர்த்திக் சகாண்ோன். தன் கணவனுக்கு எதுவும் ஆகக் கூோது என்று மாவிளக்குப் பொடுவதாக மாாியம்மதை பவண்டுகிைாள். ஊட்டிைாள். வாைம் இருட்டியது எங்கும் ஒபர அதமதி. ஊர் முழுவதும் தூங்கிக் சகாண்டிருக்க மதழ பொசவைப் செய்ததைால் மாட்டுக் சகாட்ேதக முழுவதும் தண்ணீர் செருக்சகடுத்து ஓடியது. செண் உதழப்பும் மாமியார் சகாடுதமயும் ெொெதியின் உதவிபயாடு தங்கத்தாச்ெியிேமிருந்து சகாஞ்ெம் நிலம் செற்று மிளகாய் மரவள்ளி எை விவொயம் செய்தாள். மாணிக்கம் உேல் குணமதேந்தது. கிதேக்கும் பவதலக்கு செல்லாயி சென்று ெிறுகெிறுக ெணம் பெர்த்து ஆட்டுக் குட்டிகதளயும், மாட்தேயும் வாங்கிைாள் எப்சொழுதும் ஓயாமல் பவதல செய்துசகாண்பே இருந்தாள் இரண்ோவதாக செண் ெிள்தள ெிைந்தது. செல்லாயியின் உதழப்ெிைால் வறுதமயின்ைி வாழ்ந்தார்கள். திருமணத்திற்கு முன்பு மாணிக்கம் ெக்குொய் என்னும் ஒரு செண்தணக் காதலித்தான் அம்மாவின் வற்புறுத்தலால் விருப்ெமின்ைி செல்லாயிதயத் திருமணம் செய்துசகாண்டு இரண்டு செண்ெிள்தளகளும் ெிைந்துவிட்ேை. ெக்குொய் யாதரயும் திருமணம் செய்து சகாள்ளாமல் தன்தைபய நிதைத்திருப்ெதாக பகள்விெட்ே மாணிக்கம் ெக்குொதய திருமணம் செய்து சகாள்வதாகக் கூைி தைியாக வரச்சொல்கிைான். செல்லாயி பொகவிோமல் தடுத்ததால் மாணிக்கத்தால் பொக முடியவில்தல மாணிக்கம் வராததால் ெக்குொய் தூக்கு பொட்டு இைந்துவிடுகிைாள். செல்லாயியின் கணவன் மாணிக்கம் தன்ைால் ெக்குொய் இைந்துவிட்ோள் எை நிதைத்து ஊதரவிட்பே பொய்விடுகிைான். ஒரு வருேம் ஆகியும் மாணிக்கத்திேமிருந்து எந்த தகவலும் வராதததப் ெயன்ெடுத்தி மருமகதள வீட்தே விட்டு துரத்துவதற்கு முடிவு செய்த மாமியாளும் நாத்தைார்களும் சராட்டிக்காரனுேன் செல்லாயிக்குத் தவைாை சதாேர்பு இருக்கிைது எைக் கூைி ெஞ்ொயத்து தவத்துவிடுகிைாள். இததை, “எங்க மருமவ, மவன் இல்லாத பநரத்துல அேக்க ஒடுக்கமா இருக்காம, நம்ம ஊருக்கு வர்ர சராட்டிக்காரங்கிட்ே ெிாிச்ெிெிாிச்ெி பெசுைா. அவன் புள்ளசவாளுக்கு ொக்சகட்டு ொக்சகட்ோ சராட்டி குடுக்குைான். இதசயல்லாம் ொத்துகிேகு ஒண்ணும் பகட்க முடியாம நாங்க இருக்குைம். எங்களுக்கு அவபமபல ெந்பதகமாருக்கு. அவள இழுத்துவச்ெி என்ை ஏதுன்னு நீங்கத்தான் விொாிக்கணும்”( பமலது, ெ. 178) என்று ஊர் சொியவர்களிேம் ெஞ்ொயத்து தவக்கிைாள். செல்லாயி தம் ெிள்தளகளின் மீது ெத்தியம் செய்தும் அந்த சராட்டிக்காரன் என் தம்ெிமாதிாி, நான் ெிைந்த ஊதரச்பெர்ந்தவன் கற்ெகநாதர்குளம் பொகும் பொசதல்லாம் பெரக்குழந்ததகளுக்கு 23 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES  Vol.3: Issue 1, January 2022 To Discern the Truth is Wisdom Quarterly Journal ISSN:2583-0139 23 Vol.3: Issue 1, January 2022     ISSN:2583-0139 Vol.3: Issue 1, January 2022     Vol.3: Issue 1, January 2022 ISSN:2583-0139 Vol.3: Issue 1, January 2022 சராட்டி சகாடுக்கும் ெடி என் தந்தத கூைியதைால் தான் தருகிைான் என் தந்ததபய சராட்டிக்காரனுக்கு ெணம் தருகிைார் எை எவ்வளபவா கூைியும், சராட்டி சகாடுக்கும் ெடி என் தந்தத கூைியதைால் தான் தருகிைான் என் தந்ததபய சராட்டிக்காரனுக்கு ெணம் தருகிைார் எை எவ்வளபவா கூைியும், உன்னுதேய நேப்பு ொியில்லன்னு புகார், நீ என்ை சொல்கிைாய்? என்று பகட்ோர்கள் ெஞ்ொயத்தார்கள். உன்னுதேய நேப்பு ொியில்லன்னு புகார், நீ என்ை சொல்கிைாய்? என்று பகட்ோர்கள் ெஞ்ொயத்தார்கள். தான் எந்த தவறும் செய்யயவில்தல என்று மறுத்துக் கூைி ஆபவெம் வந்தவளாய் புலம்ெிய செல்லாயிதய “….ச்ெீ வாயமூடு நாபய….. இதுக்கு பமபல ஒரு வார்த்த பெசுை செருப்ொலபய அடிச்ெி ஊரவுட்டுத் சதாரத்திடுபவன் ஆமா’ சொல்லிக்சகாண்பே குைிந்து காலிலிருந்து செருப்தெக் கழட்டிைார் ெிரெிசரண்ட் சுந்தரம்ெிள்தள”( பமலது, ெ. 179) குடும்ெத்தில் உள்ளளவர்கள் மட்டுமன்ைி ெமூகமும் செல்லாயிதய சகாடுதமப்ெடுத்தியதால் பவண்டுமளவு அழுது கதளத்து ஓய்ந்தாள். ஊதரவிட்டுப் பொகலாம் எை நிதைத்தாள். நாம் பொைால் பகவலமாகும் பொகக்கூோது எை நிதைத்து சொியமகள் வாணியிேம் ெின்ைமகள் மணிதய விட்டுவிட்டு விவொய பவதலக்கும் பொைாள் பவதல முடிந்தவுேன் விைகுசொறுக்கிக் சகாண்டு வீட்டிற்கு வந்தாள் உச்ெந்ததலமுடி எண்தணயின்ைி ெைந்தது. உண்ணாமலும் உைங்காமலும் தன் ெிள்தளகளுக்காக ஓய்வின்ைி உதழத்தாள். “ச்ெீ….. இந்தத் ததலயப் ொரு ொண்தேயாட்ேம் ெைக்குைத எை தன் ததலமுடிதய இரண்டு தககளாலும் அழுத்தி வழித்து ெின்ைால் சொரடிக் சகாண்தேக்குள் செருகிைாள். அது சொம்மிக் சகாண்டு பமபல வந்தது. சகாஞ்ெைாச்சும் எண்சணப்ெெ இருந்தாதாை எங்க சொருவுைாலும் நிக்கும் தல; எண்சணயப் ொத்து எட்டு நாளக்கி பமலாவது.”( பமலது, ெக். செண் உதழப்பும் மாமியார் சகாடுதமயும் 187, 188) இவ்வாிகள் செண்ெடும் வருதமதய சுட்டிச்செல்கிைது. திடீசரை, மாணிக்கம் வந்துவிடுகிைான். வந்ததும் வராததுமாய் ெின்ைத்தம்ெியும், தங்கத்தாச்ெியும் செல்லாயி நேத்ததக்சகட்ேவள் என்று சொல்லி விட்டு அவதள வீட்தே விட்டு துரத்தும்ெடி கூைிைார்கள்; என்ை ஏசதன்று விொாிக்காமல். திடீசரை, மாணிக்கம் வந்துவிடுகிைான். வந்ததும் வராததுமாய் ெின்ைத்தம்ெியும், தங்கத்தாச்ெியும் செல்லாயி நேத்ததக்சகட்ேவள் என்று சொல்லி விட்டு அவதள வீட்தே விட்டு துரத்தும்ெடி கூைிைார்கள்; என்ை ஏசதன்று விொாிக்காமல். புருெங்காரன் வூட்டுலயில்லன்ைா வூருபமயிை ஆெ வந்துருமா? த்தூ நீ எல்லாம் ஒரு சொம்புளயா? என்று நீண்ே நாள் கழித்து வந்த கணவன் கூை செல்லாயி யாரு சொன்ைது? தான் ஒரு தவறும் செய்யவில்தல என்று பொராடிைாள். யாருடி சொல்லனும்? யாரு சொல்லணும் இந்த வூபர தான் ஓம்பமாதரயில காாித்துப்புபத என்று கூைிய ெடி ெெியாலும் ெட்டிைியாலும் ஒட்டிப்பொயிருந்த அவளுதேய வயிற்ைில் எட்டி உததத்தான். கத்தியெடி வயிற்தைப் ெிடித்துக்சகாண்டு சுருண்டு விழுந்தாள். சதாட்டு தாலி கட்டிய கணவனுக்கு துபராகம் செய்யும் நீசயல்லாம் இருந்தா என்ை? செத்தாசயன்ை? ொவு ததலமுடிதய ெிடித்துத்தூக்கி இரண்டு கன்ைங்களிலும் ஓங்கி அதைந்து துன்ெப்ெடுத்திைான். வலிதாங்காமல் ஓடிப்பொய் கிணற்ைில் விழுந்தாள். அப்சொழுதும் விோமல் கிணற்ைில் இைங்கி மண்தேதய உதேத்ததால் கிணற்று நீர் ெிவப்ெதேந்தது. அடித்துக் அடித்து கதளத்துப் 24 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES  Vol.3: Issue 1, January 2022 To Discern the Truth is Wisdom Quarterly Journal ISSN:2583-0139 24 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES Vol.3: Issue 1, January 2022      ISSN:2583-0139 Vol.3: Issue 1, January 2022 பொய்விட்ோன். ஆத்திரம் அேங்காதவளுக்கு அழுதக பீைிட்டு வந்தது பொர்ந்து உட்கார்ந்து ஒப்ொாி தவத்தாள், ் ISSN:2583-0139 Vol.3: Issue 1, January 2022 பொய்விட்ோன். ஆத்திரம் அேங்காதவளுக்கு அழுதக பீைிட்டு வந்தது பொர்ந்து உட்கார்ந்து ஒப்ொாி தவத்தாள், ் பொய்விட்ோன். ஆத்திரம் அேங்காதவளுக்கு அழுதக பீைிட்டு வந்தது பொர்ந்து உட்கார்ந்து ஒப்ொாி தவத்தாள், பொய்விட்ோன். ஆத்திரம் அேங்காதவளுக்கு அழுதக பீைிட்டு வந்தது பொர்ந்து உட்கார்ந்து ஒப்ொாி தவத்தாள், பொய்விட்ோன். ஆத்திரம் அேங்காதவளுக்கு அழ ஒப்ொாி தவத்தாள், மாயவரஞ் ொலயிபல – நா மயிசலைங்கி பமயயிபல…. மயிலுண்ணும் ொராம மருந்து பொட்டு சுட்ோபவா மருந்து சவடிக்குதுன்னு மதுசலாரம் பொய் மைஞ்பென் மருந்து பமல மருந்து வச்பெ – எைக்கு அதிமருந்பத கூே வச்பெ… கும்ெபகாணஞ்ொலயிபல – நா குயிசலைங்கி பமயயிபல… குயிலுண்ணும் ொராம குண்டு பொட்டு சுட்ோபவா குண்டு சவடிக்குதுன்னு நா சகாளத்பதாரம் பொய் ஒளிஞ்பென் குண்டுபமல குண்ேவச்பெ – எைக்கு கூே மருந்து வச்பெ (பமலது, ெக். 191, 192) கூே மருந்து வச்பெ (பமலது, ெக். 191, 192) என்றுதன் மைக்குதைதய எல்லாம் ஒப்ொாியாக ொடி அழுதுவிட்டு ெிைந்த வீட்டிற்கு வந்து அப்ொ என்று கதைிக்சகாண்பே அவர் கால்களில் விழுந்தாள். சகட்டியாக அவருதேய கால்கதளக் கட்டிப் ெிடித்துக்சகாண்டு தன் மைச்சுதமதயக் கண்ணீராய் வடித்தாள். 25 ெிள்தளகதள விட்டு விட்டு ஒருநாள் கூே செல்லாயியால் இருக்கமுடியவில்தல ெிள்தளகதள நிதைத்து அழுதுசகாண்பே இருந்தாள். ஒருநாள் பொைது ஒரு வருேமாக இருந்தது. திருத்துதைப்பூண்டியில் இருந்து மீண்டும் கற்ெகநாதர்குளம் வந்தாள். செல்லாயியின் உதழப்பும் அவள் ெடும் துன்ெமும் சதாேர்ந்துசகாண்பே இருந்தது. செல்லாயியின் உதழப்ெிைால் மாட்டுக் சகாட்ேதகதய சுவர் தவத்து வீடு பொல் செய்து வாழ்வது, சொறுக்காத மாமியார் மாணிக்கத்திேம் இல்லாததத இருப்ெது பொல் திாித்துப் பெெி உண்தமப்பொல் நிரூெித்துவிடுவாள்; செல்லாயி என்ைச் சொன்ைாலும் அததக்பகட்காத கணவன் அடித்து உததத்து வீட்டிற்குள்பள தள்ளி சநருப்பு தவத்துவிட்ோன். தான் கஷ்ேப்ெட்டு பெர்த்து தவத்ததவகளுேன் தானும் எாிந்து ொம்ெலாகப் பொகிபைாபம எை நிதைத்தவள் தகயில் கிதேத்தவற்தை எடுத்துக்சகாண்டு சவளிபய வந்தாள். தீப்ெிடித்து எாிந்தது, வீடு ொம்ெலாகியது. மாமியார் கணவன் எை மாைிமாைி துன்ெப்ெடுத்தப்ெட்ோள் ஊர் ெைங்களுக்பக இவர்கள் செய்யும் சகாடுதம தாங்காமல், மாமியார் இருக்கும் வதர செல்லாயி நிம்மதியா வாழமுடியாது. அவள் வாழவிேமாட்ோள், சகாதலகாரப் ொவி அவளால் ஏற்ெட்ே ெிரச்ெிதைகள்தான் 25 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES  Vol.3: Issue 1, January 2022 To Discern the Truth is Wisdom Quarterly Journal ISSN:2583-0139 25 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES 26 PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES  Vol.3: Issue 1, January 2022 To Discern the Truth is Wisdom Quarterly Journal ISSN:2583-0139 26    ISSN:2583-0139 Vol.3: Issue 1, January 2022 ISSN:2583-0139 Vol.3: Issue 1, January 2022 துதணநின்ை நூல்கள் Cite this Article in English Malarvizhi. V “ Su. Tamilselvi’s Female narrators in the novel of Manickam” Pulam: International Journal of Tamilology Studies, Vol.3 Issue 1, January 2023, pp.18-27 Malarvizhi. V “ Su. Tamilselvi’s Female narrators in the novel of Manickam” Pulam: International Journal of Tamilology Studies, Vol.3 Issue 1, January 2023, pp.18-27 Malarvizhi. V “ Su. Tamilselvi’s Female narrators in the novel of Manickam” Pulam: International Journal of Tamilology Studies, Vol.3 Issue 1, January 2023, pp.18-27 27 துதணநின்ை நூல்கள் துதணநின்ை நூல்கள் 1. அகிலன், கததக்கதல, ொாி புத்தகப் ெண்தண, 1976, சென்தை. 1. அகிலன், கததக்கதல, ொாி புத்தகப் ெண்தண, 1976, சென்தை. 2. அரங்கமல்லிகா, தமிழ் இலக்கியமும் செண்ணியமும், காலச்சுவடு ெதிப்ெகம், 2002, சென்தை. 3. ெம்ெத்குமார். ொ, மு.வ. நாவல்களில் ொத்திரங்கள் ஒரு திைைாய்வு, காவ்யா சவளியீடு, செங்களூர். 4. தமிழ்ச்செல்வி. சு, மாணிக்கம், நியூ செஞ்சுாி புக் ஹவுஸ், முதற்ெதிப்பு 2007, சென்தை -98 ்்்்்்்்்் 4. தமிழ்ச்செல்வி. சு, மாணிக்கம், நியூ செஞ்சுாி புக் ஹவுஸ், முதற்ெதிப்பு 2007, சென்தை -98 5. பதவதத்தா, செண் ெதேப்ெில் செண்கள், அன்தை சதரொ ெல்கதலக்கழகம், சகாதேக்காைல்1986 5. பதவதத்தா, செண் ெதேப்ெில் செண்கள், அன்தை சதரொ ெல்கதலக்கழகம், சகாதேக்காைல், 1986. Reference Reference 1. Akilan, Kathaikkalai, Pari Book Center, 1976, Chennai. . Akilan, Kathaikkalai, Pari Book Center, 1976, Chennai. 2. Arangamallika, Thamizh Ilakkiyamum Penniyamum, Kalachuvadu Publicati 3. Sambathkumar, B, Mu.Va Navalgalil Paththirangal Oru Thiranayvu, Kavya, Bengaluru. 3. Sambathkumar, B, Mu.Va Navalgalil Paththirangal Oru Thiranayvu, Kavya, Bengaluru. 4. Tamilselvi, Su, Manickam, NCBH, Chennai, 2007. 5. Thevathatha, Pen Padaippil Pengal, Annai Therasa University, Kaodaikanal, 1986. தமிழில் இக்கட்டுதரயின் பமற்பகாள் 5. Thevathatha, Pen Padaippil Pengal, Annai Therasa University, Kaodaikanal, 1986. தமிழில் இக்கட்டுதரயின் பமற்பகாள் மலர்விழி. சவ “ சு.தமிழ்ச்செல்வி - மாணிக்கம் நாவலில் செண் கதத மாந்தர்கள்” புலம் : ென்ைாட்டுத் தமிழியல் ஆய்விதழ், சதாகுதி 3, இதழ் 1, ெைவாி 2023, ெக். 18-27 Cite this Article in English PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES ISSN:2583-0139    ISSN:2583-0139 Vol.3: Issue 1, January 2022      ISSN:2583-0139 Vol.3: Issue 1, January 2022      ISSN:2583-0139 Vol.3: Issue 1, January 2022 இவ்வளவும் என்று ஊர்மக்கள் ஒன்று பெர்ந்து ெஞ்ொயத்து தவத்து மாமியார் இருந்த வீட்தே செல்லாயிக்கு வாங்கி சகாடுத்தார்கள். மீண்டும் வீட்டில் ஒரு சொருள் கூே இல்லாமல் எப்ெடி வாழ்வது என்று மைம் தளராமல் செல்லாயி உதழத்து ஒவ்சவான்ைாய்ச் பெர்த்து குடும்ெம் நேத்திைாள் மூன்ைாவதாக ஆண்குழந்தத ெிைந்தது.    Vol.3: Issue 1, January 2022 ISSN:2583-0139 Vol.3: Issue 1, January 2022 இவ்வளவும் என்று ஊர்மக்கள் ஒன்று பெர்ந்து ெஞ்ொயத்து தவத்து மாமியார் இருந்த வீட்தே செல்லாயிக்கு வாங்கி சகாடுத்தார்கள். மீண்டும் வீட்டில் ஒரு சொருள் கூே இல்லாமல் எப்ெடி வாழ்வது என்று மைம் தளராமல் செல்லாயி உதழத்து ஒவ்சவான்ைாய்ச் பெர்த்து குடும்ெம் நேத்திைாள் மூன்ைாவதாக ஆண்குழந்தத ெிைந்தது. இவ்வளவும் என்று ஊர்மக்கள் ஒன்று பெர்ந்து ெஞ்ொயத்து தவத்து மாமியார் இருந்த வீட்தே செல்லாயிக்கு வாங்கி சகாடுத்தார்கள். மீண்டும் வீட்டில் ஒரு சொருள் கூே இல்லாமல் எப்ெடி வாழ்வது என்று மைம் தளராமல் செல்லாயி உதழத்து ஒவ்சவான்ைாய்ச் பெர்த்து குடும்ெம் நேத்திைாள் மூன்ைாவதாக ஆண்குழந்தத ெிைந்தது. அததைத் சதாேர்ந்து செண்குழந்தத ஒன்றும் ெிைந்தது, வாணி, மணி, நாஞ்ெி என்ை மூன்று குழந்ததகளுக்குப் ெின்பு ெின்ைது ெிைந்ததும் செல்லமாக வளர்க்கப்ெட்ோள். ெள்ளியில் பெர்த்து ெடிக்கதவத்தைர், நாஞ்ெியும் ஐ.டி.ஐ ெடித்திருந்தான். ெின்ைது ெள்ளிவிட்டு வீட்டிற்கு வராமல் குளத்தில் விதளயாடிக் சகாண்டிருக்கும்சொழுது பெற்ைில் மாட்டி இைந்து விடுகிைாள். துன்ெம் தாங்காமல் அழுதுசகாண்பே மயங்கிவிழுந்தாள். மற்ை மூன்று ெிள்தளகளுக்கும் திருமணம் செய்துதவத்துவிட்டு குடிகாரக் கணவன் மாணிக்கத்துேன் தைியாக வாழ்ந்த செல்லாயி, வயதாை ெின்பும் ஓயாமல் விவொய பவதலக்குச் செல்வதுேன் ஆடு, மாடு எை ெராமாித்து வந்தாள். மீன்ெிடி சதாழில் செய்து நல்ல வருமாைம் வந்தும் மாணிக்கம் குடித்பத அழிந்தான். அத்துேன் அவன் செல்லாயிதய அடிப்ெததயும் நிறுத்தவில்தல. தன் இதளய மகதள மைக்க முடியாமல் அவதளப் புததத்த அளம் பநாக்கிச் சென்ைாள். தன் ெிள்தளயின் பமல் சகாண்டிருக்கும் ொெம், மாெற்ை அவளின் அன்பு ஆவலாய் மாைி, அந்த உணர்வுகளதைத்தும் இரத்தத்பதாடு கலந்துவிட்ேது. அவள் எண்ணம் முழுவதும் அளத்ததபய சுற்ைி சுற்ைி வந்தது. தன் குழந்தததயப் ொர்த்து சநஞ்பொடு அதைத்து அவபளாடு காற்ைாய் கதரந்து பொகபவண்டும் என்று தன் குழந்தததய காணச் சென்ை செல்லாயி அளத்திபல உட்கார்ந்திருந்தாள். மிகுந்த ெலத்துேன் ஏபதா ஒன்று ெின்ைாலிருந்து அவதளத் தள்ளிவிட்ேது. குப்புை விழுந்து பெறும் ெகதியுமாக வீடு வந்து விழுந்தாள் நான்கு மாதமாகியும் ெடுத்தப்ெடுக்தகதயவிட்டு எழாமல் இருந்தாள்; ெிதததவ பநாக்கிப் ெயணித்த தைது குடும்ெத்ததக் காப்ொற்ை செல்லாயி ெட்ே துன்ெங்கள் சொல்லி மாளாது. சதாேர்ந்து பொராடி குடும்ெ அதமப்ெில் செண்ணின் இருண்ே சவளிதய சவளிச்ெமிட்டுக் காட்டும் இப்புதிைம் நவீை செண்ெமூகத்திற்கு செண்ணின் மை உறுதி, உதழப்பு, அன்பு, ொெம் முதலியவற்பைாடு, ெதளக்காமல் செண் சதாேர்ந்து பொராடும் ெக்தி சகாண்ேவள் என்ெதத எடுத்து இயம்புகிைது. முடிவுதர ஒரு புதிைம் ெிைப்ொக அதமய பவண்டுமாைால் கததயில் இேம்செறும் ொத்திரத்தின் ெதேப்பு ெிைப்புற்ைிருக்க பவண்டும். அவ்வாறு ெதேக்கும் ொத்திரங்களில் யாதரப் ெற்ைி ெிைப்ொக இேம்செைச் செய்கிைபதா அவர்கபள முதன்தம ொத்திரமாகக் கருதப்ெடுகின்ைைர். செல்லாயி தன் குடும்ெத்ததக் காப்ொற்ை மாமியார் கணவர் ெமூகம் எை அத்ததை துன்ெங்கதளயும் எதிர்சகாண்டு குடும்ெ அதமப்ெில் உயர்ந்த இேத்ததப்ெிடிப்ெதுேன் காவியத்தில் நிதலசெறுகிைாள். PULAM: INTERNATIONAL JOURNAL OF TAMILOLOGY STUDIES 022 To Discern the Truth is Wisdom Quarterly Journal ISSN:2583-0139
https://openalex.org/W4214575235	https://www.frontiersin.org/articles/10.3389/fendo.2022.843539/pdf	English	null	Acute Effects of Liothyronine Administration on Cardiovascular System and Energy Metabolism in Healthy Volunteers	Frontiers in endocrinology	2,022	cc-by	9,295	ORIGINAL RESEARCH published: 28 February 2022 doi: 10.3389/fendo.2022.843539 Shanshan Chen 1,2, George F. Wohlford 3, Alessandra Vecchie’ 4,5, Salvatore Carbone 6, Sahzene Yavuz 1, Benjamin Van Tassell 3, Antonio Abbate 4 and Francesco S. Celi 1* Shanshan Chen 1,2, George F. Wohlford 3, Alessandra Vecchie’ 4,5, Salvatore Carbone 6, Sahzene Yavuz 1, Benjamin Van Tassell 3, Antonio Abbate 4 and Francesco S. Celi 1* 1 Division of Endocrinology Diabetes and Metabolism, Virginia Commonwealth University, Richmond, VA, United States, 2 Department of Biostatistics, Virginia Commonwealth University, Richmond, VA, United States, 3 Department of Pharmacotherapy and Outcomes Sciences, Virginia Commonwealth University, Richmond, VA, United States, 4 Division of Cardiology, Virginia Commonwealth University, Richmond, VA, United States, 5 Department of Internal Medicine, ASST Sette Laghi, Varese, Italy, 6 Department of Kinesiology and Health Sciences, College of Humanities and Sciences, Virginia Commonwealth University, Richmond, VA, United States Edited by: Alessandro Antonelli, University of Pisa, Italy Context: The pharmacokinetics of liothyronine causes concerns for cardiovascular toxicity. While the effects of sustained increase in serum T3 concentrations are well described, little is known on the effects of acute changes in T3 concentrations due to rapid action of thyroid hormone. University of Pisa, Italy Reviewed by: Satish Kumar Bedada, SanoﬁIndia Limited, India Jens Mittag, University of Lübeck, Germany Correspondence: Francesco S. Celi fsceli@vcu.edu Reviewed by: Satish Kumar Bedada, SanoﬁIndia Limited, India Jens Mittag, University of Lübeck, Germany University of Lübeck, Germany Correspondence: Francesco S. Celi fsceli@vcu.edu Objective: To assess the clinical relevance of transient increase of T3 levels on cardiovascular system and energy metabolism. Correspondence: Francesco S. Celi fsceli@vcu.edu Setting: Double-blind, three arms, placebo controlled, cross-over study (ClinicalTrials.gov Identiﬁer: NCT03098433). Specialty section: This article was submitted to Thyroid Endocrinology, a section of the journal Frontiers in Endocrinology Specialty section: This article was submitted to Thyroid Endocrinology, a section of the journal Frontiers in Endocrinology Received: 26 December 2021 Accepted: 02 February 2022 Published: 28 February 2022 Study Participants: Twelve volunteers (3 females, 9 males), age 27.7 ± 5.1 years. Intervention: Oral administration of liothyronine 0.7 mcg/kg, equimolar dose of levothyroxine (0.86 mcg/kg), or placebo in three identical study visits. Blood samples for total T3, free T4 were collected at times 0’, 60’ 120’ 180’ 240’. Continuous recording of heart rate, blood pressure, and hemodynamic data was performed using the volume clamp method. Resting energy expenditure was measured by indirect calorimetry. An echocardiogram was performed on each study visit at baseline and after the last blood sampling. Received: 26 December 2021 Accepted: 02 February 2022 Published: 28 February 2022 ORIGINAL RESEARCH published: 28 February 2022 doi: 10.3389/fendo.2022.843539 INTRODUCTION observations contribute to a renewed interest in the therapeutic and pharmacologic use of LT3 (8, 12, 13), with a speciﬁc interest in developing “new well-designed adequately powered clinical trial of combination therapy” (14). The goals of the treatment of hypothyroidism are “to achieve a state of euthyroidism and normalization of the circulating levels of TSH and thyroid hormones” (1), whereby TSH is used as a powerful and reliable proxy for euthyroidism owing to its robust correlation with free T4 concentration (2). This is commonly achieved with levothyroxine (LT4) which is relatively inexpensive, available in multiple strengths and administered in single dose. This strategy relies on the conversion of the prodrug LT4 into its active metabolite T3 to correct the lack of endogenous T3 production (estimated as 15% of the total circulating pool) from the thyroid gland (3). Experimental data (4, 5) indicate that LT4 alone is not sufﬁcient to restore tissue euthyroidism, and clinical observations (6) demonstrate that circulating levels of T3 are reduced in patients receiving LT4 therapy. Clinically, a signiﬁcant percentage of patients adequately treated with LT4 complain of residual symptoms which may be attributed to hypothyroidism (7). Collectively, these observations have prompted interest in LT4/Liothyronine (synthetic T3, LT3) combination therapy or desiccated thyroid extracts as means to correct for the loss of endogenous T3, and improve symptomatology (8). The pharmacokinetic characteristics of LT4 allow for once- daily administration. In contrast, LT3 has a short distribution half-life resulting in signiﬁcant changes in serum T3 concentrations (15). Thus, a once-daily administration regimen would result in post-absorptive peaks above the normal range (15). Since extended release formulations of LT3 are not available, in LT3/LT4 combination therapy a twice daily low- dose LT3 administration is recommended to minimize ﬂuctuations of serum T3 (13, 14). The tradeoff is a cumbersome regimen which is not convenient for a lifelong therapy, as well complicating the recruitment and retention in clinical studies (16). The main concern for cardiovascular complications due to transient rise in serum T3 concentrations above the normal range stems from extrapolations from chronic exposure to supraphysiologic doses (either endogenous or exogenous) of T3, and from the in vitro evidence of rapid, non-genomic effects of T3 in vascular endothelium and cardiomyocytes (17, 18), which would be directly exposed to ﬂuctuations in the T3 concentrations as a result of LT3 administration. Citation: Chen S, Wohlford GF, Vecchie’ A, Carbone S, Yavuz S, Van Tassell B, Abbate A and Celi FS (2022) Acute Effects of Liothyronine Administration on Cardiovascular System and Energy Metabolism in Healthy Volunteers. Front. Endocrinol. 13:843539. doi: 10.3389/fendo.2022.843539 Main Outcome Measures: Changes in cardiovascular function and energy expenditure. Results: Following the administration of liothyronine, serum T3 reached a Cmax of 421 ± 57 ng/dL with an estimated Tmax of 120 ± 26 minutes. No differences between study arms February 2022 \| Volume 13 \| Article 843539 1 Frontiers in Endocrinology \| www.frontiersin.org Chen et al. Acute Effects of Liothyronine were observed in heart rate, blood pressure, hemodynamics parameters, energy expenditure, and in echocardiogram parameters. Conclusions: The absence of measurable rapid effects on the cardiovascular system following a high dose of liothyronine supports the rationale to perform long-term studies to assess its safety and effectiveness in patients affected by hypothyroidism. Keywords: liothyronine, rapid effects of thyroid hormone, cardiovascular function, energy expenditure, pharmacokinetics, pharmacodynamics INTRODUCTION We hypothesized that if there is a clinically relevant rapid action of LT3, we would be able to detect measurable and clinically- signiﬁcant, temporal changes in cardiovascular function and energy expenditure following a high, single-dose LT3 (Figure 1). Otherwise, from the clinical perspective, the action Whereas most clinical studies have showed modest to nihil improvement in symptoms and quality of life (1), a plurality of patients appeared to prefer combination therapy (9, 10). Moreover, a secondary analysis of the largest study indicated that carriers of the Ala92 allele of the type 2 deiodinase gene showed a signiﬁcant improvement in quality of life (11). These B A B FIGURE 1 \| Study hypothesis. (A) If clinically relevant, rapid action of LT3 (dashed red line), would be measurable following a pharmacological dose at or immediately after Cmax. (B) Conversely, if the thyroid action is entirely attributable to genomic signaling (continuous red line) one would expect minimal measurable effects during the distribution phase with a gradual and delayed onset. Black continuous line: pharmacokinetics of single dose LT3. The shaded area represents the reference range for serum T3. B A FIGURE 1 \| Study hypothesis. (A) If clinically relevant, rapid action of LT3 (dashed red line), would be measurable following a pharmacological dose at or immediately after Cmax. (B) Conversely, if the thyroid action is entirely attributable to genomic signaling (continuous red line) one would expect minimal measurable effects during the distribution phase with a gradual and delayed onset. Black continuous line: pharmacokinetics of single dose LT3. The shaded area represents the reference range for serum T3. February 2022 \| Volume 13 \| Article 843539 Frontiers in Endocrinology \| www.frontiersin.org 2 Acute Effects of Liothyronine Chen et al. whole-room indirect calorimeter (WRIC), were ﬁtted with the Nexﬁn ™pulse-wave monitor, and a 40’ baseline recording of Energy Expenditure (EE) was conducted before ingesting the study drugs or placebo. Blood samples were collected at times 0’, 60’ 120’ 180’ 240’. Immediately after the last blood draw the volunteers underwent echocardiogram, and were discharged from the CRSU. All studies were conducted in the morning, and each individual participant had their studies performed at the same time of the day. We selected a 4-hour observation window to capture any PD events occurring immediately after the Cmax based on our knowledge of the PK characteristics of liothyronine (1). INTRODUCTION of LT3 would be entirely attributable to long-term transcriptional effects, and transient changes in serum T3 concentrations would not be a concern, supporting single daily administration of LT3 alone or in combination with LT4. Here we present a pharmacokinetics-pharmacodynamics study designed to assess the clinical relevance of rapid effects of T3, with a particular focus on myocardial function, systemic vascular resistance, and energy metabolism. Drug Formulation and Dosing ® The study design was double blind, controlled, crossover with two active formulations (LT3 and LT4), and placebo (ClinicalTrials.gov Identiﬁer: NCT03098433). The study was approved by the Virginia Commonwealth University IRB and all study participants provided written informed consent. All study procedures were conducted in the Clinical Research Service Unit (CRSU). Oral liquid formulation LT3 (Liotir® 100 mcg/ml), LT4 (Tirosint® 100 mcg/ml), and placebo were kindly donated by IBSA Institut Biochimique Lugano (Switzerland), and utilized under a (research) FDA Investigational New Drug (IND) (n.132993, Sponsor Francesco S. Celi). Study drugs were stored and dispensed by the VCU Investigational Pharmacy which performed the randomization and maintained the study blind. LT3 was administered at a weight-based dose of 0.7 mcg/kg (equivalent to a dose of 50 mcg in a 70 Kg individual), and LT4 was administered at an equimolar dose of 0.86 mcg/kg. Inclusion criteria were age 18-45 years, BMI 20-30 kg/m2, and TSH 0.5-5.0 mIU/mL. Exclusion criteria were: thyroid autoimmunity by history or positive anti-thyroid peroxidase (TPO) antibodies; pregnancy or lactation; hypothyroidism; use of prescription drugs; diabetes mellitus; dyslipidemia; coronary artery disease; hypertension; anemia; renal insufﬁciency; liver disease or ALT >2.5x the upper laboratory reference limit; psychiatric conditions; tobacco use. Assays Serum samples were separated on the day of the study and stored in a -80°C. Free T4, total T3 and TSH assays were processed in batch on an Abbott Architect i2000SR Immunoassay analyzer by the VCU Division of Clinical Pathology. Intra-assay variability for free T4 (reference range 0.7-1.5 ng/dL) was 2.3-3.8%; inter-assay variability was 3.6-5.2%. Intra-assay variability for total T3 (reference range 60-181 ng/dL) was 1.9-2.7%; inter-assay variability was 2.3-7.3%. Intra-assay variability for total TSH was 1.2-2.0%; inter-assay variability was 1.7-3.3%. All other assays Qualifying volunteers underwent three identical study visits separated by at least 48 hours. The study scheme and procedures are reported in Figure 2. Study participants were instructed to refrain from strenuous exercise the day before the study. Upon arrival to the CRSU after an overnight fast, study volunteers were ﬁtted with an i.v. cannula, vital signs were recorded, and an echocardiogram was performed. Next, the volunteers entered the FIGURE 2 \| Study design. Top panel: timeline of the study. After enrollment, study volunteers underwent three identical visits each separated by at least 48 hours. Bottom panel: study procedures. An echocardiogram was performed upon arrival to the Clinical Research Services Unit and after completion of the stay in the whole room indirect calorimeter. Energy expenditure was recorded for at least 30’ before the administration of the study drug or placebo. Five blood samples were collected to record the LT4 and LT3 pharmacokinetics. Blood pressure, heart rate, and hemodynamics data were collected throughout the stay in the whole room calorimeter. FIGURE 2 \| Study design. Top panel: timeline of the study. After enrollment, study volunteers underwent three identical visits each separated by at least 48 hours. Bottom panel: study procedures. An echocardiogram was performed upon arrival to the Clinical Research Services Unit and after completion of the stay in the whole room indirect calorimeter. Energy expenditure was recorded for at least 30’ before the administration of the study drug or placebo. Five blood samples were collected to record the LT4 and LT3 pharmacokinetics. Blood pressure, heart rate, and hemodynamics data were collected throughout the stay in the whole room calorimeter. February 2022 \| Volume 13 \| Article 843539 Frontiers in Endocrinology \| www.frontiersin.org Acute Effects of Liothyronine Chen et al. were performed on an Abbott Architect c8000 Clinical Chemistry analyzer with the exception of TPO antibodies (LabCorp). interaction effects. Study Population The accrual occurred between July, 2017 and June 2019, and twelve volunteers (3 females, 9 males, age 27.7 ± 5.1 years, weight 75.0 ± 12.9 Kg) completed the study. LT3 and LT4 doses were 52.5 ± 9.0 mcg, and 64.5 ± 11.1 mcg, respectively. No adverse event was reported, and study volunteers did not report any subjective difference in well-being, anxiety, thermoregulation, or heart rate among the three interventions. The study population characteristics are reported in Table 1, while the recruitment details are reported in Figure 3 (CONSORT chart). Assays The ﬁrst 20~40 minutes in the WRIC were indexed as 0 and the following minutes were indexed as [1, 2, …,240], such that the baseline EE and baseline hemodynamic data were aggregated and adjusted as intercepts in the linear mixed- effects model. The serial correlations in these time series were modeled using an autoregressive correlation structure of order 1 [i.e. AR (1)]. In the mixed-effects model for analyzing echocardiogram data, we modeled the linear change of pre- and post-study, and the interaction effects of the linear time effects and the doses(LT3andLT4).Lastly, weused twoone-sided test (TOST) procedure to detect whether the estimated ﬁxed effects in these models are falling within regions that are equivalently to zero, given pre-deﬁned lower and upper bounds. Energy Expenditure The recording was conducted in a validated, small (about 5000 liters in volume) WRIC ﬁtted with airtight ports to allow for blood draws. This WRIC system allows for accurate and fast measure of energy expenditure (EE) with an error of ±45kcal/day (0.03 kcal/min on average). This error range means that a meaningful effect size in EE is above 0.06 kcal/min (2, 3). Throughout the recording the study participants were seated on a phlebotomy chair in resting position. Pharmacokinetics All the potential combinations in the randomization sequence were observed; two subjects were assigned to each combination sequence with the exception of LT4-LT3-placebo (three subjects) and LT4-placebo-LT3 (one subject). The PK parameter estimates for total T3 following oral LT3 administration and free T4 following oral LT4 administration are presented in Table 2 and Figure 4. Following a single dose of LT3, the Tmax was estimated to be 120 ± 26 minutes with a Cmax of 328 ± 57 ng/dL when corrected for baseline. The observed clearance (assuming complete oral bioavailability) was estimated to be 45 ± 13 mL/ min with an observed terminal half-life (t1/2) of 216 ± 51 minutes. The calculated AUC from baseline to 240 minutes was 54500 ± 8900 minng/dL. Repeating the analysis, without correcting baseline, the observed Cmax was estimated to be 421 ± Heart rate (HR), blood pressure (BP), and hemodynamic data cardiac output (CO), stroke volume (SV), and systemic vascular resistance (SVR), a proxy for endothelial vascular function, were measured with a ccNexﬁn system (Edwards Lifesciences Corp) (6– 8) located in the WRIC. This is a non-invasive device based on the volume clamp method, which continuously measures BP by clamping the artery to a constant volume by dynamically providing equal pressure on either side of the arterial wall, while the volume is measured by a photo-plethysmograph built into a ﬁnger cuff. This device is very precise and accurate when compared with measurement of cardiac output using pulmonary artery catheter thermodilution (6, 9). Doppler Echocardiography g Resting transthoracic Doppler echocardiography was recorded before entering in the WRIC and immediately after the completion of the study (240’). The following parameters were recorded: left ventricular (LV) end‐diastolic and LV end‐systolic volumes, LV ejection fraction (LVEF) as a measure of systolic function; early mitral annular velocities by tissue Doppler averaged between the lateral and septal (e′) annulus and the early transmitral velocity (E) on pulsed‐wave Doppler spectra. These measures were used to calculate the E/e′ ratio, a surrogate of left ventricular ﬁlling pressures. Tricuspid Annular Plane Systolic Excursion (TAPSE) was used to assess right ventricular function (4, 5). The measurements were performed by a cardiologist (AA) blinded to the treatment on an IE33 Phillips apparatus. February 2022 \| Volume 13 \| Article 843539 Frontiers in Endocrinology \| www.frontiersin.org Statistical Analysis TABLE 1 \| Study Participants characteristics. Females (3) Males (9) All Age (yr) 30.3 ± 7.7 26.8 ± 4.1 27.7 ± 5.1 Weight (Kg) 61.7 ± 7.4 79.4 ± 11.3 75.0 ± 12.9 Height (cm) 168.5 ± 4.8 177.2 ± 8.7 175.0 ± 8.6 BMI (Kg/m2) 21.5 ± 1.8 25.2 ± 2.3 24.3 ± 2.7 Systolic BP (mmHg) 106 ± 11 121 ± 7 117 ± 10 Diastolic BP (mmHg) 61 ± 5 71 ± 6 69 ± 7 Heart rate (bpm) 71 ± 14 67 ± 10 68 ± 10 TSH (mIU/mL) 0.86 ± 0.24 1.46 ± 0.84 1.30 ± 0.77 FreeT4 (ng/dL) 1.17 ± 0.06 1.02 ± 0.07 1.05 ± 0.09 Fasting Glucose (mg/dL) 85.3 ± 5.9 89.7 ± 8.5 88.6 ± 7.9 y Pharmacokinetic (PK) parameters for LT3 following LT3 administration and LT4 following LT4 administration were estimated using each individual participant’s observed concentration data (with and without background correction) as assessed by noncompartmental analysis (NCA) using PKanalix version 2020R1 (Lixsoft©) software. EE, hemodynamic metrics and echocardiogram data were preprocessed in Matlab 2020a (Mathworks Inc, Natick, Massachussetts), and analyzed using mixed-effects models in R Studio (R version 3.6.3, RStudio Inc., Boston, Massachusetts). Totest whether acute effects of study drugs occurred, we modeled the interaction effects of linear as well as quadratic time effects and the drugs (LT3 and LT4), and signiﬁcant TABLE 1 \| Study Participants characteristics. February 2022 \| Volume 13 \| Article 843539 Frontiers in Endocrinology \| www.frontiersin.org 4 Chen et al. Acute Effects of Liothyronine FIGURE 3 \| CONSORT chart. TABLE 2 \| Non-compartmental analysis pharmacokinetic parameter estimates. Pharmacokinetic Parameter Total T3 following LT3 administration (background correction) Total T3 following LT3 administration (without background correction) Free T4 following T4 administration (without background correction) N, subjects with full PK parameter estimates available 12 12 8 tmax, min (SD) 120 (26) 120 (26) 130 (80) Cmax, ng/dL (SD) 327 (57) 421 (57) 1.1 (0.15) t1/2, min (SD) 216 (51) 294 (69) 1670 (630) CL, mL/min, (SD) 45 (13) 26 (6) 2700 (940) AUC(0-240), min • ng/dL, (SD) 54500 (8900) 76900 (8600) 246 (28) FIGURE 3 \| CONSORT chart. TABLE 2 \| Non-compartmental analysis pharmacokinetic parameter estimates. TABLE 2 \| Non-compartmental analysis pharmacokinetic parameter estimates. Equivalence Tests We also tested the whether the estimated effect sizes in Tables 4, 5 are equivalence to zero using the TOST procedure. The p- values for rejecting that effect sizes fall beyond the bounded zero region are reported in Tables 6, 7 respectively. The small p- values indicate that those effect sizes are equivalent to zero, demonstrating that there was no difference in temporal trends of the physiological parameters or pre- and post- study effects of the echocardiography when comparing LT3 or LT4 against placebo. A signiﬁcant decrease in TSH from 0’ to 240’ was observed in all treatment arms (LT3 0.72 ± 0.40 LT4 0.66 ± 0.46, placebo 0.45 ± 0.49 mIU/mL, all p<0.01) (Figure 4). A mixed-effects model analysis indicates no signiﬁcant difference in TSH decrease following LT4 or LT3 administration compared with following placebo (Table 3). Statistical Analysis Pharmacokinetic Parameter Total T3 following LT3 administration (background correction) Total T3 following LT3 administration (without background correction) Free T4 following T4 administration (without background correction) N, subjects with full PK parameter estimates available 12 12 8 tmax, min (SD) 120 (26) 120 (26) 130 (80) Cmax, ng/dL (SD) 327 (57) 421 (57) 1.1 (0.15) t1/2, min (SD) 216 (51) 294 (69) 1670 (630) CL, mL/min, (SD) 45 (13) 26 (6) 2700 (940) AUC(0-240), min • ng/dL, (SD) 54500 (8900) 76900 (8600) 246 (28) Total T3 following LT3 administration (background correction) Total T3 following LT3 administration (without background correction) Free T4 following T4 administration (without background correction) systolic or diastolic BP, CO or SVR were observed following the administration of the study drugs or placebo. Similarly, no differences were observed in the Doppler echocardiography studies conducted before and after administration of the study drugs or placebo. These data are reported in Tables 4, 5. 57 ng/dL with an observed terminal t1/2 of 294 ± 69 minutes. The administration of a single dose of LT4 resulted in modest observable increases in free T4. However, when the observed concentrations were background corrected, only 2 subjects met criteria for NCA. This was attributed to the limited sampling schedule that did not capture the terminal phase. When the analysis was repeated without correction for endogenous T4, 8 of the 12 subjects data allowed for full parameter estimation by NCA. The Tmax was estimated to be 130 ± 80 minutes and Cmax 1.1 ± 0.15 ng/dL. The observed clearance was 27 ± 9 mL/min and the calculated AUC from baseline to 240 minutes was 246 ± 28 minng/dL. The reported PK parameter estimates for LT4 following LT4 administration should be considered in the setting of the limited PK sampling schedule and likely limited capture of the terminal phase. DISCUSSION LT3 therapy for hypothyroidism is associated with weight loss, decreased cholesterol concentrations and a trend toward improved diastolic function (10). Very recently, similar ﬁndings on weight and lipid parameters have been reported in elderly patients affected by subclinical hypothyroidism (11). Physiology Parameters Physiology Parameters Compared to placebo, no temporal change was observed in EE or substrate utilization. No signiﬁcant temporal changes in HR, February 2022 \| Volume 13 \| Article 843539 Frontiers in Endocrinology \| www.frontiersin.org 5 Acute Effects of Liothyronine Chen et al. TABLE 3 \| Linear mixed-effects model for changes in TSH. Serum TSH Level (mIU/mL) p Predictors Estimates CI (Intercept) 1.65 0.94 – 2.36 <0.001 Dose [LT3] -0.18 -0.77 – 0.41 0.543 Dose [LT4] 0.15 -0.44 – 0.74 0.608 Post-study -0.45 -1.04 – 0.14 0.13 Dose [LT3] Post-study -0.27 -1.10 – 0.56 0.519 Dose [LT4] * Post-study -0.21 -1.04 – 0.63 0.621 Intercept captures the baseline average of TSH level pre-study on placebo days. Dose [LT3] and Dose [LT4] captures the difference of the baseline TSH on LT3 days and LT4 days, in comparison to those on placebo days. Post-study indicates the pre- and post- study change of the outcomes on placebo days. Dose [LT3] * Post-study captures differences of the pre- and post- study change between LT3 and placebo. Dose [LT4] * Post-study captures differences of the pre- and post- study change between LT4 and placebo. Numbers for the effect sizes (Estimates) and 95% conﬁdence interval (CI) were rounded to 2 decimal places. We did not observe signiﬁcant differences of pre- and post- study changes in TSH levels from either LT3 or LT4 administration in comparison to placebo administration. Bold values indicate statistical signiﬁcance. TABLE 3 \| Linear mixed-effects model for changes in TSH. A B C FIGURE 4 \| Thyroid hormone and TSH kinetics following study drugs administration. (A) changes in total T3 concentration from baseline following LT3 administration. (B) changes in free T4 concentration from baseline following LT4 administration. (C) TSH concentrations at baseline and 240’ following administration of LT3, LT4, and placebo. Data are presented as mean ± SEM. ns, not signiﬁcant. A A B B Bold values indicate statistical signiﬁcance. was administered in single daily dose. Since LT3 was originally approved for treatment of hypothyroidism in 1956 (15), this drug did not undergo the rigorous clinical testing currently necessary for approval and has been “grandfathered” for current use. Although the package insert recommends once-a day administration, the consensus among thyroidologists is to subdivide the dose to limit the ﬂuctuations in serum T3 concentrations (16). Physiology Parameters Because of lack of long-term studies on its safety and effectiveness, the American Thyroid Association recommends against its use for the treatment of hypothyroidism (17). Speciﬁcally, the concerns for potential toxicity of supraphysiologic serum concentrations of T3 has hampered the use of LT3 alone or LT3/LT4 combination therapy in patients with hypothyroidism. C FIGURE 4 \| Thyroid hormone and TSH kinetics following study drugs administration. (A) changes in total T3 concentration from baseline following LT3 administration. (B) changes in free T4 concentration from baseline following LT4 administration. (C) TSH concentrations at baseline and 240’ following administration of LT3, LT4, and placebo. Data are presented as mean ± SEM. ns, not signiﬁcant. C C LT3 is promptly absorbed following oral administration, and doses as low as 0.25 mcg/Kg result in increase in T3 serum concentration well above the upper limit of reference, raising the concern for cardiovascular toxicity. Rapid, non-genomic actions of thyroid hormone have been demonstrated in vitro in endothelial vascular cells (18), providing additional rationale for the concern for toxicity due to the exposure of the vasculature to transient increase in T3 serum concentrations following LT3 administration. The interaction between T3 and integrins results in the production of nitric oxide, promoting vasodilation (19, 20). Despite these laboratory-based data, the clinical relevance of rapid effects of thyroid hormone is not clear. Studies performed in ex vivo models of vasculature contractility are conﬂicting: Gachkar and colleagues demonstrated a rapid response to physiologic concentrations of T3, and a decreased response to both hyper- and hypothyroid concentrations, not mediated by AKT, ERK or AMPK (21). More recently, other authors demonstrated an increase in vasodilation following exposure to T3, mediated by PI3K pathway and thyroid hormone receptor alpha (22). Experiments conducted in an animal model of hypothyroidism indicate that exposure to high dose LT3 generates a measurable transcriptional effect within 30’ reaching a maximum effect after 6 hours (23). Clinically, the FIGURE 4 \| Thyroid hormone and TSH kinetics following study drugs administration. (A) changes in total T3 concentration from baseline following LT3 administration. (B) changes in free T4 concentration from baseline following LT4 administration. (C) TSH concentrations at baseline and 240’ following administration of LT3, LT4, and placebo. Data are presented as mean ± SEM. ns, not signiﬁcant. Additionally, experimental evidences suggest that low dose supplementation of LT3 improves myocardial contractility in patients with congestive heart failure (12). Physiology Parameters Moreover, a secondary analysis of the largest LT3/LT4 combination therapy study (13) showed improved quality of life in carriers of the inactivating Thr92Ala type-2 deiodinase polymorphism receiving LT3 supplements (14). Of interest, in this trial LT3 February 2022 \| Volume 13 \| Article 843539 Frontiers in Endocrinology \| www.frontiersin.org 6 Acute Effects of Liothyronine Chen et al. TABLE 4 \| Results from linear mixed-effects models for changes in EE and hemodynamic metrics. Fixed Effects EE Heart Rate Cardiac Output Estimates SE p Estimates SE p Estimates SE p (Intercept) 1.281 0.073 <0.001 66.828 2.051 <0.001 6.427 0.348 <0.001 Dose [LT3] -0.025 0.033 0.451 1.508 1.189 0.218 0.140 0.346 0.690 Dose [LT4] 0.031 0.033 0.364 1.714 1.181 0.161 0.043 0.344 0.901 Time 0.000 0.000 0.801 -0.020 0.018 0.265 0.000 0.004 0.911 Time2 0.000 0.000 0.906 0.000 0.000 0.168 0.000 0.000 0.870 Dose [LT3] * Time 0.000 0.000 0.813 -0.004 0.022 0.867 0.002 0.005 0.623 Dose [LT4] * Time 0.000 0.000 0.475 0.018 0.021 0.409 0.002 0.005 0.679 Dose [LT3] * Time2 0.000 0.000 0.976 0.000 0.000 0.676 0.000 0.000 0.704 Dose [LT4] * Time2 0.000 0.000 0.814 0.000 0.000 0.278 0.000 0.000 0.448 Stroke Volume Systemic Vascular Resistance Estimates SE p Estimates SE p (Intercept) 97.951 5.616 <0.001 1449.916 158.203 <0.001 Dose [LT3] -0.931 4.879 0.850 -108.190 185.411 0.566 Dose [LT4] -2.704 4.871 0.584 -28.161 183.971 0.880 Time 0.038 0.034 0.265 2.498 2.898 0.389 Time2 0.000 0.000 0.137 -0.008 0.015 0.587 Dose [LT3] * Time 0.025 0.043 0.569 -3.311 3.340 0.322 Dose [LT4] * Time -0.002 0.043 0.965 -4.873 3.319 0.142 Dose [LT3] * Time2 0.000 0.000 0.833 0.021 0.017 0.197 Dose [LT4] * Time2 0.000 0.000 0.980 0.022 0.017 0.182 Intercept captures the baseline average of each outcome during the ﬁrst 20-40 minutes in the chamber before the dose administration of on placebo days. Dose[LT3] and Dose[LT4] captures the difference of the baseline averages on LT3 days and LT4 days, in comparison to that on placebo days. Time (Minute) and Time2 indicate the linear and quadratic effect of time on placebo days, capturing the temporal trends of each outcome. Dose [LT3] * Time captures the difference in linear time effects between the placebo and the LT3 dose, and Dose [LT3] * Time2 captures the difference in quartic time effects between the placebo and the LT3 dose. Physiology Parameters To control for potential non-speciﬁcity in the nongenomic effects of thyroid hormone, in addition to placebo we included a study arm of equimolar LT4 dose as additional control. By using liquid formulations of LT3 and LT4 we were able to titrate the doses to adjust for the participants’ weight, further increasing the internal validity of the study. When compared to our previous observations on tablet formulation (1), liquid LT3 showed a similar Tmax, while, as expected, the Cmax was much greater due to the higher LT3 dose employed in this study. The decrease in TSH from baseline to the end of the study, which is consistent with the hormone biorhythm (28) and possibly ampliﬁed by the prolonged fast, was not different among the active drugs and placebo. We speculated that, if rapid effects of T3 were of clinical relevance, transient exposure to T3 concentrations above normal range would result in measurable and potentially clinically signiﬁcant changes in the cardiovascular system and energy metabolism owing to the exquisite sensitivity of these end- organ targets to the action of thyroid hormone (26, 27). Moreover, since the endothelial vascular cells are exposed to changes in T3 concentrations, measurement of their function by assessing systemic vascular resistance would represent an ideal readout to assess acute effects of supraphysiologic doses of LT3. We speculated that, if rapid effects of T3 were of clinical relevance, transient exposure to T3 concentrations above normal range would result in measurable and potentially clinically signiﬁcant changes in the cardiovascular system and energy metabolism owing to the exquisite sensitivity of these end- organ targets to the action of thyroid hormone (26, 27). Moreover, since the endothelial vascular cells are exposed to changes in T3 concentrations, measurement of their function by assessing systemic vascular resistance would represent an ideal readout to assess acute effects of supraphysiologic doses of LT3. To test this hypothesis, we conducted a pharmacokinetics/ pharmacodynamic study directed to these targets of the hormonal action. In this study we chose to investigate a short timeframe based on our knowledge of liothyronine PK (1), aiming to capture events occurring at and immediately after the Cmax, i.e. at the time of maximal exposure of endothelial and myocardial cells to the peak serum T3 concentrations. Physiology Parameters Dose [LT4] * Time captures the difference in linear time effects between the placebo and the LT4 dose, and Dose [LT4] * Time2 captures the difference in quartic time effects between the placebo and the LT4 dose. Numbers for the effect sizes (Estimates) and Standard Errors (SE) were rounded to 3 decimal places. The zero values indicate that there were no temporal trends observed in the time series of the outcomes on placebo days or on the active dose days. Bold values indicate statistical signiﬁcance. TABLE 4 \| Results from linear mixed-effects models for changes in EE and hemodynamic metrics. Intercept captures the baseline average of each outcome during the ﬁrst 20-40 minutes in the chamber before the dose administration of on placebo days. Dose[LT3] and Dose[LT4] captures the difference of the baseline averages on LT3 days and LT4 days, in comparison to that on placebo days. Time (Minute) and Time2 indicate the linear and quadratic effect of time on placebo days, capturing the temporal trends of each outcome. Dose [LT3] * Time captures the difference in linear time effects between the placebo and the LT3 dose, and Dose [LT3] * Time2 captures the difference in quartic time effects between the placebo and the LT3 dose. Dose [LT4] * Time captures the difference in linear time effects between the placebo and the LT4 dose, and Dose [LT4] * Time2 captures the difference in quartic time effects between the placebo and the LT4 dose. Numbers for the effect sizes (Estimates) and Standard Errors (SE) were rounded to 3 decimal places. The zero values indicate that there were no temporal trends observed in the time series of the outcomes on placebo days or on the active dose days. Bold values indicate statistical signiﬁcance. reports of toxicity secondary to trauma with release of thyroid hormone from the gland, or following acute poisoning indicate a signiﬁcant lag time (24, 25), suggesting that the majority of the harm caused by thyrotoxicosis is ascribable to transcriptional effects of thyroid hormone. It is possible that the discrepancies between laboratory-based experiments and in vivo observations could be explained by the dynamic of acute exposure (seconds) to thyroid hormone vs. the much slower changes in thyroid hormone concentrations following oral administration. reference range. Frontiers in Endocrinology \| www.frontiersin.org Physiology Parameters We used a pharmacologic dose which would produce peak T3 concentration well above the upper Our data indicate no signiﬁcant differences or appreciable trend during our observation in the all the study endpoints despite achieving clearly supraphysiologic serum T3 concentrations. Speciﬁcally, during the four hours following the administration of LT3, we did not observe any change in HR or BP, or estimates of CO and SVR. Moreover, we independently assessed the acute effects of LT3onthemyocardiumbyperformingaDopplerechocardiography study before and after each of the studies. Similar to the ccNexﬁn continuous recordings, we observed no measurable effects on myocardial contractility. Finally, no differences were observed in energy metabolism measured by the WRIC. To test this hypothesis, we conducted a pharmacokinetics/ pharmacodynamic study directed to these targets of the hormonal action. In this study we chose to investigate a short timeframe based on our knowledge of liothyronine PK (1), aiming to capture events occurring at and immediately after the Cmax, i.e. at the time of maximal exposure of endothelial and myocardial cells to the peak serum T3 concentrations. We used a pharmacologic dose which would produce peak T3 concentration well above the upper Our data are consistent with the observations of Jonklaas et al. that showed signiﬁcant increase in heart rate and blood pressure only after ﬁve hours from the administration of a pharmacologic February 2022 \| Volume 13 \| Article 843539 7 Acute Effects of Liothyronine Chen et al. TABLE 5 \| Results from linear mixed-effects models for changes in echocardiogram outcomes. Fixed Effects LVEF E E/e’ TAPSE Estimates SE p Estimates SE p Estimates SE p Estimates SE p (Intercept) 59.182 1.701 0.000 14.327 0.521 0.000 5.493 0.315 0.000 2.450 0.105 0.000 Dose [LT3] 2.000 1.604 0.219 -0.324 0.585 0.583 0.205 0.269 0.450 0.047 0.090 0.603 Dose [LT4] 0.494 1.654 0.767 0.351 0.603 0.563 0.121 0.269 0.655 -0.102 0.093 0.285 Post-study -0.306 1.654 0.854 0.176 0.603 0.772 0.123 0.278 0.660 0.095 0.093 0.313 Dose [LT3] * Post 0.488 2.304 0.833 -0.008 0.840 0.993 -0.112 0.389 0.775 -0.122 0.127 0.344 Dose [LT4] * Post 0.906 2.359 0.703 0.089 0.860 0.918 -0.089 0.390 0.821 0.033 0.130 0.802 Intercept captures the baseline average of echocardiogram metrics pre-study on placebo days. Dose [LT3] and Dose [LT4] captures the differences of the baseline averages on LT3 days and LT4 days, in comparison to those on placebo days. Physiology Parameters Post-study indicates the pre- and post- study change of the outcomes on placebo days. Dose [LT3] * Post-study captures differences of the pre- and post- study change between LT3 and placebo. Dose [LT4] * Post-study captures differences of the pre- and post- study change between LT4 and placebo. Numbers for the effect sizes (Estimates) and standard errors (SE) were rounded to 2 decimal places. We did not observe signiﬁcant differences of pre- and post- study changes in echocardiogram outcomes from either LT3 or LT4 administration in comparison to placebo administration. TABLE 5 \| Results from linear mixed-effects models for changes in echocardiogram outcomes. Intercept captures the baseline average of echocardiogram metrics pre-study on placebo days. Dose [LT3] and Dose [LT4] captures the differences of the baseline averages on LT3 days and LT4 days, in comparison to those on placebo days. Post-study indicates the pre- and post- study change of the outcomes on placebo days. Dose [LT3] * Post-study captures differences of the pre- and post- study change between LT3 and placebo. Dose [LT4] * Post-study captures differences of the pre- and post- study change between LT4 and placebo. Numbers for the effect sizes (Estimates) and standard errors (SE) were rounded to 2 decimal places. We did not observe signiﬁcant differences of pre- and post- study changes in echocardiogram outcomes from either LT3 or LT4 administration in comparison to placebo administration. dose of LT3 (29). Of note, in their study, the subjects were given lunch at 4-hour, which per se may have resulted in post-prandial HR elevation (30); thus, the true cardiovascular response beyond 4 hours without confounding factors remains to be captured. speculate that these results provide the rationale to consider the use of single administration of LT3, rather than trying to reach stable serum concentrations by extended release formulations or by multiple daily administrations regimens. Conversely, we want to afﬁrm that absent an empirical demonstration of effectiveness and safety of once daily administration regimen, this interpretation of the data should not translate in clinical practice. Collectively, the results of our study indicate that a single administration of LT3, sufﬁcient to transiently increase the T3 concentration well above the range of reference, does not result in signiﬁcant signal in the tissues which are most sensitive to acute effects of T3, and are immediately exposed to the peak in T3 concentration. Physiology Parameters P-values presented in this table indicate the signiﬁcance of the TOST procedure, p-values < 0.05 indicate that null hypothesis that a ﬁxed effect does not fall in the region of [LB, UB] are rejected, and we thus accept the alternative hypothesis that the ﬁxed effect is zero. P-values that are less than 1e-4 are recorded as 0 in the table. TABLE 6 \| Equivalence tests (TOST procedure) of the estimated effect sizes in Table 4. EE Heart Rate Cardiac Output Stroke Volume Systemic Vascular Resistance LB=-0.1 UB=0.1 LB=-5 UB =5 LB= -1 UB=1 LB=-10 UB=10 LB= -500 UB=500 Dose [LT3] 0.004 0.003 0.000 0.004 0.002 0.011 0.038 0.018 0.023 0.002 Dose [LT4] 0.004 0.002 0.000 0.005 0.003 0.005 0.074 0.008 0.009 0.004 Time 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Time2 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Dose [LT3] * Time 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Dose [LT4] * Time 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Dose [LT3] * Time2 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Dose [LT4] * Time2 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 LB, lower bound; UB, upper bound. Two one-sided t-tests were conducted for each estimated ﬁxed effects of the models presented in Table 4. P-values presented in this table indicate the signiﬁcance of the TOST procedure, p-values < 0.05 indicate that null hypothesis that a ﬁxed effect does not fall in the region of [LB, UB] are rejected, and we thus accept the alternative hypothesis that the ﬁxed effect is zero. LB, lower bound; UB, upper bound. Two one-sided t-tests were conducted for each estimated ﬁxed effects of the models presented in Table 4. P-values presented in this table indicate the signiﬁcance of the TOST procedure, p-values < 0.05 indicate that null hypothesis that a ﬁxed effect does not fall in the region of [LB, UB] are rejected, and we thus accept the alternative hypothesis that the ﬁxed effect is zero. TABLE 7 \| Equivalence tests (TOST procedure) of the estimated effect sizes in Table 5. Physiology Parameters These ﬁndings appear to negate clinical relevance to the rapid effects of T3 which have been observed in vitro. The clinical implications are intriguing, since one could Strengths of our study reside in the rigorous study design, use of liquid formulation of the drug which allowed for a precise, weight-based dosing, and the use of state-of-the-art techniques to assess the cardiovascular system and energy metabolism. The continuous recording of BP, endothelial vascular function and TABLE 6 \| Equivalence tests (TOST procedure) of the estimated effect sizes in Table 4. EE Heart Rate Cardiac Output Stroke Volume Systemic Vascular Resistance LB=-0.1 UB=0.1 LB=-5 UB =5 LB= -1 UB=1 LB=-10 UB=10 LB= -500 UB=500 Dose [LT3] 0.004 0.003 0.000 0.004 0.002 0.011 0.038 0.018 0.023 0.002 Dose [LT4] 0.004 0.002 0.000 0.005 0.003 0.005 0.074 0.008 0.009 0.004 Time 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Time2 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Dose [LT3] * Time 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Dose [LT4] * Time 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Dose [LT3] * Time2 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Dose [LT4] * Time2 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 LB, lower bound; UB, upper bound. Two one-sided t-tests were conducted for each estimated ﬁxed effects of the models presented in Table 4. P-values presented in this table indicate the signiﬁcance of the TOST procedure, p-values < 0.05 indicate that null hypothesis that a ﬁxed effect does not fall in the region of [LB, UB] are rejected, and we thus accept the alternative hypothesis that the ﬁxed effect is zero. TABLE 7 \| Equivalence tests (TOST procedure) of the estimated effect sizes in Table 5. Predictors LVEF E E/e TAPSE LB=-5 UB=5 LB=-2 UB=2 LB=-1 UB=1 LB= -0.5 UB= 0.5 Dose [LT3] 0.000 0.034 0.003 0.000 0.000 0.002 0.000 0.000 Dose [LT4] 0.001 0.004 0.000 0.004 0.000 0.001 0.000 0.000 Post-study 0.003 0.001 0.000 0.002 0.000 0.001 0.000 0.000 Dose [LT3] * Post 0.011 0.028 0.011 0.010 0.014 0.003 0.002 0.000 Dose [LT4] * Post 0.008 0.045 0.010 0.016 0.012 0.004 0.000 0.000 LB, lower bound; UB, upper bound. Two one-sided t-tests were conducted for each estimated ﬁxed effects of the models presented in Table 5. AUTHOR CONTRIBUTIONS The relatively small sample size, non-unusual in phase I-II studies, represents a limitation. However, the absence of trends in the physiological variables indicates that it is extremely unlikely our negative ﬁndings are due to type-2 error. Some volunteers had a short, 48-hour interval between the studies raising the potential concern for carryover effects. Given the lack of measurable effects and intraindividual differences of baseline (before administration of the study drugs) data, we do not believe that this is a cofounding factor. By design, the observation was limited to a timeframe where maximum variation in serum T3 concentrations can be observed, therefore late events would not be captured. Lastly, the translational value of observation obtained in healthy individuals to patients affected by hypothyroidism, in particular elderly and with co-morbidities is unknown, and we cannot rule out the possibility that acute effects may occur in patients with lower serum T3 levels. SCh contributed to the study design, analyzed energy expenditure data, performed statistical analysis and contributed to the interpretation, and contributed to the initial draft and ﬁnal writing of the manuscript. GFW performed pharmacokinetics analysis and contributed to the writing of the manuscript. AV recorded and analyzed echocardiography data. SCa recorded and analyzed hemodynamics data. SY contributed to the study design, clinical assessment of study patients, and contributed to the editing of the manuscript. BVT contributed to the study design and interpretation of the pharmacokinetics data. AA contributed to the study design and interpretation of the echocardiography data. FSC designed the study, supervised the analysis and interpretation of the data, and wrote and edited the manuscript. All authors contributed to the article and approved the submitted version. In conclusion, in healthy individuals a single administration of LT3, able to rapidly increase the serum T3 concentration above normal range does not result in measurable changes in target organ-systems attributable to the rapid thyroid hormone action. The data support the rationale to explore the use of LT3 in single dose aiming to achieve sustained increase in tissue concentrations of T3, reducing the concern for ﬂuctuations in serum concentrations. On the other hand, patients who have longstanding signiﬁcant hypothyroidism may experience paradoxical response, and T3-based therapy should be started after achieving a euthyroid state by LT4, and residual symptoms are still present. 6. Broch O, Renner J, Gruenewald M, Meybohm P, Schottler J, Caliebe A, et al. A Comparison of the Nexﬁn(R) and Transcardiopulmonary Thermodilution to Estimate Cardiac Output During Coronary Artery Surgery. Anaesthesia (2012) 67(4):377–83. doi: 10.1111/j.1365-2044.2011.07018.x 4. Trankle C, Canada JM, Buckley L, Carbone S, Dixon D, Arena R, et al. Impaired Myocardial Relaxation With Exercise Determines Peak Aerobic Exercise Capacity in Heart Failure With Preserved Ejection Fraction. ESC Heart Fail (2017) 4(3):351–5. doi: 10.1002/ehf2.12147 ACKNOWLEDGMENTS This work was in part supported by CTSA award No. UL1TR002649 from the National Center for Advancing Translational Sciences. The authors gratefully acknowledge the help and support of Joyce Ruddley, RN, the Clinical Research Service Unit and Investigational Pharmacy of VCU Health. Liquid formulations of LT3, LT4, and placebo were a kind gift of IBSA Institut Biochimique Lugano (Switzerland). IBSA was not part of the study design and had no access to the data prior to the publication. 5. Aroditis K, Pikilidou M, Vourvouri E, Hadjistavri L, Zebekakis P, Yovos J, et al. Changes in Cardiac Function and Structure in Newly Diagnosed Graves' Disease. A Conventional and 2D-Speckle Tracking Echocardiography Study. Int J Cardiovasc Imaging (2017) 33(2):187–95. doi: 10.1007/s10554- 016-0984-z FUNDING SCa is supported by a Career Development Award 19CDA34660318 from the American Heart Association and by the Clinical and Translational Science Awards Program UL1TR002649 from National Institutes of Health to Virginia Commonwealth University. FSC is supported by the NIH-NIDDK grant 1 R21 DK122310-01A1. DATA AVAILABILITY STATEMENT The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. Physiology Parameters Predictors LVEF E E/e TAPSE LB=-5 UB=5 LB=-2 UB=2 LB=-1 UB=1 LB= -0.5 UB= 0.5 Dose [LT3] 0.000 0.034 0.003 0.000 0.000 0.002 0.000 0.000 Dose [LT4] 0.001 0.004 0.000 0.004 0.000 0.001 0.000 0.000 Post-study 0.003 0.001 0.000 0.002 0.000 0.001 0.000 0.000 Dose [LT3] * Post 0.011 0.028 0.011 0.010 0.014 0.003 0.002 0.000 Dose [LT4] * Post 0.008 0.045 0.010 0.016 0.012 0.004 0.000 0.000 LB, lower bound; UB, upper bound. Two one-sided t-tests were conducted for each estimated ﬁxed effects of the models presented in Table 5. P-values presented in this table indicate the signiﬁcance of the TOST procedure, p-values < 0.05 indicate that null hypothesis that a ﬁxed effect does not fall in the region of [LB, UB] are rejected, and we thus accept the alternative hypothesis that the ﬁxed effect is zero. P-values that are less than 1e-4 are recorded as 0 in the table. TABLE 7 \| Equivalence tests (TOST procedure) of the estimated effect sizes in Table 5. LB, lower bound; UB, upper bound. Two one-sided t-tests were conducted for each estimated ﬁxed effects of the models presented in Table 5. P-values presented in this table indicate the signiﬁcance of the TOST procedure, p-values < 0.05 indicate that null hypothesis that a ﬁxed effect does not fall in the region of [LB, UB] are rejected, and we thus accept the alternative hypothesis that the ﬁxed effect is zero. P-values that are less than 1e-4 are recorded as 0 in the table. February 2022 \| Volume 13 \| Article 843539 Frontiers in Endocrinology \| www.frontiersin.org 8 Chen et al. Acute Effects of Liothyronine ETHICS STATEMENT EE allowed us to measure even subtle changes which would not be captured by single timepoint observations. Finally, we included a treatment arm of LT4 to assess for potential non ligand-speciﬁc acute effects of thyroid hormone. To the best of our knowledge, no study has investigated the rapid effects of thyroid hormone in such details. The studies involving human participants were reviewed and approved by VCU IRB. The patients/participants provided their written informed consent to participate in this study. AUTHOR CONTRIBUTIONS Long-term studies in patients affected by hypothyroidism are necessary to assess the safety and effectiveness of single dose LT3 or LT3/LT4 combination therapy in patients affected by hypothyroidism. REFERENCES Thyroid Storm After Blunt Thyroid Injury: A Case Report. J Trauma (2007) 63(3):E85–87. doi: 10.1097/01.ta.0000232184.07317.56 25. Dahlberg PA, Karlsson FA, Wide L. Triiodothyronine Intoxication. Lancet (1979) 2(8144):700. doi: 10.1016/S0140-6736(79)92105-6 10. Celi FS, Zemskova M, Linderman JD, Smith S, Drinkard B, Sachdev V, et al. Metabolic Effects of Liothyronine Therapy in Hypothyroidism: A Randomized, Double-Blind, Crossover Trial of Liothyronine Versus Levothyroxine. J Clin Endocrinol Metab (2011) 96(11):3466–74. doi: 10.1210/jc.2011-1329 26. Razvi S, Jabbar A, Pingitore A, Danzi S, Biondi B, Klein I, et al. Thyroid Hormones and Cardiovascular Function and Diseases. J Am Coll Cardiol (2018) 71(16):1781–96. doi: 10.1016/j.jacc.2018.02.045 27. Yavuz S, Salgado Nunez Del Prado S, Celi FS. Thyroid Hormone Action and Energy Expenditure. J Endocr Soc (2019) 3(7):1345–56. doi: 10.1210/js.2018- 00423 11. Meizlik P, Cucchiara A, Kannan L, Mays J, Jolivert J, Scattergood T, et al. Physiologic Effects of Levothyroxine and Liothyronine in the in Older Individuals With Persistent Subclinical Hypothyroidism: A Randomized, Double-Blind, Cross-Over Study. ENDO (2021) 2021. doi: 10.1210/jendso/ bvab048.1996 28. Nicoloff JT, Fisher DA, Appleman MDJr. The Role of Glucocorticoids in the Regulation of Thyroid Function in Man. J Clin Invest (1970) 49(10):1922–9. doi: 10.1172/JCI106411 29. Jonklaas J, Burman KD, Wang H, Latham KR. Single-Dose T3 Administration: Kinetics and Effects on Biochemical and Physiological Parameters. Ther Drug Monit (2015) 37(1):110–8. doi: 10.1097/FTD.0000000000000113 12. Pingitore A, Galli E, Barison A, Iervasi A, Scarlattini M, Nucci D, et al. Acute Effects of Triiodothyronine (T3) Replacement Therapy in Patients With Chronic Heart Failure and Low-T3 Syndrome: A Randomized, Placebo- Controlled Study. J Clin Endocrinol Metab (2008) 93(4):1351–8. doi: 10.1210/jc.2007-2210 30. Hlebowicz J, Lindstedt S, Bjorgell O, Dencker M. Relationship Between Postprandial Changes in Cardiac Left Ventricular Function, Glucose and Insulin Concentrations, Gastric Emptying, and Satiety in Healthy Subjects. Nutr J (2011) 10:1–8. doi: 10.1186/1475-2891-10-26 13. Saravanan P, Simmons DJ, Greenwood R, Peters TJ, Dayan CM. Partial Substitution of Thyroxine (T4) With Tri-Iodothyronine in Patients on T4 Replacement Therapy: Results of a Large Community-Based Randomized Controlled Trial. J Clin Endocrinol Metab (2005) 90(2):805–12. doi: 10.1210/ jc.2004-1672 Author Disclaimer: Its contents are solely the responsibility of the authors and do not necessarily represent ofﬁcial views of the National Center for Advancing Translational Sciences or the National Institutes of Health. 14. Panicker V, Saravanan P, Vaidya B, Evans J, Hattersley AT, Frayling TM, et al. REFERENCES Common Variation in the DIO2 Gene Predicts Baseline Psychological Well- Being and Response to Combination Thyroxine Plus Triiodothyronine Therapy in Hypothyroid Patients. J Clin Endocrinol Metab (2009) 94 (5):1623–9. doi: 10.1210/jc.2008-1301 Conﬂict of Interest: The Division of Endocrinology, Diabetes and Metabolism of Virginia Commonwealth University has received an unrestricted grant from IBSA Institut Biochimique, Lugano Switzerland. Liquid formulations of LT3, LT4, and placebo were a kind gift of IBSA Institut Biochimique Lugano (Switzerland). IBSA was not part of the study design and had no access to the data prior to the publication. FC has served as consultant for IBSA, Acella and Kashiv (now merged with Amneal). j 15. FDA. Cytomel Labeling, Vol. 2021. Silver Spring MD. (2018). Available at: https://www.accessdata.fda.gov/drugsatfda_docs/label/2018/010379s054lbl. pdf 16. Wiersinga WM, Duntas L, Fadeyev V, Nygaard B, Vanderpump MP. ETA Guidelines: The Use of L-T4 + L-T3 in the Treatment of Hypothyroidism. Eur Thyroid J (2012) 1(2):55–71. doi: 10.1159/000339444 The remaining authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. 17. Jonklaas J, Bianco AC, Bauer AJ, Burman KD, Cappola AR, Celi FS, et al. Guidelines for the Treatment of Hypothyroidism: Prepared by the American Thyroid Association Task Force on Thyroid Hormone Replacement. Thyroid (2014) 24(12):1670–751. doi: 10.1089/thy.2014.0028 Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. 18. Hiroi Y, Kim HH, Ying H, Furuya F, Huang Z, Simoncini T, et al. Rapid Nongenomic Actions of Thyroid Hormone. Proc Natl Acad Sci U S A (2006) 103(38):14104–9. doi: 10.1073/pnas.0601600103 19. Cao X, Kambe F, Moeller LC, Refetoff S, Seo H. Thyroid Hormone Induces Rapid Activation of Akt/protein Kinase B-Mammalian Target of Rapamycin- P70s6k Cascade Through Phosphatidylinositol 3-Kinase in Human Fibroblasts. Mol Endocrinol (2005) 19(1):102–12. doi: 10.1210/me.2004-0093 Copyright © 2022 Chen, Wohlford, Vecchie’, Carbone, Yavuz, Van Tassell, Abbate and Celi. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). REFERENCES 1. Van Tassell B, Wohlford GFT, Linderman JD, Smith S, Yavuz S, Pucino F, et al. Pharmacokinetics of L-Triiodothyronine in Patients Undergoing Thyroid Hormone Therapy Withdrawal. Thyroid (2019) 29(10):1371–9. doi: 10.1089/thy.2019.0101 1. Van Tassell B, Wohlford GFT, Linderman JD, Smith S, Yavuz S, Pucino F, et al. Pharmacokinetics of L-Triiodothyronine in Patients Undergoing Thyroid Hormone Therapy Withdrawal. Thyroid (2019) 29(10):1371–9. doi: 10.1089/thy.2019.0101 2. Chen S, Wohlers E, Ruud E, Moon J, Ni B, Celi FS. Improving Temporal Accuracy of Human Metabolic Chambers for Dynamic Metabolic Studies. PloS One (2018) 13(4):e0193467. doi: 10.1371/journal.pone.0193467 2. Chen S, Wohlers E, Ruud E, Moon J, Ni B, Celi FS. Improving Temporal Accuracy of Human Metabolic Chambers for Dynamic Metabolic Studies. PloS One (2018) 13(4):e0193467. doi: 10.1371/journal.pone.0193467 3. Chen S, Scott C, Pearce JV, Farrar JS, Evans RK, Celi FS. An Appraisal of Whole-Room Indirect Calorimeters and a Metabolic Cart for Measuring Resting and Active Metabolic Rates. Sci Rep (2020) 10(1):14343. doi: 10.1038/ s41598-020-71001-1 February 2022 \| Volume 13 \| Article 843539 Frontiers in Endocrinology \| www.frontiersin.org 9 Chen et al. Acute Effects of Liothyronine 7. Van Tassell BW, Canada J, Carbone S, Trankle C, Buckley L, Oddi Erdle C, et al. Interleukin-1 Blockade in Recently Decompensated Systolic Heart Failure: Results From REDHART (Recently Decompensated Heart Failure Anakinra Response Trial). Circ Heart Fail (2017) 10(11):1–17. doi: 10.1161/ CIRCHEARTFAILURE.117.004373 21. Gachkar S, Nock S, Geissler C, Oelkrug R, Johann K, Resch J, et al. Aortic Effects of Thyroid Hormone in Male Mice. J Mol Endocrinol (2019) 62(3):91– 9. doi: 10.1530/JME-18-0217 22. Geist D, Hones GS, Gassen J, Kerp H, Kleinbongard P, Heusch G, et al. Noncanonical Thyroid Hormone Receptor Alpha Action Mediates Arterial Vasodilation. Endocrinology (2021) 162(7):1–10. doi: 10.1210/endocr/ bqab099 8. Van Tassell BW, Westman P, Trankle C, Johns S, Kadariya D, Buckley L, et al. Noninvasive Hemodynamic Monitoring of Cocaine-Induced Changes in Cardiac Output and Systemic Vascular Resistance in Subjects With Chronic Cocaine Use Disorder. J Cardiovasc Pharmacol (2019) 74(6):528–34. doi: 10.1097/FJC.0000000000000740 23. Danzi S, Ojamaa K, Klein I. Triiodothyronine-Mediated Myosin Heavy Chain Gene Transcription in the Heart. Am J Physiol Heart Circ Physiol (2003) 284 (6)::H2255–2262. doi: 10.1152/ajpheart.00860.2002 9. Bubenek-Turconi SI, Craciun M, Miclea I, Perel A. Noninvasive Continuous Cardiac Output by the Nexﬁn Before and After Preload-Modifying Maneuvers: A Comparison With Intermittent Thermodilution Cardiac Output. Anesth Analg (2013) 117(2):366–72. doi: 10.1213/ ANE.0b013e31829562c3 24. Hagiwara A, Murata A, Matsuda T, Sakaki S, Shimazaki S. REFERENCES The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 20. Cao X, Kambe F, Yamauchi M, Seo H. Thyroid-Hormone-Dependent Activation of the Phosphoinositide 3-Kinase/Akt Cascade Requires Src and Enhances Neuronal Survival. Biochem J (2009) 424(2):201–9. doi: 10.1042/ BJ20090643 February 2022 \| Volume 13 \| Article 843539 Frontiers in Endocrinology \| www.frontiersin.org 10
https://openalex.org/W3095897330	https://rua.ua.es/dspace/bitstream/10045/110146/1/Garcia-del-Castillo-Rodriguez_etal_2020_PLoS-ONE.pdf	English	null	Information and communication technologies and quality of life in home confinement: Development and validation of the TICO scale	PloS one	2,020	cc-by	7,835	Jose´ Antonio Garcı´a del Castillo-Rodrı´guez1☯, Irene Ramos-Soler2☯, Carmen Lo´pez- Sa´nchez2☯, Carmen Quiles-SolerID2☯* Jose´ Antonio Garcı´a del Castillo-Rodrı´guez1☯, Irene Ramos-Soler2☯, Carmen Lo´pez- Sa´nchez2☯, Carmen Quiles-SolerID2☯* 1 Department of Health Psychology, Miguel Herna´ndez University, Elche, Alicante, España, 2 Department of Communication and Social Psychology, University of Alicante, San Vicente del Raspeig, Alicante, España ☯These authors contributed equally to this work. * mc.quiles@ua.es OPEN ACCESS Citation: Garcı´a del Castillo-Rodrı´guez JA, Ramos- Soler I, Lo´pez-Sa´nchez C, Quiles-Soler C (2020) Information and communication technologies and quality of life in home confinement: Development and validation of the TICO scale. PLoS ONE 15(11): e0241948. https://doi.org/10.1371/journal. pone.0241948 Editor: Fre´de´ric Denis, Centre Hospitalier Regional Universitaire de Tours, FRANCE Editor: Fre´de´ric Denis, Centre Hospitalier Regional Universitaire de Tours, FRANCE Published: November 5, 2020 Peer Review History: PLOS recognizes the benefits of transparency in the peer review process; therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. The editorial history of this article is available here: https://doi.org/10.1371/journal.pone.0241948 Abstract The mandatory home confinement of the Spanish population, implemented in response to the COVID-19 pandemic, presents a unique opportunity to study the use and influence of Information and Communication Technologies (ICT) in people’s perception of quality of life during this exceptional situation. This article adapts and validates a psychometric scale designed to identify and measure the main dimensions of the Quality of Life construct per- ceived through ICT use. To this end, an exploratory and transversal study has been carried out in Spain on a sample of 2,346 participants. Data processing has been carried out with SPSS and EQS. The results provide evidence of the reliability and psychometric quality on the scale, which exhibits adequate consistency that facilitates its application. The confirma- tory factor analysis showed a hierarchical model of three correlated factors that account for the dimensions “Satisfaction with life”, “Emotional support” and “Social support”, which have enough correlation to measure the personal perception of quality of life associated with ICT use and are consistent with previous psychometric studies. The results of the TICO scale indicate that more than 70% of the sample feel ICT have united their family during home confinement and more than 45% experience happy feelings when they use ICT. In home confinement, ICT use has improved users’ quality of life, mainly their satisfaction with life and social and family support. PLOS ONE PLOS ONE a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 RESEARCH ARTICLE Introduction The world is facing an unprecedented crisis in human history. Previous epidemics and pan- demics, such as the 1918 influenza pandemic (also known as the Spanish flu), the HIV/AIDS pandemic of 1981, the SARS pandemic of 2002, the H1N1 influenza pandemic of 2010, the Ebola outbreak of 2014, among many others, do not measure up to the COVID-19 pandemic in terms of their impact on interpersonal relationships and, of course, on the use of ICT to mit- igate their effects. Copyright: © 2020 Garcı´a del Castillo-Rodrı´guez et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. 1 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 PLOS ONE Quality of life scale and confinement Data Availability Statement: The data for this manuscript are stored on: http://hdl.handle.net/ 10045/109886. Funding: The authors received no specific funding for this work. Competing interests: The authors have declared that no competing interests exist. As Piña-Ferrer [1] points out, at this time of great global social uncertainty, the only vaccine could be fact-checked, truthful and scientific “information”, but the reality is that the flow and, above all, the speed with which news and information is transmitted today exponentially raises fear of contagion and death. The perception of risk is directly proportional to the degree of exposure and to variables related to previous pathologies, age, and lifestyle. According to Casero-Ripolle´s [2], regarding the flow information, the media disseminate a huge volume of news, many of them fake, that generate discontent and, above all, loss of trust and credibility among the population. The author concludes that the consequences of this dynamic are very negative for the democratic system, and, in our opinion, this situation also destabilizes people’s quality of life. Many countries quickly understood the need to implement responsible policies to try to prevent the spread of the virus through the simplest system: physical distancing. In this regard, it is important to note that “social distancing” has never really been the goal [3], since in most cultures it could be understood as a punishment and not as a measure of protection, given that ICT allow people to maintain social relations without physical contact. Introduction So far ICT have become an active part in people’s lives, albeit with a series of problems asso- ciated with them, such as possible addiction, nomophobia, techno-interference, the phantom phone vibration syndrome, and the fear of missing out (FOMO), among others [4–6]. Research shows us that the use of ICT significantly increases life satisfaction in the elderly population, generating important feelings of self-sufficiency [7]. Something similar happens with young people and adults, who use ICT as a tool of social and emotional support and as a source of social interactions, which improves their psychological well-being [8–10]. In new situations, new action systems are generated. ICT have made a giant leap in this cri- sis. Many of the actions that have been enhanced, in addition to telecommuting, are psycho- logical crisis interventions and social support therapies through ICT, as reported by Inchausti, Macbeth, Hasson-Ohayon, and Dimaggio [11], as well as their new application by people who had not used them before [12]. In the face of the COVID-19 pandemic, it is necessary to distinguish between two major types of problems: to keep the population healthy against the virus and to maintain people’s quality of life. We know that the concept of quality of life is broad, controversial and varies depending on the context and the perception of each person at a given time. Therefore, it is dynamic, subjective, and adaptable. According to Schalock and Verdugo [13], there are more than 200 definitions of this con- struct, but multiple intercultural studies agree on its main dimensions: interpersonal relation ships, psychological well-being, material well-being, physical well-being, personal development, social acceptance, self-determination and rights. All these dimensions are grouped by the authors into three factors: Independence (personal development and self- determination), social participation (relationships, social acceptance and rights), and well- being (psychological, physical and material). According to Schalock and Verdugo [13], there are more than 200 definitions of this con- struct, but multiple intercultural studies agree on its main dimensions: interpersonal relation- ships, psychological well-being, material well-being, physical well-being, personal development, social acceptance, self-determination and rights. All these dimensions are grouped by the authors into three factors: Independence (personal development and self- determination), social participation (relationships, social acceptance and rights), and well- being (psychological, physical and material). For Brooks et al. PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 Introduction [14], home confinement creates a series of problems that significantly affect people’s quality of life, including the following: • Fear of contagion and death, for them and their loved ones. • Frustration and boredom, which increases without access to social media. • Perception of information provided by health care officials as deficient. • Stress and anxiety caused by shortage of protective supplies. • Impaired mental health associated with the duration of home confinement. 2 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 PLOS ONE Quality of life scale and confinement Another interesting study carried out in Spain, Italy and the United Kingdom [15] high- lights the impact that home confinement can have on some relevant aspects of quality of life: • More than 40% of the studied population are at risk of developing mental health problems. • More than 40% of the studied population are at risk of developing mental health problems. • More than 65% considers that the information provided by public health officials in the dif- ferent countries is deficient, and that what the population needs the most to improve their quality of life is to know specific plans to return to normality. • More than 60% notes that even in this pandemic, health is no more important than the econ- omy, which indicates that the economic factor is fundamental in quality of life. Considering these observations, it is pertinent to ask whether the use of ICT during home confinement can change users’ perception of quality of life. Therefore, the main objective of this work is to adapt and validate a psychometric scale that examines the influence of ICT use on quality of life during home confinement. According to its initials in Spanish, TICO is the acronym used to refer to this scale from now on. The initial hypothesis is that ICT use improves the quality of life of the general population in a state of home confinement. To this end, we will study the influence of ICT on users’ satis- faction with life in general, with active social participation, with subjective social support, and the extent to what ICT use can reduce users’ sense of personal and social isolation, improve family dynamics and increase the psychological well-being of the confined population. Introduction To measure Quality of Life, we consider three variables to be of the utmost importance in the state of confinement: loneliness and social support, psychological well-being and satisfac- tion with life. Loneliness is associated with psychological and physical discomfort. During confinement many people have been alone and the only social link was through technology. We used a Scale of Solitude validated in adult population that studies family loneliness, marital loneliness and social loneliness [16], from which those items more representative of the state of loneliness and social support were adapted to the use of technologies. We also adapted some items from the Ryff Psychological Welfare Scale, validated in Spanish populations [17]. The adaptation of the items was taken into account according to the funda- mental variables of psychological well-being: self-acceptance, positive relationships, autonomy, control of the environment, purpose in life and personal growth, all in relation to the use of technologies. Finally, from the Life Satisfaction Scale [18, 19] of the original for world population and the Spanish adaptation of the scale, items were extracted which were adapted to the situation of confinement and the use of technologies. The scale measures subjective psychological well- being, so it fits perfectly into our study for the confinement situation. Sample and data collection An exploratory and cross-sectional study has been carried out on a non-probabilistic and sequential sample (snowball). In this sampling method, which is used in cases when it is diffi- cult to access or locate the population, data collection ends when the desired quotas are reached [20]. The process to obtain the sample was as follows: special care was taken to identify the groups, organizations and people, with whom to maintain initial contact and who could pro- vide access to a sample that was as heterogeneous as possible according to the characteristics of the study (geographical scope Spain and all age groups from 18 years onwards). They were contacted and asked to participate in sharing and assisting in the dissemination of the survey. The web link to the survey was disseminated through • Social networks: Twitter, Facebook and LinkedIn. • By email. • By email. • Through WhatsApp and Telegram groups • An advertising campaign was carried out on Facebook, through the Business Manager plat- form, segmenting the population by age and geographical area. • An advertising campaign was carried out on Facebook, through the Business Manager plat- form, segmenting the population by age and geographical area. Data was collected in Spain during the home confinement implemented in response to the COVID-19 pandemic from March 29 to May 10, 2020, using Google Forms. Consent was obtained from the University Miguel Herna´ndez of Elche Project Evaluation Board, Registra- tion 2020.228.E.OIR—Reference DPS.JGR.01.20. In order to improve the response rate, invitations were sent out weekly following the ethical principles and code of conduct of the American Psychological Association [21], the first page of the online questionnaire informed participants of the objectives and importance of the research project, the estimated completion time (between 5 and 6 minutes), and the questions to answer. They were informed of their right to decline to participate and leave the online form at any time, also about the confidentiality and anonymization of the data. To participate and begin completing the form, respondents had to first state their agreement to do so. The final sample obtained from the Spanish territory is 2,346 people, of whom 51.8% are women and 48.2% are men. Procedure The first step was to select items from the original scales developed by Cardona, Villamil, Henao and Quintero [16], Dı´az et al. [17], Atienza, Balaguer and Garcı´a-Merita [18], and Diener, Emmons, Larsen and Griffin [19], for their linguistic and cultural adaptation to the Spanish social and health context derived of the home confinement implemented in response to the COVID-19 pandemic, and to ICT use at home. Three experts examined the validity of the scale’s content. The instrument, objectives and characteristics of the research were presented to the selected experts, who studied and selected items based on their clarity, importance and relevance. Following the expert validation, 14 3 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 PLOS ONE Quality of life scale and confinement items were selected to be answered on a 7-point Likert scale, where 1 was “totally disagree” and 7 “totally agree”. Items were presented as statements towards which respondents had to indicate their degree of agreement or disagreement. Finally, the 14-item scale was pilot tested among a group of 5 men and 5 women to ensure all statements were understandable. All par- ticipants stated the scale was significant to them. PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 Instrument A purpose-created questionnaire has been designed for this study to collect information. It is divided into three blocks of information: • Sociodemographic data. • ICT use: type of technological devices used at home to connect to the Internet, media, social networks or messaging and video calling applications used during home confinement, fre- quency, and main reason for connection. • ICT use: type of technological devices used at home to connect to the Internet, media, social networks or messaging and video calling applications used during home confinement, fre- quency, and main reason for connection. • Scale of quality of life and ICT use during home confinement (TICO) adapted from Car- dona, Villamil, Henao and Quintero [16], Dı´az et al. [17], Atienza, Balaguer and Garcı´a-Mer- ita [18] and Diener, Emmons, Larsen and Griffin [19]. Data analysis Data processing and analysis has been carried out with the Statistical Package for the Social Sciences (SPSS) and the Structural Equation Modeling Software (EQS). Internal consistency has been measured with Cronbach’s alpha to determine the homogeneity of the scale items. The underlying factor structure study has been identified through exploratory factor analysis (EFA), followed by principal components analysis with varimax rotation. The Bartlett’s sphe- ricity test and Kaiser-Meyer-Olkin test have been calculated to determine whether the factorial model derived from the application of the scale is appropriate. The structure of the TICO scale has been analyzed with a confirmatory factor analysis using the following goodness of fit indi- ces: the Satorra-Bentler’s scaled chi-square test (SBχ2) [22], the Comparative fit index (R-CFI; values equal to or greater than .90 indicate an acceptable model) [23] the standardized root mean squared residual (SRMR; values less than .08 indicate an acceptable model fit), and the Root Mean Square Error of Approximation (R-RMSEA; values equal to or smaller than .06 indicate a perfect fit) [24]. Sample and data collection The average age of the sample is 45 years (SD = 13.4), distributed as follows: 25% are 18 to 34 years old, 40% are 36 to 49 years old, 30% is 50 to 64 years old, and 5% are over 65. More than half of respondents are married or have a common law partner (59.3%), 29.3% are single, 10.1% are divorced and 1.3% are widowed. The average number of people living with them during the home confinement is 3 (SD = 1.28). 66% of the sample are working out- side or at home, while 17.7% are unemployed (in working age), 10.2% are students and 6.1% are retired. In terms of monthly income level, 37.5% earn from 1,000 to 2,000 euros, 30.9% earn more than 2,000 euros, 16.1% earn less than 1,000 euros, 9% have no income, and 6.4% did not answer. 97% of the sample remain in good health, with no symptoms of the coronavirus, 2.1% have mild symptoms, but have not been tested, and 0.5% claim to have had the disease but PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 4 / 14 PLOS ONE Quality of life scale and confinement experienced mild symptoms. In terms of education level, most respondents hold a bachelor’s degree (53.6%), 23.4% hold a postgraduate or doctoral degree, 20.8% hold a secondary school or vocational training degree, while 2.8% only have basic education. experienced mild symptoms. In terms of education level, most respondents hold a bachelor’s degree (53.6%), 23.4% hold a postgraduate or doctoral degree, 20.8% hold a secondary school or vocational training degree, while 2.8% only have basic education. PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 Quality of life and ICT use during the COVID-19 confinement. Differences by age and gender The results of the survey based on the TICO scale (Table 3) show that the majority of respon- dents value their lives satisfactorily, in most respects, thanks to the use of ICT in home con- finement (Mo = 5, Me = 5), and consider that ICT use improves the circumstances derived from the COVI-19 pandemic (Mo = 6, Me = 5). They feel that when they use ICT, they are more satisfied (Mo = 6, Me = 5) and achieve the things they consider important (Mo = 5, Me = 5). Most respondents believe that ICT help them fulfill their life goals during the COVID-19 pandemic (Mo = 5, Me = 5). If they had to live again in a situation of confinement they would continue to use ICT (Mo = 7, Me = 7), because they allow them to permanently communicate with other people (Mo = 7, Me = 6), to know that their personal network of con- tacts cares about them (Mo = 6, Me = 5), and to ask for help from family and/or friends (Mo = 7, Me = 6). Boredom and loneliness connect with the use of ICT during home confinement as most respondents say they use ICT when they feel bored (Mo = 7, Me = 5) or lonely (Mo = 5, Me = 4). In fact, participants state that they manage to maintain an active social and family life thanks to the use of ICT, which allow them to set up meetings and celebrations with friends and family (Mo = 7, Me = 5). Results are very heterogeneous regarding the use of ICT for the management of emotions, such as sadness or feeling loved, during confinement. Most respondents neither disagree nor agree with the statement “I use ICTs when I feel sad” (Mo = 4, Mo = 4), as shown in Table 3. The results are clearer regarding the use of ICT when respondents “do not feel loved”, as most of them disagreement with this statement (Mo = 1, Me = 3). An ANOVA (Table 1) was performed to determine whether the results obtained on the scale were related to the gender and age variables. The results of this analysis show that, in gen- eral terms, there are no differences. In all cases, the result of the independent scale of these two factors exceeds the significance level of 0.05. ICT use in home confinement The most widely used ICT during the COVID-19 confinement have been television and the Internet. The devices used to connect to social networks and the Internet have been smart- phones (96.4%), personal computers (82%), tablets (43.4%), smart TVs (35%) and video game consoles (11.3%). However, people do not get online in just one device, but in several devices that are used interchangeably throughout the day, with an average daily frequency ranging from 1 to 3 hours. In terms of the use of social networks and instant messaging applications, there is an abso- lute dominance of WhatsApp among the population, with 99.2%, followed by YouTube (74%) and Facebook (67%). Instagram (48.4%) and Twitter (31.10%) were relegated to the last posi- tions. Social media and messaging and video-calling apps have been mainly used to communi- cate with friends (83.8%) and family (82.3%), for work-related activities (71.2%), to watch movies or TV series (68.5%), to get the news (66.4%), to study (61.9%) and listen to music 5 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 PLOS ONE Quality of life scale and confinement (61.9%). Other recurring reasons for using social media are: reading (44.1%), sports or physical activities (39.9%), seek motivation or inspiration during the coronavirus crisis (36.4%), search for cooking recipes (29.3%), play games (19.9%), listen to podcasts (18.7%) and make online purchases (17.4%). Despite the intense use of the Internet during the COVID-19 confinement, 71% of the par- ticipants stated that they have not devoted all their free time to the use of ICT. Only 29% admitted doing so. Importantly, 46.2% consider they have been happy when they have used ICT, and 72% consider that the use of ICT has brought their family members together during the COVID-19 pandemic. Quality of life and ICT use during the COVID-19 confinement. Differences by age and gender However, this does not apply to the items related to sadness, loneliness, feeling loved and social participation, which do differ according to age group and, particularly, gender. Internal consistency of the TICO scale Cronbach’s Alpha Cronbach’s Alpha based on typified elements N. of elements .889 .893 14 Source: Authors’ own creation. https://doi.org/10.1371/journal.pone.0241948.t002 Table 1. ANOVA by gender and age. would be slightly higher. However, the consistency is already excellent as it is, so it was decided to keep these two statements in the scale. The same table shows the corrected homogeneity coefficient and the item-total correlation. In all cases the resulting values are not zero or nega- tive, which means all items correlate with the total. would be slightly higher. However, the consistency is already excellent as it is, so it was decided to keep these two statements in the scale. The same table shows the corrected homogeneity coefficient and the item-total correlation. In all cases the resulting values are not zero or nega- tive, which means all items correlate with the total. Table 3. Descriptive statistics and item-total correlation of the TICO scale. Min. Max. Mode Median Corrected item-total correlation Cronbach’s alpha if item is deleted 1. In most respects, ICT make my life in confinement satisfactory. 1 7 5 5 .616 .879 2. Life in confinement has improved thanks to ICT. 1 7 6 5 .603 .880 3. I am more satisfied with my life in confinement when I use ICT. 1 7 6 5 .660 .877 4. ICT help me get “important” things done in confinement. 1 7 5 5 .593 .880 5. If I had to live in confinement again, I would continue using ICT. 1 7 7 7 .577 .882 6. Thanks to ICT, I always have someone to talk to. 1 7 7 6 .600 .880 7. Thanks to ICT, I feel people care about me. 1 7 6 5 .628 .879 8. Thanks to ICT, I can ask for help from family and friends. 1 7 7 6 .613 .880 9. Whenever I feel sad, I use ICT. 1 7 4 4 .632 .878 10. Whenever I feel lonely, I use ICT. 1 7 5 4 .653 .877 11.Whenever I do not feel loved. I lean on ICT. 1 7 1 3 .537 .883 12. When I feel bored, I turn to ICT. 1 7 7 5 .595 .880 13. ICT help me settle on meetings and celebrations with friends and family. 1 7 7 5 .403 .890 14. ICT help me have a clear purpose and direction in life. Internal consistency of the TICO scale The reliability of the TICO scale has been assessed with Cronbach’s alpha, which is a measure of internal consistency, in which a value of 0.889 (Table 2) is considered excellent [25, 26]. To determine the consistency of the scale more deeply, the Cronbach’s alpha coefficient was recalculated by incorporating in the descriptive statistics the value it takes when each item is deleted independently. The objective is to determine whether the alpha goes up or down to improve its consistency. As Table 3 shows, the value of the coefficient goes down in all cases, except for the last two items, 13 and 14, so if any of them were removed, the overall consistency The reliability of the TICO scale has been assessed with Cronbach’s alpha, which is a measure of internal consistency, in which a value of 0.889 (Table 2) is considered excellent [25, 26]. To determine the consistency of the scale more deeply, the Cronbach’s alpha coefficient was recalculated by incorporating in the descriptive statistics the value it takes when each item is deleted independently. The objective is to determine whether the alpha goes up or down to The reliability of the TICO scale has been assessed with Cronbach’s alpha, which is a measure of internal consistency, in which a value of 0.889 (Table 2) is considered excellent [25, 26]. To determine the consistency of the scale more deeply, the Cronbach’s alpha coefficient was recalculated by incorporating in the descriptive statistics the value it takes when each item is deleted independently. The objective is to determine whether the alpha goes up or down to improve its consistency. As Table 3 shows, the value of the coefficient goes down in all cases, except for the last two items, 13 and 14, so if any of them were removed, the overall consistency PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 6 / 14 PLOS ONE Quality of life scale and confinement Table 1. ANOVA by gender and age. F Sig. Whenever I feel sad, I use ICT. Sex 62.321 .000 Age 11.645 .000 Whenever I feel lonely, I use ICT. Sex 26.538 .000 Age 6.907 .000 Whenever I do not feel loved, I lean on ICT. Sex 28.493 .000 Age 9.874 .000 I use ICT to settle on meetings, celebrations and parties with my friends and family. Sex 76.161 .000 Age 50.728 .000 Source: Authors’ own creation. https://doi.org/10.1371/journal.pone.0241948.t001 Table 2. Scale reliability. Internal consistency of the TICO scale 1 7 5 5 .334 .892 Source: Authors’ own creation. https://doi.org/10.1371/journal.pone.0241948.t003 Table 3. Descriptive statistics and item-total correlation of the TICO scale. PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 7 / 14 PLOS ONE Quality of life scale and confinement Table 4. KMO and Bartlett tests. KMO measure of sampling adequacy .903 Bartlett’s sphericity test Approximate Chi-square 16357.819 Degrees of freedom 91 Significance .000 Source: Authors’ own creation. https://doi.org/10.1371/journal.pone.0241948.t004 https://doi.org/10.1371/journal.pone.0241948.t005 https://doi.org/10.1371/journal.pone.0241948.t005 Factor extraction method: Principal Components Analysis. Exploratory factor analysis Matrix of component (a) Matrix of rotated components (a) Component Component Items 1 2 3 1 2 3 1 .701 -.401 .769 2 .689 -.406 .805 3 .736 -.325 .805 4 .676 .730 5 .660 -.435 .721 .324 6 .683 .475 .768 7 .703 .418 .732 8 .689 .486 .781 9 .684 .572 .865 10 .703 .565 .866 11 .593 .575 .852 12 .657 .580 .352 13 .472 .323 .527 14 .400 .397 Extraction method: Analysis of main components. A 3 extracted components Extraction method: Analysis of main components. Rotation method: Varimax with Kaiser normalization. A Rotation has converged on 5 iterations. Source: Authors’ own creation. https://doi.org/10.1371/journal.pone.0241948.t006 Table 6. Matrix of components, without rotation and with Varimax rotation. Matrix of component (a) Matrix of rotated components (a) Component Component Items 1 2 3 1 2 3 1 .701 -.401 .769 2 .689 -.406 .805 3 .736 -.325 .805 4 .676 .730 5 .660 -.435 .721 .324 6 .683 .475 .768 7 .703 .418 .732 8 .689 .486 .781 9 .684 .572 .865 10 .703 .565 .866 11 .593 .575 .852 12 .657 .580 .352 13 .472 .323 .527 14 .400 .397 Extraction method: Analysis of main components. A 3 extracted components Extraction method: Analysis of main components. Rotation method: Varimax with Kaiser normalization. A Rotation has converged on 5 iterations. Source: Authors’ own creation. https://doi.org/10.1371/journal.pone.0241948.t006 Table 6. Matrix of components, without rotation and with Varimax rotation. https://doi.org/10.1371/journal.pone.0241948.t006 The comparison, in the matrix of rotated components, of the relative saturations of each item in each of the factors shows that the first factor consists of items 1, 2, 3, 4, 5 and 14, which seem to reflect the dimension “Satisfaction with life”. The second factor includes the variables 9, 10, 11 and 12, which could represent the “Emotional support” dimension. Finally, the third component consists of items 6, 7, 8 and 13, which tell us about the “Social Support” dimension. Exploratory factor analysis A factor analysis has been carried out to explore the validity of the construct quality of life and ICT use during confinement. The objective was to study its latent structure, identifying com- mon factors and main components, and thus identify the different dimensions that could con- stitute and represent the concept appropriately. The Kaiser-Meyer-Olkin (KMO) test and the Bartlett’s sphericity test have been performed to evaluate whether the factorial model (or the extraction of factors), as a whole, is significant (see Table 4). The KMO index yielded a value greater than 0.9, which indicates the correlation between variables is strong and that the test is very good. Bartlett’s test yielded a significance (p-value) of <0.05, so the model is significant and factor analysis can be applied. For the selection of the number of components, we explored the eigenvalues obtained in Table 5 and the percentage of variance explained. As we can see, the first three components have variances (eigenvalues) greater than 1 so, following Kaiser’s rule, they are the ones to be selected. In addition, these components or factors explain almost 64% of the variance of the original variables. The matrix of the unrotated factor structure (Table 6) indicates that all items have their highest weights in the first component, with values greater than .40 in all cases, which is the one that explains the most variance. Together with the Cronbach’s alpha coefficient, this is a good indicator of the reliability and validity of the scale. Table 5. Total variance explained. Table 5. Total variance explained. Table 5. Total variance explained. Component Initial eigenvalues Rotation sums of squared saturations Total % of variance Cumulative % Total % of variance Cumulative % 1 5.963 42.592 42.592 3.467 24.761 24.761 2 1.798 12.843 55.435 2.880 20.570 45.331 3 1.115 7.962 63.397 2.529 18.066 63.397 4 .865 6.180 69.577 5 .783 5.593 75.170 6 .644 4.597 79.766 7 .483 3.453 83.220 8 .443 3.164 86.384 9 .395 2.821 89.205 10 .378 2.701 91.906 11 .354 2.530 94.436 12 .320 2.284 96.720 13 .293 2.092 98.812 14 .166 1.188 100.000 PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 8 / 14 PLOS ONE Quality of life scale and confinement Table 6. Matrix of components, without rotation and with Varimax rotation. Confirmatory factor analysis shown in Tables 7–9, the model of three correlated factors has a good fit, within all the appro- priate intervals, which does not occur with the one-factor model. shown in Tables 7–9, the model of three correlated factors has a good fit, within all the appro- priate intervals, which does not occur with the one-factor model. Confirmatory factor analysis Based on the exploratory factor analysis, we propose a model structured around three corre- lated factors: Satisfaction with Life (6 items), Emotional Support (4 items) and Social Support (4 items). To confirm this model, a confirmatory factor analysis has been performed to com- pare it with an alternative one-dimensional model, in which all items are organized under a single factor. The goodness of fit (Table 7) has been assessed with the comparative fit index (R-CFI), the Satorra-Bentler’s scaled Chi square index (SBχ2), the standardized root mean square residual (SRMR), and the Root mean square error of approximation (R-RMSEA). As Table 7. Goodness-of-fit indices for models. Models χ2 SBχ2 df R-CFI R-RMSEA (90% CI) SRMR 1 Factor 5550.40 4316.13 77 .664 .153 (.143-.157) .103 3 Correlated factors 883.84 697.83 74 .939 .060 (.056-.064) .050 Source: Authors’ own creation. https://doi.org/10.1371/journal.pone.0241948.t007 PLOS ONE \| https://doi org/10 1371/journal pone 0241948 November 5 2020 9 / 14 Table 7. Goodness-of-fit indices for models. Models χ2 SBχ2 df R-CFI R-RMSEA (90% CI) SRMR 1 Factor 5550.40 4316.13 77 .664 .153 (.143-.157) .103 3 Correlated factors 883.84 697.83 74 .939 .060 (.056-.064) .050 Source: Authors’ own creation. https://doi org/10 1371/journal pone 0241948 t007 Table 7. Goodness-of-fit indices for models. PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 9 / 14 PLOS ONE Quality of life scale and confinement shown in Tables 7–9, the model of three correlated factors has a good fit, within all the appro- Table 9. Factor correlations. Factor F1 F2 F1 F2 .467 F3 .666 .587 All correlations are significant at the p < .01 level. Source: Authors’ own creation. https://doi.org/10.1371/journal.pone.0241948.t009 Table 8. Factor loadings for the three-factor model. Items Factor loadings SV1 .775 SV2 .790 SV3 .811 SV4 .712 SV5 .707 BP6 .364 AP4 .893 AP5 .918 AP6 .753 AP7 .624 AP1 .762 AP2 .788 AP3 .773 PS1 .441 All correlations are significant at the p < .01 level. Source: Authors’ own creation. https://doi.org/10.1371/journal.pone.0241948.t008 Table 8. Factor loadings for the three-factor model. All correlations are significant at the p < .01 level. shown in Tables 7–9, the model of three correlated factors has a good fit, within all the appro- priate intervals, which does not occur with the one-factor model. shown in Tables 7–9, the model of three correlated factors has a good fit, within all the appro- priate intervals, which does not occur with the one-factor model. PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 Conclusions This article contributes to the understanding of the influence of ICT on users’ behaviors and personal perception of quality of life during such an exceptional and extreme situation as the home confinement implemented in response to the COVID-19 pandemic. Therefore, it offers a new approach to the quality of life construct in a previously unexplored area. The main objective of this work was to analyze whether the TICO scale of Quality of Life and ICT use in confinement is a valid and reliable instrument to apply in the Spanish socio- cultural context. The results allow us to conclude that there is evidence of the psychometric quality of the scale. The estimated homogeneity analysis based on the item-total correlation coefficient confirmed its adequate levels in terms of the contribution of each item to the mea- surement of the construct. The evidence for the reliability of the TICO scale shows adequate consistency that facilitates its application. On the other hand, the three factors examined in the instrument account for the following dimensions: “Satisfaction with life”, “Emotional support” and “Social support”. These dimen- sions present a sufficient correlation level to measure the personal perception of quality of life PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 10 / 14 PLOS ONE Quality of life scale and confinement associated with ICT use, as proposed in our theoretical model based on the exploratory factor analysis. This proposal is also consistent with the previous psychometric studies used as a start- ing point for the creation of this scale [16–19]. Therefore, we confirm the three-factor corre- lated model of our hypothesis. We are aware that quality of life in relation to ICT use in confinement has increased over time in many areas. Some of the most significant are the elderly population [27–30], studies related to diet and nutrition [31, 32], physical activity [33] and education [34, 35] among oth- ers. Contrary to what could be thought a priori, in a state of confinement, ICT serve as a link between family members, as confirmed by 72% of the participants of our study. Likewise, almost 50% consider they are happy when they use ICT. Only 29% admit that they have invested all their free time in ICT consumption, which suggests that the majority of the popu- lation has been making a responsible use of ICT in this context of mandatory home confinement. Conclusions It is striking to note that the highest percentages of ICT use and consumption are related to social relations, which is what has been lacking the most in this situation, surpassing 80% of the cases studied. Participants perceive their quality of life is satisfactory thanks to the use of ICT, to the point that if they had to live a similar situation again, they would resort to ICT again. In general, people perceive that using ICT has helped them to improve the circum- stances of their life in confinement, to meet many of their material needs, to stay in touch with friends and family, to realize other people care about their health, to feel accompanied and overcome boredom. Regardless of gender or age, the use of technologies during confinement has a positive influence on the assessment of quality of life. However, when we talk about emo- tions such as sadness, loneliness or the need to feel "loved", the use of technologies as a resource, support or personal relief shows a wide range of opinions that respond to age and gender differences. Moreover, 97% of all participants stated they were in good health, with no COVID-19 symptoms. Therefore, as shown by previous studies [13, 14], physical well-being together with the social support perceived through ICT, and the satisfaction with life produced by the feeling of staying safe at home, influences the personal perception of a good quality of life during home confinement. It can be concluded that ICT use in a state of confinement significantly improves the per- ception of quality of life, especially in terms of satisfaction with life and perceived emotional and social support. The study is based on a large sample, representative of all Spanish regions and sociodemo- graphic groups. However, it is not without limitations because sampling was not random. Thus, results should be considered as an initial approach with a large sample. From the work carried out, the following future lines of research emerge: Study of the per- ception of Quality of Life in the post-Covid-19 /post confinement era; influence of ICTs on the modification of behavior and motivational levels; and perception of the time of use of ICTs and real use of them, differentiating the different social networks by use/consumption, taking into account age and gender. Supporting information Supporting information S1 File. (DOC) S2 File. (DOC) 11 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 PLOS ONE Quality of life scale and confinement Author Contributions Conceptualization: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Carmen Lo´pez-Sa´nchez, Carmen Quiles-Soler. Data curation: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Carmen Lo´pez-Sa´nchez, Carmen Quiles-Soler. Formal analysis: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Carmen Lo´pez-Sa´nchez, Carmen Quiles-Soler. Investigation: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Carmen Lo´pez-Sa´nchez, Carmen Quiles-Soler. Methodology: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Carmen Lo´pez-Sa´nchez, Carmen Quiles-Soler. Project administration: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Car- men Lo´pez-Sa´nchez, Carmen Quiles-Soler. Resources: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Carmen Lo´pez- Sa´nchez, Carmen Quiles-Soler. Software: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Carmen Lo´pez- Sa´nchez, Carmen Quiles-Soler. Supervision: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Carmen Lo´pez- Sa´nchez, Carmen Quiles-Soler. Validation: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Carmen Lo´pez- Sa´nchez, Carmen Quiles-Soler. Visualization: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Carmen Lo´pez-Sa´nchez, Carmen Quiles-Soler. Writing – original draft: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Car- men Lo´pez-Sa´nchez, Carmen Quiles-Soler. Writing – review & editing: Jose´ Antonio Garcı´a del Castillo-Rodrı´guez, Irene Ramos-Soler, Carmen Lo´pez-Sa´nchez, Carmen Quiles-Soler. PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 References 1. Ferrer L. El COVID 19: Impacto psicolo´gico en los seres humanos. Revista Arbitrada Interdisciplinaria de Ciencias de la Salud. SALUD Y VIDA. 2020; 4(7):188–199. 2. Casero-Ripolle´s A. Impact of Covid-19 on the media system. Communicative and democratic conse- quences of news consumption during the outbreak. El profesional de la Informacio´n. 2020; 29(2): e290223. https://doi.org/10.3145/epi.2020.mar.23 3. Garcı´a del Castillo J A. De problemas y prioridades. Diario Informacio´n de Alicante. 2020, April 25. https://www.diarioinformacion.com/opinion/2020/04/26/problemas-prioridades/2258965.html 4. Garcı´a del Castillo JA. Adicciones tecnolo´gicas: El auge de las redes sociales. Health and Addictions/ Salud y Drogas. 2013; 13(1):5–14. 5. Ramos-Soler I, Lo´pez-Sa´nchez C, & Torrecillas-Lacave T. (2018). Percepcio´n de riesgo online en jo´venes y su efecto en el comportamiento digital. Comunicar. 2018; 26(56):71–79. https://doi.org/10. 3916/C56-2018-07 6. Rodrı´guez O, & Estrada L. Comunicacio´n familiar en la era del vacı´o: nuevas patologı´as asociadas al uso problema´tico del mo´vil. Revista de Psicologı´a. 2019; 18(2):105–123. https://doi.org/10.24215/ 2422572Xe040 1. Ferrer L. El COVID 19: Impacto psicolo´gico en los seres humanos. Revista Arbitrada Interdisciplinaria de Ciencias de la Salud. SALUD Y VIDA. 2020; 4(7):188–199. 2. Casero-Ripolle´s A. Impact of Covid-19 on the media system. Communicative and democratic conse- quences of news consumption during the outbreak. El profesional de la Informacio´n. 2020; 29(2): e290223. https://doi.org/10.3145/epi.2020.mar.23 3. Garcı´a del Castillo J A. De problemas y prioridades. Diario Informacio´n de Alicante. 2020, April 25. https://www.diarioinformacion.com/opinion/2020/04/26/problemas-prioridades/2258965.html 4. Garcı´a del Castillo JA. Adicciones tecnolo´gicas: El auge de las redes sociales. Health and Addictions/ Salud y Drogas. 2013; 13(1):5–14. 4. Garcıa del Castillo JA. Adicciones tecnologicas: El auge de las redes sociales. Health and Addictions/ Salud y Drogas. 2013; 13(1):5–14. 5. Ramos-Soler I, Lo´pez-Sa´nchez C, & Torrecillas-Lacave T. (2018). Percepcio´n de riesgo online en jo´venes y su efecto en el comportamiento digital. Comunicar. 2018; 26(56):71–79. https://doi.org/10. 3916/C56-2018-07 6. Rodrı´guez O, & Estrada L. Comunicacio´n familiar en la era del vacı´o: nuevas patologı´as asociadas al uso problema´tico del mo´vil. Revista de Psicologı´a. 2019; 18(2):105–123. https://doi.org/10.24215/ 2422572Xe040 12 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 PLOS ONE Quality of life scale and confinement 7. Klein A. El creciente uso de las nuevas tecnologı´as en los adultos mayores. Estudos Interdisciplinares sobre o Envelhecimento. 2017; 22(2):95–111. Available from: https://seer.ufrgs.br/RevEnvelhecer/ issue/archive 8. Meng J, Martı´nez L, Holmstrom A, Chung M, & Cox J. Research on social networking sites and social support from 2004 to 2005: A narrative review and directions for future research. Cyberpsychology, Behavior and Social Networking. References 2016; 20(1):44–51. https://doi.org/10.1089/cyber.2016.0325 PMID: 28002686 9. Pe´rez V. Aproximacio´n a la investigacio´n psicolo´gica en Internet y redes sociales. Summa Psicolo´gica. 2018; 15(1):98–105. https://doi.org/10.18774/0719-448x.2018.15.vp 10. Verduyn P, Ybarra O, Re´sibois M, Jonides J, & Kross E. Do Social Network Sites Enhance or Under- mine Subjective Well-Being? A Critical Review. Social Issues and Policy Review. 2017; 11(1):274– 302. https://doi.org/10.1111/sipr.12033 11. Inchausti F, Macbeth A, Hasson-Ohayon I, & Dimaggio G. Psychological intervention and COVID-19: What we know so far and what we can do. PsyArtXiv. 2020. https://doi.org/10.1007/s10879-020- 09460-w PMID: 32836375 12. Lluch-Canut M. (comp.). Deca´logos de Salud Mental Positiva adaptados a diferentes situaciones y/o personas afectadas por circunstancias de la pandemia de Coronavirus COVID-19. Coleccio´n OMADO; 2020. http://diposit.ub.edu/dspace/handle/2445/155018 13. Schalock R, & Verdugo M. Handbook on quality of life for human service practitioners. American Asso- ciation on Mental Retardation; 2002. 14. Brooks S, Webster R, Smith L, Woodland L, Wessely S, Greenberg N, et al. The psychological impact of quarantine and how to reduce it: rapid review of the evidence. The Lancet. 2020; 395(10227):912– 920. https://doi.org/10.1016/S0140-6736(20)30460-8 PMID: 32112714 15. Codagnone C, Lupiañez-Villanueva F, Liva G, Folkvord F, Bogliacino F, Charris R, et al. Estudio sobre los efectos del COVID-19 y el confinamiento en Italia, España y Reino Unido; 2020. https://open- evidence.com/wp-content/uploads/2020/05/20-05-10-COVID19-Open-Evidence-1-wave-ES.pdf 16. Cardona JL, Villamil MM, Henao E, & Quintero A. Validacio´n de la escala ESTE para medir la soledad de la poblacio´n adulta. Investigacio´n y Educacio´n en Enfermerı´a. 2010; 28(3):416–27. Available from: http://www.scielo.org.co/scielo.php?script=sci_serial&pid=0120-5307&lng=en&nrm=iso 17. Dı´az D, Rodrı´guez-Carvajal R, Blanco A, Moreno-Jime´nez B, Gallardo I, & van Dierendonck D. Adapta- cio´n española de las escalas de bienestar psicolo´gico de Ryff. Psicothema. 2006; 18(3):572–577. Available from: http://www.psicothema.com/ 18. Atienza F, Balaguer I, & Garcı´a-Merita M. Satisfaction with Life Scale: analysis of factorial invariance across sexes. Personality and Individual Differences. 2003; 35(6):1255–1260. https://doi.org/10.1016/ S0191-8869(02)00332-X 19. Diener E, Emmons R, Larsen R, & Griffin S. The Satisfaction with Life Scale. Journal of Personality Assessment. 1985; 49(1):71–75. https://doi.org/10.1207/s15327752jpa4901_13 PMID: 16367493 20. Grande I, & Abascal E. Fundamentos y te´cnicas de investigacio´n comercial. Madrid, Spain: ESIC; 2007. https://www.esic.edu/editorial/editorial_producto.php?t=Fundamentos+y+t%E9cnicas+de +investigaci%F3n+comercial&isbn=9788417024406 21. American Psychological Association, APA. Ethical Principles of Psychologists and Code of Conduct; 2017. https://www.apa.org/ethics/code/index 22. Satorra A, & Bentler P. A scaled difference chi-square test statistic for moment structure analysis. Psy- chometrika. 2001; 66(4):507–514. http://dx.doi.org/10.1007/BF02296192 23. Schumacher R, & Lomax R. A beginner’s guide to structural equation modeling. New Jersey, NJ: Law- rence Erlbaum; 1996. 24. PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 References Browne M, & Cudeck R. Alternative ways of assessing model fit. In Bollen K., & Long J. (Eds.), Testing structural equation models. Beverly Hills, CA: Sage, 1993. 25. George D, & Mallery P. SPSS for Windows step by step: a simple guide and reference 18.0 update; 2011. Pearson. https://wps.ablongman.com/wps/media/objects/385/394732/george4answers.pdf 26. Grande I, & Abascal E. Fundamentos y te´cnicas de investigacio´n comecial. Madrid, Spain: ESIC; 2011. 27. Casamayou A, & Morales MJ. Personas mayores y tecnologı´as digitales: desafı´os de un binomio. Psi- cologı´a, Conocimiento y Sociedad. 2018; 7(2):199–226. Available from: https://revista.psico.edu.uy/ index.php/revpsicologia 28. Dı´az J, Pe´rez A, & Florido R. Impacto de las tecnologı´as de la informacio´n y las comunicaciones (tic) para disminuir la brecha digital en la sociedad actual. Cultivos Tropicales. 2011; 32(1):81–90. Available from: http://scielo.sld.cu/scielo.php?script=sci_serial&pid=0258-5936&lng=es&nrm=iso 13 / 14 PLOS ONE \| https://doi.org/10.1371/journal.pone.0241948 November 5, 2020 PLOS ONE Quality of life scale and confinement 29. Pavo´n F. Tecnologı´as avanzadas: nuevos retos de comunicacio´n para los mayores. Comunicar. 2000; 15:133–139. Available from: https://www.revistacomunicar.com/index.php?contenido= detalles&numero=15&articulo=15-2000-21 30. Rodrı´guez R. Personas mayores y aprendizaje a lo largo de la vida de las tecnologı´as de la informacio´n y la comunicacio´n. Pixel-Bit. Revista de Medios de Comunicacio´n. 2007; 41–48. https://recyt.fecyt.es/ index.php/pixel/index 31. Dı´az R, & Aladro M. Relacio´n entre uso de las nuevas tecnologı´as y sobrepeso infantil, como problema de salud pu´blica. Enfermerı´a Comunitaria. 2016; 4(1):46–51. Available from: https://www.seapaonline. org/revistas/44-rqr-enfermeria-comunitaria-vol-4-n-1-invierno-2016 32. Pe´rez C, & Gianzo M. Alimentacio´n 3.0: aspectos de intere´s en la pra´ctica diaria. Nutricio´n Hospitalaria. 2018; 35(4):85–95. https://doi.org/10.20960/nh.2132 PMID: 30070129 33. Barahona J. La enseñanza de la Educacio´n Fı´sica implementada con TIC. Revista Educacio´n Fı´sica y Deporte. 2012; 31(2):1047–1056. Available from: https://revistas.udea.edu.co/index.php/ educacionfisicaydeporte/article/view/14409 34. Garcı´a M, Reyes J, & Godı´nez G. Las TIC en la educacio´n superior, innovaciones y retos. Revista Iber- oamericana de las Ciencias Sociales y Humanı´sticas. 2017; 6(12). https://doi.org/10.23913/ricsh. v6i12.135 35. Pe´rez R, Mercado P, Martı´nez M, Mena E, & Partida J. La sociedad del conocimiento y la sociedad de la informacio´n como la piedra angular en la innovacio´n tecnolo´gica educativa. Revista Iberoamericana para la Investigacio´n y el Desarrollo Educativo. 2018; 8(16). https://doi.org/10.23913/ride.v8i16.371 14 / 14

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.